Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. A STUDY OF THE EFFECTS OF DEFOLIATION AND WATER STRESS ON GROWTH AND DEVELOPMENT OF STYLOSANTHES HAMATA ( L . ) TAUB . CV VERANO A thesis presented in partial fulfilment of the requirements for the Degree of Doctor of Philosophy in Agronomy at Massey University SAYAN TUDSRI 1 9 8 6 i ABSTRACT Verano stylo ( Stylosanthes hamata ( L • ) Taub . ) i s an important pioneer legume in the trop ics and its potential as a pasture .l egume under grazing appears to be promis ing in Thai land . This thesis was carried out in two parts - the first part was conducted in the Controll ed Cl imate Rooms at the P lant Physiology Divi sion , DSIR , Palmerston North , New Zealand . The a im of these studies was to obtain basic information on growth patterns and the response of Verano stylo to cutting at different intensities , frequencies and stages of growth and at two levels of water stress in terms of quantity and quality of herbage produced . The second part was a graz ing trial conducted at Muaklek , Thailand , to test the graz ing management hypothesis derived from the Controlled Climate Room studies . The results from the Controlled Cl imate Room studies showed that the growth and development of intact Verano stylo was slow at the pre-flowering stage and increased rapidly after the onset of flowering . Maximum growth rate of 2 . 0 4 grams/plant/day was recorded between 7 0 and 8 0 days and maximum dry weight of 1 0 5 grams/plant was achieved approximately 1 0 8 days after seedling emergence . During this post-flowering stage , plant growth in terms of plant dry weight , branch development , leaf number and leaf area increased rapidly . Flowering commenced 3 5 days a fter . seedling emergence and continued throughout the experimental period . Stem was the ma jor plant component , followed by the inflorescence and leaf fractions . In terms of the response to various cutting regimes , the results showed that the more s evere the cutting the more deleterious was the effect on regrowth . Cutting the primary branches had a greater effect on plant regrowth in terms of i i plant dry weight , branch number , leaf number and leaf area than defoliating the main stem . Severe cutting of primary branches ( i . e . to node 0 ) plus hard cutting of the main stem ( i . e . to node 3 ) resulted in the death of the plant after two cuts . When defoliation was delayed to the later stage o f growth ( near maximum growth rate ) , severe cutting of the primary branches ( i . e . to node 0 ) caused extensive plant death following only a s ingle cut . Al l growth parameters recorded were markedly reduced when the interval between cutting was decreased . It i s suggested that the response of Verano stylo to defol iation i s dependent upon the number and especially the s i z e of the primary branches , the number of growing points , the amount of stubble reserves and the residual leaf area immediately after cutting . The differences in yields were largely due to changes in the stem and to a lesser extent the inflorescence and leaf fractions . Growth of the plant in terms of plant dry weight , branch "' . . number , l eaf number and leaf area were reduced to a greater extent under severe than under mild water stress . The differences in plant dry weight between the two level s of stress were largely due to the s i ze of the stem fraction . After rewatering there was a rapid increase in growth by both the previously mild and severe water stressed plants , resulting in a marked increase of a l l the variables recorded . However , growth of plants previously under severe water stres s was less than those previously under mi ld water stres s . The increase in total plant dry weight was due to an increase in a l l plan� components , espe�ial ly leaf and inf lorescence fractions . Severity of cutting had less ef fect on plant variables than water stress . The ef fect of cutting was more apparent under mild water stress than under severe water stress in terms of plant dry weight , branch number and leaf area , and continued to show this ef fect on rewatering with respect to leaf number and leaf area . iii Verano stylo herbage quality , as measured by crude protein concentration , was re latively high even in the uncut control plants . Defoliation i nc reased the protein concentra­ tion , but within the cutting treatments there was l ittle e f fect of cutting intens ities and frequencies on the crude protein concentrations of all p lant components , except the stern fraction which was s light ly superior under frequent than infrequent cutting . The protein �entration was higher in the le af and inflorescence and lower in the stem at a l l cutting intensitie s and frequencies . Severe moisture stres s i ncreased the crude protein content in the leaves , stems and inflorescence s compared with mild moisture stress and continued to show this effect on rewatering with re spect to the l eaf and stubble f ractions . Hard cutting in the drought period also increased protein concentrations in the leaves , sterns and inf lore scences compared wit h l ax cutting and continued to show this effect on rewat ering with respect to the stubble and stern fractions . Al though the crude protein concentrations in di f ferent plant parts and for different cutting intens ities , frequencies and stages of cutting and for d i f ferent water regimes were relatively smal l , the amounts per plant were large due to the substant i al and significant dif ferences obtained in dry weight between treatments . The increase in crude p rotein was largely due to the inflorescence fraction , especia l ly under lax cutting. Crude protein yields were also serious ly reduced under frequent and hard cutting of the primary branches . Previously stressed plants at either mi ld or severe levels greatly i ncreased their crude protein yield after rewater ing , and this was largely due to the crude protein yield of the leaf and inf l orescence components . iv In terms of carbohydrate reserves , the results of this study clearly showed that the concentration of these reserves in the residual top and roots o f Verano stylo were low (< 3 % of dry weight ) , were comprised mainly of sugar and were independent of the stage , intensity and frequency o f cutting . However , carbohydrate concentrations were substantially increased by severe and especially mild water stress . Starch was the ma jor component and accumulated in a l l plant parts espe cial ly the stubble , stem and tap root fractions . The e f fects of cutting during the drought period were only evident in the stubble , inflorescence and tap root fractions - the level s declining with increasing intens ity o f defolia­ tion , particularly of the starch fraction . However , these carbohydrates , especially the starch fraction in the stubble , stem and tap root , almost totally disappeared during the rapid recovery phase , suggesting it was used for regrowth . In terms of the amounts of carbohydrates , the results showed that the di fferences between cutting intensity were largely due to the di f ferences in the residual dry weights especial ly in the stubble . Generally the more severe the cutting , the lower the amount of carbohydrates in the stubble . However , cutting frequency had no s ignificant e f f ect on carbohydrate accumulation . Severe water stressed plants accumulated only hal f the reserves of the mild water stressed plants during the drought period . Under both mild and severe water stress , the stem was the ma jor accumulator of these reserves , particularly of the starch fraction . On rewatering , there was a marked increase in the accumulation o f sugar akin to the increase in dry matter yields . However , starch yields in the stem and tap root showed a substantial drop during this period . During the drought period , hard cutting s igni ficantly depressed the accumulation of sugar and starch especia l ly under mi ld water stress . In the roots only the starch fraction was affected . On rewatering , previous hard cutting continued to depress carbohydrate yield but only of the starch fraction of those plants under previous s evere water stress . v: The results f rom the field experiment confirmed the importance of residual leaf and branch numbers on plant regrowth in terms of dry matter production , branch development , leaf number and leaf area and thei r persistence . Under cl imate room conditions , 6 weekly cutting produced signi ficantly higher yields of a l l growth components than did 3 weekly cutting . However , under field graz ing conditions frequent grazing ( every 4 weeks ) produced s ignificantly higher yields than infrequent graz ing ( every 8 weeks ) . Frequent gr�z ing also maintained a higher dens ity of Verano stylo plants and a lower weed content . The results are discussed in relation to the poss ible gra z ing management of Verano stylo in Thailand . vi ACKNOWLEDGEMENTS I am indebted to Professor B . R . Watkin , my chief supervisor , for his patience and understanding , for correcting my English , for constructive criticisms in preparation of this thesis and for advice , encouragement and guidance throughout this pro j ect . Without h i s strong recommendation to the Thai Government and Ministry of Foreign Affairs , New Zealand , this pro ject would not have been undertaken . I am also indebted to Dr . A . C . P . Chu and Dr . B . J . Forde , my eo-supervisors , for their advice , encouragement and dis­ cus s ion dur ing the experimental work and thesis preparation . I would like to express my appreciation to Dr . Sanan Jankam , local supervisor , Head of Agronomy Department , Kasetsart University , for his support , help and guidance during my f ield experiment in Thailand . I am grateful for the use of the Control led Climate facilities at the Plant Phys iology Division , DSIR , Palmerston North , and would like to thank Mr . I . Warrington , Mr . L . Ford and El izabeth Halligan for their help and interest . I am also grateful to : - Dr . R . M . Haslemore , D . S . I . R . , for teaching and providing an enzyme for carbohydrate determination ; - M . D . Hare , D . S . I . R . , for providing some useful reference s ; - Dr . M . J . Hill , Seed Technology Centre , for pro- - viding Verano stylo seed and suggestion ; - Dr . I . L . Gordon , Department of Agronomy , for his statistical help ; - Dr . A . Robertson , Department of Agronomy , for providing equipment for carbohydrate determination ; - F . J . Brown and D . T . Sol l itt , Technicians in the Agronomy Department , for helping in the carbohydrate determination ; vii - Dr . P . Waikakul , Khon Kaen University , Tha iland , for providing Verano style seed ; - Mr . Manut Hongsaprung , Deputy Director , Dairy Promotion and organi sation of Thailand , for support and kind provision of land , office and laboratory faci lities to undertake the field experiment in Thailand ; - Staff at the Dairy Promotion and organization of Thailand , Sumran Somkasem , Somkian Prasampanit and Narongrit Wonsuwat for their a s s istance with the field experimental programmes . Special thanks are extended to Kaset Wittaynuparpyuenyoung for providing accommodation and his office facilities ; - Jiradet and Apirat Sakulneeya for their hospitality while I did my f ield experiment in Thai land ; - Mr . Sawad Utamong , head of the Packchong Forage Crop Station , and his staf f , Miss Pensri Sornprasit for providing a mowing machine and Verano style seed ; - Mrs . Griselda Bla zey for typing the manuscript and making helpful suggestions regarding presentation ; - G . Hal l igan for drawing the figure s ; I a lso wish to express my sincere appreciation to Mrs . Watkin and Mrs . Jenny Chu for their hospital ity while I was in New Zea land . Mr . T . Na Nagara and his family are a l so acknowledged . The help of fel low graduate students , D . Smith and V . Suruprasert , is greatly appreciated . Fina l ly , my grateful thanks to my mother , s i ster and brother ( Kumpong , Pensri , Samlee , Dr . Sithichai and Sopa Tudsri ) for their mora l support while I was in New Zealand and also Pranee Tudsri , my wife , for encouragement , understanding and patience . I gratefully acknowledge the New Zealand Government for the opportunity to undertake this study and Kasetsart Univers ity for al lowing me to carry on with this study . Abstract Acknowledgements List of Tables List of Figures List of Plates List of Appendices TABLE OF CONTENTS CHAPTER 1 . INTRODUCTION AND OBJECTIVES viii Page i vi xi i i xvii i xxiv xxviii CHAPTER 2 . LITERATURE REVIEW 5 2 . 1 Stylosanthes hamata cv Verano 5 2 . 1 . 1 Origin , Taxonomy , Distribution and Ecology 5 2 . 1 . 2 Morphological and Growth Habit 7 2 . 1 . 3 Fertil i zer , Nodulation and Yield 8 2 . 2 Effect of Defoliation on Tropical Legumes 1 0 2 . 2 . 1 Morphology and Growth Habit 1 1 2 . 2 . 2 Regrowth Following Cutting 1 2 2 . 2 . 3 Dry Matter Yield 1 8 2 . 2 . 4 Underground Plant Organs 1 9 2 . 2 . 5 Herbage Qual ity 2 1 2 . 3 Ef fect of Defoliation on Stylosanthes hamata cv Verano 2 2 2 . 3 . 1 Morphology and growth habit 2 2 2 . 3 . 2 Dry matter yield 2 3 2 . 3 . 3 Underground plant organs 2 4 2 . 3 . 4 Herbage qual ity 2 4 2 . 4 Ef fect of Water Stress on Tropical Legumes 2 4 2 . 4 . 1 Germination , Establishment and Survival 2 5 2 . 4 . 2 Morphological Effects 2 8 2 . 4 . 3 Phys iological Ef fects 2 9 2 . 4 . 4 Dry Matter Yield 2 . 5 Management of Tropical Forage Legumes 2 . 6 Management of Stylosanthes hamata cv Verano ( Verano style ) 3 0 3 4 4 2 CHAPTER 3 . EXPERIMENT 1 : GROWTH PATTERN OF STYLOSANTHES HAMATA CV VERANO UNDER CONTROLLED ENVIRONMENT CONDITIONS I . Introduction I I . Materials and Methods A . Environmental Conditions and Planting Procedures B . Harvesting Schedule C . Plant Measurements I I I . Results and Discus sion A . Morphological Features B . Growth and Development CHAPTER 4 . THE EFFECT OF DEFOLIATION ON THE GROWTH AND REGROWTH CHARACTERISTICS OF STYLOSANTHES 4 6 4 6 4 7 4 7 4 8 4 8 4 8 4 8 5 1 HAMATA CV VERANO 5 6 EXPERIMENT 2 : STAGE OF GROWTH AND INTENS ITY 5 6 I . Introduction 5 6 I I . Materials and Methods 5 7 A . Environmental Conditions and Planting Procedures 5 7 B . Treatments 5 7 C . Harvesting Schedule 5 8 D . Plant Measurements 5 9 E . Chemical Measurements 6 0 F . Statistical Analysis 6 0 I I I . Results 6 2 A . Phenological Observations 6 2 B . Plant Regrowth 6 2 C . Chemical Composition 6 6 D . Rel ationship Between Regrowth and Res idual ( stubble ) Plant Variables 6 8 E . Rel ationship Between Total Plant Dry Weight and Growth Parameters ( Branch Number , Leaf Number and Leaf Area per Plant ) 6 8 IV . Discuss ion 6 9 ix EXPERIMENT 3 : INTENSITY OF DEFOLIATION I . Introduction I I . Materials and Methods A. Environmental Conditions and Planting Procedures B . Treatments C . Number of Cutting Occasions D . Harvesting Schedule E . Plant Measurements F . Chemical Measurements G. Statistical Analysis I I I . Results A. Phenological Observation B . Plant Regrowth C . Chemical Composition D . Relation Between Regrowth Yield and Residual Plant Variables E . Relationship Between Tota l Plant Dry Weight and the Main Growth Parameters ( Branch Number , Leaf Number and Leaf Area ) IV . Discussion EXPERIMENT 4 : FREQUENCY OF DEFOLIATION I . Introduction I I . Materials and Methods A. Environmental Conditions and Planting Procedures B . Treatments C . Plant Measurements D . Chemical Measurements E . Statistical Analysis I I I . Results A. Phenological Observations B . Plant Regrowth C . Chemical Compos ition D . Relationship Between Plant Dry Weight and the Main Growth Parameter s ( Branch Number , Leaf Number and Leaf Area ) IV . Discussion 7 5 7 5 7 6 7 6 7 6 7 7 7 7 7 7 7 8 7 8 7 8 7 8 8 0 8 2 8 4 8 4 8 5 9 0 9 0 9 1 9 1 9 1 9 2 9 3 9 3 9 3 9 3 9 4 9 6 9 7 9 8 X CHAPTER 5 . EXPERIMENT 5 : EFFECT OF WATER STRESS AT DIFFERENT DEFOLIATION LEVELS ON REGROWTH CHARACTERISTICS OF STYLOSANTHES HAMATA ( VERANO STYLO ) 1 0 1 I . Introduction 1 0 1 I I . Materials and Methods 1 0 2 A . Environmental Conditions and Planting Procedures 1 0 2 B . Treatment 1 0 3 C . Measurements 1 0 4 D . Statistical Analysis 1 0 8 I I I . Results 1 0 8 A . Phenological Observations 1 0 8 B . Plant Regrowth 1 0 9 C . Chemical Composition 1 1 6 D . Relationship Between Total Plant Dry Weight and the Main Growth Parameters ( Branch Number , Leaf Number and Leaf Area ) 1 2 1 IV . Discuss ion 1 2 1 CHAPTER 6 . EXPERIMENT 6 : EFFECT OF GRAZING MANAGEMENT ON STYLOSANTHES HAMATA PRODUCTION AND SURVIVAL I . Introduction I I . Materials and Methods A . Environmental Conditions and Planting Procedures B . Treatments C . Measurements I I I . Results A . Seedling Establishment B . Plant Regrowth C . Weed Content IV . Discus sion 1 2 8 1 2 8 1 2 9 1 2 9 1 3 0 1 3 1 1 3 2 1 3 2 1 3 2 1 3 4 1 3 7 xi CHAPTER 7 . GENERAL DISCUSSION A . Growth and Development B . Defol iation and Plant Growth C . Water Stress and Plant Growth D . Herbage Quality E . Management Recommendations E . 1 Cutting E . 2 Grazing APPENDICES BIBLIOGRAPHY xii 1 3 9 1 3 9 1 4 0 1 4 3 1 4 6 1 5 0 1 5 0 1 5 1 1 5 4 1 6 0 Table 1 . 1 Table 1 . 2 Table 1 . 3 Table 1 . 4 Table 2 . 1 Table 2 . 2 Table 2 . 3 Table 2 . 4 Table 2 . 5 Table 2 . 6 LIST OF TABLES Number of branches per plant Plant height of Verano style ( cm ) Leaf number , leaf s i z e , dry weight per leaf and specific leaf area Inflorescence number and dry weight Plant components analysed Ef fect of stage and intensity of x i i i Page 5 0 a 5 1 b 5 3b 5 3 c 6 0 a defoliation on branch number per plant 6 4b Ef fect of stage and intensity of defoliation on leaf area ( cm2 ) ( LA ) and leaf number ( LNO ) per plant Ef fect of stage and intensity of defoliation on number o f " growing points " on the stubble 1 0 days a fter defoliation ( no/plant ) Ef fect of stage and intensity of defoliation on crude protein concentration ( % of dry matter ) in leaf ( L ) , stem ( S ) , inflorescence ( I ) and root ( tap ( T ) and f ibrous ( F ) ) components and in the 6 5b 6 6 a res idual stubble ( R ) 6 6b Correlations o f residual plant variables with net regrowth yield 6 8a Table 2 . 7 Table 3 . 1 Table 3 . 2 Table 3 . 3 Table 3 . 4 Table 3 . 5 Table 3 . 6 Table 3 . 7 Table 3 . 8 Linear correlation coefficients between plant dry weight ( DM ) and various main growth parameters ( branch number ( B ) , leaf xiv number ( LNO ) and leaf area ( LA ) 6 8b Plant components analysed Ef fect of defoliation intens ity on net regrowth yield over the experimental period ( g/plant ) Ef fect of defol iation intens ity on res idual leaf area ( crn2 ) ( LA ) and leaf number per 7 8a SO a plant remaining immediately after each cut 8 1 e Ef fect of defoliation intensity on number of "growing points " on the stubble 1 0 days after cutting Ef fect of defoliation intensity on crude protein concentration ( % of dry matter ) in leaf ( L ) , stem ( S ) , inf lorescence ( I ) , root ( tap ( T ) and f ibrous ( F ) components and in the residual stubble ( R ) Ef fect o f defoliation intensity on sugar concentration ( % of dry matter ) in the residual above ground ( stubble ) and below ground ( roots ) Ef fect of defoliation intensity on starch concentration ( % of dry matter ) in the residual above ground ( stubble ) and below ground ( roots ) Correlations of residual plant variable s with net regrowth yield 8 2 a 8 2b 8 3b 8 3d 8 4a Table 3 . 9 Linear correlation coefficients between plant dry weight ( DM ) and main growth parameters ( branch number ( B ) , leaf number ( LNO ) and leaf area ( LA ) 8 4b Table 3 . 1 0 Effect of defoliation intensity on number of residual branches per plant 8 5 a Table 4 . 1 Table 4 . 2 Table 4 . 3 Table 4 . 4 Table 4 . 5 Table 4 . 6 Table 4 . 7 Plant components analysed Ef fect of intensities ( A ) and frequencies ( B ) of defoliation on absolute growth rate ( mg/day ) Ef fect of intensities ( A ) and frequencies ( B ) of defoliation on res idual leaf area 2 ( cm /plant ) Ef fect of intensities ( A ) and frequencies ( B ) of defoliation on res idual leaf number ( no/plant ) Ef fect of intensities ( A ) and frequencies ( B ) o f defoliation on the number of " growing points " on the stubble ( no/plant ) 1 0 days after cutting 9 3b 9 4e 9 5d 9 5e 9 6a Ef fect of intensities and frequencies of defoliation on crude protein concentration in the leaf ( L ) , stem ( S ) , inf lorescence ( I ) , root ( tap ( T ) and fibrous ( F ) ) components and in the residual stubble ( R ) Ef fect of intensities and frequencies o f defoliation on sugar concentration ( % of dry matter ) in residual above ground ( stubble ) and below ground ( roots ) 9Gb 9 7a XV Table 4 . 8 Table 4 . 9 Table 5 . 1 Effect of intensities and frequencies o f defoliation on starch concentration ( % o f dry matter ) i n residual above ground ( stubble ) and below ground ( roots ) Linear correlation coeff icients between plant dry weight ( DM ) and main growth parameters ( branch number ( B ) , leaf number xvi 9 7b ( LNO ) and leaf area ( LA ) ) 9 7d Soil moisture content ( % of dry weight ) during drought and recovery period 104a Table 5 . 2 .Plant components analysed 1 0 8a Table 5 . 3 Table 5 . 4 Table 5 . 5 Table 5 . 6 Table 5 . 7 Table 5 . 8 Main ef fect of water stress and defoliation on total plant dry weight during drought and recovery period ( g/plant ) Effect of water stres s and defoliation on absolute growth rate and relative growth rate during drought and recovery period Main ef fect of water stress and defoliation on rate of branching during drought and recovery period ( no/plant/day ) Effect of water stres s and defol iation on leaf appearance rate ( no/plant/day ) Effect of water stre s s and defol iation on 1 0 9a 1 1 0a 1 1 2a 1 1 4a leaf s i ze ( cm2/leaf ) 1 1 4a Effect of water stres s and defoliation on specific leaf area ( cm2/g ) 1 1 5a xvi i Table 5 . 9 Main e ffect of water stres s and defoliation on crude protein concentration in the leaf , stem , inflorescence , stubble and tap and fibrous root components ( % of dry matter ) 1 1 6a Table 5 . 1 0 Main effect of water stress and defoli ation on total crude protein ( leaves + stem + stubble + inflorescence ) during drought and recovery period ( mg/plant ) Table 5 . 1 1 Linear correlation coefficients between plant dry weight ( OM ) and main growth parameters ( branch number { B ) , leaf number 1 1 7b ( LNO ) and leaf area ( LA ) ) 1 2 1 a Table 6 . 1 Plant density on the day before first graz ing and at the end of the experiment ( plants/m2 ) 1 3 2a Figure 1 . 1 Figure 1 . 2 Figure 1 . 3 Figure 1 . 4 Figure 1 . 5 Figure 1 . 6 Figure 1 . 7 Figure 2 . 1 Figure 2 . 2 Figure 2 . 3 xvi i i LIST OF F IGURES Page Length of primary branches along the main stem ( cm ) 5 1 a Changes in total dry weight with time : A . Actual data ; B . From fitted data ; 1 and 2 represent line of best fit between each phase of growth Relative growth rate of Verano stylo calculated from f itted growth model ( Figure 1 . 2 ) Relative growth rate of Verano stylo calculated from fitted growth model ( Figure 1 . 2 ) Dry matter yield of leaf , stem , inf lorescence and roots of Verano sty lo ( g/plant ) Distribution of dry matter yield on the primary branches of the main stem B = lower primary branches ( 1 6 ) T = upper primary branches ( 7 1 2 ) Changes in total leaf area of Verano s tylo 5 1 c 5 2a 5 2 a 5 3a 5 4a with time ( cm2/plant ) 5 5a Diagrammatic i l lustration of the appearance of the plants after dif ferent intensities of defoliation . Harvesting schedules over the experimental 5 8a period 5 8b Total dry weight ( cut at ground leve l ) A . Cut at early stage of growth ( g/plant ) B . Cut at late stage of growth ( g/plant ) 6 2 a Figure 2 . 4 Figure 2 . 5 Figure 2 . 6 Figure 2 . 7 Figure 2 . 8 Figure 2 . 9 Effect of stage and intensity of defoliation on the components of plant xix dry weight ( g/plant ) 6 3a Ef fect of stage and intensity of defoliation on the absolute growth rate ( g/day ) 63b Ef fect of stage and intensity of defoliation on the relative growth rate ( mg/mg/day ) 6 3c Ef fect of stage and intensity of defoliation on total net regrowth yield over ten weeks with two successive defoliations ( g/plant ) Effect of stage and intensity of 6 4a defoliation on rate of branching ( no/day ) 6 5a Effect of stage and intensity of defoliation on crude protein yield in l ea f , stem ( including stubble ) , inf lorescence and roots components ( g/plant ) 6 6c Figure 2 . 1 0 Ef fect of stage and intensity of defoliation on sugar and starch concentration in the res idual top ( stubble ) and below ground ( % of dry matter ) Figure 2 . 1 1 Ef fect of stage and intensity of defoliation on sugar yield in the residual top ( stubble ) and below ground 6 7 a ( mg/plant ) 6 7b Figure 3 . 1 P lanning of Experiment 3 7 7 a Figure 3 . 2 Figure 3 . 3 Figure 3 . 4 Figure 3 . 5 Figure 3 . 6 Figure 3 . 7 Figure 3 . 8 Figure 3 . 9 Main ef fect of defoliating the main stem ( A ) and the primary branches ( B ) on total net regrowth yield over the four regrowth cycles 8 0b Ef fect of defoliation i ntensity on total plant dry weight at various cutting cycles ( g/plant ) Ef fect of defoliation i ntensity on absolute ( A ) and relative ( B ) growth rate at various regrowth cycles Ef fect of defoliation i ntensity on total net regrowth yield over the four regrowth 8 0 c 8 0d cycles ( g/plant ) 8 1 a Ef fect of defoliation i ntens ity on the components of plant dry weight ( g/plant ) E f fect of defoliation i ntensity on branch number per plant Effect of defoliation i ntensity on leaf 2 area ( cm ) and leaf number per plant at all regrowth cycles Effect of defoliation i ntensity on tota l crude protein yield a t cycle 1 and 4 ( g/plant ) 8 1 b 8 1 c 8 1 d 8 2 c Figure 3 . 1 0 E f fect o f defoliation intensity on crude protein yield in the plant components ( g/plant ) at cycle 1 and 4 8 3a Figure 3 . 1 1 E f fect of defoliation intens ity on sugar yield in the residual top ( stubble ) and below ground ( roots ) ( mg/plant ) 8 3c Figure 4 . 1 Planning of Experiment 4 9 2a XX Figure 4 . 2 Figure 4 . 3 Figure 4 . 4 Figure 4 . 5 Figure 4 . 6 Figure 4 . 7 Figure 4 . 8 Figure 4 . 9 Figure 5 . 1 Figure 5 . 2 xxi Ef fect of cutting intensities and f requencies on the net regrowth yields of Verano stylo over the experimental period 9 4 a Total dry weight fol lowing each cutting : A . cutting every 3 weeks at dif ferent inten�itiesi B . cutting every 3 weeks vs cutting every 6 weeks at the same intensity Ef fec� of cutting intensities ( A ) and cutting frequencies ( B ) on regrowth 9 4b components ( g/plant ) 9 4 c and 9 4d Branch number per plant : A . cutting every 3 weeks at dif ferent intensitiesi B . cutting every 3 weeks vs 6 weeks at the same intensity Ef fect of cutting intensities ( A ) and cutting frequencies ( B ) on l eaf area per 9 5a plant 9 5b Effect of cutting intensities ( A} and f requencies ( B ) on leaf number per plant 9 5 c Effect o f cutting intens ities ( A ) and frequencies ( B ) on crude protein yield per plant 9 6c Effect of cutting intensities ( A} and cutting frequencies ( B ) on total non­ structural carbohydrate ( TNC ) yield ( mg/plant ) 9 7 c Mean percentage o f relative water content ( % RWC ) 1 0 S a Effect of water stre s s and defol iation on total plant dry weight ( g/plant} 1 0 9 b Figure 5 . 3 Figure 5 . 4 Figure 5 . 5 Figure 5 . 6 Main effect of water stres s ( A ) and defoliation ( B ) on plant components dry weight ( g/plant ) Effect of water stress and defoliation on plant components dry weight ( g/plant ) Effect of water stress and defoliation on total branch number per plant during drought and recovery period Ef fect of water stress and defoliation on leaf area ( cm2 ) and leaf number per xxii 1 1 Ob 1 1 0 c 1 1 1 a plant during drought and recovery period 1 1 3a Figure 5 . 7 Figure 5 . 8 Figure 5 . 9 Effect of water stress and defoliation on crude protein concentration in the plant components ( % of dry matter ) . Note : dif ferent letters above column and in the same shade differ at the 5 % level Effect of water stress and defoliation on crude protein yield in the plant components ( �g/plant ) Main effect of water stress ( A ) and defoliation ( B ) on carbohydrate concentration in the plant components ( % of dry matter ) Figure 5 . 1 0 Main ef fect of water stress ( A ) and defoliation ( B ) on non-structural carbohydrate concentration in the tap and fibrous roots ( % of dry matter ) Figure 5 . 1 1 Ef fect of water stres s and defoliation on total non-structural carbohydrate ( TNC ) yields ( mg/plant ) 1 1 7 a 1 1 7c 1 1 8a 1 1 9 a 1 1 9b xxi i i Figure 5 . 1 2 Ef fect o f water stress and defoliation on sugar yields of the plant components ( mg/plant ) 1 1 9 c F igure 5 . 1 3 Effect of water stress and defoliation on starch yields of the plant components ( mg/plant ) 1 1 9d F igure 5 . 1 4 Effect of water stress and defoliation on total non-structural carbohydrate yield in the roots ( tap + fibrous ) ( mg/plant ) 1 2 0 a F igure 6 . 1 Figure 6 . 2 Figure 6 . 3 Figure 6 . 4 Figure 6 . 6 Figure 6 . 7 Figure 6 . 8 Figure 6 . 9 Monthly rainfall and maximum and minimum temperatures at Muaklek , Thai land 1 3 0 a Timing of graz ing and regrowth period throughout the experiment Changes in legume plant density with time 1 3 0b ( number/m2 ) 1 3 2b Total net regrowth yield over the exper imental period ( ton/ha ) 1 3 2c Effect of gra z ing management on dry matter production ( ton/ha ) 1 3 3a Total plant dry weight ( g/plant ) 1 3 3b Ef fect of graz ing management on the plant components dry weight ( g/plant ) 1 3 3c Ef fect of gra zing management on branch number per plant 1 3 3d Ef fect of graz ing management on leaf area 2 ( cm ) and leaf number per plant 1 3 4a Figure 6 . 1 0 Percentage of weeds during the experimental period 1 3 4b Plate 1 . 1 P late 1 . 2 Plate 2 . 1 Plate 2 . 2 Plate 2 . 3 Plate 2 . 4 Plate 2 . 5 Plate 3 . 1 Plate 3 . 2 xxiv LIST OF PLATES Page General view of the Controlled Climate Room Growth and development of Verano style a fter seedling emergence up to first f lowering appearance Uncut control treatment at early stage ( 3 9 days after seedling emergence ) Immediately after lax defol iation of the primary branches ( E-5-4 ) at early stage ( 3 9 days after seedling emergence ) Immediately after hard defoliation of the primary branches ( E-5-0 ) at early stage ( 3 9 days after seedling emergence ) Immediately after lax defoliation of the primary branches ( L-5-4 ) at late stage of growth ( 8 8 days after seedling emergence ) . Note : L- 5-4 ( forward ) and uncut control ( backward ) Immediately after hard defoliation of the primary branches ( L-5-0 ) at late stage of growth ( 8 8 days after seedling emergence ) The cutting treatments immediately after the first cut ( LH = E- 3- 0 ; LL = E- 3- 4 ; HH = E-7-0 ; HL = E-7-4 ; C = uncut control ) Cutting treatments immediately after the second cut ( for the symbols see Plate 3 . 1 ) showing a greater number of leaves on plant 4 7a 4 9 a 6 1 a 6 1 a 6 1 a 6J 6 1 7 9 a HL ( l ightest cutting treatment : E-7 - 4 ) 7 9 a P late 3 . 3 P late 3 . 4 Plate 4 . 1 Plate 5 . 1 Plate 5 . 2 Plate 5 . 3 Plate 5 . 4 The cutting treatments immediately after the third cut ( for the symbols see Plate 3 . 1 ) showing that only a small number of leaves and branches remain under all treatments but the s i ze of the residual stubble differs greatly between the E- 3 - 0 ( LH ) and the E-7- 4 ( HL ) treatment 7 9 The cutting treatments immediately a fter the fourth cut ( for the symbol s see Plate 3 . 1 ) showing the negl igible number of leaves or active branches along the primary branches 7 9 Comparison o f cutting intensities and frequencies taken immediately prior to the third 3 weekly cutting and second 6 weekly cutting ( Note : greater production under 6 weekly versus 3 weekly cutting ; greater production under lax versus severe cutting XXV of the main stem ) 9 3a Plant under mild water stress compared with non-stressed plant immediately before the cutting treatment was imposed ( Note : plant is slightly wilted under mild water stress ) 1 0 3 a Plant under severe water stress compared with non-stres sed plant immediately before the cutting treatment was imposed ( Note : death·and signif icant wilting of leaves on the main stem under severe water stres s A general view o f the plants under the mild water stress six weeks after cutting ( Note : signif icant retention of leaves ) A general view of the plants under severe water stress s ix weeks after cutting ( Note : large number of leaves senescent and fall ing of f ) 1 0 3 a 1 0 6 a 1 0 6 a Plate 5 . 5 Plate 5 . 6 Plate 5 . 7 P lants under mild water stres s nine weeks a fter cutting Plants under severe water stress nine weeks a fter cutting { Note : maintenance of fresh green leaves and stems under mild water stress compared with grey-green leaves and reddish-purple stems under severe water stress xxvi 1 0 6 1 0 6 Cutting treatments of plants under mild water stress twelve weeks after cutting { Note : clusters of small branches along the primary branches ) . E-6-4 = W 1 C 1 , E-4-4 = W 1 C2 and E-2-4 = W 1 C 3 1 0 7 a Plate 5 . 8 Plate 5 . 9 Cutting treatments of plants under severe water stress twelve weeks after cutting { Note : fewer c lusters of branches along the primary branches compared with plants under mild water stress , Plate 5 . 7 ) . E-6-4 = W2C 1 , E- 4-4 = W2C2 and E-2- 4 = W2C3 Cutting treatments of plants under previous mild water st-ress three weeks after re-watering ( Note : s igni ficant increas e in plant si ze and in number of branches and 1 0 7 a leaves ) 1 0 7 Plate 5 . 1 0 Cutting treatments of plants under previous severe water stress three weeks after re-watering ( Note : s ignificant increase in plant s i ze and in number of branches and Plate 6 . 1 leaves ) 1 0 7 A general view of the experimental s ite , showing the well prepared seed-bed before sowing 1 2 9 a Plate 6 . 2 Plate 6 . 3 Plate 6 . 4 The area was irrigated a fter sowing to ensure good germination A general view showing good Verano style establishment xxvii 1 2 9 a 1 3 5 a Immediately before first graz ing when approximately 5 0 % of plants begin to f lower - approximately 30 - 3 5 cm in height 1 3 5 a Plate 6 . 5 Plate 6 . 6 Plate 6 . 7 Plate 6 . 8 Plate 6 . 9 Dry cows on Verano style at first graz ing ( Note : ready acceptance by grazing anima l s ) 1 3 5 Immediately after first grazing - plant height approximately 1 2 - 1 5 cm Immediately before third graz ing of four weekly interval s ( Note : l arge number of green leaves and branches close to ground ) Immediately before second grazing at eight weekly intervals ( Note : negligible number of green leaves and branches close to the ground ) During second graz ing at eight weekly intervals and third graz ing at four weekly intervals ( Note : Ease o f prehension of four weekly grazing and severe wastage of eight 1 3 5 1 36a 1 36a weekly graz ing ) 1 36 Plate 6 . 1 0 Four weekly grazing treatment at the end of experimental period ( Note : maintenance of relatively dense and c lean stand of Verano style ) 1 3 7 a Plate 6 . 1 1 Eight weekly graz ing treatment at the end of experimental period ( Note : relatively l ow dens ity of Verano style and subsequent ingression of weeds ) 1 3 7 a LIST OF APPENDICES Controlled environment conditions 2 Cl imate Lab - N . C . U . S . Phytotron Nutrient 3 Changes in total dry weight with time : A . Fitted with logistic growth model B . Fitted with polynomial growth model xxviii Page 1 5 4 1 5 5 ( 0 - 1 3 1 days after seedl ing emergence ) 1 56 4 Ef fect of defoliation on total dry matter ( g/plant ) 1 5 7 5 Effect of defoliation on total plant dry weight ( g/plant ) 1 5 8 6 Effect of water stress and defol iation on total non-structural carbohydrate concentration of the plant components ( % of dry matter ) 1 5 9 1 CHAPTER 1 INTRODUCTION AND OBJECTIVES More than 5 0 % of the world ' s ruminant animal population i s in the tropics and subtropics ( 3 0°S to 3 0°N ) ( Whiteman , 1 9 8 0 ) . These animals are mostly raised on poor natural grassland ( Whiteman , 1 9 8 0 ) and less than 5 % of these pastures have been improved ( Mannet j e , 1 9 7 8 ) . The yields of these natural grasslands are generally low and o f poor quality due ·to rapid maturity producing feed of low protein content and resulting in low animal intake of digestible nutrients ( Shaw and Bisset , 1 9 5 5 ; Norman , 1 9 6 2 ) . Even with cultivated gra s s species , the qual ity i s still relatively low when compared with the temperate species ( Minson and McLeod , 1 9 7 0 ) and i s a lso highly variable between genera and species ( Reid e t a l , 1 9 7 3 ) . Hence , productivity i s poor compared with that of temperate regions ( Whiteman , 1 9 8 0 ) . To overcome the nutritional limitations of natural and cultivated tropical gras sland , the use of supplementary feeding , of nitrogen in con j unction with other fertilisers and of nodulated legumes , has been proposed by several workers ( Jones , 1 9 7 2 ; Humphreys , 1 9 7 8a ; Whiteman , 1 9 8 0 ; Murray , 1 9 8 3 ) . Of these alternatives , the first two approaches are rather expensive and in the case of nitrogen fertilisers , are too dependent and variable according to the f luctuating environments ( Jones , 1 9 7 2 ; Whiteman , 1 9 8 0 ) . The most feasible and economical approach on a large scale bas i s is through the u s e o f suitable legumes . The role of legumes in temperate pastures to increase the productivity through symbiotic nitrogen fixation and increased feed quality , i s well known . Much research over the years has centred on an effort to f ind suitable legumes for the tropics ( Mannet j e 1 9 7 8 ; Whiteman , 1 9 7 8 ) . Austral ian scientists have played a ma jor role in this area of research and development with the succes s ful introduction of Townsville style ( Stylosanthes humi lis kunth . ) in th� early 1 9 0 0 ' s ( Humphreys , 1 9 6 7 ) to improve the natural grassland in 2 the northern part of Australia . Improvement of these natural grasslands encouraged scientists to col lect and introduce new species from many parts of the tropics and subtropics for breeding and selec­ tion and subsequently the release , as commercial species , of a number of important legume species such as Siratro ( Macroptilium atropurpureum ( DC ) Urb . ) , Stylo , Desmodium , etc . ( Burt et al , 1 9 7 1 ; Edye et al , 1 9 7 4 ; Hutton and Beal , 1 9 7 7 ; Whiteman , 1 9 7 8 ; Imrie et al 1 9 8 3 ; Edye and Grof , 1 9 8 3 ) . However , the search for species suitable for particular areas is difficult s ince a wide range of climatic and edaphic conditions in the tropics and subtropics i s encountered . These include low rainfall , high variabil ity of precipitation ( Skerman , 1 9 7 7 ; Whiteman , 1 9 8 0 ) , low soil ferti l ity and in some areas very low soil pH ( Edye and Grof , 1 9 8 3 ) . A short rainy season and a long dry spell is the common occurrence in the monsoonal countries such as Northern Australia , Thailand , Burma , Laos and India , whereas in parts of Latin America and Indonesia , heavy rainfall occurs throughout the year ( Crowder and Chheda , 1 9 8 2 ; Edye and Grof , 1 9 8 3 ) . Species that have a high genetic variation tend to adapt wel l in the former conditions ( Staples , 1 9 8 1 ) and are illustrated by the annual pasture legume , Townsvi lle stylo , in the northern part of Austral ia 1 967 ; Whiteman , 1 9 7 8 ) . which has been recognised since 1 9 0 0 ( Humphreys , Townsvi l le stylo has proven successful for many years and in many tropical countries such as Thailand ( Humphreys , 1 9 7 8 a , 1 9 8 4 ) . However , it is a poor competitor in associa­ tion with the vigorous native grasses ( Ritson et al , . 1 9 7 1 ; Tors sel l , 1 9 7 3 ) and i s easily attacked by the fungus disease Anthracnose ( Colletotrichum gloeosporioides ) , resulting in a marked decl ine in yield and seed production ( Robertson , 1 9 7 8 ; Humphreys , 1 9 8 4 ) . However , with the introduction and testing of a large number of lines of Stylosanthes species into Australia in 1 965 ( Burt et al , 1 9 7 1 , 1 9 7 4 ; Edye et al , 1 9 7 4 , 1 9 7 5a , 1 9 7 5 b , 1 9 76 ) improved cultivars have been selected and released , such as Stylosanthes hamata ( L . ) Taub . 3 cv Verano , S . scabra Vog . cv Seca and cv Fitzroys in 1 9 7 3 ( Edye e t al , 1 9 7 5 b , 1 9 76 ; Edye and Grof , 1 9 8 3 ) . These three perennial cultivars appear to be more productive than the annual Townsville stylo because of their abil ity to produce herbage more quickly fol lowing the opening of the rainy season . This appl ies in pa�ticular to S . hamata cv Verano ( Gi l lard et a l , 1 9 8 0 ) . They are also able to respond to unseasonal rain s . With their longer growing season and greater height and yield , these perennial legumes maintain a higher content in the past-ure and a greater botanical stability ( Gil lard et al ' 1 9 8 0 ) . In addition , s. -- hamata is more resistant to Anthracnose ( Humphreys , 1 9 7 8b ; Whiteman , 1 9 7 8 ; Vini j sanond and Topark-Ngarm , 1 9 7 8 ) . This pattern of development is ideal for regions with a highly variable cl imate such as Australia and Thailand , and it appears to be adapted to a wider range of soil type s , and establishes better under lower rainfall than Townsville stylo ( Burt et al 1 9 7 4 ; Edye et a l , 1 9 7 Sa , 1 9 76 ; Gil lard et al , 1 9 8 0 ) . The better adaptability of S . hamata to low soil fertility and rainfall conditions , has led to a rapid and growing interest in S . hamata as a tropical pasture particularly throughout the dry tropics ( Humphreys , Lenne and Sonoda , 1 9 7 9 ; Whiteman , 1 9 8 0 ) . l egume 1 9 7 8a ; Over the past decade , there has been considerable research on the behaviour of S . hamata cv Verano under varying cultural and managerial conditions ( Gil lard et al , 1 9 8 0 ; Gardener , 1 9 8 1 ) . These studies have revealed that in spite of its promise it sti ll has s imi lar problems of variable establishment and persistence in a mixed , grazed sward ( Gil lard et a l , 1 9 8 0 ; Gardener , 1 9 8 1 ) a s does Townsvi lle stylo , reflecting the common dependence of dry matter production on seedling regeneration from seed reserves in the soil ( Gardener , 1 9 8 1 ) . In a long term study over 9 years at Landsdown , Austral ia , Gardener ( 1 9 8 1 ) found that the ma j ority of � hamata plants died in their seedl ing years and only 0 . 0 3 % survived to the end of the third year . In most year s , S . hamata had to re-establish almost entirely from seed . Therefore , low legume yields were expected when seed- 4 lings of S . hamata competed poorly with the annual gras ses . A high rate of shoot replacement is neces sary for maintenance of growth under repeated defoliation and numerous axi llary buds close to the crown of a plant the the wil l enhance its ability to regrow under these conditions . This has been noted in S . guianens is ( Aubl . ) Sw . ( Grof et a l , 1 9 7 0 ) , Desmodium intortum C . P . I . 2 3 1 8 9 ( Imrie , 1 9 7 1 ) and Psoral ea eriantha ( Gutteridge and Whiteman , 1 9 7 5 ) . Most of the research carried out on S . hamata has emphasized the e f fects of gra z ing or cutting on seed production ( Wilaipon and Humphreys , 1 9 7 6 , 1 9 8 1 ; Wilaipon et al , 1 9 7 9 and Waikakul , 1 9 8 3 ) . Little information is avai lable on the factors which affect the regrowth of this legume . It seems that the criteria to utili ze thi s l egume must be defined in order to achieve maximum persi stence and/or production . Therefore , an understanding of the morphological and physio­ logical development of thi s l egume is needed to develop better cutting or grazing practices . With thi s in mind , a series of experiments was designed to determine the e f fects of defoliation on S . hamata cv Verano . The ob jectives of this work were : - 1 . To study the growth pattern of S . hamata under controlled conditions . 2 . To define the morphological and physiological characteristics of S . hamata in terms of leaf , stem , inf lorescence and branch growth in response to di f ferent defoliation regimes . 3 . To study the effect of defoliation and water stress on growth characteristics of S . hamata . 4 . To study the effects of graz i ng management on S . hamata production and survival . 5 CHAPTER 2 LITERATURE REVIEW 2 . 1 STYLOSANTHES HAMATA CV VERANO 2 . 1 . 1 Origin , Taxonomy, Distribution and Ecology The genus was establi shed in 1 7 8 8 by 0. Swart z ( Mohlenbrock , 1 9 5 8 ) and most of the detailed systematic work on the genus was done in the 1 8th and 1 9th centuries . In recent years , Mohlenbrock ( 1 9 5 8 ) has described 30 species of Stylosanthes including Stylosanthes hamata ( L . ) Taub . whi le Burt et al ( 1 9 7 1 ) and Edye et al ( 1 9 7 3 , 1 9 7 4 ) have described a number of morphological and agronomic characters based on numerical classif ication . The most recent review of thi s genus i s that o f Mannet j e ( 1 9 8 4 ) . Stylosanthes species occur typically on savannahs and s imi lar areas in Central America , South America , Southern Africa , Southeast Asia , India and in the east of the United State s ( Mohlenbrock , 1 9 5 8 ) . The distribution of species including hamata was closely related to the climate in which the species originated . Burt and Reid ( 1 9 7 6 ) have divided Stylosanthes into six c limatic types within nine phyto­ geographic regions of the tropical parts of Central and South America . Stylosanthes belongs to the tribe Aeschynomeneae and comprises 2 5 species ( Mannet j e , 1 9 8 4 ) . Of these , f ive were considered by Whiteman ( 1 9 8 0 ) to be of agricultural importance , namely the annual : S . humilis and the perennial s : � guianensi s ( syn . �gracil is ) , S . hamata , S . scabra and S . subsericea . The popularity of the genus a s a tropical forage legume has increased markedly in the pas t 2 0 years , partly due to its remarkable hardiness and the adaptability of some of its species to various climatic and soil conditions ( Burt et a l , 1 9 8 3 ; Edye and Grof , 1 9 8 3 ) . Australia was the first country to r ecognise the value of Stylosanthes species for improving tropical grassland particularly� humilis ( Humphreys , 1 9 6 7 ) . Because of the limitations of the natural pastures i n 6 northern- Australia , most pasture improvement has been based on introduced pasture species . Among these , a large number of Stylosanthes species were introduced from tropical America . Burt et al ( 1 9 7 1 ) and Edye et al ( 1 9 7 4 ) described and di scussed a great number of these introduced Stylosanthes s pecies by us ing agronomical and morphological character­ i stics ( M-A group ) . As a result , Stylosanthes hamata cv Verano ( Verano stylo ) emerged a s an outstanding acces s ion and the most promis ing for a wide range of dry environments among the groups studied ( Burt et a l , 1 9 7 1 ; Edye et al 1 9 7 5b , 1 9 76 ) . Thus , it was recognised as a valuable forage legume i n 1 9 7 3 ( Mackay , 1 9 7 5 ) and commercial seed production began in 1 9 7 4 ( McKeague et al , 1 9 7 8 ) . This cultivar ( Verano ) has been tested throughout a number of countries in the tropics and subtropics including Thai land ( Humphreys , 1 9 7 8a ) , Malaysia ( Humphreys , 1 9 8 4 ) and USA ( Lenne and Sonoda , 1 9 7 9 ) . During the last decade , the ecology of Verano style in pasture systems has been reported by several workers ( e . g . Tors sell et al , 1 9 76 ; Gil lard et al , 1 9 8 0 ; Gardener , 1 9 8 1 ) . The results of these studies indicated that the main factors a f fecting the balance of species in the mixed pasture are : i ) the grazing pressure ii ) the soil fertility and ferti li zer application i i i ) the nitrogen supply from the legume iv ) the other species in the pasture v ) the time of ·opening rain . Gardener ( 1 9 8 1 ) studied population dynamics and stabi lity of Verano stylo in grazed pasture s . He found that the ma j ority o f plants died in the seedling year , and in most years Verano stylo had to re-establi sh a lmost entirely from seed . Therefore , a large carry over o f seed from season to season was essential to maintain the soil seed reserve for the l ong term pers istence of Verano style in the pasture system . In add ition to these f indings , he developed a conceptual model of the plant ecological proces ses i n the grazed pasture . Bas ic processes in the system , such as : the amount of seed i n the diet , the intake of feed , the percentage of intact s eed pas sing through the digestive tract , the amount and rate of hard-seed breakdown and the survival of seedlings , were 7 considered . He a l so showed that stocking rate had l ittle ef fect on seed input . A seed reserve , suf ficient to supply the 30 - 1 0 0 seedlings/rn2 , was needed each year for the maximum production of the legume . This gave Verano style the potential to coloni ze any site in the community left vacant by previous deaths . Whether this potential is reali zed depends on the timing and relative growth of the other species competing for the s ite . Gardener ( 1 9 8 1 ) concluded that germination , establi shment , growth , competition and seed production were important for seedl ing survival as previously reported in Stylosanthes hurnilis ( Torssel l , 1 9 7 3 , 1 9 7 6 ; Torssell and Mckeon , 1 9 7 6 ) . Drought avoidance mechanisms such as faster rate of root growth , rapid seed germination and radicle elongation , are important factors for survival during the germination/establ i shment phase . However , heavy grazing after seed germination favoured the survival of Verano stylo . Gillard et al , ( 1 9 8 0 ) also reported that the stabil ity of a pasture mixture with Verano stylo is greatly l imited initially by increasing pres sure from the companion grasses following increased soil fertil ity . Torssell et a l ( 1 9 7 6 ) studied the competition and dynamics of the popula- tions of three mixtures in l egume-based pastures . Mixtures studied were S . humilis and Digitaria cil iaris ; S . hamata and D . cil iaris ; S . hamata and Urochloa mosarnbicensis . They reported that S . hamata yielded more than the other species and its competition with D . c i liaris was stronger than that of S . humilis . The changes in species composition were studied us ing the conceptual model of Torssell ( 1 9 7 3 ) . They reported that the legume proportion was always higher in the mixture S . hamata and D . c i l iaris than in any of the other mixture s studied . 2 . 1 . 2 Morphological and Growth Habit Burt et al ( 1 9 7 1 ) classi fied Verano style as a herbaceous semi-erect or prostrate non-determinate plant with profuse branching and s lender sterns and smal l l eaves . The average number of primary branches up to 5 cm on the main axis i s 5 . 8 - 6 . 2 . Verano stylo i s a syrnpodial branching type and flowering occurs at an early stage of growth ( Cameron and Mannet j e , 1 9 7 7 ) . However , growth is continued 8 by branches developed from axil l ary buds ari sing below the f loral apex . These branches soon cease growth as a result of conversion of the vegetative apex to the reproductive state . The axil lary bud below the f loral apex again produces a new branch and develops in a s imi lar manner . As a result , canopy growth of Verano stylo i s of a non-determinate , branching habit with a number of primary branches which are l imited and vary according to the environment ( Carneron and Mannet j e , 1 9 7 7 ) . The plant may grow up to one metre in height in good conditions , but more typical ly 0 . 60 metre ( Clernents , 1 9 8 0 ) . Many growing buds have been reported close to the ground and a crown may develop under grazing or cutting conditions ( Hurnphreys , 1 9 7 8b ; Burt et a l 1 9 8 3 ) . The stems have short white hairs down one side ( Edye et al , 1 9 7 4 ) . Leaves are tri foliate , leaflets lanceolate , acute glabrous with 4 - 6 pairs of veins and length-breadth ratio of 4 . 5 - 5 . 3 , rachis 4 - 5 mm long and the bidentate stipules adnate to the base of the petiole with hairs on the sheath and teeth . The inflorescence i s an oblong spike with 8 - 1 4 f lower s on a long stern . The loment has two articulations whi ch are usua l ly both fertile . The upper one i s usua l ly hairy and has a moderately thick , coiled beak which is from more than hal f to equal length to that of the pod ( Edye et al , 1 9 7 4 ) . Verano stylo has a tap root system , with rapid root development at the seedling stage ( Gardener , 1 9 7 8 ) . Roots of Verano stylo can penetrate into the soil up to 75 - 80 cm . These roots are very fine at such depth , and at a depth of 5 0 - 60 cm roots begin to divide progressively into small roots of approximately the same si ze ( Gutteridge , 1 9 8 2 ) . 2 . 1 . 3 Fertili zer , Nodulation and Yield Verano stylo is adapted to a wide range of soi l types but prefers sand and sandy loam ( Hurnphreys , 1 9 7 8 b , 1 9 8 0a ; Hare , 1 9 8 5 ) . It exhibits super ior adaptation to acid and infertile soi l s compared with other tropical l egumes ( Bogdan , 1 9 7 7 ; Hurnphreys , 1 9 8 0 a ; Edye and Grof , 1 9 8 3 ) . Andrew and Norri s ( 1 96 1 ) attributed that feature of adaptation to the high abil ity of Stylosanthes species to extract calcium from 9 soil . Furthermore , Verano stylo exhibited greater tolerance to high aluminium concentration than other Stylosanthes species (� guianensi s cv Cook , S . humilis c v Paterson , � scabra cv Fitzroy and S . vi scosa ( Carvalho et a l , 1 9 8 1 ) . Wilaipon et a l ( 1 9 7 8 ) reported that Verano style and Siratro were more salt-tolerant than Pueraria phaseoloides and Calopogonium mucunoides . Carvalho et a l ( 1 9 8 0 ) also reported a remarkably e f f icient symbiotic as sociation with Rhizobium at their highest aluminium concentration . Similar to other species in the genus Stylosanthes , Verano style has a very low phosphorus requirement for growth compared with Siratro and Centrosema , and has a high abi lity to extract nutrients from soi l s compared with these l egumes ( Robinson , 1 9 8 3 ) . Jehne ( 1 9 8 4 ) suggested that the presence of endotrophic mycorrhiza found on the roots of Verano style could account f or its apparent low requirement for soil phosphorus . He also noted that mycorrhi za may aid the tolerance of the species to environmental and soil stress . Although some reports show the adaptation of this species to low phosphorus f ertilizer , other reports have also shown strong responses to phosphorus fertil i zer ( Topark-Ngarm et a l , 1 9 7 9 ; Hall , 1 9 7 9 ; Shelton et al , 1 9 7 9 ) . Verano stylo nodulates f reely and can be e f fectively nodulated with a broad spectrum of cowpea inoculant ( Bogdan , 1 9 7 7 ) . Nodulation begins within 8 days after germination ( Date et al , 1 9 8 0 ) . - Verano stylo grows wel l under an annual rainfall between 5 0 0 and 1 2 7 0 mm and with a pronounced dry season ( Burt et al 1 9 7 4 ; Edye et a l , 1 9 7 5b ; G i l lard et al , 1 9 8 0 ; Gilbert and Shaw , 1 9 8 0 ) . I t has been r eported to be drought tolerant ( Burt and Miller , 1 9 7 5 ) but has no tolerance to water-logging ( Humphreys , 1 9 8 0b ) . Verano stylo f lowers early and vegetative growth continues while conditions are favourable ( Cameron and Mannet je , 1 9 7 7 ; Edye and Grof , 1 9 8 3 ) . The potential yields , therefore , are closely as sociated with the length of growing season as demonstrated by Edye et al ( 1 9 7 7 ) . They showed that the growth of one accession of s . hamata and 2 5 seasonal , and 1 0 accessions of S . guianensi s were highly that the yield differences among accessions were greatest at the beginning and end of the growing season . Plants capabl e of producing feed under highly variable cl imates could diminish the problem of forage availabi lity whi ch is considered to be a ma jor l imiting factor in animal production . Topark-Ngarm and Akkasaang ( 1 9 7 8 ) reported yields of 4 . 3 to 5 . 8 ton/ha/year under 4 and 6 week cutting in Thailand . Edye et al ( 1 9 7 5b ) reported from the dry tropics of Australia that in a three-year trial , S . hamata CPI 3 8 8 4 2 ( cv Verano ) was superior in dry matter- yield to the other accessions at a number of di f ferent sites . Bishop et al ( 1 9 8 0 ) a l so demonstrated its outstanding performance in terms of dry matter yield , persistence and seedl ing regenera­ tion under the high rainfal l conditions of the Central Queensland wet coast . A mixture of Guinea grass ( Panicum maximum ) and Verano stylo was also shown to produce the total yield of 8 . 5 ton/ha in Thailand ( Wilaipon and Humphreys , 1 9 8 3 ) . In thi s trial , Verano stylo yield and density in the second year was superior to that of Siratro under grazing conditions . 2 . 2 Effect of Defoliation on Tropical Legume Defol iation is defined as the process of partial removal of the above-ground parts of plants by grazing animal s or cutting machines ( Hodgson , 1 9 7 9 ) . Defoliation by hand or machine cutting often involves more complete and sudden removal of plant tis sues than occurs under graz ing ( Humphreys , 1 9 7 8b , 1 9 8 1 ) . However , grazing by animals i s characteri sed by a spatial heterogeneity of removal in both the vertical and hori zontal dimension . Grazing also involves animal treading , selecting and the return of nutrients to pasture a s dung and urine ( Watkin and elements , 1 9 7 8 ) . Defoliation has often been considered in terms of frequency , intensity and t iming ( Harris , 1 9 7 8 ; Humphreys , 1 9 7 8b ) . The e ff ects of defoliation on pastures have been reviewed by many workers ( e . g . Jameson , 1 9 6 3 ; Humphreys , 1 966a ; Harr i s , 1 9 7 8 ; Watkin and Clements , 1 9 7 8 ) . These 1 1 authors pointed out that defoliation either by gra zing an imals or by cutting can influence morphology , regrowth abil ity , dry matter yield , root systems and herbage qual ity o f the plant . Thus , the e f fects of defol iation on tropical l egumes wi l l be discus sed in relation to the effects on these parameters . 2 . 2 . 1 Morphology and Growth Habit Graz i ng or cutting modifies the growth form of many tropical legume species , and through phenotypic adaptation h igh rates of herbage production or persistence may still occur under these conditions . Hare ( 1 9 8 6 ) observed that when Townsvil l e stylo was grazed heavily from the start of the wet season , i t developed a prostrate habit with many f ine branches growing close to the ground . In areas reserved for s eed production which were uncut throughout the wet season , i t grew tal l and erect with only a few branches close to the ground . Jones , R . M . ( 1 9 7 3 a ) found that after three years of continuous grazing of Siratro ( Macroptilium atropurpureum ) , the s i ze of crown of Siratro was greatly affected . At the l ightest stocking rate , most crowns were large ( > 1 . 0 cm diameter ) with appreciable rhi zome development , whereas at the heaviest stocking rate there was negl igible rhi zome development and most crowns were small ( < 0 . 5 cm i n d iameter ) . Jones ( 1 9 7 4 ) also showed that in Siratro repeated defoliation to remove a l l l eaves will result in reduced stolen development , reduced vigour and even death of plants . Thi s e f f ect could be even greater when the legume i s growing in competition with grass . He found a close correlation between dry matter yield and stolen development . The death of plants weakened by excessive defoliation has also been reported , particularly under late cutting with species such as Townsville stylo ( Fi sher , 1 9 7 3 ) , Crotalaria juncea ( Kess ler and Shelton , 1 9 8 0 ) and Psoralea eriantha ( Gutteridge and Whiteman , 1 9 7 5 ) . Treading by grazing anima l s may a l s o a f fect growth and survival o f plants ( Watkin and elements , 1 9 7 8 ) . 1 2 2 . 2 . 2 Regrowth Following Cutting According to Hurnphreys ( 1 9 8 1 ) , successful pasture plants must be able to tolerate defoliation by cutting or graz ing , by making adequate regrowth after defoliation and by s urviving long enough to reproduce . Initial ly , regrowth i s determined by the number o f growing points ( Mannet j e e t a l , 1 9 8 0 ) . The rate of r egrowth i s determined by the leaf area r emaining on the plant , the energy reserves and the water and nutrient uptake by roots ( Ward and Blaser , 1 9 6 1 ; Hurnphreys , 1 9 6 6 a , 1 9 6 6b ; Hurnphreys and Robinson , 1 9 6 6 ; Grof et al , 1 9 7 0 ; Jones , 1 9 74 ; Yarnada , 1 9 7 5 ; Harris , 1 9 7 8 ; Ludlow and Charles­ Edwards , 1 9 8 0 ; Mannet j e et al , 1 9 8 0 ; Hodgkinson and Wi lliarns , 1 9 8 3 ) . 2 . 2 . 2 . 1 Location and number of growing points The location and number of growing points are determined by the plant habit and growth form of the species , the stage of plant development and the previous pasture management . Tropical legumes are more variable in growth habit and location of growing points than grasses ( Hodgkinson and Wil liarns , 1 9 8 3 ) . The growing points of stoloniferous , prostrate l egumes such as Lotononis , Townsville stylo , Desrnodiurn tri folurn and D . heterocarpon are located close to the ground and are wel l-protected by small leaves ( Whiternan , 1 9 6 9 ; Mannet je et al , 1 9 8 0 ; Hurnphreys , 1 9 8 1 ) . This i s in contrast to cl imbing or trail ing or to an erect growing l egume such as Siratro ( Whiternan , 1 9 6 9 ; Jones , 1 9 7 4 ) and Desrnodiurn ( Jones , 1 97 3 ) with the growing points located a long - the trai l ing stern or higher above the ground such as in one accession of Psoralea eriantha ( Gutteridge and Whiternan , 1 9 7 5 ) . As such , close defoliation of these latter legumes removes the ma j or proportion of the young active buds , leading to a reduced rate of regrowth and abil ity to compete with companion grasses . For example , Grof et al ( 1 9 7 0 ) found poor survival and regrowth owing to poor basal branching and def iciency of basal bud sites in an erect genotype of Stylosanthes guianensis . A s imilar e ffect was found in an erect l ine of Psoral ea eriantha ( Gutteridge and Whiternan , 1 9 7 5 ) , Desrnodium and Siratro ( Whiternan , 1 9 6 9 ; Jones , 1 9 6 7 , 1 3 1 9 7 3 , 1 9 7 4 ) . Imrie ( 1 9 7 1 ) sub j ected five l ines of D . intortum to de foliation 1 0 or 3 0 cm from the crown , at 4-week intervals . The severe 1 0 cm treatment substantially reduced yield in f our l ines , but not in CPI 2 3 1 8 9 . The superior performance of this legume under severe cutting was due to defol iation stimulating growth from the axillary bud close to the crown o f the plant . The other l ine initiated growth only from one or two buds immediately below the point of stern cutting . These f indings illustrated the importance of growing points on regrowth . During the vegetative stage , it i s unl ikely in most common tropical legumes for the growing point to be damaged by graz ing or cutting . The position of the growing point above the cutting height late in the growing season , or with long cutting intervals , was reported to be the main factor contributing to low yield and mortal ity of Townsville stylo ( Robertson et al , 1 9 76 ) . In some legumes , l eaf senescence plus twining habit caused most of the growing points to be r emoved ( Whiternan , 1 969 ; Jones , 1 967 , 1 9 7 3 , 1 9 7 4 ; Gutteridge and Whiternan , 1 9 7 5 ) . 2 . 2 . 2 . 2 . Leaf area remaining Under conditions of adequate water and nutrient supply , pasture growth depends on leaf area and light interception ( Watson , 1 9 5 2 ; Donald and Black , 1 9 5 8 ; Brown and Blaser , 1 968 ) . Brougharn ( 1 9 60 ) found that productivity was a l so - c losely related to the amount of chlorophyll in leaf and non- leaf components . Although stern , petiole , leaf sheath and inf lorescence intercepted light and participated to varying degrees in photosynthesis , the ma jor contribution came from the leaf and hence l ead Brougham ( 1 960 ) to determine and use l ea f area as a mea sure of productivity ( leaf area being commonl y expressed as leaf area index ( LAI ) , i . e . the ratio of leaf area to ground area ( Watson , 1 9 4 7 ) . The s ignificance of l eaf area in pasture growth has been reviewed by several workers ( e . g . Donald and Black , 1 9 5 8 ; 1 4 Brown and Blaser , 1 968 ) . In a young stand or in defoliated pasture , the quantity of light available is greater than can be intercepted and uti li sed by the plants and hence much of it is " lost " to the ground . As the plant grows and new l eaves are produced , more and more of the incident l ight is i ntercepted by the leaf canopy until interception i s c omplete . Competition for l ight then develops among the p lants and among the leaves on the plant . As the density of the leaf canopy increases , older leaves receive insuf ficient l ight and may become parasitic on the plant - although some authors ( Vickery , 1 9 8 1 ) question this parasitic state . Finally , lower leaves begin to die and , in time , the rate of death of older leaves is equal to the rate of appearance of new leaves . A maximum or ceiling herbage growth rate has now been reached . Leaf area index at this point was c lassified as the optimum or critical leaf area index ( Brougham , 1 9 5 8 ) . The critical and optimum LAI value varies among species because of the dif ference in their growth habit ( Black , 1 9 5 7 ; B rougham , 1 9 5 8 ; Humphreys , 1 966b ) . As LAI increases above the optimum , leaf loss and shedding cause the growth rate to drop ( Donald and B lack , 1 9 5 8 ) . The relationship between LAI and pasture growth rate explains the general response of pasture to defoliation . Reduction in growth rate under low LAI is explained on the bas i s of inadequate light interception , while that occurring under a high LAI value is a result of accelerated senescence and increased respiratory load of leaves near or below their light compensation point ( Donald and Black , 1 9 5 8 ) . In terms of pasture practice , Harri s ( 1 9 7 8 ) suggested that : 1 ) defoliation frequency should be such that the regrowth interval is extended until pasture growth rate begins to decline from its maximum 2 ) intensity of defol iation should be to the level that leaves the amount of biomas s at which maximum growth rate i s f irst attained . However , there are many factors affecting the application of the LAI concept to defoliation practice . Thi s i s because the critical LAI value appears to vary with the environment ( Brougham , 1 9 5 8 ; Davidson and Donald , 1 9 5 8 ; B lack , 1 963 ) , 1 5 with the angle of sun elevation ( Brougharn , 1 9 5 8 ) with light intensity ( Stern and Donald , 1 9 6 2 ; Black , 1 9 6 3 ) and with the qual ity of the res idual leaves ( Hurnphreys , 1 9 7 8b ) . The significance of LAI on subsequent regrowth has been demonstrated in both tropical and temperate pasture species . With a tropical legume , Jones ( 1 9 7 4 ) found that the regrowth yield of Siratro was linearly related to res idual LAI on the stubble at the start of growth up to 7 0 days . This is in contrast to l ucerne which showed the beneficial ef fect of residual l eaf area for only a short period of regrowth ( Leach , 1 9 6 7 ) . Ludlow and Charles-Edwards ( 1 9 8 0 ) reported that residual leaf area index is a ma jor determinant of Setaria/Desmodiurn sward regrowth . Grof et al ( 1 9 7 0 ) reported that poor regrowth of S . guianensis was associated with low lea f number and area remaining at low cutting frequency . Because of low leaf areas in the lower part of the plant , cutting at the late stage or late in the growing season resulted in poor regrowth ( Fisher , 1 9 7 3 ; Kessler and Shelton , 1 9 8 0 ) . 2 . 2 . 2 . 3 Reserves The significance of energy reserves including carbo­ hydrates , protein and other labi le plant f ractions for regrowth , have been reviewed by several researchers ( e . g . May , 1 9 6 0 ; Humphreys , 1 9 6 6 a ; Mcl lroy , 1 9 6 7 ; Smith , 1 9 7 3 ; Yamada , 1 9 7 5 ; Harris , 1 9 7 8 ) . These authors concluded that such energy reserves are util i zed as a substrate for respira­ tion and for new shoot and root growth when photosynthesis is insufficient to meet the demand . The main reserve organs are the basal stern ( crown ) , stolen or rhi zome and tap root ( Yarnada , 1 9 7 5 ; Harris , 1 9 7 8 ) . Subsequent regrowth i s dependent o n leaf area and photosynthesis , and the rate of nutrient uptake by the roots . The main energy reserves involved are non-structural carbohydrates , namely glucose , fructosan , dextrin and starch ( Mcllroy , 1 9 6 7 ) . The accumulation of these reserves appears to be characteristic of the species and is inf luenced by climatic condi tions and cutting management . These aspects are 1 6 di scussed in Section 2 . 2 . 4 . Immediately fol lowing intense cutting or gra z ing , root growth may cease for several days ( Troughton , 1 9 5 7 ) . Substrates for the respiration proces s are drawn from carbohydrate reserves including nitrogenous compounds ( Davidson and Milthorpe , 1 966 ) , and to provide energy for initiation of new root and shoot growth . Respira­ tion requirements and structural compounds of this new growth are supported by export of reserves from root and stubble until photosynthesis i s suf f icient to meet the demand for respiration and growth requirements of the plant ( Davidson and Milthorpe , 1 966 ) . Recent work has shown that the reduction in carbohydrate of both root and stubble following defol iation was largely accounted for , by respiration , particularly of roots . The direct contribution of reserves to new growth tis sue was small and transitory ( Davidson and Milthorpe , 1 966 ) . For example , in Dactylis glomerata whi ch was severely defoliated , the carbohydrate reserves contributed only 5 0 mg compared with 700 mg from photo­ synthesi s during the eight days following defoliation . Regrowth rate of this grass depended on carbohydrate reserves for only the first 2 - 4 days . Ueno and Smith ( 1 9 7 0 ) a lso reported that in alfalfa only one third of the initial total non-structural carbohydrate in roots was utili zed during the period from cutting to the time of greatest reduction of total non-structural carbohydrate ( 1 - 3 weeks a fter cutting ) . These demonstrated that not all reserves were uti l i zed for regrowth . However , the length of time to become independent i s greater after more severe cutting . While numerous experiments have been carried out with temperate pasture species , l imited information is available for tropical pasture species , particularly legumes . Adegbola ( 1 966 ) reported a positive correlation between the actual carbohydrate content of the stolons and rhizomes of giant star grass and the regrowth potential . Dovrat and Cohen ( 1 9 7 0 ) working with Rhodes grass � Chloris gayana kunth . ) found that the amount of shoot weight produced was positively correlated with the percentage of total soluble carbohydrate in the root . 1 7 The accumulation of carbohydrates in tropical legumes is discussed in Section 2 . 2 . 4 . but the role of reserves in recovery of tropical legumes after defoliation i s not considered s ignif icant . This was illustrated by Jones ( 1 9 7 4 ) with Siratro . He found a pos itive relationship between regrowth and leaf area rather than with the carbohydrate l eve l . The poor persistence of thi s l egume was due to the l ow number of active growing points close to the crown . No f urther i n formation is available for other tropical l egumes . Thus , the role of carbohydrate in tropical legumes demands more attention . 2 . 2 . 2 . 4 Water and Nutr ient Uptake . The rate of regrowth of pasture species i s also markedly dependent on water movement to and within the plant and f rom the plant to the surrounding atmosphere . Restriction of water movement wil l inf luence nutrient uptake and distribu­ tion within the plant . Defoliation can inf luence nutrient and water uptake through the cessation or reduction of root growth . Reduced root growth prevents the extension of root into regions not a lready depleted of soil moi sture ( Jacques and Edmond , 1 9 5 2 ; Oswalt et a l , 1 9 5 9 ; Davidson , 1 9 7 8 ) . This i s important . particularly in dry soil as demonstrated by Jantti and Kramer ( 1 9 5 7 ) . Lack of water may prevent the resumption of growth by many buds which would quickly resume growth in plants supplied with adequate water . Defoliation can also influence nutrient and water uptake through the r eduction in water absorption by r educing transpiration ( Jantti and Kramer , 1 9 5 6 ) . Restriction of water movement leads to l imitation of the f low of photosynthate or reserves from top to root for respiratory substrate ( Hatr ick and Bowling , 1 9 7 3 ) . Thi s wil l also l imit active ion uptake ( Milthorpe and Davidson , 1 9 6 6 ) . I n summary it residual leaf area appears of the that carbohydrate reserve s , stubble , the availabil ity of growing points and the extent of water and mineral uptake by 1 8 roots , a l l inf luence the rate of regrowth of pasture a fter cutting or grazing . The relative importance of these parameters on regrowth wil l vary between species , with environments and with grazing or cutting management . Although the quantity of carbohydrate reserves required may be smal l , it can be highly important under severe cutting when leaf area is severely l imited and when the abil ity of the root to take up water and nutrients is restricted . Effects of defoliation on these parameters may be modif ied by plant morphology and geneti c make-up and defoliation management . 2 . 2 . 3 Dry Matter Yield The ef fects of defol iation on total dry matter production of · tropical l egumes can be interpreted as an e f f ect derived from the inf luence of defol iation on the components of that yield . Among these are the rate of branch or stolon and leaf development . Jones ( 1 9 7 4 ) reported that total stolon production ( number/plant ) of Siratro increased with increasing leaf number remaining after cutting ( 2 3 . 2 - 67 . 3 I plant ) . indicates that under severe and frequent cutting which to l ow numbers of leaves rema ining , a low dry matter results . This was also demonstrated by Whiteman This l eads yield ( 1 969 ) . Gutteridge and Whiteman ( 1 9 7 5 ) reported that in P soralea eriantha mean shoot number was markedly reduced with three week cutting intervals . Hence , low dry matter was obtained in thi s treatment . Loch and Humphreys ( 1 9 7 0 ) reported that cutting at floral initiation increased the rate of branching , whi ch l ead to an increase in dry matter yield of Townsvil l e stylo . The increase in branch number when cut at an early stage may explain the increased accumulated y ield o f Townsville stylo in Fisher ' s ( 1 9 7 3 ) experiment . The number of expanding l eaves and the rate of leaf appearance has been reported to increase at all stages of cutting , particularly early cutting ( Loch and Humphreys , 1 9 7 0 ) . Although there was no report directly on rate of leaf 1 9 appearance , a high rate of net as s imilation may be ref lected in the high rate of leaf appearance following cutting , as shown in Townsville stylo ( Fisher , 1 9 7 3 ) . As a result , leaf dry weight was not sensitive to cutting when compared with stem dry weight The latter fraction increased substantially following long cutting interval s ( Mufandaedza , 1 9 76 ) . Delayed flowering due to cutting has been frequently reported ( Loch and Humphreys , 1 9 7 0 ; Kes sler and Shelton , 1 9 8 0 ; Humphreys , 1 9 8 1 ) and intense and frequent cutting can l ead to low flowering fractions , in contrast to long cutting intervals ( Mufandaedza , 1 9 76 ) . 2 . 2 . 4 Underground Plant Organs There is little information on the effect of defoliation on underground plant organs of tropical legumes . Agronomic experiments seldom include root data and graz ing exper iments s e ldom include a measure of roots ( Davidson , 1 9 7 8 ) . In terms of the information available , various studies have shown that root growth is retarded by partial defolia­ tion . The lower the height of cutting , or the more f requent the defoliation , the greater wil l be the retardation of root growth such as in Crotalaria juncia ( Kessler and Shelton , 1 9 8 0 ) and Centrosema pubescens ( Bowen , 1 9 5 9 ) . Root system reduction in Centrosema pubescens is usually evident both in weight and length and defoliation also reduces the number of roots initiated ( Bowen , 1 9 5 9 ) . Thus , the defoliated plant wil l produce a shallow and sparse root system . Kessler and Shelton ( 1 9 8 0 ) have shown that a s ingle cutting of Crotalaria juncea above node 4 or 1 0 at 6 , 8 or 1 0 weeks after planting , significantly reduced root growth . In the severely defoliated treatment , root weight was reduced in the immediate post-defoliation period , fol lowed by recovery except after late defoliation at node 4 , when plants died . Late defoliation was more damaging than early defoliation , and the difference s in root weight between node 4 and node 1 0 treatments increased with time . Crider ( 1 9 5 5 ) has also shown that removal in a s ingle cutting of 4 0 % or more of the top 2 0 growth o f several species � _ stopped root elongation . The larger the percentage removed , the longer the period of root growth stoppage . Mufandaedza ( 1 9 76 ) reported that the dry matter yield of the roots of Stylosanthes guianensis decreased under more frequent cutting at 4 cm cutting height , but not at 1 0 cm cutting height . Defoliation also reduces the growth of rhi zomes and the number of rhi zomes initiated · in Siratro under graz ing conditions ( Jone s , R . M . , 1 9 7 3a ) . Los ses of root and root nodules have been reported in many tropical legumes . Bowen ( 1 9 5 9 ) reported that cutting to s imulate heavy gra zing resulted in a los s of two thirds of the Centrosema pubescens roots by weight , and inactivation and. s loughing of a ma jor proportion of the nodules . Whiteman ( 1 9 7 0 ) and Whiteman and Lulham ( 1 9 7 0 ) have shown that the population of nodules of Siratro , Greenleaf desmodium and S i lverleaf desmodium , decreased fol lowing defoliation . In Desmodium , the reduction in nodule population was greater with more severe defoliation and greater when young active leaves were removed , than when older , less active leaves were removed . This indicates that a heavy defoliation or graz ing can lead to poor nitrogen fixation . Loss of roots and nodules after severe defoliation has also been reported in temperate legumes ( Butler et al , 1 9 5 9 ) . These e f fects of defoliation on the root system may be modi f ied by the growth habit of the species . Shaw ( 1 96 1 ) , Robertson et al ( 1 9 76 ) and Fi sher ( 1 9 7 3 ) reported that Townsville stylo developed a prostrate habit when it had been gra zed . This also occurred with Verano style ( Wilaipon and Humphreys , 1 9 76 ) and Lotononi s bainesii ( Whiteman 1 969 ) . These legumes develop a substantial amount of the res idual l eaf close to the crown . Thus , the ef fect of defoliation on root growth should be les s than in the climbing legumes such as Siratro , Desmodium and Glycine ( Jones , 1 967 , 1 9 7 3 , 1 9 7 4 and Whiteman , 1 969 ) . A s imilar reaction has also been reported in some grass species . For example , Weinmann and Goldsmith ( 1 9 4 8 ) demonstrated that the root growth of bermuda 2 1 grass was not restricted by repeated cutting due to its dense prostrate growth habit preventing severe defol iation . It s e ems likely , therefore , that a prostrate legume wil l have better regrowth than a twining type under heavy gra z ing . The regrowth ( Harris , s igni ficance of carbohydrate has been well recognised 1 9 7 8 ) and in some tropical reserves in legume in temperate species grasses ( Humphreys , 1 9 6 6b ; Humphreys and Robinson , 1 9 6 6 ) . By comparison , few studies have centred on tropical legumes . Hunter et al ( 1 9 7 0 ) reported that at the vegetative stage , Glycine wightii ( leaf ) , Greenleaf desmodi um ( leaf ) and S i lverleaf desmodium ( stem ) contained negl igible quantities of starch , while the more mature plants ( at f lowering ) c ontained 1 - 2 % starch in dry matter . Noble and Lowe ( 1 9 7 4 ) r eported a high a lcohol-soluble carbohydrate content in autumn for Desmodium intortum and Glycine wightii . There i s little information available on the e f fect of defol iation on the carbohydrate content in the root of tropical legumes . Jones ( 1 9 7 4 ) found that increasing intervals between cutting did not increase root s i ze and concentration of soluble carbohydrates . This was unl i ke the response obtained with lucerne , where increasing the interval between cutting increased the s i ze of the root system and greatly increased the available carbohydrates ( Graber et al , 1 9 2 7 ) . Mufandaedza ( 1 9 7 6 ) also reported that cutting i ntensity and frequency had no s ignif i cant effect on total non-structural carbohydrates in the roots and stubble of S . guianensis . It seems that the ef fect of defoliation on the l evel of the avai l able carbohydrates in tropical l egumes warrants more atte ntion . 2 . 2 . 5 Herbage Quality Nutritive value of herbage may be expressed i n terms of chemical composition and digestibility . Protein content i s a common and measurable parameter of the nutritive value o f the plant . Decrease in the percentage o f crude protein in forage plants with advancing maturity is wel l documented ( Milford , 1 9 6 � ; 1 9 7 0 ) . Frequent percentage of Mi l ford and Minson , cutting general ly protein in forages . 1 9 6 6 ; Fisher , results in a Mufandaedza 2 2 1 9 6 9 , higher ( 1 9 7 6 ) reported that nitrogen concentration of several strains of S . guianensis increased with more f requent cutting . S imilar results were reported for S . humil i s ( Robertson et al , 1 9 7 6 ; Bendy , 1 9 7 1 ; Ive , 1 9 7 4 ) , Pigeon pea ( Akinola and Whiteman , 1 9 7 5 ) Crotalaria juncea ( Kessler and Shelton , 1 9 8 0 ) and Desmodium ( Jones , 1 9 6 7 ) . Under grazing and cutting conditions , Whiteman ( 1 9 6 9 ) reported s imilar levels between the two types of defoliation and both were higher than in the control plants . For phosphorus concentration , Wilaipon et al ( 1 9 8 1 ) r eported that cutting Verano stylo every three weeks produced h igher phosphorus in the apex compared with uncut plant s . However , Robertson et al ( 1 9 7 6 ) and F i sher ( 1 9 7 3 ) reported that in Townsvil l e stylo phosphorus concentration in the apex was unaffected by cutting interval . Dry matter digestibility is generally higher under f requent cutting , with long cutting intervals markedly reducing dry matter digestibility because of the high stem fraction ( Mufandaedza , 1 9 7 6 ) . 2 . 3 EFFECT OF DEFOLIATION ON STYLOSANTHES HAMATA CV VERANO 2 . 3 . 1 Morphology and growth habit Verano stylo acts as an annual if it is left ungra zed or uncut for too l ong . The plant wil l grow tal l and erect and may drop its leaves and die ( McKeague et al , 1 9 7 8 ) . At Landsdown , Australia , Gardener ( 1 9 8 1 ) observed that f lowering and vegetative growth occurred together in the wet season . He then recorded heavy flowering and seed set in April/May and subsequent death of all above-ground parts and many plants despite the presence of adequate soil water . Little new growth arose from the crown of the surviving plants in response to winter rains . Plants occas ionally survived through the dry season , only to die immediately after the 2 3 opening " rains . No explanations have been offered by these workers . However , plant death may be due to the absence of new growing points or buds when plants are allowed to reach such an advanced stage of maturity . This is supported by Wilaipon and Humphreys ( 1 9 8 1 ) who found that re j uvenation of the plant by grazing late in the wet season increased the number of perennating plants . Gillard et al ( 1 9 8 0 ) a lso reported a rapid regrowth of a s tand of Verano stylo when heavily graz ed at the onset of rains ( early stage ) ; it made more growth than a stand re-established from reserves of seed in the soil . They suggested that the rapid regrowth abi lity of this s pecies seems to be one of the main advantages compared with the annual Stylosanthes humilis . Wilaipon and Humphreys ( 1 9 7 6 ) also observed that the number of branches of Verano stylo increased fol lowing a s ingle heavy graz ing at the early stage of growth , resulting in increased s eed production . Many workers have pointed out that early gra zing encourages the plants to become prostrate or creeping as wel l as checking the companion grasses and improving l ight penetration . 2 . 3 . 2 Dry matter yield Dry matter yield of tropical l egumes cut at the early and late s tage appears to be correlated positively with the number of buds or growing points left , and persistence i s a l so related to the severity of defoliation . However , f ew experiments have recorded the number of growing points at dif ferent growth stages . Wilaipon and Humphreys ( 1 9 7 6 ) found that a s ingle grazing or mowing at an early stage ( first f lower ) produced superior dry matter yield to i ntermittent graz ing at later stages . Dry matter yield was also reduced by graz ing at the late stage due to the increase in voluntary grasses . Thi s indicates that dif ferences i n the timing o f defoliation appeared to alter the competitive balance between the legume and the gras s , with grass dominance being accentuated by l ate defoliation (Wi laipon and Humphreys , 1 9 7 6 ) . 2 4 - - As with many other tropical l egumes , a s previously reviewed , frequent cutting or gra z ing generally leads to an overall reduction in dry matter yield . Topark-Ngarm and Akkasaeng ( 1 9 7 8 ) reported that under a four-week cutting frequency , Verano style yields were s ignificantly lower than a s ix week cutting frequency . The yield reduction was more pronounced in the establishment year than in the second year . 2 . 3 . 3 Underground plant organs There are no data available for the ef fect of defol iation on the root size , root nodule and the energy reserves in the roots of Verano style . However , in general , s imi lar effects may apply to Verano style as those which have been discussed in the previous section . 2 . 3 . 4 Herbage quality Both nitrogen and phosphorus concentration have been reported to increase under frequent cutting ( Wilaipon et al , 1 9 8 1 ) . The above review clearly shows that the response of Verano style to defol iation in terms of growth , yield and chemical composition warrant more attention . 2 . 4 EFFECT OF WATER STRESS ON TROPICAL LEGUMES The previous section has considered the ef fect of defoliation on plant growth and yield of tropical l egumes . Thi s section considers the ef fect of water stress on plant growth and development of tropical forage legume species , where data are available . Where data are not ava ilable , other tropical species are used to i llustrate the point . 2 5 2 . 4 . 1 Germination , Establ ishment and Survival According to Pearen and Koeghan ( 1 9 8 2 ) , the development of pasture seedl ings may be divided into three distinct phases : 1 ) Germination 2 ) Establ ishment 3 ) Growth . The success or fai lure of seedling establishment wi l l in large measure depend on the extent to which favourable conditions Germination provision Winkworth for promotion of these phases can be met . requirements include a permeable seed coat , and of adequate air and moisture ( Humphreys , 1 9 7 8b ) . ( 1 9 6 9 ) showed that seeds of Townsvi lle s tylo ( Stylosanthes humilis ) required a moi st soil above -0 . 1 MPa ( - 1 bar ) for field germination to occur . Simi lar results have also been reported for Stylosanthes hamata , S . s cabra and S . vi scosa ( Mott et al 1 9 7 6 ) . It has been shown that at an osmotic stress of - 9 bars ( - 0 . 9 MPa ) , the germination of S . guianensi s , � scabra and S . vi scosa decl ined rapidly but there was much less effect on S . hamata , S . humilis and S . subsericea ( Mcivor , 1 9 7 6 ) , indicating a difference in the rate of water uptake by seeds of di f ferent species . Under hot arid cl imates where the soil dries out rapidly after rainfall , the presence of vegetative cover or a l ayer of litter that reduces the evaporative load on the seedl ing is advantageous (Murtagh , 1 9 6 3 ; Mi l ler and Perry , 1 9 6 8 ; Keya et a l , 1 9 7 2 ) . - Once the germination process i s underway , the rate of seedl ing root development and seedl ing vigour are often very important in conditions of rapid soil drying . These are related to seed size ( Tudsri and Whiteman , 1 9 7 7 b ; Gardener , 1 9 7 8 ) and dif fer between species ( Gardener , 1 9 7 8 ) . For example , S . hamata and S . humil i s exhibit rapid rates of germination and root elongation compared with S . scabra , S . vi s cosa and S . frutica ( Mc lvor , 1 9 7 6 ; Mott et al , 1 9 7 6 ; Gardener , 1 9 7 8 ) . This a l lows the former to survive under rapid soil drying conditions . 2 6 I n most pasture regimes of the world , a period o f soil moi sture stress , either short term or long term ( drought ) is common during the growing s eason ( Whiteman , 1 9 8 0 ) . Thus , during the establ i shment and growth phases , plants may experience water stress . In order to adapt and survive , plants have to develop both morphological and physiological mechanisms to enable them to withstand and adapt to such water stress . Thi s aspect has been recently reviewed by Mannet je et al ( ·1 9 8 0 ) ; Ludlow ( 1 9 8 0a ) and Fi sher and Ludlow ( 1 9 8 4 ) . These authors suggested that the adaptation of these plants could take three forms : escape from drought , avoidance of low lea f water potential , and tolerance of low leaf water potential . The first morphological adaptation is a reduction in transpiration by the presence of a cuticle capable of protecting the plant against excess moisture loss (Whiteman , 1 9 8 0 ) . Another abi l ity to drought-avoidance and survival mechani sm 1 s the shed l eaves at relatively high leaf water potentials . For example , Siratro leaves are shed at leaf water potential s of - 1 MPa but some plants retain sma l l , thick , turgid l eaves near the apex . I f stress continues , whole branches die until only the crown remains , most of which is under the soil surface ( Peake et al , 1 9 7 5 ; F i sher , 1 9 8 3 ) . This phenomenon has also been observed in Townsville style ( Stylosanthes humil i s ) , which appears to be an avoider of drought by dropping its leaves early and surviving via seed ( Fisher , 1 9 6 9 ; Gil lard and Fi sher , 1 9 7 8 ) � Many tropical legume species also avoid water stres s by a reduction in the exposed leaf area during drought . For exampl e L . l .eucocephala folds its l eaves · during drought ( National Academy of Science , 1 9 7 7 ) . In Townsville style , leaves are orientated parallel towards the sun ( para­ hel ionasty ) ( Begg and Tors sel l , 1 9 7 4 ; Fi sher and Campbel l , 1 9 7 7 ) . This phenomenon has also been observed in S i ratro , whi ch also forms smaller l eaves when water-stressed than when water i s well-supplied ( Fi sher , 1 9 8 3 ; Jantakool , 1 9 8 3 ) . 2 7 Reduction in the size of new leaves caused a reduction in leaf temperature and water loss ( Fisher and Ludlow , Fisher , 1 9 8 3 ) . 1 9 8 2 ; Several plants have also adapted to drought survival by developing deep and extensive root systems to fully exploit the available soil moi sture , such as Siratro ( Macroptilium atropurpureum ) ( Peake et al , 1 9 7 5 ; Mannetje et al , 1 9 8 0 ; Sheri f f and Ludlow , 1 9 8 4 ) . A rapid development of a deep tap root was also reported in Townsvi lle stylo ( Torssell et al , 1 9 6 8 ) which also showed rapid root development as a seedling growing under limited soil moi sture conditions , viz 5 0 - 6 0 cm of root growth being recorded a fter only 3 8 mm of rain . An annual pasture species avoids the drought period by growing and setting seed over the wet season , and by passing through the dry season as seed , e . g . Townsvil l e stylo . Seeds are also protected by embryo dormancy and impermeable seed coat ( Mott et al , 1 9 8 1 ) . These seeds are able to survive water potentials in equ i librium with air of 2 0 to 5 0 % relative humidity , which i s equivalent to water potentials of approximately - 95 to - 2 2 0 MPa ( Gaff , 1 9 8 0 ) . Thus , the seeds will not germinate a fter light out- of- season rain ( Mott et a l , 1 9 8 1 ) . C losure of the stomata during drought i s an important physiological mechanism to reduce water loss and maintain a high level of water potential . This has been well demon­ strated in Siratro (Macroptil ium atropurpureum ) . In this l egume , stomata close at h igh leaf water potentials ( - 1 . 2 to 1 . 5 MPa ) and respond to f a l ling atmospheric humidity indepen­ dent of water potential ( Ludlow and Ibaraki , 1 9 7 9 ) . This contrasts with Desmodium uncinatum cv Silverleaf , which lacks stomatal control mechanisms that respond to humidity ( Sheriff and Kaye , 1 9 7 7 ) and to leaf water potential ( Ludlow and Ibaraki , 1 9 7 9 ) . Many dif ferences i n the leaf water potential for stomatal closure can be accounted for by differences in the osmotic potential or osmotic ad justment of the plant ( Turner , 2 8 1 9 7 4 ; Begg and Turner , 1 9 7 6 ; Ludlow , 1 9 8 0a ) . Osmotic ad j ustment i s the main mechani sm of stomatal ad justment which a llows stomata to remain partially open at progressively lower water potentials as water stress increases . Siratro has small osmotic and stomatal ad justments and therefore the survival of this legume depends mainly on avoidance mechanisms ( Ludlow , 1 9 8 0 a , 1 9 8 0b ; Fi sher and Ludlow , 1 9 8 4 ) . Stylosanthes species , however , exhibit both osmotic and stomatal ad j ustment ( Ludlow 1 9 8 0a ; Fisher and Ludlow , 1 9 8 4 ) but the variation in osmotic ad justment between them i s small ( 0 . 8 - 1 . 1 MPa ) . Fisher and Ludlow ( 1 9 8 4 ) suggested that osmotic ad justment inf luencing growth may not be an and survival important characteristic in di fferent moisture environments in these species . However , the dif ferences in sensitivity to desiccation among Stylosanthes species are considerable , and appear to be correlated wel l with the moisture availabi lity in the areas to which they are adapted ( Fisher and Ludlow , 1 9 8 4 ) . This suggests that the high tolerance to drought of these species may be due to the insensitivity to desiccation - desiccation sensitivity being the leaf water potential at which the last vi sible leaf dies ( Fi sher and Ludlow , 1 9 8 4 ) . 2 . 4 . 2 Morphological Effects Cell enlargement depends on turgor ( Hsiao , 1 9 7 3 ) . As a consequence , restriction in water supply is likely to affect leaf expansion . Ludlow and Ng ( 1 9 7 6 ) showed an 8 0 % reduction in leaf elongation rate of Panicum maximum var . trichoglume , when the leaf water potential fell from - 4 to - 7 bars ( - 0 . 4 to - 0 . 7 MPa ) , and elongation ceased at - 1 MPa . With tropical legumes , Wil l i ams and Gardener ( 1 9 8 4 ) demonstrated that leaves of the annual tropical legume , s. humilis , wi l l lose all turgor and d ie at the leaf water potential o f - 1 . 3 MPa . Death of S . hamata does not take place unti l - 3 . 0 MPa , whilst S . s cabra leaves are able to reduce their leaf water potential to less than - 4 . 0 MPa and survive through the drought with l iving tis sue . Low leaf potentials also influence leaf production through their e f fects on l eaf initiation in meristems and 2 9 subsequent rate of cell division ( Hsiao , 1 9 7 3 ; Slatyer , 1 9 7 3 ) . The rate of leaf initiation may become slower or even cease as stress continues . The desiccation sensitivity of Stylosanthes species varies between - 6 and - 1 2 MPa ( Fi sher and Ludlow , 1 9 8 4 ) , which is in contrast to Siratro ( - 2 . 3 MPa ) ( Wi l son et al , 1 9 8 0 ) . Although low leaf water potential has a large effect on the rate of production of new leaves , it a lso causes a reduction in the existing leaves . For example , Fi sher ( 1 9 8 3 ) found that Siratro lost old leaves from the base to the top as the water stress continued , as did Townsvi lle stylo ( Fi sher , 1 9 6 9 ; Fi sher and Campbell , 1 9 7 7 ) . One of the most important consequences of the sensi­ tivity o f cel l enlargement to a small water defi�it i s a marked reduction in leaf area . Leaf area may be reduced through an increase in rate of leaf senescence , which has been discussed earlier . A reduction in leaf area will reduce crop growth rate particularly during the early stages of growth when there is incomplete light interception . 2 . 4 . 3 Physiological Effects As available soil moisture declines , an internal water deficit develops in the plant , leading to loss of turgor , until the plant reaches a state of wi lting when suf f icient moisture can no longer be extracted from the soil to maintain turgor . As plant turgor decline s , progressive stomatal closure causes reduction in the rate of co2 uptake ( Begg and Turner , 1 9 7 6 ) . As a consequence , restriction in water supply i s l ikely to a f f ect the rate of the photosynthetic process . This aspect was reviewed by Hsiao ( 1 9 7 3 ) , Begg and Turner ( 1 9 7 6 ) , and Turner and Begg ( 1 9 7 8 ) . Mostly these reviews concentrate on crops and temperate pasture species . Therefore , emphas i s in this review wil l be placed on stomatal behaviour and photosynthesis . Since stomata have an important role in regulating the pathway for gaseous exchange between the plant and atmosphere ( Turner and Begg , 1 9 7 8 ) , the behaviour of stomata in response 3 0 to environmental influences i s o f considerable physiological importance . In terms of tropical pasture species , Ludlow ( 1 9 8 0a , 1 9 8 0 b ) and Wilson et a l , ( 1 9 8 0 ) have been notably active in thi s field of research . Stomatal opening depends on turgor ( Hsiao , 1 9 7 3 ) , and the main mechanism of turgor maintenance is osmotic ad justment ( Ludlow , 1 9 8 0 a , 1 9 8 0b ) . As such , many of the differences in l eaf water potential for stomatal closure can be accounted for by differences in the o smotic ad justment ( Turner and Begg , 1 9 7 8 ; Ludlow , 1 9 8 0a , 1 9 8 0b ; Fi sher and Ludlow , 1 9 8 4 ) . These have been discussed in Section 2 . 4 . 1 . Reduction in net photosynthes i s , resulting from closure of stomata , was demonstrated by Ludlow and Ng ( 1 9 7 6 ) with the tropical gra s s , Panicum maximum var . trichoglume . They showed that stomatal closure causing net photosynthesis to cease occurs at leaf water potential s of - 1 2 bars ( - 1 . 2 MPa ) in control led environments . However , in the field where water deficits u sual ly build up more slowly , and roots exploit a greater depth for moisture , photosynthesis wi l l cease at 1 . 9 MPa ( Ludlow and Ng , 1 9 7 6 ) . For tropical legumes , no such data could be found related to photosynthesis . 2 . 4 . 4 Dry Matter Yield Shortage of soil moisture , resulting from variation in rainfall , is closely correlated with dif ferences in herbage yield from year to year according to McCown et al ( 1 9 7 4 ) . They demonstrated that cumulative yields of native spear grass plus � humilis , or buffel grass ( Cenchrus cil iaris ) plus S . humilis varied widely over 5 years , but yield was closely related to evapo-transpiration . They used a s imple water analysis to predict the effect of soil water storage on pasture yield at Townsville , Australia . Wil l iams and Gardener ( 1 9 8 4 ) us ing rainfal l , evaporation and temperature in a weekly balanced analysis , calculated a multiplicative growth index for each of the interval s over which stem e longation was measured on Verano stylo . They found that extension growth was substantially reduced when the growth index fell below 0 . 5 . They concluded that for 3 1 conditions in which water and temperature are the only factors limiting dry matter yield of Verano stylo , it i s possible to estimate dry matter yield at a given site from rainfall , evaporation and temperature . The ef fects of water stres s on total yield can be inter­ preted as an effect derived from the inf luence of water stress on the components of that yield . rate of branch and leaf development . Among these are the Evidence suggests that both rate and number of branches decrease with increasing stress . Waikakul ( 1 9 8 3 ) reported that the number of branches and rate of branching in Verano stylo when under stress were directly related to the water supply . At the end of thi s experiment , the plant that had been stressed at the vegetative phase produced only 3 5 % of the control . Similar results have been reported in Siratro ( Peak et al , 1 9 7 5 and Juntakool , 1 9 8 3 ) . The ef fects of water stress were not only to decrease the branch number but a l so to reduce stern elongation of these branches . For example , Wi l l iarns and Gardener ( 1 9 8 4 ) reported that when the relative leaf water content at 1 5 . 0 0 hr fell below approximately 7 5 % , the stern elongation rate of Verano stylo was small - usually less than 2 mm/day . As a result , total dry matter yield was reduced through a reduction in stern elongation . Water stress can inf luence total yield through the reduction in leaf area by restricting the rate of leaf expansion and by accelerated senescence of older leaves ( Fisher , 1 9 6 9 , 1 9 8 3 ; Fi sher and Carnpbell , 1 9 7 7 ; Carvalho , 1 9 7 8 ; Waikakul , 1 9 8 3 ; Juntakool , 1 9 8 3 ) . Reduction in leaf area leads to reduced photosynthesis and yield ( Hs iao , 1 9 7 3 ) . Leaf number and leaf s i ze were also reduced by water def icit . Fisher ( 1 9 8 3 ) reported that in Siratro the new leaves that were produced during stress were small in s i z e and fewer i n number ( Juntakool , 1 9 8 3 ) . A s a result o f reduction i n leaf number and leaf s i ze , leaf dry weight was a l so reduced ( Fi sher , 1 9 6 9 ; Carvalho , 1 9 78 ) . 32 The stage in plant growth at which the stress occurs is another important e f fect of water stre s s on plant perform­ ance . Fisher and Campbell ( 1 9 7 7 ) reported that the yield of Townsville stylo was low if stress occurred at the vegetative stage , whereas water stress during and after flowering had l ittle ef fect on herbage yield . This response seemed to be related to the fact that growth rate decreases with matura­ tion . They noted that water stress during vegetative growth hastened f lowering by two weeks . Pod dry matter yields were also reduced when water stress was applied late in the f lowering period . Similar results were also reported in Verano stylo and Siratro ( Waikakul , 1 9 8 3 and Juntakool , 1 9 8 3 ) . Carvalho ( 1 9 7 8 ) also reported a delay in flowering of about 8 days in the water stres sed S . hamata plants when compared with the control . Yields were also reduced in both � guianensi s and S . hamata . 2 . 4 . 5 . Herbage Quality It must be emphasized that water stress inf luences not only dry matter yield but also forage quality . Pasture quality generally refers to the nutritive value and inf luences forage consumption . The former i s characterized by chemical composition , digestibility and the nature of the digested product whereas the latter i s the rate of forage intake by animals . Chemical composition is as sociated with only the plant and its environments while the latter two are as sociated with both plant and animal . The e f fect of water stres s , therefore , wil l be di scussed in terms of chemical composition and digestibil ity . Wi l son and Ng ( 1 9 7 5 ) demonstrated that l ignin content in green dif fer age . l eaves of Panicum maximum var . trichoglume did not between stressed and non-stressed plants at any plant With tropical legumes , Carvalho ( 1 9 7 8 ) showed that Stylosanthes hamata cv Verano and S . guianensis exposed to a cycle of water stress always had low cel l wal l percentages in both stem and leaves . This contrasted with Macroptilium atropurpureum cv Siratro in which the small , stress adapted leaves were higher in l ignin and cel lulose content . But hemicellulose was markedly lower than that in non-stressed 3 3 plants (Wil son , 1 9 8 3 ) . The accumulation of some soluble carbohydrates such as inositol in stressed leaves of Vigna species and Siratro , was reported by Ford ( 1 9 8 2 ) and Ford and Wilson ( 1 9 8 1 ) . However , glucose and sucrose did not accumulate in stressed leaves of Siratro whereas sucrose increased substantially in tropical gras ses ( green panic , buffel and spear grass ) ( Ford and Wil son , 1 9 8 1 ) . Water stress decreases the phosphorus concentration in many tropical pasture and grain legumes . Fi sher ( 1 9 8 0 ) showed that phosphorus concentration in the plant top ( above ground ) of Stylosanthes hurnilis was s ignif icantly reduced by water stress ( 0 . 0 8 vs 0 . 2 3 % phosphorus ) . This ef fect was most marked when the stress occurred at an early stage of growth . Similar results have been reported for Sec a and s . vi scosa CPI 3 4 9 0 4 ( Probert , 1 9 84 ) and the tropical grain l egumes ( Glycine rnax CV Buchanan and Durack , Lab lab atropurpureus cv Highworth , Cajanus cajan cv Royes , Vigna unguiculata cv Red Caloona , � rnungo cv Regur and V . radiata cv Berken ( Wil son and Muchow , 1 9 8 3 ) . Water stress had only sma l l e f fects on nitrogen concentration . In fact , an increasing nitrogen concentration in the plant top was frequently reported under water stress ( Carvalho , 1 9 7 8 ; F i sher , 1 9 8 0 ; Wi l son and Muchow , 1 9 8 3 ) In grass species , the increase in nitrogen concentration arose largely from the delayed development of the stern fraction under stress ( Gate , 1 9 6 8 ; Wi l son and Ng , 1 9 7 5 ) . Water stress frequently increases the dry matter digest­ ibi lity of tropical pasture species ( Wil son and Ng , 1 9 7 5 ; Carvalho , 1 9 7 8 ; Wi l son , 1 9 8 1 ; Wil son and Munchow , 1 9 8 3 ) . In terms of the limited information on tropical legumes , Carval ­ ho ( 1 9 7 8 ) showed that water stress cycles increased i n vitro organic matter d igestibility in stern and in leaves of S . hamata and S . guianensi s . Wi l son and Muchow ( 1 9 8 3 ) a lso reported an · increase in dry matter digestibil ity of leaf , stern and whole tops of several tropical grain legumes . For the whole tops , such increases ranged from 2 to 3 . 7 % units . 3 4 2 . 5 MANAGEMENT OF TROPICAL FORAGE LEGUMES For some years , tropical pasture legumes have been used to improve production from tropical pasture as well as soil fertility ( Skerman , 1 9 7 7 ) . The success of these legumes has depended on the cl imatic conditions and agronomic techniques of pasture establishment such as adequate land preparation , correcting any nutrient deficiency , the selection of the r ight species and proper seed treatment , the appropriate depth and method of sowing and the correct ear ly management of the establ ishing sward . A particular problem in many subtropical and tropical pasture regions is that of moisture stress which can be very severe for part of the year . Under such conditions , legume growth can be affected , particularly when grown with- grasses . For example , Jones et al ( 1 9 6 7 ) showed that the growth of S iratro ( Macroptilium atropurpureum ) in a mixture with Paspalum plicatulum was linearly related to the ef fective rainfall over the growing season ( r = 0 . 9 8 7 ) . Variation between legumes in this regard has also shown that Townsville s tylo , for example , was unable to persist under 800 mm of annual rainfall , whereas Verano stylo , Seca and Fitzroy , performed quite well under these conditions ( Burt et a l , 1 9 7 4 ; Edye et al , 1 9 7 5b , 1 9 7 6 ; Gil lard et al , 1 9 8 0 ) . As a result of many years of evaluation and in some cases considerable breeding , selection and testing , a range of forage legume species and cultivars has been released to the f armer as being highly suitable under particular conditions of improved tropical grassland production . The agronomic practices favourable to these species have been presented by O ' Reil ly , ( 1 9 7 5 ) ; Bogdan , ( 1 9 7 7 ) ; Skerman , ( 1 9 7 7 ) ; Humphreys , ( 1 9 8 0a ) ; Whiteman , ( 1 9 8 0 ) . The techniques of pasture establishment vary from place to place and depend on the area of land to be improved . In a dense forest , clearing the whole area needs to be done before seed preparation . However , in the open grassland or savannah woodland , there are s everal possibilities available instead of the complete land preparation . The cheapest and poss ibly 3 5 the most economical method i s to sow the legume seed into the existing native pasture directly ( Jones , 1 9 7 2 ; Humphreys , 1 9 7 8 a , 1 9 7 8 b , 1 9 82 ; Whiteman , 1 9 8 0 ; Murray , 1 9 8 3 ) . However , in most cases the use of se lected strategies including burning , spraying with herbicide , heavy gra zing or cutting to reduce the competition f rom the existing grasses or weeds , have been employed succes sful ly throughout the tropics . In Northern Austral ia , burning in the early wet season has been recommended for the establishment of Townsvi lle stylo in the spear grass country ( Norman , 1 9 6 5 ; Mi ller and Perry , 1 9 6 8 ) . In Kenya , burning in the dry season and planting in the early wet season have been recommended for oversowing Silverleaf desmodium into Hyparrhenia rufa pasture ( Keya et al , 1 9 7 2 ) . In the highland areas of Thai land , Falvey et al , ( 1 9 8 5 ) s uccessfully introduced ��crotyloma axi l lare and Greenleaf desmodium into a blady grass pasture a fter burning late in the dry season . Doughlas ( 1 9 6 5 ) also • successfully established Siratro and Stylosanthes guianens i s b y oversowing after burning blady grass ( Imperata cylindrica ) but the establi shment of Lotononis , Silverleaf desmodium and G lycine was less successful under these conditions . The success of legume establishment fol lowing burning , particularly for Townsville stylo , appears more dependent on the rainfall and soi l moi s ture than on the pre-treatment ( Mi l ler , 1 9 6 7 ) and in some cases the reduction of grass cover gave poorer establishment due to less favourable surface moi sture conditions ( Mi l ler and Perry , 1 9 6 8 ) . In fact , pre- - treatment may not be necessary if the competition from the existing grasses or natural grasses is not excessive . Robertson ( 1 9 7 8 ) reported a good establishment of Townsvil l e stylo , Siratro , and Stylosanthes into native bamboo grass and the communal land areas which consisted of Arundinaria pus i l la , Dactyloctinium aegyptium and Brachiaria reptans pas ture in Northeastern Thailand . Chemical spraying to reduce competition has been s uggested in both tropical and temperate regions ( McLeod , 1 9 6 2 ; Murtagh , 1 9 6 3 ; Evans , 1 9 6 4 ; Campbell , 1 9 6 9 ; White , 3 6 1 9 7 2a } . The advantages c laimed for us ing herbicide are that it ensures rapid f il ling of the existing pasture and improving soil fertility , reduces weed competition , prevents soil erosion , and spraying can be carried out in areas where normal cultivation cannot be done ( e . g . steep areas } . In areas where the rainfall is low , the dead material remaining can reduce soil moisture loss and minimise the adverse e f fects of cold weather on plant growth ( Dowl ing et a l , 1 9 7 1 } . Keya et al ( 1 9 7 2 } reported that the density of S ilverleaf desmodium was much higher under sprayed plot than burning or cutting prior to sowing . However , in relatively dense vegetation , Tudsri and Whiteman ( 1 9 7 7 a } found l e s s advantage with chemi cal spraying than with burning or cutting before oversowing Siratro . This was because much of the Siratro seed oversown on the sprayed plot failed to reach ground level and hence establishment was - poor . Re-establishment of legumes into depleted sward of cult ivated tropical grasses has generally been less success ful than oversowing o f legumes into natural grasse s , as previously r eviewed . Under favourable conditions of soil moi sture and fertil ity , competition from the a lready established grasses is a ma jor factor a f fecting succes s ful establishment . Tudsri and Whiteman ( 1 9 7 7 a ) demonstrated that pretreatments of cutting , burning or spraying with " Paraquat " of a Setaria anceps sward did not sufficiently reduce t i l ler dens ity to a l low oversown Siratro to establish . Light culti­ vati on was a l so unsuccessful ( Middleton , 1 9 7 3 ) . Only r igorous cultivation of the Setaria stand allowed successful e stablishment of Siratro ( Tudsri and Whiteman , 1 9 7 7a } . However , Lotononis bainesii was successful by merely over­ sowing into Setaria anceps ( Tudsri and Whiteman , 1 9 7 7b ) . With ful l or complete cultivation , there i s generally no problem with establishment and growth . This i s probably due to the complete removal of competition and the provis io� of suitable soil conditions for germination ( Tudsri and Whi teman , 1 9 7 7a ; Wilaipon and Pangskul , 1 9 8 3 ; Cook , 1 9 8 4 ) . Furthermore , with tropical legumes that are slow to nodulate ( e . g . Glyc ine wightii ) , cultivation can increase the rate of 3 7 minerali zation which can assist the legume to establish a� _ an early stage of growth ( Henzell , 1 9 7 7 ) . The problem with ful l cultivation is one of cost and topography . Pastures are usually grown on lower fertil ity soi l s which are not well suited t o cropping ( Will iams and Andrew , 1 9 7 0 ) . Thus , soil fertility problems are found in most tropical pasture situations . The ma jor limiting nutrients for growth on most soi l in the tropics and subtropics are nitrogen and phosphorus . In some areas many are a l so de f icient in potassium and sulphur and notably some of the trace elements . Tropical pasture legumes have the ability to f ix nitrogen through nitrogen-fixing bacteria living in the root nodules . This al lows them to grow and reproduce in soil whi ch is low in nitrogen ( Humphreys , 1 9 7 8b ) . Therefore , in general terms , the application of large amounts of nitrogen ferti l i zer is not common practice with such mixed pastures . The disadvantage of applying nitrogen fertili zer at a high rate i s the stimulation of grass growth and subsequent increased competition for light and nutrients , leading to reduction in the legume content . Furthermore , nitrogen ferti l i zer wil l inhibit nodulation and reduce the rate of nitrogen fixation (Whiteman , 1 9 8 0 ) . Small amounts of nitrogen , however , are suggested for ( e . g . Glycine wighti i ) to assist ( Skerman , 1 9 7 7 ) . some tropical l egumes their establi shment Phosphorus plays a very important role in tropical forage legume growth . Numerous studies have shown that the app l ication of phosphorus fertil i zer at the time of sowing stimulates legume seedl ing growth ( Blunt and Humphreys , 1 9 7 0 ; Keya et a l , 1 9 7 1 ; Ol sen and Moe , 1 9 7 1 ; Tudsri and Whiteman , 1 9 7 7 b ) and results in good legume establishment . I n addi­ tion , Gate ( 1 9 7 4 ) showed that phosphorus fertil i zer enhanced early nodule formation and nitrogen fixation . The nodule dry wei ght , number and density increased with increasing l evel of pho sphorus up to at least 25 kg phosphorus I hectare . 3 8 The beneficial ef fects of phosphorus on- · legume establishment and nitrogen fixation , as mentioned above , and the response in terms of dry matter yields , have been reported by many workers ( Keya et al , 1 9 7 1 ; Olsen and Moe , 1 9 7 1 ; Tudsri and Whiteman , 1 9 7 7b ) . There are dif ferences in phosphorus responses between and within legume species . Stylosanthes humilis and Lotononis bainesii were the least responsive when compared with other legumes ( Andrew and Robins , 1 9 6 9 ) . The ability of S . humilis to absorb greater quantities of phosphorus from a soil with low avai lable phosphorus was shown by Andrew ( 1 9 6 6 ) . White ( 1 9 7 2b ) suggested that the better growth of S . humi lis in a phosphorus deficient situation may be related to the trans­ l ocation of phosphorus from older to younger tissues . This r esults in better growth of this legume under poor soil fertility , as reported by Robinson ( 1 9 8 3 ) and Robertson et al ( 1 9 7 6 ) . Jones , R . K . ( 1 9 7 4 ) described differences in phos­ phorus requirement of a range of Stylosanthes species . In general , the basal level of phosphorus application varies f rom place to place . In Queensland , Australia , a basal dres s ing of 250 - 5 0 0 kg/ha of single Mo superphosphate plus 1 2 5 kg/ha superphosphate f or maintenance is recommended ( Jones and Jones , 1 9 7 1 ) . In Thailand , a basal dressing of 2 0 0 - 2 5 0 kg/ha of superphosphate is recommended ( Tudsri , 1 9 8 0 ) . In terms of potas sium , Andrew and Pieter ( 1 9 7 0 ) showed that Si lverleaf desmodium ( Desmodium uncinatum ) , Perennial stylo ( Stylosanthes guianensis ) and Townsville stylo ( Stylo­ santhes humilis ) were sensitive to potassium deficiency . Hall ( 1 9 7 1 ) also demonstrated that the dry matter yield of S ilverleaf desmodium was s ignificantly depressed when a l lowed to compete with Setaria in the ni l potassium treatment . By contrast , either the addition of potas sium fertil i zer or the s eparation of the roots of the two species , resulted in good l egume growth comparable to that of the companion setaria ( Setaria anceps ) . This indicated that the grass species was stronger in competition for available potassium than the legume . Thus , the botanical competition of the mixed pasture can vary under high and low potassium supply . Jones ( 1 9 6 5 a ) 3 9 reported that the dry matter yield of Siratro was only 1 3 % of total dry matter yield in a mixed sward under no potas sium fertili zer . This i s in contrast to the plot that received pota ssium ferti li zer at a rate of 1 9 0 kg/ha , which produced a Si ratro dry matter yield of 2 9 % of total dry matter yield . Ha l l ( 1 9 7 1 ) pointed out that potassium was the key fertili zer for the establishment of Siratro when mixed with green panic gra s s on poor soil . In this experiment , the application of potassium alone increased Siratro yield substantial ly . Accordingly the use of potassium ferti l i zers is now generally recommended - according to Jones and Jones ( 1 9 7 1 ) it should be 6 6 2 - 2 5 0 kg/ha KCl annually , while Teitzel ( 1 9 7 5 ) considers that 5 0 - 2 0 0 kg/ha KCl annually i s adequate in high rainfall areas . It should be noted however that the appl ication of KCl at rates of 1 0 0 - 2 0 0 kg/ha can cause the death of seedl ings of Desmodium intortum ( Jones , R . M . , 1 9 7 3b ) . Eyles et al ( 1 9 7 3 ) also reported KCl toxicity with Townsville style and Centrosema under higher rates of appl ication and recommended the alternative use of potassium sulphate . Many of the tropical legumes such as Siratro , Glycine , and Calopogo are readily nodulated by Rhizobium strain CB 7 5 6 ( cowpea group ) which occurs naturally in tropical soils . Therefore , �hi zobium inoculation for these legumes i s not required ( Humphreys , 1 9 7 8b , 1 9 8 0a ) . However , some genera have specific Rhizobium requirements , which may dif fer between species and even between cultivars within species . For example , within Stylosanthes guianensi s , cultivars such as Schofield , Cook and Endeavour are nodulated by the general cowpea inoculant CB 7 5 6 . However , the cultivar ' Oxley ' was shown to have a specific Rhi zobium requirement (Whiteman , 1 9 7 8 ; 1 9 8 0 ) . Lotononis , Leucaena , Centrosema and Kenya white clover a l so had specific Rhi zobium requirements ( Humphreys , 1 9 8 0a ) . Mannet j e ( 1 9 6 9 ) a l so demonstrated that Stylosanthes contains species and forms within the specie s , which have specific Rhi zobium requirements . Breaking s eed dormancy or overcoming hard-seededness 4 0 prior to sowing i s recommended for . -Siratro , Lotononis , Stylosanthes species , Centrosema and Leucaena . The common method i s the hot water treatment at 8 0°C for ten minutes ( Skerman , 1 9 7 7 ; Whiteman , 1 9 8 0 ) . Many tropical pasture legume seeds are small and have l imited stored reserves . Therefore , depth of sowing is a critical factor in promoting success ful establishment . Seeds that are left uncovered are subject to rapid changes in soil moi s ture conditions and are readily ki lled by drought fol lowing light showers that promote germination ( e . g . Stylosanthes humili s ) . Seeds that are placed too deep may f a i l to emerge due to insuf ficient energy reserves . Al so fol lowing germination , a crusted soil surface may prevent emergence , especially on fine-textured soils . Therefore , for sma ll seeded legumes such as Lotononis , Stylosanthes and Kenya white clover , only light cover at sowing is recom­ mended . However , with large seeds such as Siratro and Lablab , a depth of 0 . 5 to 1 . 5 cm is preferred ( Smith , 1 9 6 7 ; Stonard , 1 9 6 9 ) . Insect pests and diseases are also very important in l egume establishment . One of the most noticeable - the fungal di sease Anthracnose - was found on Townsvi lle stylo in Australia in 1 9 7 3 and was subsequently reported in Thailand in 1 9 7 6 ( Humphreys , 1 9 7 8a ) . This s igni ficantly reduced the potential of thi s species for improving natural grassland . Rhi zoctonia solani was also recorded in Siratro in the high rainfall tropics , making this species l ess productive . The l egume " l ittle leaf virus " caused by mycoplasma , was also found in Lotononis ( Bryan , 1 9 6 9 ; Hutton , 1 9 7 0 ) , and a long with the Arnnemus weevil , a lso attacks Silverleaf desmodium ( Skerman , 1 9 7 7 ) . Ants can cause substantial losses of seed through their " food-gathering " habit and often require control by dieldrin spraying . Siratro i s also readily attacked by bean fly ( Me l anagromyza phaseoli ) during seedl ing stages ( 3 - 4 weeks ) ( Skerman , 1 9 7 7 ) , but thi s can be prevented by seed treatment with dieldrin , aldrin or endrin at 2 - 4 g . a . i . /kg of seed 4 1 ( Jones , 1 9 6 5b ) . Leptopius weevil . - ( Leptopius spp . ) was found to damage Greenlea f desmodium ( Clements et al , 1 9 8 3 ) and in the northern part of Queensland , many tropical legumes have been destroyed by caterpi llars , Heliothis and Lamprosema species ( Skerman , 1 9 7 7 ) . It is therefore obvious that this period of establish­ ment is a critical stage in the l ife of a pasture if one i s to achieve a high yielding sward free of weeds and capable of high stocking capacity . In a mixed sward , after the legume germinates , it soon starts to compete with gra sses or weeds for water , nutrients and light . The correct early management at this stage can have a very marked ef fect on the l egume content . and on total legume yield and survival . However , tropical legumes appear to respond in d i f ferent ways to the s ame management treat­ ments . For instance , in the case of Townsville stylo , heavy graz ing or cutting at an early stage of growth can enhance legume survival and encourage branching . This i s attributed to the reduction in competition from grasses allowing more light to the legume ( Shaw , 1 9 6 1 ; Norman and Phil l ips , 1 9 7 0 ; Fisher , 1 9 7 3 ; Torssell et al , 1 9 7 6 ; Torssell and McKeon , 1 9 7 6 ; Skerman , 1 9 7 7 ) . Gra zing or cutting at a late stage of growth or late in the growing season , can be di sadvantageous in terms of l egume yield and survival . However , in the case of twining legumes , such as Siratro , competition for light i s l e ss intense owing to its twining habit ( Whiteman 1 9 6 9 ) . Therefore , graz ing should be carried out at an advanced stage of growth or following twining and a l lowing the legume to set seed ( Skerman , 1 9 7 7 ) . Siratro prefers l ight graz ing , as frequent and hard graz ing can lead to rapid reduction in yield ( Jones , 1 9 6 7 ; Whiteman , 1 9 6 9 ) . S imilar responses were found in Desmodium intortum ( Jones , 1 9 7 3 ) and Glycine wightii ( Wh iteman , 1 9 6 9 ) . 4 2 2 . 6 MANAGEMENT FOR STYLOSANTHES HAMATA CV VERANO ( VERANO STYLO ) Although Verano stylo has been released for commercial use for the last decade , detai led information on the behaviour of the plant under varying cultural and managerial conditions is lacking . However , its resemblance to Townsvil l e s tylo has caused several workers to believe that the same type of general management can be appl ied to both ( McKeague et al , 1 9 7 8 ; Humphreys , 1 9 8 0a ) . Unlike most of the tropical legumes , Verano stylo i s better s uited to areas o f erratic rainfall because it can f l ower at any time and vegetative growth conti nues after the onset of f lowering ( Cameron and Mannet je , 1 9 7 7 ; Gillard et al , 1 9 8 0 ) . It also persi sts in areas receiving an annual rainfal l as low as 6 0 0 mm and usually produces more forage yield than Townsville stylo under such conditions ( Burt e t a l , 1 9 7 4 ; Edye e t al , 1 9 7 5 b ; Gillard e t al , 1 9 8 0 ) . Verano stylo can be sown in exactly the same way a s Townsvi l le stylo or other tropical legumes that have been previously reviewed . It i s well adapted to a wide range o f soil types and cl imatic conditions and shows a greater resistance to anthracnose disease than Townsville s tylo ( Vini j ts anond and Topark-Ngarm , 1 9 78 ) . Use of this l egume to improve natural and cultivated gras ses is commonly employed in many parts of the tropics , such as Thailand ( Humphreys , 1 9 7 8 a ; Gutteridge , 1 9 8 2 ) and Malaysia ( Whiteman , 1 9 7 8 ) . Seed-bed preparation may or weeds are not be short essential ( Robertson , i f the 1 9 7 8 ; existing grasses Humphreys , 1 9 8 2 , 1 9 8 4 ; McKeague and Holmes , 1 9 7 9 ) . However , cul tivation of the land prior to sowing general ly results in better germination , establishment and dry matter yield ( Wi l aipon , 1 9 8 0 ; Wilaipon and Pongskul , 1 9 8 3 ) . Verano stylo can nodulate freely with native soi l Rhizobium . Work in Thailand indicates that Verano stylo can produce high yields on contrasting soil types ( Korat and Nam­ Pong Series ) . Nitrogen f ixation also appears to be 4 3 ef f icient , as no difference was recorded in terms of both dry matter and nitrogen yields in the presence or absence of added nitrogen fertili zer in the field , conducted by Ruaysoongnern ( 1 9 7 8 ) . Although Verano stylo i s easily e stablished , poor seed germination can be a problem reducing effective establishment ( Gardener , 1 9 7 5 ) . Unless Verano stylo seed is treated to reduce hard-seedednes s , only about 1 0 % f ield germination can be expected in the establishment year ( Mott et al , 1 9 8 1 ) . Therefore , seed treated with hot water at 8 0°C for 5 - 1 0 minutes is recommended ( Whiteman , 1 9 8 0 ; Crowder and Chheda , 1 9 8 2 ) . However , to cope more e f fectively with large quantities of seed , Mott ( 1 9 7 9 ) developed a rotating drum in whi ch to treat such seed for 1 5 - 2 0 seconds at 1 5 5°C . The result showed that such a heat treatment increased Verano s ty lo germination 1 0 - fold compared with the control at Katherine , Austral ia . Legume yields from the resulting pas ture were 1 6 times those of the control ( Mott , 1 9 7 9 ) . In common with the other tropical legumes previous ly reviewed , planting date had a marked ef fect on the establ ish­ ment of Verano stylo , with early sowing being strongly recommended in Thailand ( Wilaipon and Pongskul , 1 9 8 3 ; Humphreys , 1 9 8 4 ) . In Australia , it is recommended Verano stylo be sown just before the wet season , into or short grass . Aircraft or ground equipment can be but markers are needed for accurate aircraft sowing standing timber ( McKeague et al , 1 9 7 8 ) . that burnt u sed , into Basal fertili zers are also recommended prior to sowing . The addition of phosphorus ferti lizer increased dry matter yields and percentage Verano stylo in a number of experiments ( Hal l , 1 9 7 9 ) . Hal l { 1 9 7 9 ) also reported that the basal superphosphate of 3 0 0 kg/ha on a deep sandy yellow earth increased the dry matter yield of Verano stylo from 1 ton to 2 . 5 ton/ha . However , in Thai land , the response of Verano stylo to phosphorus and sulphur fertil i zers has been variable , and in some cases non-existent { Gutteridge , 1 9 7 8 ; Wil aipon and Pongskul , 1 9 8 3 ) . These authors suggested that 4 4 the native level s o f soi l phosphorus and sulphur were often adequate for Verano style growth . Nitrogen ferti li zer however was not recommended in a mixed pasture , but in a pure s tand of Verano style , Brolman and Sonoda ( 1 9 8 1 ) reported a better dry matter yield when nitrogen , at the equivalent of 1 8 0 kg N/ha , was added . Verano style cult ivated tropical is not grasses only sown but also with natural and with upland crops . These include rice , cassava and kenaf . Work at Khan Kaen University , Thailand , has shown that Verano style can be established in stands of cassava and kenaf without seriously a f f ecting the yield of the companion crop . However , the time of sowing is important . Wilaipon et a l ( 1 9 8 1 ) found that when Verano style and cas sava were planted at the same t ime , there was a s igni ficant reduction in cassava yield compared with the cassava alone treatment ( 7 . 6 vs 2 6 . 8 ton/ha ) . But when Verano style was sown six weeks after the cassava , there was a very small and insigni f icant reduction in cas sava yields ( 2 4 . 5 vs 2 6 . 8 ton/ha ) . Pongskul et al ( 1 9 8 2 ) found that kenaf yields when planted alone were 1 2 . 1 ton/ha . When sown in combination with Verano style , yields were reduced to 8 . 9 ton/ha . However , Verano s tyle yields were reduced f ive­ fold by delaying the sowing in the kenaf . Although there were no beneficial effects from sowing Verano style in these crops in terms of yields , it provided a substantial amount of additional forage for l ivestock . This forage can be uti l i z ed before harvesting kenaf or cas sava . Newly sown Verano style needs early graz ing to check gra s s growth and to encourage the legume to branch . It i s des irable to encourage the plant to act a s a perennial and j udicious grazing may be inf luential . If Verano style grows and f lower s unchecked for too long , it may drop its l eaves and die . Regrowth from seed is usually reliable but weeds are more l ikely to invade when the l egume is allowed to act as an annual . However , the l ong term pers is tence o f the l egume in heavily grazed situations is nevertheless dependent on the a ppropriate form of grazing management to ensure adequate seeding and seedl ing regeneration in subsequent seasons (Wilaipon and Pongskul , 1 9 8 3 ) . 4 5 I n communal land areas where the Verano style was lightly s tocked in the wet season and only moderately stocked during the dry season , the l egume content at the end of the second wet season increased f rom 3 4 % in 1 9 7 8 to 5 6 % in 1 9 7 9 . This i s in contrast to the treatment that was heavily grazed which decl ined in l egume c ontent from 3 1 % in 1 9 7 8 to 1 9 % in 1 9 7 9 . In Australia , McKeague et al ( 1 9 7 8 ) have suggested that leaving a 1 0 cm stubble in July and August makes the best use of wet and dry season growth . I n Thai land , early heavy grazing in the wet sea son and deferment in the late wet season is recommended to a l low seed setting ( Wi laipon and Humphreys , 1 9 8 1 ; Humphreys , 1 9 8 2 ) . CHAPTER 3 EXPERIMENT I : GROWTH PATTERN OF STYLOSANTHES HAMATA CV VERANO UNDER CONTROLLED ENVIRONMENT CONDITIONS I . INTRODUCTION 4 6 There are few publications on the growth and development of Verano stylo ( Stylosanthes hamata cv Verano ) . Gardener ( 1 9 7 8 ) measured the relative growth rate and dry weight increments of a number of Stylosanthes species including Verano stylo over a period of six weeks after sowing . He found that relative growth rate and dry weight increments in Verano stylo were higher than those of perennial stylo s pecies l ike Stylosanthes scabra but similar to that of S . hurni l is . However , the relative growth rate during this period was still low compared with that of C4 grasses ( e . g . Panicurn maximum ) ( Ludlow and Wi lson , 1 9 7 0 ) . This could be a problem when Verano s tylo i s grown with such vigorous companion grasses . Gardener et al ( 1 9 8 2 ) studied the changes in total dry weight and its components from a pure stand of Verano stylo over two growing seasons . Total dry weight production reached a peak at about 1 2 8 days after sowing . The contribution of the leaf fraction to dry matter yields was high during the early stages of growth , before decl ining throughout the wet season , with the subsequent ma j or contribution coming from stern tis sue . This introductory experiment reports the uninterrupted growth and development of Verano stylo under conditions of adequate soil moisture and nutrients over a period of 1 3 1 days in the control led environment of a growth room and hence provides a desirable background and understanding of the species in ques tion . It includes various aspects of plant morphological development , including branch numbers and plant height , and describes the increase and distribution of dry matter yield and leaf area development . 4 7 II . MATERIAL AND METHODS A . ENVIRONMENTAL CONDITIONS AND PLANTING PROCEDURES . The experiment was carried out in the Controlled Cl imate Rooms at the Plant Physiology Divi sion , DSIR , Palmerston North , New Zealand ( Plate 1 . 1 ) . P lastic pots of 1 6 cm diameter and 1 6 cm depth , with drainage holes at the base were f i l l ed with 5 . 5 kg of a mixture of peat , vermiculite and gravel ( 1 5 : 1 5 7 0 ) . Stylosanthes hamata cultivar Verano ( Verano stylo ) seeds col lected from North-Ea stern Tha i land were selected for s imi lar s i z e and scarified with concentrated sulphuric acid for 30 minutes , then washed thoroughly and dried . Seeds were incubated ln a germinator for 4 8 hour s at 2 0 - 3 0°C before planting . Ten germinated seeds were sown in each pot , covered with a thin layer of soi l , and watered . Pots were placed directly in the controlled environment room . Environmental imposed ( detai led in Appendix 1 ) were : 1 2 hours conditions Photoperiod Light intens ity Temperature Humidity 1 6 0 Wm- 2 ( 4 0 0 - 7 0 0 nm ) 3 0°C day and 2 4°C night 70 day and 90 night Carbon dioxide level - 2 9 0 - 3 4 0 ppm These environmental conditions were chosen , fol lowing reference to cl imatic records for the Central Plain region of Thai l and , to s imulate as closely as possible the cl imate prevai l ing during the growing season . A complete nutrient solution ( NCSU - Appendix 2 ) was app l i ed twice daily for the first three weeks of pl anting . Watering frequency was then increased to four times per day unti l the end of the experiment . After three weeks of growth , some ' tip burn ' was observed . To minimi se thi s condition , pots were f lushed through with deioni zed water every week to remove nutrients accumulating in the pot . Rhi zobial bacteria ( in agar form ) were applied to the soi l surface one week after sowing . Seedl ings were thinned to three plants per pot , after two weeks and finally to one plant per pot . 47a P l3te 1 . 1 : General vj ew of the Control led C l j mate Room . 4 8 B . HARVESTING SCHEDULE Harvests of the entire plant were made on days 7 , 1 4 , 2 1 , 2 8 , 3 9 , 6 0 , 6 7 , 8 1 , 8 8 , 1 1 7 and 1 3 1 days after seedling emergence . occasion . Four plants were removed for dissection on each C . PLANT MEASUREMENTS At each harvest , plants were separated into leaf , stern and inf lorescence . The samples were dried in a vacuum air oven for 4 8 - 72 hours and weighed . The number of l eaves , branches and inflorescences were recorded . Leaf area was also measured us ing an electronic Leaf Area Meter ( Model 3 1 0 0 Area Meter ) . Soil was washed from the root system and the tap root ( including large woody root ) and fibrous roots were separated and weighed after drying . I l l RESULTS AND DI SCUSSION Before presenting the data on the growth and development of Verano style , it i s helpful to describe and explain its ma j or morphological features . The fol lowing section descr ibes the growth and development of intact plants of Verano stylo , 1 3 1 days . from seedling emergence to full flowering at A . MORPHOLOGICAL FEATURES A . 1 Stage of Growth Baldos and Javier ( 1 9 7 6 ) recognised three sharply def i ned development phases in the growth of Townsvi lle s tylo ( Stylosanthes hurnili s ) : juven i le , vegetative and repro­ ductive . In contrast , Verano style , a quantitative short-day plant , had no sharp destinction between the vegetative and reproductive phase of development , as shown by Carneron and Mannet je ( 1 9 7 7 ) . The commencement of first flowering was a l so early compared with that of other tropical legumes l ike S . guianensi s ( Loch et al , 1 9 7 6 ) and S . hurnil i s ( Ba ldos and Javier , 1 9 7 6 ) . Similar results were observed in the present 4 9 work and also showed that the juveni le and vegetative stages were diff icult to distinguish . Thus , on morphological grounds , growth in thi s study was divided into two stages , vi z pre-flowering and flowering . Pre-f lowering covered both the so-called juvenile and vegetative stages while the f lowering stage covered the period after the onset of f loweri ng . In thi s study , the pre-flowering stage extended over the f irst 3 5 days of growth , which is shorter than tha t r ecorded under field conditions in Queensland ( Skerman , 1 9 7 7 ; Wi laipon and Humphreys , 1 9 7 6 ) . These authors reported the pre-f lowering period to be about 6 0 - 67 days . However , under f ield conditions in Thailand Wilaipon and Humphreys ( 1 9 8 1 ) found that Verano style had a somewhat shorter pre­ f lowering period of approximately 40 days from seedl ing emergence . Such variation may in part be explained by the d i f ferences in photoperiod . Under· field conditions , day­ length varied from month to month while in the control led room a constant 1 2 hrs daylength was employed . Cameron and Mannet je ( 1 9 7 7 ) found that under control led conditions , Verano style flowered at a l l photoperiods tested , but it occurred much earl ier in the shorter photoperiods of 1 0 and 1 1 . 5 hrs . In their exper iment , Verano style flowered 3 2 days a fter sowing at 1 1 . 5 hrs photoperiod and 3 2°/ 2 4°C ( day/night ) - this i s s imilar to the present experiment . Plants at the pre-flowering stage had a s lower rate of growth and produced fewer branches and leaves , accompanied by only small increments in plant height compared with the f lowering stage . However seed germination was rapid and the production of the f irst trifol iate leaf was observed within 7 days ( Plate 1 . 2A ) . The ma jor portion of dry weight accumulation during the l i fe of the plant occurred after the onset of flowering and included vegetative and reproductive growth . Thi s was a lso observed by Cameron and Mannet je ( 1 9 7 7 ) and Wilaipon et a l ( 1 9 7 9 ) . In this study , the f lowering stage commenced approximately 3 5 days after seedling emergence and continued throughout the experimental per iod . Plate 1 . 2 : Growth and development of Verano stylo a f ter seed l i ng emergence up to f i rst f lO\•!ering appearance . A = 1 week; B = 2 weeks ; C = 3 weeks ; D = 4 weeks ; E = 5 weeks . � ...0 Q 5 0 A . 2 BRANCH NUMBER Branching of the main s tem commenced in the axi l s of the cotyl edons approximately 2 1 days a fter emergence ( Plate 1 . 2C ) . Over a subsequent 2 - 3 wee k period , branch numbers increas ed only s l owly at about 3 branches per week . At the onset of f lowering , when secondary branch formation commenced , the rate of branch produc tion ros e sharply ( Table 1 . 1 ) . Branching progres sed up the main stem and i n thi s study , the terminal growing point developed an inf lorescence after the formation of 1 1 nodes . This was s lar to the f igure g iven by Cameron and Manne t j e ( 1 9 7 7 ) . Branch numbers rapidly increased as primary , s econdary and h i gher order branches appeared . As a resu l t , the rate o f branch a ppearance was highest between day 6 0 and 6 7 at 3 2 branches per day . Total branches per p lant reached a maximum of 1 8 2 3 a t day 1 1 7 . The dec l ine in the number of branches at a later s tage ( day 1 3 1 ) was due to s hedding o f the sma l l and older reproductive branches . The results demonstrate that a high rate of branch dif ferentiation continues during f l owering in Verano stylo . It a l so shows that this cultivar i s capabl e o f producing large number s of branche s and greater , than some other tropical l egume s . For example , with Townsvi l le stylo , Baldos and Javier ( 1 9 7 6 ) reported 8 4 5 /plant 1 6 8 days a fter sowing , with a rate o f branch appearance of 1 6 branches per day between day 1 1 2 and 1 4 0 . Loch and Humphreys ( 1 9 7 0 ) , working with l le s tylo , a l s o reported a maximum rate of branch appearance of approximate ly 1 . 1 branches per day between days 8 9 and 1 0 4 . For Verano stylo , Wil a ipon et a l ( 1 9 7 9 ) reported that maximum total branch den s i ty was�9 8 0 branches/m2 at 9 6 days after sowing , under g lasshous e conditions . The high number of branches produced indicates that Verano stylo has a h i gh potential for bud production for subsequent growth . A c lose relationship ( R = 0 . 9 5 2 * * ) between branch numbers and p l ant dry wei ght , suggests that graz ing or cutting management should a at encouraging branch development . 50 a Tabl e 1 . 1 Number o f branches per plant Days a fter Number of branches s eed l i ng per ant emergence 7 0 1 4 0 2 1 2 + 0 - 2 8 6 + 0 . 5 - 3 9 1 4 6 + 1 2 . 6 - 6 0 3 1 7 + 1 3 0 . 1 - 6 7 5 9 6 + 1 4 8 . 5 - 8 1 1 0 4 0 + 1 9 0 . 6 - 8 8 1 1 4 7 + 2 0 6 . 8 - 1 1 7 1 8 2 3 + 6 6 6 . 7 - 1 3 1 1 5 0 5 + 3 4 8 . 0 - 1 5 0 % of the pl ant population f lowering . 5 1 A . 3 LENGTH OF PRIMARY BRANCH The f i r s t primary branches i n the axi l s of the cotyledon were obse rved three weeks after seed l ing emergence . By the fourth week , another three branches ( 1 . 8 0 , 3 . 9 0 and 1 . 8 3 cm i n l ength respectively ) had a r i sen f rom the axi l o f the f irst , s econd and third true l e aves . Al l primary branches increas ed in length with the mos t marked increas e s occurring in the lower primary branches ( Figure 1 . 1 ) . A . 4 PLANT HEIGHT The s l ow increase in p l ant height during the pre- f lowering s tage is shown in Tabl e 1 . 2 . However , there was a s teep increase in plant he ight j u s t before the plant f l owered and a maximum height of 1 . 2 m was mea sured on day 1 3 1 . Thi s was s imi l a r to f ield values obtained for Verano s ty l o in Queens l and by elements ( 1 9 8 0 ) under good growing conditions . The i n i t ia l s l ow increase of plant height in Verano stylo compared with the more rapid increas e in ight of a s s ociated tropical grasses at the s ame s tage of growth confers s ometh ing o f a competitive d i sadvantage on Verano s ty lo under mixed p a sture conditions . ( Gardener , 1 9 7 8 and Udchachon , 1 9 8 5 ) . B . GROWTH AND DEVELOPMENT B . 1 Growth Analys i s The relationship between total plant dry we and t was tested us ing both the log i st i c and the c polynomial model ( Evans , 1 9 7 2 ) . The results suggest that nei ther mode l describes the entire growth pattern of the accuratel y . The l o g i stic growth model over- e s t total ant dry weight during the early s tages of growth ( Appendi x 3A ) whi l e at t he l ate stage ( days 1 1 7 - 1 3 1 ) the model icted a n increas e in plant dry weight with time . T h i s contrasts with actual p lant growth data ( Fi gure 1 . 2A ) where tot a l plant weight dropped sharply at that time . Thus , maximum yield wa s unable to be obtained from the logi stic curve . However , with 51 a 0 60 50 40 30 20 21 DAYS POST SEEDLING EMERGENCE 39 DAYS 1- 28 DAYS 60 DAYS E u :r. (.) 0 o������������U-- � 100 la 90 >- � 80 ::i g: 70 � 60 :r. 1- 50 (!) z UJ ...J 40 30 20 01- 8 1 DAYS 131 DAYS - o���������----�--�������4---I Z 3 4 5 6 7 8 9 JO II 1 2 3 4 5 6 7 8 9 10 11 PRIMARY BRANCH Figure 1 . 1 : Length of primary branches a long the ma in stem ( cm ) 5lb s after s eedl ing emergence Plant height ( cm ) 7 2 . 6 + 0 . 3 9 - 1 4 3 . 4 + 0 . 8 5 - 2 1 4 . 7 + 0 . 2 9 - 2 8 1 6 . 4 + 0 . 5 4 - 3 9 1 2 8 . 9 + 0 . 8 3 - 6 0 6 4 . 2 + 0 . 7 9 - 6 7 7 2 . 7 + 2 . 3 2 - 8 1 9 0 . 5 + 6 . 8 1 - 8 8 9 2 . 6 + 4 . 1 0 - 1 1 7 1 1 2 . 2 + 4 . 7 8 - 1 3 1 1 1 4 . 9 + 3 . 4 0 - 1 5 0 % o f t h e plant population f lowering 5 1 C 120 1 00 c � 80 Q. ... • Q. do - 60 1- � >-� 40 ..J - a:: -2 0 ..J � 0 t- -4 -5 A: ACTUAL DATA 50•4 FLOWERiNG 8 20 FITTED 40 60 80 iOO 120 140 DATA 1 : log• ty ) : 0 1 92 1 6 ·74 l M f< = 0 · 980 2 . Y ' 75 16 - 5 t5 I + 0 !2 1 2- 0 00048 t3 rf. 0 927 .. Y • DRY WGT. !g . per PLANT ) t • DA.YS AFTER SEEDLING EMERGENCE 20 40 60 80 100 120 140 DAYS AFTER SEE DLING EMERGENCE F igure 1 . 2 : Changes i n tota l dry we 1 ght wi th t 1 me : A . Actua l da ta ; B . From f 1 t ted data ; 1 and 2 r epresent l i ne of best f 1 t between each phase o f growth 1 00 -1 0 --; J> r 80 0 :u -< � 60 (;) .-1 .0 't) 40 • 't) 0 � 20 5 2 the polynomia l model ( Appendix 3 B ) , the growth o f the p lant was under-estimated during the ear ly s tages of growth ( 0 - 2 8 days ) . Thes e l ed to the conclus ion that the entire g rowth pattern of thi s s pecies in the pres en t experiment could not be described u s ing onl y one model . Hunt and Par s ons ( 1 9 7 7 } sugge sted that a l engthy and compl ex s er ie s of data shou ld be approached by the segmentation o f the growth curve . The relationship between plant dry weight and time was therefore described us ing two equations . The f ir s t section of the data , up to 5 0 % f lowering ( 0 - 39 days ) , was bes t fitted with a l inear equation ( Figure 1 . 2B ) whi le the remainder gave the bes t f it with a cubic pol ynomial ( Figure 1 . 2 B ) . Thi s latter equation was able to describe the rapid growth phase as wel l a s the later dec l ine in tota l p lant dry weight . To obtain more information on the growth pattern o f Verano stylo , absolute growth rate was calcu lated at 1 0 day interva l s u s in g the above equation s . The pattern of growth rate at the pre-f lower ing s tage was s low and rapid ly increased a fter the onset of f l ower ing . The highes t absolute growth rate occurred between 7 0 and 8 0 days a f t e r seed l ing emergence at 2 . 0 4 g per day ( Figure 1 . 3 ) . Max imum total ant dry weight of 1 0 5 g per plant was obtained 1 0 8 days after seed l ing emergence ( Figure 1 . 2 B ) us ing the pred icted growth curve . Abs olute growth rate decreased rapidly to zero between 1 1 0 and 1 2 0 days . The relative growth rate calcul us ing the two equations , i s s hown in F 1 . 4 . I n common with the obs ervation for most p lant s pecies , re lative rate decl ined a s the aged . s i s related to the decreas ing proportion of active l y growing t i s sue s . 5 2 a 2· 30 2·00 1 ·50 g 1 ·00 :. ()-5 Q. 0 FLOWERING 40 60 80 1 0 120 1 40 DAYS AFTER SEEDLING EMERGENCE Figure 1 . 3 : Absolute growth ra te of Ve rano sty lo ca lcu l a t�d f rom f i tted growth mode l ( Fi gure 1 . 2B ) -3 0 ·200 ... & 0 · 1 80 c,;. l 0 · 160 0· 1 40 0 · 1 20 0 ·020 0·040 �----�----._--�----�----�------ 20 40 60 eo 100 120 140 DAYS AFTER SEEDLING EMERGENCE Figure 1 . 4 : Re lat ive growth rate of Verano stylo ca l cula ted f rom f i t ted growth model ( Fi gure 1 . 2B ) 5 3 B . 2 DRY MATTER YIELD The growth and devel opment of p lant components are presented in Figure 1 . 5 . Leaf and s tem fraction s i ncreased s lowly during the pre-f l owering s tage ( 0 - 3 9 days ) . Then a l l components , including i n f l orescences , s igni f i cantly increased with the ons et of f lowering , des pite ma j or difference s in the rate o f growth for each component . For example , during the first four weeks fol lowing the onset o f f lowering ( i . e . to day 6 3 approximately ) , a l l components increased s ubstantial ly with the maximum rate of increase occurring in the stem fraction , which continued to increase signi f icantly . The leaf fracti on increased at a much s l ower rate during the 4 weeks from flowering , as d id the root fraction . After thi s time , both the leaf and root fractions remained re latively constant unti l the end of the experiment . All components reached a max imum dry weight by day 1 1 7 , and then e i ther remained constant or decl ined to the end of the experiment . Figure 1 . 5 c learly shows that stem i s the main plant component , fo l lowed by the inf lorescence fraction . The two contributed more than 5 0 % o f total plant dry wei ght . Although Verano stylo had a high rate o f l ea f and branch dif ferentiation a fter the onset of the f lowering , tota l leaf dry weight ( Figure 1 . 5 ) did not show the same trend a s leaf number ( Table 1 . 3 ) and branch number ( Table 1 . 1 ) beyond day 6 7 . Thi s could be explained by reduced dry per leaf f rom day 3 9 , by reduced spec i f ic leaf area from 8 8 and through an observed se in lea f senescence ( leaf number ) near the of the exper 1 per iod ( e 1 . 3 . Branch number was a l so found to be c losely carrel with inf l orescence dry weight ( R = 0 . 9 7 8 * * ) , th s tem weight ( R = 0 . 9 3 9 * * ) and with total leaf dry wei ght ( R = 0 . 8 8 7 * * ) . The inf l orescence fraction dry weight was s trong l y re lated t o inflorescence number ( R = 0 . 9 6 2 * * ) rather to the s i z e o f the individua l inf lorescence ( R = 0 . 4 5 3 * ) e 1 . 4 ) . The decl ine in inf lorescence dry weight over the f inal 100 90 80 1- 70 z - 20 0:: 0 1 0 5 3a STEM LEAF FIRST F LOWERING � 20 80 100 120 140 DAYS AFTE R SEEDL I N G E M E RGENCE Fioure 1 . 5 : Dry matter yield of l ea f , s tem, inf l orescence and roots of Verano styl o ( g /plant ) 5 3b Leaf number, leaf size , ---- weight .rer leaf specific leaf area . after Number LE:af size LE:af S.recific leaf of 2 dry 2 (mt /leaf } area an /g ) 1 (m:j/leaf ) emergence .rer plant 7 0 . 66 5 . 0 1 4 3 0 . 86 5 1 84 . 3 2 1 5 1 . 28 7 205 . 8 28 1 8 1 . 7 1 9 1 97 . 5 3 92 1 07 3 . 1 5 1 5 2 1 0 . 1 60 538 2 . 78 1 2 227 . 6 6 7 984 2 . 39 1 1 227 . 0 8 1 1 42 0 2 . 1 8 9 24 1 . 0 8 8 1 424 2 . 1 0 8 254 . 3 1 1 7 1 946 1 . 58 7 224 . 4 1 3 1 1 634 1 . 32 7 202 . 2 1 A trifoliate leaf . 2 50% of the plant population flowering . 5 3c Tab l e 1 . 4 In f lor e s cence number and dry we ight Days a f ter I n f l orescence Dry weight ( mg ) seedl ing number per inflorescence emergence per plant 3 9 2 1 0 . 2 6 0 1 9 0 1 9 . 7 6 7 4 6 5 2 2 . 7 8 1 9 1 4 2 1 • 9 8 8 9 9 0 2 2 . 4 1 1 7 1 7 7 7 1 5 . 2 1 3 1 1 4 9 9 1 4 . 4 5 4 1 6 days i s explained by the reduction i n inf l orescence number over thi s period ( Table 1 . 4 ) . In add it ion , s hedding o f older i nf lorescences coul d a l so be important , a s demonstrated by Gardener et al ( 1 9 8 2 ) . I n the pre s ent study , the dry matter yield o f a l l the components mentioned above arose mainly from the l ower primary branches ( branches 1 - 6 ) on the main stem ( Figure 1 . 6 ) . Only 3 0 % o f total dry matter yield came from the upper primary branches ( br anches 7 - 1 2 ) which originated from the s ixth to eleventh nodes on the main stem . Once again , dry matter yield cons i sted mainly of s tem t i s sue , part icular l y the lower primary branches ( Figure 1 . 6 ) , with a n increa s ing mas s up to day 1 1 7 . Beyond day 1 1 7 a l l components dec l i ned in weight , on both upper and lowerr branches , a lthough the proportion of s tem c ontinued to r i s e . Unl ike the s ituation described by Tor s se l l e t a l ( 1 9 6 8 ) who showed that tota l leaf dry weight o f the lower branches dec l ined as stand height increased , total l eaf dry wei gh t of the lower branches i n this experiment actual ly increased with plant height up to day 1 1 7 and on ly dec l ined over the l a st 1 4 days . I n the present study the proportion o f leaf and s tem was s imi lar in both the upper and lower branches a nd sugges t s that h i gher rates o f l eaf dif ferentiation on the l ower branches may compensate for loss o f when the plant i s grown under pos s s through ageing , ideal condi tions . Growth o f roots during the seed l i ng stage ( 0 - 2 8 days ) was mainly due to e l ongation of tap root , but with f ew branch root s . A more rapid g rowth rate fol l owed immediately a f ter the onset of f l owering ( day 3 5 ) a nd s i gni f icant l y increa s ed unt i l day 6 7 . Thereafter , root dry weight changed l itt le throughout the experiment ( Figure 1 . 5 ) . A s a result , the shoot to root ratio increased from 3 on day 7 to 2 2 on day 1 1 7 • 54 a - 25 1-z <( 20 ...J Q.. .... 1 5 Q) Q. cia 10 :r u z <( a:: m >-0:: <( 65 � w :r 1- o 4 ...J � 40 >- 0:: 35 w � 30 ­ � 20 15 10 39 DAYS POST SEEDLING EMERGENCE 8 T 81 DAYS IJJ 8 (.) z IJJ (.) (/) IJJ 0:: 0 li z IJ... <( � :E IJJ 1:­(/) 8 60 DAYS 67 DAYS 8 8 1 1 7 DAYS 13 1 DAYS 8 Figure 1 . 6 : Distribut ion of dry matter y i eld on the primary branches of the ma in stem B = l ower pr imary branches ( 1 - 6 ) T upper pr i mary branches ( 7 - 1 2 ) B . 3 LEAF AREA ( CM2/PLANT ) , NUMBER OF LEAVES AND SPECIFIC LEAF AREA { CM2/G ) Lea f characteri s t i c s o f Verano s tylo plants 5 5 are presented in Table 1 . 3 and F igure 1 . 7 . The numbe r o f leaves and total leaf area did not increase sign i ficantly during the pre-f lowering stage of growth . However , from the onset of f lowering , the number of l eaves increased rapidly from day 2 8 up to day 1 1 7 . Leaf area a l s o increased rapid from day 2 8 but reached a maximum o n day 8 1 - showing that attainment of maximum leaf area does not nece s sari ly depend on leaf number , as leaf s i z e and spe c i f ic leaf area are a l so important in determining l eaf area production . The reduction in leaf number f rom day 1 1 7 , due presumably to leaf sene s cence , was probably the main contributor to the reduction l eaf area at that stage . Although leaf area index could not be c a l c ulated , the maximum leaf area of 3 0 9 7 cm/plant was noted on day 8 1 . However , the maximum rate of leaf appearance o f 6 3 . 2 leaves per p lant per day was obtained between day 6 0 a nd day 6 7 . The results o f thi s s tudy demonstrate that Verano s tylo had a high rate of l ea f di f ferentiation following the onset of the f lowering . Wi laipon et al ( 1 9 7 9 ) a l s o showed that a leaf . -2 -dens1ty of 3 5 5 0 0 per m o ccurred 9 6 days after s owing g la s shouse conditions . However , these results wer e not s hown on a per plant bas is . ..... z <{ .J a.. ..... Q) a.. "'E 0 <{ w a:: <{ lL <{ w .J 5 5 a 3000 2000 1 000 I " S. D. DAYS AFTE R SEEDL ING EMERGENCE Figure 1 . 7 : Changes in total leaf area of Verano stylo with t i me ( cm 2 /plant ) 5 6 CHAPTER 4 THE EFFECT OF DEFOLIATI ON ON THE GROWTH AND REGROWTH CHARACTERI STICS OF STYLOSANTHES HAMATA CV VERANO . EXPERIMENT 2 : STAGE OF GROWTH AND INTENSITY ! . INTRODUCTION A necessary s tep in the development o f management strategi e s f or the long term per s i s tence of legumes in grazed and cut pastures involves r esearch into the e f fect o f defoliation at critical per iods i n the plant ' s development . With Verano stylo under f i e ld condi tions , evidence that de fol iation at an early s tage of growth ( fi r s t does not a f fect the regrowth abi l ity o f the plant . s uggest s f lower ) However , gra z ing or cutting late i n the g rowing s ea son depre s s e s yields ( Wi laipon and Humphreys , 1 9 7 6 , 1 9 8 1 ) . I n thes e experiments , Verano styl o was grown with volunteer gra s ses . In addition , the ef fects o f l ate cutting were confounded with the unfavourable conditions at the end of the growing s eason . Thus , the e f fect of cutting per s e could not be quant i f ied . Experiment 1 was des i gned to s tudy the growth pattern o f Verano s tylo under control led cl imatic conditions s imulating those found in Thai land . The results cated a worthwhi le ld potent ial of thi s resulting from a branch and leaf dif ferenti ation over 1 3 1 rate o f o f the ex per t . The next was to examine the growth re sponse pattern of the var ious plant components when the p l ant was sub j ected to dif ferent ities of i ation at different stages of growth and devel opment . 5 7 I I . MATERIAL AND METHODS A . ENVIRONMENTAL CONDITIONS AND PLANTING PROCEDURES Environmental condit ions and plant ing procedures were s imi lar t o the previous exper iment ( Experiment 1 ) . After thinning t o one plant per pot , a l l plants wer e b locked into four rep l ica tions based on the i r uni formity of e s tabl i s hment . After treatments were imposed , pots o f s imi lar treatments were located together so that between treatment inter f erence was minimis ed during regrowth . B . TREATMENTS There were two cutting intens it i e s and two s tages o f growth i n t h i s experiment , a s fo l l ows : - Stage of growth : Cutting intens ity : Control : 1 . Early s tage - when 5 0 % o f the p lants commenced f lower ing ( approximately 3 9 days after seed l ing emergence ) . 2 . Late s tage - a t a pproximate ly 8 8 days a fter seed l ing emergence ( cl o s e to maximum growth rate ) . 1 . Cut the mai n s t em above node 5 ( between nodes 5 and 6 ) and pr ry above node 4 ( between nodes 4 and 5 a long the branch ) . 2 . Cut the mai n s tem above 5 ( between node s 5 and 6 } and pr branch immediate below node 1 ( along the branch ) . Uncut . 5 8 Therefore the f ive treatment s ( Figure 2 . 1 ) can be l i s ted a s fol l ows : 1 . Control : 2 . E - 5 - 4 : 3 . E- 5 - 0 : 4 . L- 5 - 4 : 5 . L - 5- 0 : Uncut . Cut the main stem above node 5 ( be tween nodes 5 and 6 ) and primary branch above node 4 ( between nodes 4 and 5 a long the branch ) when 5 0 % of the plants commenced f lowering . Cut the main stem above node 5 ( between nodes 5 and 6 ) and pr imary branch immediately be low node 1 ( al ong the branch ) when 5 0 % of the plants commenced f lowering . Cut the main s tem above node 5 ( between nodes 5 and 6 ) and primary branch above node 4 ( between nodes 4 and 5 a long the branch ) at 8 8 days a fter seedl ing emergence . Cut the mai n stem above node 5 ( be tween nodes 5 and 6 ) and primary branch immediately be l ow node 1 ( along the branch ) at 8 8 days after s eed l ing emergence . The f i r s t cutting was according to the s tage of growth and the s econd cut was sed after a regrowth per iod of 4 2 days . A f inal cut was after a further regrowth period of 2 8 days , except the uncut control which was l ly sampled at 1 3 1 days a fter seedl ing emergence . C . HARVESTING The harvesting s chedule i s given in Figure 2 . 2 . 58 a A: EARLY STAGE OF GROWTH ( EARLY ) MAIN STEM L EAF PRIMARY BRANCH TREATMENT E- 5- 4 I TREATMENT : E-5-0 8 : LATE STAGE OF GROWTH ( LATE ) TREATMENT : L-5-4 PRIMARY BRANCH SECONDARY BRANCH MAIN STEM NODE 5 NODE 3 NODE I ON MAIN STEM TREATMENT : L-5-0 --'-----.;__c·-=-1: ::Ji aqrarc1a t J c i l l ustra t i on o f the appearance of the p l ants a f ter di f ferent i ntensi t i es of de fol i a t i on . 58b TREATMENT: CONTROL I H I E -5- 4 E - 5 - 0 L -5- 4 I L - 5 - 0 l lo H2 H3 H4 I H5 IH6 �r iHe IH9 I H IO r D2 r t, IH 2 H I H4 01 r r fH I IH2 H3 H4 01 02 r t H I tH3 IH2 H4 F _P!,_A�T-�·IH_ - - - - - H I 20 40 60 80 100 120 140 1 6 0 DAYS AFTER S E EDL ING EMERGENCE KEY : H • HARVEST D • DEFOLIATION Figure 2 . 2 : Harve st i ng schedules over the experimenta l per iod 5 9 D . PLANT MEASUREMENTS D . 1 Dry Matter Yield Plant dry matter yields wer e mea sured by random s ampl ing of the pl ant populations on week 0 , 4 , 6 and 1 0 a fter the f irst cut . On each harves t occas ions , 4 plants per treatment were taken and separated into leaf , stem and inf lorescence on each primary branch . Soil was washed from the root system and the tap ( including large woody root ) and f i brous roots separated . Dry matter yields wer e obtained by drying a l l samples in a vacuum oven f o r 7 2 hour s . D . 2 Lea f Area and Leaf Number Lea f area and numbe r wer e mea sured at the s ame t ime a s plant dry weight measurement . Lea f area was determined us ing the Electronic Leaf Area Metre { Model 3 1 0 0 Area Meter ) . D . 3 Branch Number Branch number was recorded on weeks 0 , 1 , 2 , 3 , 4 , 6 , 7 , 8 , 9 and 1 0 a fter the f i r s t cut . The number s on control plants were determined on days 3 9 , 6 7 , 8 1 , 8 8 , 1 1 7 and 1 3 1 after seedling emergence . A branch i s def ined in this s as a shoot with at least one l ly expanded tr i f o l iate l ea f . D . 4 The number of v i s ible growing s aris ing from the stubble of the n stem and primary branches was determi 1 0 days after each cutting . A " growing " is de fi this s tudy a s a bud f rom which a young leaf has s tarted to emerge . E . CHEMICAL MEASUREMENTS E . l Crude Protein Nitrogen concentrations in each component were mined after semi-micro K je ldahl digestion , using a Auto 1 0 3 0 Ana lyser . Total nitrogen ( including 6 0 deter­ K j el t i c nitrate nitrogen ) was determined according to K jeltic Auto 1 0 3 0 Analyse r Manual and crude protein val ues calculated ( N x 6 . 2 5 ) . E . 2 Tota l Non-struc tural Carbohydrate ( TNC ) The ana lytical procedure used for determining s oluble sugars and s tarch was that des cr ibed by Has lemore and Roughan ( 1 9 7 6 ) . The l evel s of these two constituents were summed to g ive total non - structural carbohydrates ( TNC ) . E . 3 Plant Materia l Plant material anal ysed i s s hown i n Table 2 . 1 . F . STATISTICAL ANALYSI S Data were analysed according t o the common procedure o f a randorni z ed complete b lock des ign for a l l plant characters ( Li ttle and H i l l s , 1 9 7 5 ) . The ana lys i s was done by Gens tat prograllh'tte ( Alvey et a l 1 9 7 7 ) . lea st s i gn i f icant d i f ference at the 5 % l eve l wa s used to identi f y s tati st ica l di f ferences . The symbol s u s ed to des ignate s tat i s t i ca l s igni ficance are * ( P = 0 . 0 5 ) , * * ( P = 0 . 0 1 ) and n s ( no t s ig­ n i f icant ) . 6 0a Tab l e 2 . 1 P l ant componen t s a na l y s ed T ime Chem i ca l s - -- - - - -- P l a n t compone n t s Stubble Leaves Stem I n f l or . Root ---- -- --- -- --- -- 1 • At 1 s t c u t P r ot e i n + + TNC + + 2 . 4 wks a f t e r P r ot e i n + + + + + 1 s t c u t TNC 3 . 6 wk s a f t e r Prote i n + + + + + 1 s t c u t TNC + + 4 . At 2 nd c u t P r o t e i n + + + + + TNC + + 5 . 4 wk s a f t e r Prot e i n + + + + + 2 n d c u t TNC + + ( Fi n a l h a r ve s t ) - ------ ------ U ld Plate 2 . 1 : Uncut contro l treatment at ear ly stage ( 39 days a f ter seed­ l i ng emergence ) . P l ate 2 . 2 : Immedi a te l after l a x defo l i a t ion of the pri mary branches ( E- 5 -4 ) at early stage ( 39 days after seedl i ng emergence ) P late 2 . 3 : Immedi ately after hard defo l i ation of the pr imary branches ( E- 5 -0 ) at ear ly stage ( 39 days a fter seed­ l ing emergence ) Plate 2 . 4 : Immediate ly a f ter lax de fol i at i on of the primary branches ( L-5-4 ) at late stage of growth ( 88 days a f ter seedl i ng emergence ) . Note : L-5-4 ( forward ) and uncut control ( backwards ) P late 2 . 5 : Immed i a te l y a fter hard cutt i ng o f the pr i mary branches ( L- 5 -0 ) at l ate stage of growth ( 88 days a fter seedl i ng emergence ) . 6 1 6 2 I I I RESULTS A. PHENOLOGICAL OBSERVATIONS The plants grew very s lowly dur ing the e st ab l i s hment phas e ( 0 - 4 weeks ) , a s observed by many other worker s ( e . g . Wi l a ipon and Humphreys , 1 9 8 3 ) . Dur ing that per iod the p l an t s devel oped noticeable " leaf t i p burn " e specia l ly o n the l ower l eave s , but by 5 to 6 weeks they had fu l ly recovered and thereafter devel oped healthy and rapid growth . The s tage o f " f i r st f l ower " occurred 3 5 days f rom seedl ing emergence which was cons iderably ear l ier than the f ie ld ( Wi laipon and Humphreys , 1 9 7 6 ; Skerman , 1 9 7 7 ) . With severe de foliat ion ( L- 5 - 0 ) at the late s tage o f growth , 8 5 % of the plants died , whereas with ear ly and lax defol iation there was no l ethal e f fect on plant s . B . PLANT REGROWTH B . 1 Tota l P l ant Dry Weight Total dry wei ght per plant { cut to ground l eve l ) i s presented i n Figure 2 . 3 and Appendi x 4 . The e ar ly , l ax cutting treatment ( E- 5 - 4 ) achieved a lmos t as h a produc ­ tion level a s the uncut control over the f i r s t 4 weeks o f regrowth . However , the ef fect o f the ear ly de f o l i at ion became in the subs equent 2 weeks re sulting in a s igni f icant reduction in dry matter yield over 6 weeks regrowth compared th the uncut contro l . A greater detr 1 ef fect result f rom severe cutting ( E- 5 - 0 ) of the pr imary branches was evident by the 4th week after defo l iation and the continued to the end of the 6 weeks of ear l y , c learly e f However , there was a r e lat ive improvement i n thes e severe l y f o l branches i n the 4 t o 6 week period ( Figure 2 . 3 ) . Regrowth o f the late and l ax cut treatment ( L- 5 - 4 ) was relatively s low at first in spite o f the greater amount o f 6 2a �05 <( .J Q.. 90 ..... ! 275 ....: �60 � 45 1- � 30 ..J Q.. ..J � � 1- z �05 �90 0> A . 20 B . 20 40 60 80 100 120 140 )---�I �� I l 1 I I I I I I I I I : I : t:.------6 CONTROL • • E ·5 ·4 G-----O E ·5 ·0 � DEFOLIAT ION 160 t:r------6 CONTROL •e--.e L· 5 ·4 o----o L·5 ·0 � DEFOLIAT ION A ! e !/ / i /PlANT � DEATH • • ! .I. --o .. ..L.- ..0 40 60 80 100 120 140 160 DAYS AFTER SEEDL ING EMERGENCE Figure 2 . 3 : Tota l dry we i ( cut at ground l eve l ) A . Cut a t early stage o f growth ( g/pl ant ) B . Cut at late stage of growth ( g /plant ) 6 3 res idual s tubble ( Figure 2 . 4 ) , but showed remarkably rapid growth in the las t 2 weeks ( 4 - 6 weeks ) and reac hed a yield l i ttle d i f f erent f rom the contro l treatments a nd s igni f i ­ cantl y higher than the early cut treatments ( Appendix 4 ) . The late and hard cut treatment was l ethal to mos t o f the p lants so defol iated . The d i f f erent treatment e f fects on regrowth f o l l owing the second cut s i x weeks later , were s imi l a r t o fo l lowing the f i r s t cut . I ntense and early cutting o f those the primary branches had a ma jor depres s ing effect on production whi le intense late cutting caused further plant morta l i ty . In terms of the var ious p l ant component response s - of stem , l ea f and i n f l orescence - to treatments , they genera l ly showed trends very s imi lar to total yield . The ma j or exception was the root frac t i on which showed an early depress ion in response to hard cutting but after 6 weeks this e f fect was barely evident . Regrowth in absolute terms { g/day ) i s p re sented in Figure 2 . 5 f or the two success ive g rowth period s . Dur ing the f i rst four weeks a f ter the f i r s t cut , plants in the early , lax cut treatment were superior to those in the other two cutt ing treatments and not s i gni ficant d i f ferent the uncut control . However , dur ing the subsequent 2 weeks ( 4 - 6 weeks ) , p l ants in this early , lax cut treatment showed the s lowes t growth rate , l e those i n the l a te , l ax cut treatment of the s ame nsity had the fastest rate . Four exposed weeks a f ter the second cut ( F igure 2 . 5 ) , to an ear l y hard cut o f pr ry branches plants the lowest growth rate , whi l e tho s e in the other two treatment s were s l a r in response . The data on the growth and development o f the plants ( control ) in thi s exper iment have a l ready presented in Experiment 1 . Thus , the relative which was obtained by u s ing the predicted g rowth ( Experiment 1 ) is presented here to compar e with uncut been rate model the L+ f -- � =� '::! lioO .. }; )' • 0 ... � + � 10 � I. ._.... l a V tor � --- <\ ,.. � � I I � ;;.AIILY 2S -;;;- ,. .. _, a • r- ';; u - � ,.. � � ... .... I � I Il l I I I I LA1"W. .._...._..,.. 42 DAYS AFTER DEFOLIATION � " � a .. L". I I L S I n " � ... I I l i LATE ,, a.tr 11 INFLORUC!:NCE l I I LSO !5% 0 LEAF f l I D STEM I S ) • ROOT ( R l � � 1 � I �1� ..,. � 0 ,;, .j "' "' 4.2 f ffiiiMI£f)(ArEO' 70 AFTFR I�WCI/TJ Figure 2 . 4 : Effect of stage and i ntensi ty of defo l i at i on on the components of plant dry we i ght ( g /plant ) 0" "'' [1) 63b � 4 Cl !... GJ 3 a. 0 2 w � a: A: FIRST CUT I.IJ1 0 - 28 DAYS I ..J 2 ..J 0 V 0 I a: ..J � I I 10 10 0 V 8 I I a: I I.IJ 1.1.1 t- 10 z 0 _j 0 a b EARLY LATE B : SECOND CUT 0 - 28 DAYS ( REGROWTH > V 0 I 1 10 10 I I I.IJ I.IJ EARLY V I 10 I ..J a LATE 28- 42 DAYS IJJ ..J I I ..J ..J «;t 0 0 I a: V 0 a: 10 1- I I t- I z 10 10 z ..J 0 I I 0 0 UJ UJ 0 a EARLY cd !:.ill NOTES : � BETWEEN DAY 39- 67 BETWEEN DAY 88 - 1 1 7 Figure 2 . 5 : Ef fect of stage and i ntensity of defol iation on the absolute growth rate ( g /day ) 63c � 0 & 0. E .... !. 0. E w � a: :I: 1- � 0 a: � .J IJJ a: F isure 2 . 6 : CUT I 200 1 80 160 140 A . 1 20 · 100 · 80 . 60 40 · 20 · 0 · -20 • -40 · 20 4() 200 180 · 160 · 140 B . 1 20 · 100 80 60 40 20 0 - 20 .. 40 zo 40 f EARLY STAGE : 39 DAYS ) I E ·5-0 I I I , I A: EARLY STAGE I \ I I \ I \ \ \ I I \ , \ ,cuT ,' . \ 2 , \ \ f E-5-4 60 60 \ I , \ �,' I 80 100 120 140 8 : LATE STAGE CUT I ( LAT� t,"' 80 100 CUT 2 l ,. ... .. L·!r4 , .. ... , ... ... STAGE - 88 DAYS ) 120 140 DAYS AFTER SEEDLING EMERGENCE E f f ect of s tage and i ntens i t y o f defol iati on the relat ive grwoth rate ( mg/mg/day ) on 6 4 defol iated treatments at the s ame time ( Fi gure 2 . 6 ) . P lants in defol iated treatment s genera l ly had higher r e lative growth rates than contro l plants , particularly with early , hard cutting . Initial l y , defoliation reduced r e la tive growth rates in the early cut treatments compared with the contro l , but after 5 0 days they general ly decl ined l e s s rapidly than the control , particularly the ear ly hard cut treatment , ( E- 5 - 0 ) through t o the second cut ( Figure 2 . 6A ) . I n contrast , the l ate lax cut treatment ( L- 5 - 4 ) s howed a marked increase i n relative growth rate over the six weeks of regrowth f o l lowing the f ir s t defoliation ( F igure 2 . 6 B l � Fol lowing the s econd cut , the early lax and early hard cut treatment produced a s ubstantial increase relative growth rate ( Figure 2 . 6A ) . I n contras t , p lants in the late and lax growth s imi lar expected 2 . 6 8 ) . cut treatment showed a s l i ght dec l ine i n r e lative rate after the second cut , a lthough the l evel was to that i n the ear l y , lax cut treatment and a s much higher than that o f the uncut control ( Figure Net regrowth yields over the exper imental per iod a re presented in Figure 2 . 7 . There wa s no s igni f icant dif f erence net regrowth of leaf , s t em or lores cence for the two stages of �efol iation at the same intens ity ( E- 5 - 4 - vs L- 5 - 4 ) . However , both intens e cutting treatments s i f icantly s s ed the yield of al l components . B . 2 Branch Number The marked e f fects of the s tages and intens i t i e s o f cutting on 2 . 2 . Lax the branch number per t are shown and e specially hard de fol iat ion caused in a s igni f icant depre s s branch number . Howeve r , the rap e ly bui ld up in number of branches , fol l owi early l ax cutting , over the f i r st 4 weeks and especial l y lax cutting over the 4 - 6 weeks per i od , was very l owing late The more rapid increa se in branch number in ear ly l ax cut treatment during the f i r st four weeks of regrowth , 64a 1 00 - 80 t- z - :r t- � 0 Q: (!) 20 w Q: t- w z v 0 v I I l 10 10 10 I & I w w ..J a A B A 0 I 10 I ..J :::r t- w .:_, 0 .;, � Fi aure 2 . 8 : Ef fect o f stage and intensity of defol i a t i on on rate of branching ( no/day) 6 5b Ta E f f e c t o f s t age and i n t en s i ty d e f o l i a t i on on l e a f a r e a ( cm2 ) ( LA ) and l e a f n umb e r ( LNO ) per plant . T r e a tmen t E a r l y Con t ro l - E ( Da y 3 9 ) E- 5 - 4 E- 5 - 0 Late Contr o l - L ( Day 8 8 ) L - 5 - 4 L - 5 - 0 S i g . Ear l y E- 5 - 4 E - 5 - 0 Late L- 5 - 4 L- 5 - 0 S i g . Immed i a t e l y a f t e r 1 s t c u t LNO LA 1 0 7 b 1 3 3 7 b 6 4 b 7 c 1 8 6 c 2 4 d 1 4 2 4 a 2 9 9 6 a 6 c 4 c * * 5 d Sd * * Immed i at e l y a f t e r 2 nd c u t LNO 3 a S a 1 2 a n s LA 3a 8 a 2 9 a n s 4 weeks 6 wee k s a f t e r 1 s t c u t a f ter 1 st c u t LNO LA 9 8 4 b 2 3 6 1 b 8 4 8 b 2 0 4 1 b 3 2 9 c 8 6 2 c 1 9 2 3 a 3 0 7 7 a 5 0 2 c 1 1 9 2 c * * * * 4 weeks a f te r 2 nd c u t LNO LA 9 0 7 a 1 9 7 5 a 2 8 1 b 7 4 9 b 9 3 2 a 2 0 2 6 a * * * * LNO LA 1 4 2 0 ab 3 0 9 4 a 9 6 5 b c 2 2 2 a 7 7 5 c 2 1 2 3 a 1 6 4 2 a 2 1 2 7 a 1 3 7 1 ab 3 4 9 7 a * * n s 1 va l u e s i n the s ame v e r t i ca l c o l um n n o t f o l lowed by the s ame l et t e r d i f fe r at P = 0 . 0 5 6 6 early , hard cut p lants remained s igni f icantl y depres s ed compared with mos t treatments ( Table 2 . 3 ) . B . 4 Number o f Growing Points The number o f growing points on the s tubbl e a re s hown in Table 2 . 4 . Mos t of the growing points were l oc ated on the pr imary branch . ei ther a t early Thu s , hard cut ting of the pr imary branche s , or late s tages of growth , greatly and s igni f icantly reduced growing point numbers . C . CHEMICAL COMPOS ITION C . 1 Crude Protein Crude protein concentrations are presented in Tabl e 2 . 5 . All cutt ing treatments caused a greater increa se i n crude protein than the uncut control for most o f the p lant component s . I n contrast , among the cutting treatments , the differences in c rude protein were smal l , except in the residual s tubb le and roots f o l lowing the f i r s t defol iation . However , there were large d i f ferences in the crude protein l eve l s between p lant parts . Crude protein conce ntration was highes t i n the l ea f , a l s o h igh in inflorescences , but much lower in the s tem and part icul arly the tap root . Apart from the exception mentioned above , c rude protein concentrations in the s le ( ma inly s tem ) were cons istent ly l ow . When converted to a protein yield bas i s , the f ferences between treatments were h l y s i f icant and f o l l owed very much the plant we responses presented ear l ier , i . e . prote in y ie ld s were most s ed when both the ma s tem and the primary branches wer e defol iated ( Figure 2 . 9 ) . The effects of s tage of cutting were s hown onl y a fter 4 regrowth fo l lowing the f i r s t defo l iation . The crude protein l d s o f the plant simi l ar ly a f f ected by the d i f ferent intens it i e s were defo l i a- tion a nd s tage of growth a s for component dry weights ( Figure 2 . 9 ) . H owever , it is noteworthy that the y i e ld of crude 66a Table 2 . 4 feet of stage and inten s ity of def iat number of " growing point s " on the stubble 1 0 a f te r cutting ( no/plant ) on Treatment E-5 - 4 E- 5 - 0 L- 5- 4 L- 5 - 0 S i g . F i r s t c u t Main s tem 6 . 3a 1 6 . 0a 6 . 3a 2 n s P r imary branch 2 7 . 5a 1 4 . 5b * Total 3 3 . 8a 6 . 0 c 2 0 . 8b * * cut Main s tem 6 . 0a 5 . 7a 5 . 7a ns P r imary branch 1 6 . 3 a 2 0 . 0a ns Total 2 2 . 3a 5 . 7b 2 5 . 7a * * Values in the same hor i z ontal l ine not followed by the same l etter d i f f e r at p = 0 . 0 5 . 2 Plant death 66b �a b l e 2 . S E f f ect of stage and intens i t y o f defoliat i on on crude protei n concentration I % of d r y matter ) in leaf L , s tem ( S ) , i n f lorescence ( 1 ) and root ( tap ( T l and f i brous ( F ) ) components and i n the residua l stubble ( R ) . Trea tments Control - E E- 5 - 4 E-5-0 Con trol-L L- 5 - 4 L- 5 - 0 { 39 days ) ( 88 days ) ------- R 5 . 3b 5 . 6b T + F' 2 1 . 4 a 2 1 . S a 1 0 . 1 b 8 . 9b R 6 . 5a 7 . l a 6 . l a L 2 9 . 4 a 2 9 . 4a 3 0 . 7a 2 5 . 1 b 2 9 . 9a s 1 2 .. 8c 1 5 . 9b 1 6 . 5b 1 1 . Se 1 8 . 2a 1 2 3 . 2 5 . 7 a 2 5 . 8a 2 2 . 5 b 2 5 . 2a T 8 . 0a B . Oa 7 . 3a 6 . 7a B . 1 a F 1 4 . 0a 1 4 . 6 a 1 4 . 8a 1 2 . 0a 1 4 . 8a R 6 . l a 6 . 8a 6 . 2a L 2 6 . 2b 2 7 . Ba 2 8 . 6a 2 6 . 2b 2 8 . 6a s 9 . 5c 1 1 . 4b 1 3 . 8a 7 . 5d 1 3 . 8a I 2 2 . 7 b 2 3 . 4b 2 5 . 2a 2 5 . 3a 2 2 . 5b T 7 . 6a 9 . 0a 7 . 2a 8 . 3a 6 . 9a F 1 4 . 8 1 4 . 7 1 6 . 6 1 3 . 7 1 5 . 6 B . Second cut Trea tments E-5-4 E-5-0 L- 5-4 L-5 -0 S ig . Immed iately a f ter second cut R 6 . 1 a 6 . 8a 6 . 2a n s T 9 . 0a 7 . 2a 6 . 9a ns F 1 4 . 7a 1 6 . 6a 1 5 . 6a n s R 6 . 8a 6 . 6a 7 . 5a ns L 3 1 . 4a 30 . 6a 3 1 . l a ns s 1 5 . 9b 1 7 . 2a 1 5 . 5b * I 2 4 . 0a 2 3 . 5a 2 3 . 7a n s T 7 . 2a 7 . 8a S . Oa ns F 1 4 . 5a 1 6 . 4a 1 6 . 7a ns Values in the same horizontal line not followed by the same letter differ at P = 0 . 0 5 S i g . ,. . * * n s * * * * * * n s ns ns * * * * * n s n s ,_ , z - 10 a:: <( (!] ::::> 0 (/) 1 0 67b TOP ( STUBBLE } SECOND C U T A F T E R FOUR F I R S T C U T ( 00 ) S E CO N D CUT ( Do ) W E E K S ( D 2 8 ) "t 0 'lt 0 V 0 V I I I I I I 10 10 10 I{) I{) I{) I{) I I I w I I I w w ...J b lLl ...J ...J c a a b a 0 0 0 a Roots ( TAP PLUS F I BROUS a ROOTS ) Figur2 2 . 1 1 : Effect of stage and i ntensity o f defol iation on sugar y ie ld in the residual top ( stubble ) and be low ground ( mg /p lant ) D . RELATION SH IP BETWEEN REGROWTH AND RES IDUAL ( STUBBLE ) PLANT VARIABLES 6 8 The r e lationship between regrowth yield and re s idual plant var i ab le s is presented ln Tabl e 2 . 6 . S ignif icant pos i t ive correlation was found for both cutting treatments . Fol lowing the f ir s t cut , the relationships of regrowth yield to res idual var iable o f branch number , l eaf a rea , leaf number , s t ubbl e to root rat io , t he number of growing points and the amount o f TNC in s tubble , were pos i tive and s igni f icant during the f i r s t four weeks of regrowth . When examined for the total 6 week regrowth period , h owever , no s igni f icant re lationships for any of the variab le s with y ield were recorded . Fol l owing the s econd cutting , the relationships between regrowth y i e l d and the res idual vari ables of s tubble dry weight , the number of growing points , s tubble to root ratio , and the amount of non-structural carbohydrate were pos it ive and s igni f icant . E . RELAT IONSHIP BETWEEN TOTAL PLANT DRY WEIGHT AND GROWTH PARAMETERS ( BRANCH NUMBER , LEAF NUMBER AND LEAF AREA PER PLANT . High l y s igni f icant and pos i tive correlations were f between total plant dry weight and the ma growth parameters { branch number , l eaf number and leaf area ) when compared at the 4 and 6 week sampl ing time s , s uggesting that the s e para­ meters are important in determin ing f inal y ld ( Table 2 . 7 ) . The s e results are in contr a s t to the lack of correlation between thes e parameters mea su red immediately a fter the second cut and the s ubsequent y ie ld , a s shown e 2 . 6 . It appea r s that repeated cutt ing tends to s i z e the importance of other parameters , s uch as the number of growing po ts rema i ng a fter cutting , the s tubble to root ratio , the res idual dry weight and par ticularly the res idual amount of carbohydrate in stimulating ear l y recovery - and hence enable the development of the se ma j or contributor s to subsequent y ield , l ea f area and branch number . 68a . 6 C o r r e l a t i o n s o f r e s i du a l p l ant var i ab l e s w i t h net ...;._c....;._c....;._c._.;:;.._;__ r e g r owth y i e l d . Res i du a l p l a n t 4 wks a f ter 6 wks af ter 4 wk s a f ter var i ab l e s 1 s t c u t 1 s t c u t 2 nd cut Branch numbe r 0 . 8 5 2 * * - 0 . 0 5 5n s 0 . 4 1 0 ns Lea f 2 0 . 8 8 1 * * - 0 . 0 6 7 n s 0 . 2 4 7 ns a r e a ( cm / p l a n t ) L e a f n umb e r ( no/pl ant ) 0 . 8 3 4 * * 0 . 0 3 3 n s 0 . 2 6 2 n s Grow i n g po i n t ( no/p l a nt ) 1 0 . 8 9 5 * * 0 . 3 5 0 n s 0 . 9 1 1 * * S t u bb l e to r oo t r a t i o 0 . 9 0 0 * * 0 . 0 5 6 n s 0 . 7 6 3 * Stubbl e dry we i gh t ( g /pl an t ) 0 . 0 5 4 ns 0 . 6 6 3 n s 0 . 8 0 6 * * Root dry w e i gh t - 0 . 3 0 4 n s 0 . 5 4 1 n s 0 . 6 0 9 ns % TNC i n e 0 . 5 4 1 ns - 0 . 4 1 7 n s 0 . 4 5 3 n s % TNC i n r o o t s 0 . 5 2 7 ns - 0 . 1 0 0n s - 0 . 1 8 6 ns TNC y i e l d s tubbl e ( mg/p l a n t ) 0 . 9 1 2 * * - 0 . 1 4 2n s 0 . 9 3 5 * * TNC y i e l d r o o t s ( mg/pl a n t ) - 0 . 3 8 n s 0 . 1 4 6n s 0 . 8 2 0 * * 1 Det e rmi n e d 1 0 d a y s a fter c u t t i n g 68b e 2 . 7 near correlation coe f f ic ients between p lant dry wei ght ( DM ) and main growth parameters ( branch ( B ) , leaf number ( LNO ) and leaf area ( LA ) ) 4 weeks a fter f i r s t cut 6 weeks a fter f i r s t cut 4 weeks after second cut Parameter s of growth B 0 . 9 6 0 * * 0 . 8 4 7 * * 0 . 9 4 6 * * LNO 0 . 9 8 0 * * 0 . 9 0 7 * * 0 . 9 9 3 * * LA 0 . 9 3 6 * * 0 . 6 8 8 * * 0 . 9 7 6 * * 6 9 IV DI SCUS S I ON S ignificant p lant mortal ity a f ter l ate and intense cutting has been reported in many tropical legumes such a s Stylosanthes h umi l i s ( Fi sher , 1 9 7 3 ) , Crotal a r ia j unci a ( Ke s s ler and Shel ton , 1 9 8 0 ) . Thi s a l so occurred i n the present experiment ( L- 5 - 0 ) . Howeve r , intens e defol iation provided it i s done at an ear ly s tage o f growth ( E- 5 - 0 ) i s not lethal but does result in a s ignificant del ay i n regrowth . S imi l a r ly late defoliat ion , provided i t i s lax ( L- 5- 4 ) , wi l l not - ki l l plants but wil l cause an appreciab le delay i n recovery . In contrast ear l y lax defol i at ion r e su lt s i n rapid recovery al though a s igni f ic ant reduct i on in y ie ld sti l l occurs compared with no defo l iation . Factor s which are known to a f fect the rate o f in itial regrowth a fter defol iation are u s ual ly related to : 1 ) The carbohydrate reserves above and be l ow ground . 2 ) The number of branches capable of regrowth . 3 ) The s i z e o f res idual leaf area . 4 ) The number of growing point s . The importance o f carbohydrate reserves to r egrowth has been recogni s ed by many workers ( e . g . Yamada , 1 9 7 5 ; Hurnphreys , 1 9 7 8 b ) . These authors sugges ted that reserve organic regrowth compounds are uti l i zed dur ing the early and that later regrowth is dependent on s tages of l ea f area and photosynthes i s . The rna reserves playing thi s r o le are suggested to be s ugars and starch which are s tored ma in the s tern base and roots ( Yamada , 1 9 7 5 ) . In thi s exper iment , the TNC ( suga r s and s tarch ) concentrations were relat low in both stubble and roots i n a l l treatments , and showed no d i f f er ence between cutting treatments at a s l a r stage of growth . However , despite the low concentrat of TNC , the amount s ( mainly s ugar ) were higher under lax cutting than under s evere cutting of the p r imary branches ( F igure 2 . 1 1 ) s umably ref lecting the greater s tubbl e y ie ld i n former treatment . 7 0 A s ignificant positive corre lation w a s a l so found between regrowth y ie ld and the amount of TNC in the s tubble , s uggesting that the init ia l regrowth a fter cutting wa s dependent upon the amount of these reserves i n the s tubbl e ( Table 2 . 6 ) . Thi s h i gh l ights the need t o retain a greater s i z e of primary branch after defoliation and thereby provide a larger supply o f TNC for regrowth - and hence compensate for the l ow TNC concentration in the s tubble and root s . Thi s a l s o may explain why Verano stylo was able to recover under late l ax cutting d e s pite lack of res idual photosynthet ic t i s sue . Low amounts o f res idual TNC under severe cutting o f t h e pr imary branch , and low res idual leaf area due t o natural f a l ling of the l ower leaves , may have accounted for the mortal ity of thes e plants soon after the f i r st cut . The number of branches capable of regrowth after cutting was a l s o found to be pos itively related to regrowth in s ome tropical legumes ( Ke s s ler and She lton , 1 9 8 0 ) . The resul t s o f thi s s tudy support thi s finding . In the ear ly cut treatment ( E- 5 - 4 ) , many branches capable of regrowth had a l ready e longated on the pr imary branches , whi l e i n the later cut t reatment , the branches capable of regrowth were l ow i n number immediatel y a fter cutting ( Table 2 . 2 ) . Ther efore , s l ow initial regrowth was observed in the s e late cut treatments due to s l ow bui ld up in branch production . The importance o f res idual leaf area and the number o f growing points are s upported results obta with other tropical legumes a s i ndicated by gh corre lat ion between regrowth yield and the s e parameters ( Grof e t a l , 1 9 7 0 ; Jones , 1 9 7 4 ; Kes s ler and Shelton , 1 9 8 0 ; ow and Char s - Edwards , 1 9 8 0 ) . I t i s a l s o relevant that plants whi ch were cut a t late r s tages o f g rowth had lost a l l their l ower l eaves t hrough natural l e a f f a l l and were therefore lacking in photosynthetic t i s sue . The number o f growing points was a l s o s igni f i cantly r under these treatments ( E- 5 - 0 and L - 5 - 0 ) . With the marked reduction in TNC amounts , a s d i s c u s s earl ier , plants h ad s low regrowth and many s oon d i ed , resulting from the photosynthate be ing insu f f ic i ent to mee t the demands for r e spiration and growth ( Yamada , 1 9 7 5 ) . Thus 7 1 the high correlation obtai ned between r egrowth y ield and s tubbl e to roots rati o was not s urpri s ing s ince defol iation is known to a f fect root g rowth ( Ke s s ler and Shelton , 1 9 8 0 ) and death of root has frequently been reported ( Bowen , 1 9 5 9 ; Whi teman , 1 9 7 0 ; Whiteman and Lulham , 1 9 7 0 ) . Furthermore , re stricted death may plant s . s uppl ies o f water and nutrients because o f a l s o account f or s l ow regrowth and death of root the From the above res u l ts , i t appea r s that regrowth of Verano stylo fol lowing de f ol iation i s dependent on the s i z e o f pr imary branches a s soc iated with s everal r e s idual p lant characteristics v i z . level of carbohydrate reserves the s tubble and root s , the res idual number of branches , the res idual number of growing points and the res idual leaf area . The carbohydrate reserves although l ow , appea red to be o f s ome importance in thi s r espect as d i d the number of growing poi nt s , whi l e residual l eaf and branch number were only important after the f i r s t cutting . I n contras t the total non-structural carbohydrate contained in the roots was only of value when the plants were older and bigge r . The response to cutting in terms of total plant dry weight was a l s o related to the above factors as d i scus s ed earl r . Plant dry we ight was not reduced under 1 t cutt ing . Thi s res e was ma ly through increase ln s tem , i n f l orescence , and to a l es ser extent , increa se in leaf dry we i ght ( F igure 2 . 4 ) . The result of th i s demon­ s trated that a l l components of dry matter yield can a f fected by defol iation , the s tage at wh ich the plant is cut , and the sever ity of cutting . The se results are c ontrary to the resu l t s of l a and Humphreys ( 1 9 7 6 , 1 9 8 1 ) who reported that Verano s l o y i e ld s were reduced when gra z ing or cutting was de layed to late i n the growing s ea son . However , t he reduction yie ld s in s uch an experiment appeared t o be related to c l i c condi t ions a s we l l a s t o cutting . S the cutting t was near the end o f the growing season , temperature and moi s ture may have been restricting regrowth . I t s eems that lax 7 2 cutting o r graz ing a t l ater s tages o f growth , a l though caus ing s low initial recovery , wi l l not reduce regrowth ability under f avourable c ondi t i ons . Plant dry weight was a l s o related to the re sponse o f branch number t o defoliation a s ind icated by the s igni f i cant correlation between branch number and total dry weigh t . Branch production was markedly reduced when the primary branches were cut hard , particul ar ly when cutting o ccurred at the late s tage o f development . Thi s reduction 1n branch number corre sponded to a reduction in dry matter yield . However , lax cutting o f the primary branches at an ear l y stage of growth rapidly increas ed branch production and hence the potential to achieve a higher p lant dry weight than hard cutting of the pr imary branche s . S imilar e f fe c t s were a l s o found in another tropical l e gume , Townsvi l l e stylo ( Fi sher , 1 9 7 3 ) and in the temperate l egume Tri fol ium s ubterraneum ( Ross iter , 1 9 6 1 ) . An increase in b ranch number was due to an increase in branching rate under lax cutting . The present experiment a l so showed that lax cutting at the late stage o f growth s t imulated branch product ion over the six weeks of r egrowth . The initia l s low development o f branches over the first four weeks was more than compensated for by the explos ive development f rom the fourth to s ixth Although not stat i s t ic a l ly s igni f icant , l eaf number and l ea f area showed s lar Thi s i s interesting find s ince it has a l so been s a mos t that increa sed branch deve can result in i nc reased s eed produc t i on ( Ros s iter , 1 9 6 1 ; sher , 1 9 7 3 ; W i la and Humphreys , 1 9 7 6 ) . Thi s may explain in part why Verano stylo can produce a s imi lar dry matter yield to the c ontrol even under l ate defol iation . I t suggests that when ng i s nece s sari l y late , a per iod o f a t l e a s t 6 weeks s hould be l e f t t o a l l ow the plant t o produce new branches t o c ompensate . The number o f l eaves and leaf area wer e a l s o by defol iation . The complete removal of the prima ry s especi a l l y at the later s tage of growth caus ed the s l ow ld up in both number and area o f leaves and hence a reduction in 7 3 growth rate , y i e l d and a l so death o f plant s . I n contra s t , lax cutt ing of the p rimary branches e spec ia l l y at the ear l y s tage of growth enab led a rapid recovery of both number and area of l eaves during the f i r s t four weeks - thi s recovery occurring 4 weeks l ater ( during the 4 th to 6 th week ) in the l a t e l ax cut treatment . The r e s u l t s o f this s tudy demonstrate that Verano s ty lo has a high qua l ity in terms of nutritive value for an imal feeding . There was no s igni f icant ef fect of s tage o f growth and intensity of cutting on the crude protein content in the regrowth components . These l evel s of protein were h igher than the c r i t ical l eve l suggested by Mi l ford and Minson { 1 9 6 6 ) at a l l s tages of growth . Norman and Phi l l i p ( 1 9 7 0 ) have shown that animal production i s proportional to the nitrogen content o f the pasture . Thus , animal production from pasture containing th i s legume should b e capab le of achieving a r elative l y high l evel of per formance ( Gi l l a rd et a l , 1 9 8 0 i G i l l a rd , 1 9 8 3 ) . As expected , l eaf and inf l ore s cence were higher i n crude protein content than the s tem f raction at a l l s tages of growth . The s e crude protein l evel s decreas ed w ith prolonged regrowth condition s , but the decrea s e was le s s rapid i n l ea f and i n f lore s cence f ractions than in the s t em f raction . This 1 9 6 9 ) , ( Lee 1 9 8 2 } . has been f ound in Townsvi l le s t y lo ( Fi sher , S ty l o santhe s ( Mufandaed za , 1 9 7 6 ) , Lucerne and th , Verano s tylo ( e t a l , Although the c rude protein content was not g reatly a f fected by cutting , there was a marked difference protein y i e l d between treatments . Hard cutt pr imary branches produced the l owes t crude pr reflected the l ow dry matter yie lds recorded . o f the y i e and Thi s i s of some importance in terms of pasture management a s the l of animal s i s governed to a l arge extent by the amount of herbage pre s ent ( Stobb , 1 9 7 4 ) . 7 4 Total crude protein y ield in l ea f plus inf l orescence i n the l ax cutting , f or both stages of growth , was h igh when compared with s evere cutting ( E- 5 - 0 ) . Observat i ons in the f ield i nd icate that the i n f lorescence a s wel l as the leaf i s read i l y acceptabl e t o the anima l , even when the plants are mature ( Gardener , 1 9 8 0 ) , and therefore an important source o f protein . The leve l s o f non-structural carbohydrates ( suga r and s tarch ) were low and did not increase a s the t ime of cutting • was · delayed . Thi s i s unl ike the response obtained with l ucerne , where advancing stage o f growth great ly i ncreas ed the avai lable carbohydrate l evel ( Ne l s on and Smith , 1 9 6 8 ; 1 9 6 9 ) . With tropi cal legume s , Hunter et a l { 1 9 7 0 ) r eported that Glycine wighti i ( lea f ) , Greenleaf desrnodi um { le a f ) and S i lverleaf desrnodium ( stern ) contained negl igible quanti t i e s of s tarch at an early s tage o f devel opment , but the mor e mature plants { at f lowering ) contained 1 - 2 % starch i n the dry matter . Jones ( 1 9 7 4 ) reported that l ong cutting interval s ( 1 6 weeks ) did not increase the carbohydrate in the roots of Siratro . However , thes e l evel s of c arbohydrates reported were much h igher than those in the present e xper i ­ ment . Unfortunately there a r e no s peci f ic data ava i l ab le o n the concentration of carbohydrates i n Verano stylo w i th which to compare the present results . However , the present data do indicate that mos t o f the photos s were ut i l i z for g rowth and not accumulated in the root and crown o f the plant . Ne lson and th ( 1 9 6 8 ) a l so showed that in B irds foot trefo i l , a temperate legume , the total non-structural carbo­ ate was l ow duri the growi s eason , and that mos t of the photosynthate was used for top growth . 7 5 EXPERIMENT 3 : I NTENSITY OF DEFOLIATI ON I . INTRODUCTION on I ntens ity yield and of cutting or gra z ing can have a ma j or e f fect survival of the l egume component o f the pastur e , e s pecial l y when carbohydrate reserves are l ow f r om frequent cutting or o ther cause s . Lax cutting can l eave a greater photosynthetic sur face and provide energy for initial regrowth a fter cutti ng , by compar i s on with c l o s e cutting ( Smi th and Ne lson , 1 9 6 7 ; Jones , 1 9 7 4 ; Kes s le r and She l ton , 1 9 8 0 ; Grof et al , 1 9 7 0 ) . Experiment 2 clear ly showed that s evere cutting of the pr imary branche s , particularly at a late stage of growth , greatly reduced yield and even c au sed s igni f icant mortal ity of Verano styl o . However , lax cutting of the pr imary branches , at either early or late s tages o f growth , was advantageous i n terms o f d ry matter yield and p lant s urvival . Carbohydrate ana l y s i s indicated that thi s spec ie s contained very low l evels i n the res idual both above and b e l ow ground . H owever , with components the l a rger residual from l ax cutting , the greater amount o f carbohydrate therefore avai lable was cons idered to b e important f o r initial regrowth , particularly fol l owing late de l iation . I n Exper iment 2 , only two cutting cyc l e s were stud and the long term e f f ects of s evere and lax cutti d ned . wer e not The present e xper was conducted to in f ur ther information on the response of Verano s l o to d i f f erent cutting intens i e s in terms of regrowth abil ity and eh cal compos ition over 4 r egrowth cycle s . I I . MATERIAL AND METHODS A . ENVIRONMENTAL CONDITIONS AND PLANTING PROCEDURES 7 6 Environmental conditions and planting procedures wer e the s ame a s reported for Chapter 3 . After thinning to one plant per pot , al l plants wer e blocked into thre e rep l icates based on plant uni formity . After cutting , pots conta ining plants undergoing simi l ar treatments were located together s o that interference between treatments was minimi sed duri ng regrowth . B . TREATMENTS There were five cutti ng i ntens ities as fol l ows : Treatment Detai led 1 . E- 7- 4 : Cut the main s tem above node 7 ( be tween nodes 7 and 8 ) and primary branches above node 4 ( between nodes 4 and 5 a long the branche s ) 2 . E-7 - 0 : Cut the main stem above node 7 ( be tween nodes 7 and 8 ) and primary branches immediate ly below node 1 ( a long the branch ) 3 . E- 3 - 4 : Cut the s tem above node 3 ( 3 and 4 } and pr branche s tween node s node 4 ( between nodes 4 and 5 along the b ranch ) 4 . E- 3 -0 : Cut main stem above node 3 ( be tween node s 3 and 4 ) and p r ry branches below node 1 ( a long the branch ) 5 . Control : Non-defo l iated plant ately 7 7 C . NUMBER O F CUTTING OCCASIONS The f i r s t cutting occurred when 5 0 % of the p lant popula­ tion commenced f l owering and was repeated a f ter s ix weeks ' regrowth . Cutting was repeated 4 times over approximately 2 0 0 days giving four regrowth cycl e s ( Figure 3 . 1 ) . D . HARVESTING SCHEDULE Nine harves t occa s ions were taken as shown in Figure 3 . 1 . E . PLANT MEASUREMENTS E . 1 P lant Dry Weight Meas urements o f dry matter yields and other phenological observations were obtained f rom 3 plants per treatment . Times o f harve st s are shown i n Figure 3 . 1 . P l ant s were separated for meas urements as described in Exper 1 . Dry matter yields were obtained by drying a l l s amples in a vacuum oven for 7 2 hour s . E . 2 Leaf Area and Leaf Number Leaf area ( cm2 per plant ) and leaf number wer e recorded at the same as dry matter yield was determined . Leaf area wa s mea sured u s ing E troni c Leaf Area �eter ( Mode l 3 1 0 0 Area Meter ) . E . 3 Branch Branc h number in each regrowth cycle was recorded a t weekly interva l s up to 4 week s of regrowth and a t 6 week s iately before cutting was repeated . E . 4 Number o f Growing Points number of v i s ib le growing points ar i s i ng from stubble o f the main s tem and primary branches was recorded af ter ten days of regrowth in e ach cyc le . PH��t; I : ESTABLISHMENT UP TO !50-t. FLOWERING EMERGENCE 0 0 : DEFOLIATION H : HARVEST CYCLE I . D J ....... � 0� 0- 25 J: 200 1- � � (.!) w � 1- LIJ z _J 1 75 1 50 1 25 1 00 g 75 � A : MAIN STEM 8 : PRIMARY BRANCHES TOTAL ** L * * s * * * * I N FLORESCENCE ( I ) � LEAF { L ) D STEM ( S ) TOTA L ** L ** s * * I * * 0�------_. __ .__. __________________ __ 4 0 Figure 3 . 2 : Main effect of de fol iating the mai n stem ( A ) and the primary branches ( B ) on tota l net regrowth yield over the four regrowth cycles � CUT I e� eo ...J Cl. 7 01 CYCLE t !_ ( 42 DAYS ) ...J � � 0 14 28 CUT 2 42 CYCLE 2 J CYCL E 3 _:f CYCLE 4 f- ( 42 DAYS ) ( 42 DAYS ) ( 52 DAYS ) I 56 70 98 1 12 126 140 DAYS AFTER FIRST CUT Figure 3 . 3 : E ffect of defol iat i on i ntens i ty on tota l pl ant dry weight at var i ous cutting cycles ( g/plant ) 182 (l) 0 () CYCLE 1 30001 O - 28 DAYS A: 2500r- , : E-7- 4 .... 2000 2: E- 7- 0 � 0 l 0 1 5001- E 1 000� a:: ci I Figure 3 . 4 : Effect of de fol iation intensity on absol ute { A ) and re lative { B ) growth rate at various regrowth cycles . (X) 0 0.. 8 1 cutting , as shown in Figure 3 . 5 and 3 . 6 . The more depress ing effect of primary branch removal compared with main stem removal on a l l components including roots ; the initial slow rate of regrowth of all these components over the first 2 to 4 weeks fol lowed by rapid regrowth ; and the increasing proportion of stem developing over the experimental period , are a l l worthy of note . B . 2 Branch Number The ef fects of defoliation on branch number per plant are presented in Figure 3 . 7 . In the regrowth period of 6 weeks fol lowing the 1 st cut , lax defoliation of both the main stem and primary branches ( E-7- 4 ) did not a f f ect branch development ( number of branches ) compared with the uncut control . However , more intense defoliation , particularly of the primary branches , signi ficantly depressed branch numbers . Of particular note was the relatively s low recovery , particularly of those plants suf fering severe primary branch removal ( E- 7 - 0 and E- 3- 0 ) during the first 28 days fol lowed by a general and substantial increase in branch numbers during the fol lowing 1 4 days ( 2 8 - 4 2 days ) - and evident in all regrowth cycles . B . 3 Lea f Area and Number Leaf area and leaf number per plant between treatments fol lowed very s imilar patterns of development after each defoliation , with lax defoliation of the main stem and especially the primary branches encouraging s igni ficantly more leaf area and leaf numbers than intense defol iation ( Figure 3 . 8 ) . Although the res idual leaf areas following the 1 st , 2 nd and 3 rd cut were signi f icantly dif ferent , reflecting the defoliation intensities , these differences were sma l l in actual area and number ( Table 3 . 3 ) . Nevertheles s , the regrowth achieved over the 1 s t , 2 nd and 3 rd regrowth cycles did tend to ref lect the res idual leaf area per treatment and this may have contr ibuted to such dif ferences . However , in Bla - ..... z <{ ...J CL .... Q) 0. 200 0 - 0 _J w >- I ..... � 0 er (.!) w er ..... w z ...J � 0 ..... 100 2 3 4 TREATMENT ::r: I I I L S I T LSD ( P = 0 · 05 ) KEY : - I NFLORESC ENCE 0 LEAF ( L ) D STEM ( S ) T TOTAL I : E-7- 4 2 : E- 7 - 0 3 : E-3-4 4 : E -3-0 Figure 3 . 5 : Effect of defol iation intensi ty on tota l net regrowth yi eld over the four regrowth cycles ( g/plant ) . Blb r-z 20 Cl. 0 r- I <..? 40 w 3: r 30 � 20 r-z --o £ - 3-0 ,..__... cONTROL IfCOHTAOLI DEATH IC'OHTROLI • Effect of defoliation intensity on branch number per plant . I A: LEAF AREA C cm per PLANT ) CYCLE 1 CYCLE 2 �-_;;C....;_Y....;_C,;;;;,;LE 3 __ _ CYCLE 4 __ I <[ E-3-4 E-7-0 I I � 1000 <[ lL � ..J 0� £:� � � .� I 8 : LEAF NUMBER per PLANT L.S. D. C 0 · 0 5 ) a: � :E ::> z � LIJ ..J DAYS AFTER F I RST CUT Figure 3 . 8 : Ef fect of defoliation intensity on leaf area ( cm 2 ) and leaf number per plant at all regrowth cycles . I 8 00 f-' Q, Ble Table 3 . 3 Effect of defoliation intensity on residual leaf area (LA) ( an2 ) and leaf number ( LNO) per plant remaining immediately after each cut . Treatrrent Cycle 1 Cycle 2 Cycle 3 Cycle 4 LA LNO LA LOO LA LNO LA LNO E-7-4 1 97a 1 68a 53a 37a 1 9a 34a E-7-0 24c 1 1 c 1 3b 1 0c Oc Ob E-3-4 1 43b 54b 25b 24b 6b 1 0b E-3-0 1 3c 1 0c 4b 6c 1 c 4b Sig . ** * * * * ** ** ** Values in the same vertical column not followed by the same letter differ at P = 0 . 05 . 8 2 the 4th cycle , s imilar and s ignificant treatment differences in regrowth occurred ( Table 3 . 2 ) but purely from stem residual s ( Table 3 . 3 ) B . 4 Number of Growing Points As shown in Table 3 . 4 1 for a l l regrowth cycles , there was a marked effect of cutting on the number of growing points . Defol iation of the primary branches caused a much greater reduction in growing points than defoliation of the main stern . In fact , once the s i ze of the primary branches had been severely reduced ( i . e . from 4 to 0 nodes ) , the more intense defol iation of the main stern ( i . e . from 7 to 3 nodes ) made no further impact on the number of growing points present . However , the most severe defoliation treatment ( E- 3-0 ) did lead to total plant death by the 4th regrowth cycle . It was clearly evident from the results that the ma jority of growing points were on the primary branches and hence signi ficantly a f fected by defoliation . C . CHEMICAL COMPOSITION C . 1 Crude Protein As shown in Table 3 . 5 , for the two cycles analysed ( 1 st and 4th ) , there was no signi ficant effect of intensity of defoliation on the crude protein concentration of any of the plant components measured and di f f erences were generally small . There were large dif ferences , however , in the crude protein level s between the plant components . Crude protein concentration was highest in the leaf , also high in the inflorescences , but much lower in the stem and very low in the tap root and stubble . When converted to a protein yield basis , the differences between treatments were highly signi f icant and followed very much the plant weight responses presented earlier i . e . protein yields were depressed more by primary branch removal than by mai n stern removal and most by defol iation of both primary branches and the main stem ( Figure 3 . 9 ) . 82a Table 3 . 4 Ef fect of intensity of defol iation on number of " growing points " on the stubble 1 0 days after cutting . Regrowth cycles Treatment 2 3 4 E-7 - 4 4 3a 1 3 9 a 3 0a 2 9 a E- 7 - 0 8c 8c 9b 1 Sb E-3 - 4 2 7 b 2 3b 2 8a 3 1 a E-3- 0 Se Se 7b S igni f icance * * * * * * * * 1 Values in the vertical column same not fol lowed by the s ame letter differ at p = 0 . o s . 82b Table 3 . 5 E f f ect of defol iation intensity on crude protein concentration ( \ of dry matter ) in leaf ( L ) ' stem ( S ) , i n f lorescence ( I ) ' root ( tap ( T ) and f ibrous ( F ) ) c0111ponents and in the residual s tubble ( R ) . Treatments Control 2 E-7-4 E-7-0 E-3-4 E-3-0 Sig . C:tcle 1 � R 24 . 0 3a 1 1 9 . 82b 2 3 . 9 9a 2 2 . 28a * T + F 2 3 . 7 3 2 1 . 9 8 2 1 . 98 2 1 . 6 8 2 1 . 6 8 ns Dai' 2 8 R 7 . 1 6 7 . 4 7 6 . 3 8 5 . 5 8 ns L 2 6 . 7 2 27 . 3 2 2 8 . 3 5 27 . 5 1 3 1 . 0 4 ns s 1 2 . 98 1 3 . 4 3 1 6 . 4 1 1 2 . 9 5 1 5 . 88 ns I 2 3 . 4 1 22 . 6 2 25 . 00 2 4 . 36 2 2 . 8 6 ns T 7 . 8 5 7 . 2 4 9 . 1 5 7 . 0 8 7 . 80 ns F 1 5 . 86 1 6 . 9 9 1 7 . 57 1 4 . 3 8 1 6 . 0 8 ns Dai' 42 R 6 . 96 5 . 9 8 7 . 1 9 7 . 2 4 ns L 26 . 9 2 28 . 67 26 . 6 2 28 . 8 0 ns s 1 2 . 6 5 1 3 . 7 9 1 1 . 86 1 4 . 4 4 ns I 22 . 8 7 2 3 . 1 7 2 4 . 7 2 2 4 . 9 5 ns T 6 . 2 8 6 . 4 9 5 . 8 3 6 . 0 0 ns F 1 4 . 6 9 1 5 . 9 9 1 4 . 3 6 1 8 . 0 1 ns Ci'cle 4 � R 5 . 5 5 5 . 30 5 . 1 9 6 . 1 6 ns T 7 . 6 0 5 . 8 1 6 . 6 8 6 . 6 6 7 . 1 0 ns F 1 3 . 3 5 1 5 . 9 1 1 3 . 66 1 5 . 66 1 5 . 37 ns Dai' 1 4 R 5 . 2 7 4 . 7 4 5 . 3 6 ns L 3 2 . 4 4 3 1 . 5 1 3 2 . 4 6 ns s 1 8 . 1 0 1 9 . 2 2 2 1 . 0 1 ns I na na na T 5 . 54 5 . 8 5 6 . 1 5 ns F 1 2 . 4 7 1 3 . 38 1 1 . 7 7 ns Dai' 2 8 R 5 . 2 9 5 . 5 9 5 . 4 6 ns L 3 0 . 4 0 2 9 . 3 1 3 0 . 1 2 ns s 1 6 . 57 1 7 . 6 1 1 6 . 60 ns I 2 2 . 4 7 2 2 . 0 3 ns T 7 . 2 1 7 . 1 4 6 . 7 9 ns F 1 4 . 2 5 1 6 . 0 3 1 4 . 3 1 ns Dai' 4 2 R 4 . 59 5 . 54 5 . 9 7 ns L 20 . 40 26 . 3 8 3 0 . 1 8 2 6 . 4 2 ns s 8 . 37 1 2 . 90 1 4 . 8 5 1 3 . 28 ns I 20 . 56 2 3 . 4 4 2 3 . 78 2 2 . 6 1 ns T 6 . 4 1 7 . 3 9 6 . 1 6 7 . 6 5 n s F 1 1 . 7 4 1 3 . 53 1 1 • 6 9 1 2 . 7 1 ns Values in the same horizontal l ine not fol l owed by the same letter dif fer at p = 0 . 05 . 2 Not included in statistical analysis . � 4 � l Ot 0 .J w > z UJ b f 14 CYCLE 1 .., CYCLE 2 CYCLE 3 4 CYCLE 4 ..,1 ( 42 DAYS >* ( 42 DAYS ) ( 42 DAYS ) ( 42 DAYS ) 28 4l E - 7- 4 L S D 5% E - 3- 4 E - 7- 0 E-3- 0 N . A. 56 7 0 DAYS AFTER F IRST CUT * PERIOD OF REGROWTH N . A . : NOT AVA I L AB L E N . A . NS 84 §a 1 1 2 126 140 Figure 3 . 9 : Ef fect of defoliation intensity on total crude protein yield at cycle l and 4 ( g/plant ) (JJ IV () 8 3 The crude protein yields o f the plant components were · similarly affected by the different intensities of defol iation as were component dry weights ( Figure 3 . 1 0 ) . However , it is important to note that the yield of crude protein from the leaf plus inflorescence fraction represented a s ignificant amount , especially in those treatments and plants receiving lax primary branch defoliation only . . In contrast , only small quantities of crude protein were found in the roots . C . 2 Non structural Carbohydrate Concentration and Yield Level s of non-structural carbohydrates ( sugars and starch ) were measured in the res idual plants ( above and below ground ) immediately after the 1 st and 2nd defoliation and after 52 days ' regrowth in the 4th cycle . As presented in Table 3 . 6 , sugar concentrations in both the tops and roots were very low in all samples and showed non-signif icant di fferences between defoliation treatments . Sugar concentrations were generally much higher in the tap root than in the fibrous roots , but not greatly dif ferent between tap root and tops ( above ground ) . In terms of sugar yields ( Figure 3 . 1 1 ) , the dif ferences recorded tended to reflect the dif ferent residual cutting treatments imposed . The exception , however , was the E- 3 - 4 treatment at the end o f the 4th cycle , suggesting that the sugar in the residual primary branches and even roots of those older plants , was s ignif icant in amount and greater than that in the residual main stern . Starch level s are presented in Table 3 . 7 and show that concentrations were extremely low in all residual parts and in all treatments . 1-z <( _, a.. '- ID a. 0 (/) 1-z w z 0 a.. ::?! 0 u 1-z <( _, a.. z c _, w >= z w 1-0 0:: 6 a.. w 5 c � 0:: 0 83a CYCLE 01 ,. I L s STUBBLE '" • ROOT. ( R I Wiil INFLORESCENCE ( I I r2L::J LEAF ( L I O srEM ( s ) I LSD ( P • 0·05 ) na ns L s STUBBLE 0 ns ns ns I L s 14 2B DAYS A F T E R D E FO L I AT I O N I I I L s I r i L s I Figure 3 . 10 : Effect of defol iation intensity on crude protein yield in the plant components ( g/plant ) at cycle 1 and 4 ns 52 8 3b Table 3 . 6 E f f ect of defoliation intensity on sugar concentration ( % of dry rnatter ) in the residual above ground ( stubble ) and below ground ( roots ) Treatments Cycle 1 Cycle 2 Cycle 4 ( day 0 ) ( day 0 ) ( day 5 2 ) A . Above ground residual - stubble E- 7 - 4 1 • 1 4 0 . 8 4 0 . -7 0 E- 7 - 0 1 . 1 0 0 . 9 2 0 . 8 2 E- 3 - 4 1 . 0 0 0 . 5 9 1 . 1 5 E- 3 - 0 1 . 6 8 0 . 8 0 Significance ns ns ns B . Below ground residual - roots Tap+ F ibrous Tap Fibrous Tap Fibrous E-7 - 4 0 . 7 3 0 . 7 2 0 . 0 4 1 . 3 2 0 . 0 4 E-7 - 0 0 . 7 3 0 . 6 0 0 . 0 3 0 . 7 1 0 . 0 4 E - 3- 4 0 . 7 9 0 . 4 4 0 . 0 3 1 . 4 1 0 . 0 4 E- 3 - 0 0 . 7 9 0 . 5 9 0 . 0 4 Significance ns ns ns ns ns 83c 1-z Cl _, Q. ... !. r A: RESIDUAL ABOYE GROUND CYCLE t ( DAY 0 ) • I � I L&J 0 8: ROOTS a a a a CYCLE 2 ( DAY 0 ) a a a a a a CYCLE 1 DAY 0 CYCLE 2 DAY 0 CYCLE 4 DAY 5 2 REGROWTH CYCLES Figure 3 . 1 1 : Ef fect of defol iation intensity on sugar yield in the residual top ( stubble ) and below ground ( roots ) (mg/plant ) 83d · Table 3 . 7 Ef fect o f defoliation intensity on starch concentration ( % of dry matter ) in the residual above ground ( stubble ) and below ground ( roots ) Treatments Cycle 1 Cycle 2 Cycle 4 ( day 0 ) ( day 0 ) ( day 5 2 ) A . Above ground Residual - stubble E- 7 - 4 0 . 0 6 0 . 0 2 0 . 1 1 E - 7 - 0 0 . 0 3 0 . 0 4 0 . 1 0 E- 3 - 4 0 . 0 2 0 . 0 3 0 . 1 4 E - 3 - 0 0 . 0 1 0 . 0 2 S igni ficance ns ns ns B . Below ground residual - roots Tap+ Fibrous Tap Fibrous Tap Fibrous E- 7 - 4 0 . 0 5 0 . 0 6 1 0 . 0 9 neg . neg . E-7 - 0 0 . 0 5 0 . 0 4 neg . 0 . 0 5 neg . E - 3 - 4 0 . 0 2 0 . 0 4 neg . 0 . 1 1 neg . E- 3 - 0 0 . 0 2 0 . 0 5 neg . S ignificance ns ns ns 1 Neg . negligible . = D . RELATION BETWEEN REGROWTH YIELD AND RESIDUAL PLANT VARIABLES The relationships between regrowth yield and residual plant variables are presented in Table Significant and positive correl ations were found for 8 4 the 3 • 8 • a l l cuttings . Branch number , leaf area , l eaf number , the number of growing points , res idual top dry weight ( stubble ) and shoot to root ratio were all signif icantly and pos itive ly correlated with regrowth yield for most regrowth cycles . There was no s igni ficant correlation between regrowth yield and the concentration of TNC in stubble and roots for the regrowth cycles examined , except for a negative correlation for the 6 week period in cycle 1 . However , the relationships between regrowth yield and the amount of TNC , particularly in the stubble , was significant and positive . E . RELATIONSHIP BETWEEN TOTAL PLANT DRY WEIGHT AND THE MAIN GROWTH PARAMETERS ( BRANCH NUMBER , LEAF NUMBER AND LEAF AREA ) Highly ( Table 3 . 9 ) s ignificant and positive correlations were found between plant dry weight and the main growth parameters ( branch number , leaf number and leaf area ) for almost all regrowth cycles , suggesting that these parameters are important in determining yield , a s found in Experiment 2 . 84a Table 3 . 8 Correl � t i un o f rf•s i dual (' J ant var i ftb J ts with net regrowth y ie ld . --------- - - -- -- -·- - - --- ----- -------------- Res idual plant variables immediately after cutt ing Weeks a f ter f i rst cut 6 ------- ----------� - ---------- -------- Branch number 0 . 8 52 * * 0 . 7 3 5 * * Lea f number 0 . 7 88 * * 0 . 6 80 * * Lea f area 0 . 780 * * 0 . 7 1 5* * ( cm2/plant ) Growinq points 0 . 898 * * o . 8o o • • ( no. /plant l 1 Stubble : root rat i o 0 . 7 85 * * 0 . 8 5 2 * * Residual stubble 0 . 7 7 4 * * 0 . 7 69 * * dry wt . ( g/plant ) Root dry weight - 0 . 0 3 8 -0 . 1 80 ' TNC in stubb le 0 0 308 0 . 5 86 ' TNC i n roots -0 . 6 7 9 -0 . 908 * * TNC yield in 0 . 8 4 1 * * 0 0 773 * stubble (mg/plant ) TNC yield in 0 . 096 - 0 0 4 4 7 roots ( mg/plant ) 1 Determined 1 0 days after cutt ing Weeks a fter second cut 4 6 0 . 9 1 4 * * 0 . 7 1 8 * * 0 . 890 ** 0 . 702 * * 0 . 88 1 * * 0 . 567 0 . 979 * * 0 . 8 1 4 * * 0 . 926 * * 0 . 620 * 0 . 92 6 * * 0 . 6 20* o . 8n•• o . 4 38 -0 0 1 1 8 - 0 0 4 4 2 0 . 2 1 4 -0 . 1 4 6 0 . 90 7 * * 0 . 4 2 9 0 . 7 85 * 0 . 24 1 Weeks after third cut 6 Weeks a fter fourth cut 2 6 ------------- ------------------ 0 . 6 1 5 * 0 . 90 9 * * o . 7 98 0 . 86 1 * * 0 . 7 2 9* * 0 . 852 ** 0 . 776 * * 0 . 7 8 2 0 . 9 2 5 * * 0 . 820* * 0 . 7 6 4 * * 0 . 882 * * 0 . 7 95 0 0 773 * 0 . 828 ** 0 . 888 * 0 0 4 8 5 0 . 8 30 * * 0 . 7 47 * * 0 0 58 1 -0 . 0 1 0 0 . 605 84b Table 3 . 9 Linear correlation coe f f icients between plant dry wei ght ( DM ) and main growth parameters ( branch number ( B ) , leaf number ( LNO ) and leaf area ( LA ) ) Parameters of growth B LNO LA Cycle 1 4 weeks 0 . 9 7 3 * * 0 . 9 8 6 * * 0 . 9 7 3 * * 6 weeks 0 . 9 8 7 * * 0 . 9 0 1 * * 0 . 8 2 2 * * Cycle 2 4 weeks 0 . 9 9 4 * * 0 . 9 9 8 * * 0 . 9 2 8 * * 6 weeks 0 . 9 1 6 * * 0 . 9 5 2 * * 0 . 8 5 8 * * Cycle 3 6 weeks 0 . 9 6 9 * * 0 . 7 4 4 * * 0 . 6 3 5 * * Cycle 4 2 weeks 0 . 7 5 4ns 0 . 8 8 4 * * 0 . 8 9 3 * * 4 weeks 0 . 8 9 7 * * 0 . 8 7 6 * * 0 . 8 8 0 * * 6 weeks 0 . 8 2 9 * * 0 . 8 5 4 * * 0 . 9 4 2 * * 8 5 IV DISCUSSION Results of this experiment showed that s evere and repeated cutting of Verano style was extremely deleterious to the plant as reflected in the death of 1 9 % of the plants following the second cutting in treatment E- 3 - 0 . Yields from this treatment were greatly reduced and no regrowth occurred during the 4th cycle of regrowth ( Table 3 . 2 ) . Even when defoliation of the main stern was lax ( E-7-0 ) , the complete removal of the primary branch sti l l had a dominant and detrimental ef fect on yield . Thi s is in contrast to the effect of lax cutting of the primary branches ( E- 7 - 4 and E- 3- 4 ) which was much less harmful to yield even when the main stern was cut hard ( Table 3 . 2 ) . These results demonstrated that lax defoliation of the primary branches i s of ma jor importance in maintaining high forage production even if the main stern is severely defoliated . The beneficial ef fects of lax cutting of the primary branches were related to the greater s i ze and levels of residual plant variables fol lowing cutting , such as residual leaf area ( Table 3 . 3 ) , the number of branches capable of regrowth ( Table 3 . 1 0 ) , the number of growing points ( Table 3 . 4 ) and the amount of stubbl e reserves ( Figure 3 . 1 1 ) . These parameters were all closely r elated to regrowth yields ( Table - 3 . 8 ) a s a l so shown in Experiment 2 . The importance of residual leaf area in regrowth of Verano style i s s imilar to results obtained with other tropical legumes ( Grof et al , 1 9 7 0 ; Jones , 1 9 7 4 ; Akinola and Whiternan , 1 9 7 5 ; Ludlow and Charles-Edwards , 1 9 8 0 ) . Jones ( 1 9 6 7 ) , Whiternan ( 1 9 6 9 ) and Jones ( 1 9 7 4 ) found that frequent cutting reduced yield and persistence of Siratro and this effect was attributed to a low leaf area remain ing on the stubble . In another experiment with individual plants , when 0 , 5 or 1 0 leaves were left on the stubble a fter cutting every four weeks , Jones ( 1 9 7 4 ) found that dry weight of regrowth and stolen development were greatest when most leaves were left . There were no plant deaths when 5 or 1 0 leaves were left but two thirds of the plants died under 85a Table 3 . 1 0 Ef fect o f defol iation intensity on number of res idual branches per plant Treatment Cycle 1 Cycle 2 Cycle 3 Cycle 4 ( day 0 ) ( day 0 ) ( day 0 ) ( day 0 ) E- 7- 4 2 1 a 1 1 4a 1 1 a 1 1 a E- 7 - 0 4 c 4b 3 b Oc E-3-4 1 7b 7b 6 b S b E- 3 - 0 4 c 3 b 2 b Signi ficance * * * * * * * Values in the s ame hori zontal l ine not followed by the same letter dif fer at P = 0 . 0 5 . 8 6 complete defoliation . A s imilar effect occurred i n this experiment with Verano stylo , particularly under severe cutting when very little leaf was left after cutting and stubble and root reserves were low in both concentration and amount . In thi s experiment , residual leaf areas were much greater under lax cutting than under severe cutting of the primary branches , when cutting was repeated at s ix weekly intervals ( Table 3 . 3 ) . Thus , under severe cutting of the primary branche s , energy for initial regrowth must have come from the l imited reserves in the main stem and roots , and hence lead to slow initial regrowth and , in the extreme treatment ( E- 3 - 0 ) , even to death of the plant under repeated cutting . Although TNC concentrations were general ly low the actual amounts of reserve carbohydrates in the stubble and roots varied considerably according particularly of the primary branches plant dry weight . The severe cutting markedly reduced the plant ' s abil ity to and of to cutting intens ity , hence to res idual the primary branches rebuild its reserves of carbohydrates between cuts and when also sub j ected to severe cutting of the main stem lead to " carbohydrate exhaustion " and eventual death . There has been considerable controversy concerning the relative importance of reserve carbohydrates and res idual leaf areas in determining the rate of regrowth in pasture species . Ward and Blaser ( 1 9 6 1 ) measured the regrowth rate of cocks foot grass tillers previously subjected to short-term shading and to varying intensities of defol iation . They concluded that the regrowth of the apex blades was inf luenced by both carbohydrate reserves and res idual L . A . I . The former predominated until day 25 and the latter was more inf luential thereafter . In their study it also appeared that the response to higher residual L . A . I . values in terms of dry matter production by apex blade , new tillers , and increase in apex blade length , was greater in the higher carbohydrate reserve treatments . Thi s may a l so apply to the present work 8 7 as indicated by the higher growth rate under lax cutting ( E- 7-4 and E-3-4 ) during the first 4 weeks of the f irst two regrowth cycles examined ( Figure 3 . 4 ) . In contrast , i n the l ast cycle in which initial regrowth was probably more dependent upon the reserve energy , due to an absence of residual leaves , regrowth was very s low during the first two weeks but increased steadily when the plants produced more leaves . This may suggest that in the situation where there are no leaves remaining on the plant a fter cutting , carbo­ hydrate reserves do play an important role in early regrowth . The number of branches capable of regrowth was h igh under lax cutting of the primary branches even when the main stem was severely defoliated ( Table 3 . 1 0 ) . The c l ose association of the number of branches with regrowth yield confirmed the importance of thi s parameter in regrowth abil ity . This has also been reported in species ( Kess ler and Shelton , 1 9 8 0 ) . determin ing many legume The advantage of lax cutting of primary branches was also related to the number of growing points remaining for regrowth . Regrowth of Verano stylo after cutting arose mainly from axillary buds located on the res idual primary branches , with only a small number of growing points located on the main stem - as also reported in Experiment 2 . Thus , intense cutting of the primary branches removed a lmost a l l of the growing points for regrowth , l eading to s low initial regrowth in these treatments . The importance of growing points for regrowth has also been demonstrated in Siratro ( Whiteman , 1 9 6 9 ) , Stylosanthes guianensis ( Grof et al , 1 9 7 0 ) , Desmodium intortum ( Imrie , ( Kess ler and Shelton , 1 B 8 0 ) . 1 9 7 1 ) and Crotarlaria juncia From the above discuss ion , it appears that the response of Verano stylo to defoliation is dependent on the number and especially the s i ze of primary branches , the number of growing points and the amount of stubble reserves immediately after defol iation . Although it was not possible to determine thei r relative importance , it does appear that these parameters assume dif fering level s o f importance depending on the state and condition of the plant . 8 8 The response of total plant dry weight to l a x cutting of the primary branches ( i . e . to node 4 ) was mainly through an increase in the stem , and to a lesser extent , the inf lorescence and leaf components ( Figure 3 . 5 and 3 . 6 ) . The other main growth parameters such as branch leaf number and leaf area were also a f fected by Intense cutting of the primary branches ( i . e . to number , cutting . node 0 ) especially in con junction with the severe cutting of the main stem markedly depressed or delayed development of these parameters . This resulted in slow recovery , reduced vigour and even death of the plants . In contrast , lax cutting of the primary branches ( i . e . to node 4 ) enabled a rapid recovery of branch number , leaf number and leaf area and hence high dry matter production , as reflected by a strong relationship between plant dry weight and these main growth parameters ( Table 3 . 9 ) . Although one must be cautious in extrapolating the results of growth room studies into field s ituations , the data on crude protein concentrations from thi s experiment do , however , indicate the relatively high nutritive value of this species . The crude protein levels recorded were sl ightly higher than those reported by Gardener ( 1 9 8 0 ) for Verano stylo and other tropical legumes by Hendy ( 1 9 7 1 ) , Ive ( 1 9 7 4 ) , Mufandaedz a ( 1 9 7 6 ) and Robertson et al ( 1 9 7 6 ) and well above - - the level that would be expected to impair animal production - s ince Mil ford and Minson ( 1 9 6 6 ) showed that intake is only reduced if crude protein of the forage is less than 8 % . As shown for other tropical pasture species ( Fi sher , 1 9 6 9 , 1 9 7 0 ; Robinson and Jones , 1 9 7 2 ; Mc ivor , 1 9 7 9 ; Gardener et al , 1 9 8 2 ) leaves contained higher concentrations of crude protein than stem , and inflorescence generally had level s s imilar to leaves . By comparison concentrations in the stubble and tap root were consistently low . Cutting treatments did not af fect the crude protein concentrations in any components , a s a l so obs erved by Mufandaedza ( 1 9 7 6 ) in � guianensis and by Olsen ( 1 9 7 3 ) in Desmodium intortum . However , a contrasting result was noted for S . humilis ( Fi sher , 1 9 7 3 ) . 8 9 Although the cutting intens ity did not affect the crude protein l evels , the total crude protein yield dif fered markedly between treatments ( Figure 3 . 1 0 ) . This was due to the difference in total plant dry weight . Crude protein yields under severe cutting of the primary branches was s ignif icantly lower than under lax cutting . Thi s shows that such cutti ng intensity not only depresses net regrowth available for animal intake , but also depresses total protein available to the grazing animal . The high crude protein yield p er plant recorded under l ax cutting of both primary branches and main stem resulted from an increase in crude protein yield in the inf lorescence fraction and to a l esser extent in the stem and leaf ( Figure 3 . 1 0 ) . This is of some importance in terms of its feeding value s ince the inf lorescence f raction is acceptabl e to stock even when the plants are mature and hence can be utili zed l ate in the season ( Gardener , 1 9 8 0 ) . Experiment 2 showed that total non-structural carbo­ hydrates ( sugar and starch ) in the residual top and roots of Verano stylo were low ( < 2 % of dry weight ) and compri sed mainly of sugar . The present work confirmed this finding and also showed that carbohydrate levels were independent of cutting intensity ( total non-structural carbohydrate present , < 2 % ) . Unfortunately no reports could be found in the literature on the effects of defoliation on the carbohydrate status of Verano stylo for comparative purposes . However , the results of the present experiment were s imilar to those obtained with another tropical legume S . guianensis ( Mufandaedza , 1 9 7 6 ) where cutting intensity and frequency had no s ignificant ef fect on total non-structural carbohydrates in the roots and stubble . In addition , he noted that the roots had a consistently higher total non-structural carbohydrate content than s tubble , which was not the case in this study . 9 0 EXPERIMENT 4 : FREQUENCY OF DEFOLIATION I . INTRODUCT ION Numerous studies with other perennial legumes and grasses have shown that . both intens ity and frequency of cutting inf luence the yield and quality of forage production , root growth and survival of species ( Jones , 1 9 7 3 , 1 9 7 4 ; Murphy et a l , 1 9 7 7 ; Jones and Carabaly , 1 9 8 1 ) . These ef fects were also observed in Verano stylo in Experiment 3 . In that experiment , the complete removal of the primary branches ( i . e . to node 0 ) caused a signifi cant reduction in yield and this ef fect was more accentuated when the main stem was a l so severely defol iated , and even lead to plant mortality . An examination of some of the f actors affecting regrowth indicated the importance of photosynthetic areas , carbo­ hydrate reserves , the number and s i z e of primary branches and the number of growing points remaining after cutting ( Experiment 3 ) . While the e f fects of dif ferent cutting intensities were examined in Experiment 3 , no asses sment was made of the e f fects of cutting frequencies on plant productivity and survival . Under field conditions in Thailand , Topark-Ngarm and Akkasaeng ( 1 9 7 8 ) found that the yield of Verano stylo under a four week cutting interval was signi ficantly lower than under a six week cutting frequency . However , no asses sment was made of the possible phenological responses , such as leaf area , carbohydrate levels etc . The present experiment , therefore , was conducted to gain further information on the response of Verano stylo to 3 and 6 weekly cutting at three intensities of cutting of th€ main stern , in terms of regrowth and chemical composition . 9 1 II . MATERIAL AND METHODS A. ENVIRONMENTAL CONDITIONS AND PLANTING PROCEDURES Environmental conditions and planting procedures were the same as reported for Chapter 3 ( Experiment 1 ) . After treatments were imposed , pots of similar treatments were located together to minimise the interference effects during regrowth . B . TREATMENTS Treatments replicated four times in a randomised block design were as fol lows : Treatment Detailed 1 . E-7- 4-3 : Cut the main stem above node 7 ( between nodes 7 and 8 ) and primary branches above node 4 ( between nodes 4 and 5 along the branch ) every 3 weeks . 2 0 E-5-4-3 : Cut the main stem above node 5 ( between nodes 5 and 6 ) and primary branches above node 4 ( between nodes 4 and 5 along the branch ) every 3 weeks . 3 . E- 3-4-3 : Cut the main stem above node 3 ( between nodes 3 and 4 ) and primary branches above node 4 ( between nodes 4 and 5 along the branch ) every 3 weeks . 4 . E- 5-4-6 : Cut the main stem above node 5 ( between nodes 5 and 6 ) and primary branches above node 4 ( between nodes 4 and 5 along the branch ) every 6 weeks . 9 2 The first defoliation was carried out when 5 0 % of the plant population had commenced flowering and was repeated after three weeks ' regrowth except in treatment E-5- 4 - 6 . Cutting was repeated 8 times with 8 regrowth cycles over approximately 2 0 0 days under 3-weekly defoliation and 4 times with 4 regrowth cycles under 6 -weekly defoliation ( Figure 4 . 1 ) • Nine harvests were taken in all treatments except treatment E- 5- 4- 6 where only five harvests occurred . C . PLANT MEASUREMENTS C . 1 Plant Dry Weight Measurements of plant dry weights and other phenological observations were obtained from 4 plants per treatment . Plants were sampled at the end of each regrowth cycle ( Figure 4 . 1 ) and separated for measurement as described in Experiment 1 . At each sampl ing , roots were also washed and separated into tap and fibrous roots . C . 2 Leaf Area ( cm 2/plant ) and Leaf Number Leaf area and leaf number were recorded at the same time as plant dry weights were determined . Leaf areas were measured us ing the Electronic Leaf Area Meter ( Model 3 1 0 0 Area Meter ) . C . 3 Branch Number per Plant Branch number per plant was recorded immediately after each regrowth cycle and before cutting was repeated , as described in Experiment 2 . C . 4 Number of Growing Points The number of vi sible growing points was recorded after 1 0 days of regrowth , as described in Experiment 2 . TREATMENT E -7-4 -3 E- 5- 4-3 E- 3-4- 3 E- 5-4-6 ESTABLISHMENT PERIODS : UP TO 50 % FLOWERING DEFOLIAT I O N REGROWTH PE RIODS ( DAYS ) Figure 4 . 1 : Planning of Experiment 4 1..0 1\.l Ill 9 3 D . CHEMICAL MEASUREMENTS Crude protein and total non-structural carbohydrate ( sugar and starch ) concentrations were determined as described in Experiment 2 . The plant fractions analysed are shown in Table 4 . 1 . E . STATISTICAL ANALYS IS Data were analysed according to the common procedure of a randorni zed complete block design for all plant characters ( Little and Hi lls , 1 9 7 5 ) . The analysis was done by Genstat programme ( Alvey et al , 1 9 7 7 ) . The least s ignif icant difference at the 5 % level was used to identify statistical differences . The symbol s used to designate stati stical s igni ficance were * ( P = 0 . 0 5 ) , * * ( P = 0 . 0 1 ) and ns ( not s igni ficant ) . I I I . RESULTS A . PHENOLOGICAL OBSERVATION Growth and development dur·ing the establishment phase were the same as those reported in Experiment 1 . The cutting treatments were first imposed when 5 0 % of the plant population had commenced flowering and repeated at three-week intervals for treatment E- 7- 4 - 3 , E- 5- 4- 3 and E- 3 - 4 - 3 , and at six weeks for treatment E- 5 - 4 - 6 . Under three weekly cutting , all treatments produced a dense mat of small new branches on the lower part of the main stern following the f i fth cut . Flowering time was a l so delayed by at least 2 weeks in these treatments . 9 3a Plate 4 . 1 : Compari son of Cutting Intens ities and Frequency taken immediate l y prior to the third 3 weekl y cutting and second 6 weekly cutting . ( Note : Greater production under 6 weekly versus 3 weekly cutting ; greater production unde r lax vers us severe cutting of the main s tem ) . 93b Table 4 . 1 P l ant components analysed . Time and Treatment Chemical Stubble A . E- 7 - 4 - 3 , E - 5 - 4 - 3 and E- 3 - 3 - 4 . 1 . Immediately a fter Protein + First Cut TNC + 2 . Immediately a fter Protein + Third Cut TNC + 3 • 3 Weeks a f ter Protein + Eighth Cut TNC + B . E- 5 - 4 - 6 1 . Immediately a fter Protein + First Cut TNC + 2 . Immediatel y a fter Protein + Second Cut TNC + 3 . 6 Weeks a fter Protein + Fourth Cut TNC + Plant Components Leaves Stem Inf lor . Root + + ""'"" + + + + + + + + + + 9 4 B . PLANT REGROWTH B . 1 Plant Dry Weight The total net regrowth for each treatment over the ful l exper imental period i s presented i n F igure 4 . 2 . There was a marked ef fect of cutting intensity and frequency on the total net regrowth . With a cutting interval of three weeks , net regrowth yields obtained under the heavy ( E- 3 - 4 - 3 ) and moderate ( E- 5- 4- 3 ) cutting treatments were similar , but both lower in yield than that achieved under lax cutting ( E- 7 - 4 - 3 ) Total net regrowth yield under 6 weekly cutting was doubl e that obtained under 3 weekly cutting at the moderate intensity of cutting ( E-5-4-6 ) . It was also c lear that cutting frequency had a relatively greater effect on plant growth than cutting intensity . The regrowth patterns fol lowing each defol i ation are presented in Figure 4 . 3 . Al l treatments under 3 weekly cutting showed a decl ine in yield at the 3rd and 4th cutting particularly under hard ( E- 3 -4 - 3 ) and moderate ( E- 5 - 4- 3 ) intensities of defoliation . Thereafter , the e f fects of defoliation intensity remained relatively constant throughout subsequent cuttings . The regrowth pattern of the plant� defol iated every 6 weeks showed a s imi lar decline in yield to the 4th cut . Al l regrowth components , of inf lorescence , leaf , stem and root were simi larly and in most cases s igni f i cantly depressed by cutting intensity and frequency as shown in Figure 4 . 4 . However the depress ion resulting from the most intense defol iation ( E- 3 -4-3 ) was not s ignificantly dif ferent from that resulting from moderate defoliation ( E- 5 - 4 - 3 ) . As might be expected , the more frequent cutting treatment ( 3 weekly ) markedly depres sed inf lorescence development compared with less frequent cutting ( 6 weekly ) . The absolute growth rates ( mg/day ) were also depressed by cutting intensity but not always s ignif icantly and to a much lesser degree than by cutting frequency ( Table 4 . 2 ) . 9 5 B . 2 Branch Number The ef fects of defoliation on branch number per plant are presented in Figure 4 . 5 . Moderate and hard cutting s ignificantly decreased branch numbers compared with lax cutting , whi le the difference between moderate and hard cutting was not significant . Cutting more frequently ( 3 weekly ) depres sed branch numbers significantly compared with cutting less frequently ( 6 weekly ) . B . 3 Lea f Area and Number Leaf area ( Figure 4 . 6 ) and leaf number ( Figure 4 . 7 ) per plant followed very simi lar patterns of development a fter each cut , with the most lax or infrequent cutting generally encouraging significantly more leaf area and leaf number than intense and frequent cutting . Although there was no s igni ficant ef fect of the res idual leaf area on leaf development following the f irst cutting , leaf area and leaf number of plants undergoing dif ferent cutting intensities increased in proportion to the leaf area and number remaining in the subsequent regrowth cycles ( Table 4 . 3 and 4 . 4 ) . Delaying cutting to six weeks in the moderate intensity treatment ( E-5- 4 - 6 ) had a nil or depress ing ef fect on res idual leaf area . Neverthe less , the regrowth achieved over 4 regrowth cycles in this treatment was s ignif icantly greater than in the frequent cutt ing treatment . B . 4 Number of Growing Points The number of growing points was significantly reduced by cutting intensity but only during the first four cycles of regrowth . Again , there was no s ignif icant difference in thi s respect between moderate and hard defoliation . However , a l l cutting treatments showed a n increase i n the number o f growing points with time , particularly under moderate and severe cutting , leading to a s imilar number between treat- 9 6 ments i n subsequent cycles ( Table 4 . 5 ) . The reason for this lack of response to lax defoliation over the later cuttings is difficult to explain , but observations i ndicated an increasing concentration of new shoots arising from the lower primary branches on the main stem with repeated cutting . Delaying cutting to six weeks signi ficantly depressed the number of growing points , whi le frequent cutting led to a substantial increase as the experiment progres sed . C . CHEMICAL COMPOSITION C . 1 Crude Protein As shown in Table 4 . 6 for the four cycles ( 1 st , 2nd , 7th and 8th ) analysed , there was virtually no s igni ficant effect of intensity and frequency of defoliation on the crude protein concentration of any of the plant components measured and dif ferences were generally small . However , there were large dif ferences between the plant components . Crude protein concentration was highest in the leaf , a l so high in the inf lorescences , but lower in the stem and very low in the roots , especially the tap root ( Table 4 . 6 ) . Total yields of crude protein were markedly a f fected by cutting frequency { Figure 4 . 8 ) and to a lesser extent by cutting intens ity and tended to ref lect the same trends as shown in regrowth yields . It i s interesting to note that leaves tended to be the ma jor contributor to total protein yield particularly under frequent cutting . 9 7 C . 2 Non-structural Carbohydrate Concentration and Yield As presented in Table 4 . 7 , sugar concentrations in both the stubble and roots were very low in all samples and showed non-significant dif ferences between defoliation treatments . Sugar concentrations were generally much higher in the stubble than in the tap root and virtually absent in the fibrous roots . Starch l evel s are presented in Table 4 . 8 and show that concentrations were extremely low in all residual parts . In terms of TNC yields in the stubble ( Figure 4 . 9 ) the treatment dif ferences tended to reflect the different cutting treatments imposed . For example , sugar in the residual of laxly defol iated plants was significantly greater than that of severely defoliated plants . This was very evident by the end of the 8th cycle ( day 2 1 ) ( Figure 4 . 9 ) . Frequency of cutting had no signif icant ef fect on either sugar or starch yields . . The amount of starch present in the stubble of a l l treatments and on each measurement occasion was consistently low and much less than sugar yields . D . RELATIONSHIP BETWEEN PLANT DRY WEIGHT AND THE MAIN GROWTH PARAMETERS ( BRANCH NUMBER , LEAF NUMBER AND LEAF AREA ) Highly s igni f icant , positive correlations were found between plant dry weight and the main growth parameters for almost all regrowth cycles ( Table 4 . 9 ) , suggesting that these parameters are important �n determining yield , as found in Experiments 2 and 3 . 9 8 IV . DISCUSSION This experiment showed that moderate ( E- 5 - 4- 3 ) and severe ( E- 3 - 4 - 3 ) defoliation of the main stem reduced yield by approximately 2 8 % compared with lax defoliation over a period of approximately 2 8 weeks and involving 8 cuts . This difference occurred within the first 2 cuts and thereafter remained relatively constant . Severe defol iation however was no more detrimental to regrowth than moderate defoliation . It i s suggested that the superior regrowth abi lity of the lax defol iation ( E- 7- 4- 3 ) treatment was related to the greater amount of stubble ( Figure 4 . 4 ) and hence the amount of TNC available for regrowth ( Figure 4 . 9 ) greater and the greater residual lea f area . The response in terms of total plant dry weight was mainly through an increase in the stem and to a les ser extent in the leaf fractions ( Figure 4 . 2 ) . Total net regrowth yield f rom 6 weekly cutting was twice that achieved from 3 weekly cutting at the moderate intensity of defol iation imposed . The increase in dry matter yield recorded under less frequent cutting was also reported with the same species by Topark-Ngarm and Akkasaeng ( 1 9 7 8 ) in their f ield experiment and with other pasture legumes such as Siratro ( Jones , 1 9 6 7 ) , Desmodium ( Jones , 1 9 7 3 ; Ludlow and Charles-Edward , 1 9 8 0 ) and Psoralea eriantha ( Gutteridge and Whiteman , 1 9 7 5 ) . Much of this increase in yield was due to a substantial deve l opment of the stem fraction and to a lesser extent the inflorescence and leaf components . The other main growth parameters such as branch number , leaf number a nd leaf area were also af fected by cutting intensity and f requency . The severe and moderate removal of the main stem s ignif icantly depressed or delayed development of these parameters and hence lead to lower dry matter production . These growth parameters were also markedly depressed under frequent cutting in spite of a greater number of growing points present . This resulted in low dry matter yield as ref lected by a strong relationship between total dry matter and these main growth parameters ( Table 4 . 9 ) . Further - 9 9 more , frequent cutting may reduce yield due to the inability of the legume to build up these growth parameters over such short interval s between cutting when compared with longer cutting intervals . From a practical point of view , it i s important to graze or cut Verano stylo laxly when under frequent defoliation . I f defol iation is severe then stubble i s reduced and regrowth severely impaired , especially i f the removal of the primary branches is al so severe as demonstrated in Experiment 3 . Furthermore , if the legume i s in association with a grass , as is normally the case , then the deleterious competitive e f fect can lead to further restriction on l egume regrowth and hence low yield . The percentage of crude protein in Verano stylo did not change markedly as a result of varying the cutting intens ity . Simi larly there was little ef fect on crude protein concentra­ tion from cutting frequently or infrequently , a lthough there was some evidence of depress ion in the stern f raction under infrequent cutting . Dif ferences in crude protein yields were largely a ref lection of the changes in dry matter yield . With the severe ( E- 3 - 4 - 3 ) and the moderate ( E- 5 - 4 - 3 ) removal of the main stern , the crude protein yield response was s imi lar but dif fered signi f icantly from lax cutting ( E- 7 - 4 - 3 ) of the main stern ( Figure 4 . 8 ) . The ma jor contributor to this response was the crude protein yield ln the stem and to a lesser extent in the leaf components . Delayed cutting to 6 weeks greatly increased the crude protein yield of the stem and to a lesser extent the leaf and inf lorescence fractions . There is little information in the literature on the effect of cutting frequency on carbohydrate reserves in Verano stylo . However , in Stylosanthes guianensis , Al ferez ( 1 9 7 4 ) found that under infrequent cutting ( 9 0 day interva l ) there was greater accumulation of total non-structural carbo- hydrates in the roots than under interval . In addition , total root by frequent cutting . In contrast , nitrogen and soluble carbohydrates };fASSEY UN :VERSI'['l LIBRARY, a 3 0 or 6 0 day cutting dry weights were reduced Jones ( 1 9 7 4 ) found that in the roots of S iratro 1 0 0 ( Macroptil ium atropurpureum ) were reduced under infrequent cutting ( 1 6 weeks ) compared with frequent cutting ( 4 and 8 weeks ) . Young and Robinson ( 1 9 6 3 ) also found that under l ight defoliation ( 2 0 % of top growth removed every two weeks ) and heavy defoliation ( 6 0 % of top growth removed every two weeks ) there was a large increase in the percentage of carbohydrates in Siratro compared with the undefoliated control . In the present study , cutting frequency had no signi ficant e f fect on either concentrations or amounts . However , the l evel of total non-structural carbohydrate in the roots and stubble under both frequent and infrequent cutting were lower than that recorded in Siratro ( Jones , 1 9 7 4 ) . The specific growth conditions such as optimum temperature for growth ( Ludlow and Wilson , 1 9 7 0 ) , and the unlimited water and nutrient supply , are cons idered to be largely responsible . On the other hand , cutting intens ity did a f fect carbohydrate yields , ref lecting dry matter responses , with lax defoliation providing a s ignif icantly greater residual of reserves in the stubble than moderate or severe defoliation . CHAPTER 5 EXPERIMENT 5 : EFFECT OF WATER STRESS AT DIFFERENT DEFOLIATION LEVELS ON REGROWTH CHARACTERISTICS OF STYLOSANTHES HAMATA ( VERANO ) I . INTRODUCTION 1 0 1 The performance and yield of a pasture species i s the genotypic expression as modulated by the continuous interactions with the environment . Among the environmental factors , one of the most widely l imiting for pasture growth i s water . Within tropical area s , a period of drought , either short term or long term , commonly occurs during the growing season . This is the ma jor determinant of pasture yield . Although there is an increas ing understanding of drought response in many crop plants , few attempts have been made to measure these responses in pasture plants , specif icially in tropical legumes . Much ef fort has been expended in order to maintain a grass-legume as sociation without knowing the water requirements and relationships . The plants ' response to this sort of hazard should be investigated in order to understand the dynamics of the associated plants in sown pastures . 1 0 2 II . MATERIALS AND METHODS A . ENVIRONMENTAL CONDITIONS AND PLANTING PROCEDURES The experiment was carried out in the Climate Controlled Rooms of the Plant Physiology Department , DSIR , Palmerston North , New Zealand . Verano stylo was grown in plastic pots ( 1 5 cm diameter by 40 cm deep ) f i l led with 7 kg of " Opiki loam " and sand mixture ( 7 0 : 30 v/v ) . Appropriate ferti­ li zers were also added . Sowing procedures were similar to the previous experiment . The conditions in the controlled environment room were : temperature humidity photoperiod - 30°/24°C ( day/night ) , - 7 0/ 9 0 RH ( day/night ) - 1 2 hours co2 level was monitored during the experiment and it ranged from 2 9 0 - 3 5 0 ppm during day conditions , and from 3 2 0 - 3 9 0 during night conditions . A complete nutrient solution ( NCSU - Appendix 1 ) was applied four times dai ly until 5 0 % of the plant population commenced f lowering . The solution was added to each pot to completely saturate the soil . At least 2 0 0 ml of the solution added drained out of the bottom of each pot over­ night and the pot was weighed at the start o f the light period . This was recorded as the field capacity . Pots were left until the water reached stress 1 and 2 , ( see below ) when the cutting treatments were then imposed . soil to day and syringe were used to replace the water ( deioni zed ) that was lost through transpiration . Correction for plant weight was made every time plants were harvested and watering ad j usted to maintain the two levels of water stress achieved , as shown in Table s . t After 8 4 days of water stress , all plants then received f ul l watering over the recovery period of three weeks to the end of the experiment . At the start of the water stress period , the surface in the pot was covered with plastic and gravel reduce water loss . Pots were weighted at random every before the beginning of the light period . Plastic tube B . TREATMENT The two l evel s of water stre s s imposed were : Stres s 1 ( Mi ld Stress - W 1 ) : Leaves on the main s tem become yel low and plant s l ightly wil ted ( 1 1 days after water withdrawn - Plate 5 . 1 ) . S tre s s 2 ( Severe Stress - W2 ) : Leaves on the main s tem 1 0 3 s tart to fa l l o f f , l ower leaves become yel low and ant s everel y wilted ( 1 8 days after water withdrawn - P late 5 . 2 ) . As s een in Table 5 . 1 the two leve l s o f water s tres s proposed were satis factor i ly achieved and maintai ned . When the p lants reached Stres s and 2 - 1 1 and 1 8 day s , r e spe ctive ly a fter ces s ation o f wateri ng - defol iation was imposed as fol lows : - 1 . Cut the main stem above node 7 ( between nodes 7 and 8 ) and primary branches above node 4 ( between nodes 4 and 5 a long the branch ) ( C 1 ) 2 . Cut the main stem above node 5 ( between nodes 5 and 6 ) and primary branches above node 4 ( between nodes 4 and 5 a long branch C 2 ) 3 . Cut the main stem node 3 ( nodes 3 and 4 ) and pr imary branches above node 4 ( between nodes 4 and 5 a long the branch ) ( C 3 ) 10 3a P l ate 5 . 1 : Plant under mi ld water stres s compared with non­ s tressed plant immed i a te l y before the cutting treatment was imposed ( Note : Plant i s s l i ghtly wi l ted under mi ld water stres s ) . P late 5 . 2 : Plant under s evere water stress compared with non­ stressed plant immedi a te ly before the cutting . ( Note : Death and s igni f i cant wi l ting of l eaves on the ma in s tem under severe water stres s ) . 1 0 4 The experiment therefore compri sed two leve l s o f water stre s s , with three intensities of defoli ation and can be l i s ted as f o l lows : - W 1 C 1 : Cut the main s tem above node 7 ( between nodes 7 and 8 ) and primary branches above node 4 ( between nodes 4 and 5 a long the branch ) at mi ld stres s l eve l . W 1 C 2 : Cut the main stem above node 5 ( between nodes 5 and 6 ) and pr imary branches above node 4 ( between nodes 4 and 5 along the branch ) at mi ld stre s s l eve l . W 1 C 3 : Cut the main stem above node 3 ( between node s 3 and 4 ) and pr imary branche s above node 4 ( between nodes 4 and 5 along the branch ) at mild stres s l eve l . W2C 1 : Cut the main s tem above node 7 ( between nodes 7 and 8 ) and pr imary branches above node 4 ( between nodes 4 and 5 a long the branch ) at s evere stres s leve l . W2C2 : Cut the main stem above node 5 ( between nodes 5 and 6 ) and pr imary branches above node 4 ( between nodes 4 and 5 along the branch ) at severe stres s leve l . W2C 3 : Cut the ma s tem above node 3 ( between nodes · 3 and 4 ) and primary branches above node 4 ( tween nodes 4 and 5 a long the branch ) at severe s tres s l eve l . C . MEASUREMENTS C . 1 Water Status C . 1 . 1 . Soi l water s tatus Soil moi sture content was measured at the day of harves ting . As previously s tated , the two level s of water stress proposed were s at i s factorily achieved and maintained ( Table 5 . 1 ) . 1 0 5 C . 1 . 2 . Plant water s tatus During the water stre s s period and a fter r e-watering , leaf relative water content ( %RWC ) was determined on day 0 , 1 4 , 2 8 , 4 2 , 6 3 , 7 0 , 8 4 and 1 0 5 a fter cutting . On each occa s i on , f ive newly expanded l eaves were taken in order to provide leaves o f a s imi l ar phys io logical age . The leaf sample s were then weighted to determine fres h wei ght and f loated in di stil led water for three hours unt i l they became ful l y turgid . Turgid wei ght was determined and dry weight was recorded a fter vacuum air oven drying for 6 hours . Relative water content of leaf was calculated a s f o l l ows : - %RWC = Fresh weight dry weight x 1 0 0 turgid weight - dry weight As seen in Figure 5 . 1 , the two leve l s of water s tres s propos ed were general l y achi eved and maintained . C . 2 P l ant Measurements C . 2 . 1 . Plant dry matter yield Measurements of dry wei ght and phenological observations were obtained from a total o f 3 plants per treatment . P lants were s eparated for mea surements a s described in Experiment 1 . Dry matter y ields were obtained by drying a l l s s in a vacuum oven for 7 2 s . C . 2 . 1 . Leaf area and 2 Leaf area ( cm /plant ) were mea day 0 , 4 2 , 6 3 , 8 4 and 1 0 5 a r cutting . The area o f was determined i n the s ame manner a s reported experiment ( Experiment 1 ) . a on f us C . 2 . 3 . Branch number ( no/plant ) Branch numbers were r ecorded on day 0 , 8 4 , 9 1 and 1 0 5 after cutti ng . 1 4 , 2 8 , 4 2 , 6 3 , Plate 5 . 5 : P lants under mi ld water stress ni ne weeks a fter cutti ng . Pl ate 5 . 6 : Plants under severe water stress n j ne weeks a fter cut t j ng . ( Note : Maj ntenance of fresh green l eaves and stems under mj ld water stress compared wj th grey-green leaves and reddj sh­ purple stems under severe water stress ) . 106 Plate 5 . 3 : A genera l view of the plants under the mi ld water stress s i x weeks a fter cutti ng . ( Note : s i gn i f i cant reta i n of leaves ) P late 5 . 4 : A genera l v iew o f the plants under severe water stress six weeks a fter cutting . ( Note : Large number of leaves senescenced and fal l of f ) . Plate 5 . 9 : The cutting treatments o f plants under previous mi l d water stress three weeks after re-watering ( Note : Significant i ncrease in plant s ize and i n number of branches a nd l eaves ) . ( E-6-4 = WlCl , E-4-4 = WlC2 and E-2-4 = WlC3 ) . P late 5 . 1 0 : The cutting treatments of plants under previous severe water stress three weeks a fter re-watering ( Note : Significant increase in plant s ize and in number of branches a nd l eaves ) . ( E-6-4 = W2Cl , E-4-4 = W2C2 and E- 2-4 = W2C3 ) . 107 i u t a Pl ate 5 . 7 : The cutti ng treatments of plants under mi ld water stress twe lve weeks a fter cutting . ( Note : Cl usters of sma l l branches a long the primary branches ) . ( E- 6 - 4 = WlCl , E- 4-4 = WlC2 and E-2-4 = WlC3 ) . Plate s . a : The cutting treatments of plants under severe water stress twe lve weeks a fter cutting . ( Note : Fewer clusters of branches a long the primary branches compared with p lants under mi ld water stres s , Plate 5 . 7 ) . ( E- 6 - 4 = W2Cl , E-4-4 = W2C2 and E-2-4 = W2C3 ) . 1 0 8 C . 3 Chemical Compositions The chemical compositions ( crude protein and TNC ) were determined in the same manner as reported in Experiment 2 . Plant parts and time of harvesting are shown in Table 5 . 2 . D . STATISTICAL ANALYSIS Data were analysed according to the procedure of a factorial design ( Little and Hills , 1 9 7 5 ) . The analysi s was done by Genstat programme ( Alvey et al , 1 9 7 7 ) . The least s ignificant dif ference at the 5 % level was used to identify statistical dif ferences . The symbols used to designate statistical significance were * ( P = 0 . 0 5 ) , * * ( P = 0 . 0 1 ) and ns ( not s ignif icant ) . I l l RESULTS A . PHENOLOGICAL OBSERVATIONS After a relatively s low establishment phase of 28 days , the plants grew rapidly and 5 0 % of the plant population were f lowering by day 3 5 . During the drought period , plants under mild stress maintained green leaves and stem surfaces ( Plate 5 . 5 ) , whereas under severe stress the plant stems developed a reddish purple colour ( Plate 5 . 6 ) . All plants in a l l treat­ ments shed their leaves , particularly those under severe stre s s , and the new leaves produced were smaller , thicker and a darker green colour . - New branches were a l so smaller and had very short internodes and tended to display a c luster of branches ( Pl ate 5 . 7 and 5 . 8 ) . During the recovery period , plants previous ly under both mild and severe stres s displayed leaf turgor within 24 hours . New leaves were noticeably bigger in s i ze and the plants grew very rapidly during the recovery period . New branches also had longer internodes ( Plate 5 . 9 and 5 . 1 0 ) . 1 08a Table 5 . 2 P lant components analysed . Weeks Chemicals Plant components a fter cutting Stubble Leaves Stem Inflor . 0 Protein + + + + TNC + + + + 6 Protein + + + + TNC + + + + 1 2 Protein + + + + TNC + + + + 1 5 Protein + + + + TNC + + + + Roots + + + + + + + + 1 0 9 B . PLANT REGROWTH B . 1 . Plant Dry Weight The results presented i n Table 5 . 3 show the main effects of water stres s and cutting on plant dry weight during the drought and recovery periods . Under severe water stres s total plant dry weight was significantly lower than under mild water stress throughout the drought period . Al so hard cutting of the main stem s ignificantly depres sed yield compared with the moderate and light cutting ( Table 5 . 3 ) . The response of Verano style to cutting under both mild and severe water stress showed s imilar trends ( Figure 5 . 2 ) . Under both levels of water stress , hard cutting general ly depressed plant dry weight compared with lax and moderate cutting , this ef fect being more apparent under mi ld water stres s . Whi le plants under mild water stress showed a steady increase in dry weight from the commencement o f regrowth , plants under severe water stress showed little i ncrease for the first 42 days and then only a small increase in weight to the end of the drought period . During the recovery period ( 1 2 - 1 5 weeks ) , in a l l cutting treatments under previous mi ld and severe stress , the growth of plants increased substantially . As a result , at the end of the recovery period , all cutting treatments , now at the same moisture regime , showed no signi ficant dif ferences in total plant dry weight ( F igure 5 . 2 ) . However , mean total plant dry weight ( over a l l cuttings ) in the s evere water stress treatment remained signif icantly lower than that in the mild water stress treatment ( Table 5 . } ) . Although the absolute growth rates were similar between cutting treatments , plants under s evere water stress were lower in absolute growth rate during both drought and recovery periods ( Table 5 . 4 ) . 1 09a Table 5 . 3 Main e f fect of water stress and defoliation on total plant dry weight during drought and recovery periods ( g/plant ) . During drought period Recovery Treatments from drought Days after cutting 0 4 2 6 3 8 4 2 1 days Water stress W 1 4 . 0 7 7 . 5 7 9 . 2 6 1 2 . 4 0 4 3 . 3 0 W2 4 . 0 9 4 . 4 8 5 . 4 9 5 . 7 9 3 0 . 2 0 S igni f icance ns * * * * * * * * Defoliation C 1 4 . 8 4 a 1 7 . 0 0 a 8 . 4 0a 9 . 7 4a 3 6 . 6 0 C 2 4 .. 2 7 a 6 . 4 4b 7 . 9 5a 9 . 8 0 a 3 9 . 6 0 C3 3 . 1 2b 4 . 6 3c 5 . 7 7b 7 . 7 4b 3 4 . 2 0 S igni f icance * * * * * * * * ns Interaction ns * * ns ns ns (W X C ) 1 Values in the vertical column not followed by the same same letter di f fer at p = 0 . 0 5 40 35 30 I- 25 z -a:: 0 I- 1 0 z - a:: 0 l l Ob A: WATER STRESS DROUGHT PERIOD 42 84 .. 0 63 DAYS AFT�R DAYS AFTER DAYS AFTER DAYS AFTER DEFOL IATION DEFOLIATION DEFOLIAT ION DEFOLIATION na na u •• •• IHI •• •11 a •• IHI .. L S I L S I L S I L S I RT : ns RT : na RT : na RT• • ns ns ns 8: QEFQI..lif[!ON L S I L s I L s L s L s - = - nl - na = nl nl RT: ns RT : na Figure 5 . 3 : Main e f fect of water stress ( A ) and defoliation ( B ) on plant components dry weight ( g/plant ) ..... z cl � �0 ... � c;. (/) !z w z 0 DAYS DROUGHT PERIOD RECOVERY 10� DAYS FOLLOWING DEFOLIATION I I �5 1 L S I 0 a. � 0 30t- AFTER DEFOLIATION 42 DAYS AFTER 63 DAYS AFTE R 84 DAYS AFTER T u ..... 20 � _J a. IJ.. 0 ..... :I: Q w � >- � LSD �-;. � .., ,!; � .L :r - 1.> T :t � u � :3 OTlU.. (T) ·� � t\1 3.: ROOT : NS KEY - D fZJ - DEFOLIATION L S I T - - - 27 INFLORESCENCE ( I l STEM ( S l LEAF ( L l ROOT ( R l DEFOL IATION L S I T • .J: . :r DEFOLIATION L S I T % %. - :z """ - I ROOT Figure 5 . 4 : Effect of water stress and defol iation on plant components dry weight ( g/plant ) LSD �-,. ...... ...... 0 0 1 1 1 B . 2 Branch Number During the first 1 4 days after cutting , plants under mi ld and severe water stress showed s imilar branch numbers . Thereafter , the increase in branch number was more rapid in the mild water stress treatment ( Figure 5 . 5 ) There was no s ignificant difference due to cutting under severe water stres s ( Figure 5 . 5 ) . However , under mild water stress , the W 1 C 1 and W 1 C2 treatments maintained s igni f icantly higher branch numbers than the W 1 C 3 treatment up to day 6 3 . The greater depressive ef fect of hard cutting on branch number in the mild water stress treatment compared with that in the severe water stress caused the significant interaction recorded between water stress and cutting at this stage . However , this e f fect had disappeared by the end of the drought period on day 8 4 , by which time the number of branches in the hard defol iation treatment ( W 1 C3 ) had rapidly increased and attained a level that was not s igni ficantly dif ferent from the other two treatments . During the recovery period , branching was stimulated in all cutting treatments under previous mild and severe water stress conditions ( Figure 5 . 5 ) . However , the main ef fect of the earl ier water stress was still apparent at the end of the recovery period of 2 1 days i . e . plants previously under severe water stres s were signif icantly lower in branch number than those previously under mild water stres s . In contrast , the ef fect of previous cutting was apparent only during the first 7 days of the recovery in the mild water-stres sed treatment as shown by a lower branch number under the hard cutting compared with the moderate and light cutting . This effect fai led to reach significance during the last 1 4 days of recovery as shown by the non-s ignif icant dif ference in branch number between cutting treatments at the final har­ vest . I n the severe water-stressed treatments there were no significant effects due to cutting intensity ( Figure 5 . 5 ) . In terms of branching rates , the main ef fect of cutting had no s ignif icant inf luence during the drought period ( Table l l la � z z I 0 z 0: 20 u � 15 10 5 0 40 35 1 1 7a .. DROUGHT PERIOD A: STUBBLE � ,.a � ... � r!- ( ' " " B: LEAF c b a ob ab ob .-- r- C: STEM b b ob o a a r- 0: INFLORESCENCE c b b b b .---r-r-- E: ROOTS .-r- 1- r-I 2 .J • , 6 1-- f-- B A 0 0 c 0 r-r- b � b b b c .-- 1--f--..-1-- b b b ..L b ,..!. 1- r- a a a a a a -_r- bc a o c ob a -r- r-r- .-- ,_. r-- r-r-1--r-1- A A A A A A . � RECOVERY c b b b � b rill n b b b O a ,!... r-r-- c b ,.2_ c ab� _- _r- • • • • •.,!!.. ,_ r- r-1- -1- 1--..- A B C B c c • .--r-1-- 1-- FIBROUS A A A A A A � t7 � � � � � v � � � p � t/. F7 [7� t7. � [/ V I/ [/ V r/ 1/ 0 0 0 �� 0 0 a ob b b b 0 0 V 1/ V / V V l/ 1/ V [/ V 0 42 DAYS AFTER DEFOLIATION t/ 17 � � V V 0 0 V v 84 [/ � I/ [/ l/ o a v l/ t/ 17 � [/ 0 0 1/ l/ 1/ 1/ 1/ I/ 0 0 1/ 1/ 105 � 1/ 1/ 0 0 1/ 1/ Figure 5 . 7 : Effect of water stress and defol iationon crude protein concentration in the plant components TAP ( % o f dry matter ) . Note : 1 = W 1 C 1 , 2 = W 1 C2 , 3 = W 1 C3 , 4 = W2C 1 , 5 = W2C2 and 6 = W2C3 . 1 1 7b Table 5 . 1 0 Main effects of water stress and defoliation on total crude protein yield ( leaves + s tern + stubble + inf lorescence ) during drought and recovery period ( rng/plant ) . Treatments Water stre s s W 1 W 2 Sign i ficance Defoliation C 1 C 2 C 3 Significance Interaction ( W X C ) During drought period 4 2 days a fter 8 4 days after cutting 9 5 5 5 7 2 * * 8 4 2 a 1 8 3 0a 6 1 8b * * * cutting 1 6 8 4 7 5 9 * * 1 2 7 0a 1 3 2 7 a 1 0 6 8b * * ns Recovery from drought 2 1 days 8 3 1 8 6 0 4 7 * * 6 6 5 6 7 8 5 9 7 0 3 2 ns ns 1 Values in the same vertical column not followed by the s ame letter di ffer at P = 0 . 0 5 ..... z :3 a.. ... Q) a. tit .s a _J w >- z w b 0: a.. w a :::> 0: u 1 17c 9000 8000 7000 6000 5000 4000 3000 2000 0 1111 INFLORESCENCE ( I ) � LEAF ( L ) c::=J STEM ( S ) .. ROOT ( R ) DROUGHT PERIOD 42 84 DAYS AFTER DAYS AFTER DEFOLIATION 1 I I :::1: L .X s I R r :::1: I t DEFOLIATION � ::c = I L S T 1oooL I I I L S I RECOVERY 2 1 DAYS AFTE R EFOL IATION Figure 5 . 8 : Ef fect o f water stress and defoliation on crude protein yield in the plant components ( g/plant ) 1 1 8 C . 2 Non-structural Carbohydrate Concentration and Yield Both the sugar and starch , and hence the TNC concentra­ tions of the plant components markedly increased during the drought period . However , there was no significant e f fect of level of water stress on the sugar and starch concentration of the stem and stubble ( except immediately after cutting ) throughout the entire drought period ( Figure 5 . 9 ) . This contrasted with the sugar and starch concentrations in the inf lorescence where severe water stress s ignificantly depressed sugar and starch concentration relative to mild water stress by the end of the drought period . In the leaf , this ef fect was recorded only in the starch fraction . The main ef fects of cutting intensity on starch of the plant components were observed drought period but were only apparent in the sugar during stubble and the and inf lorescence fractions ( Figure 5 . 9 ) . Severe and moderate defoliation ( C 2 and C3 ) depressed sugar concentrations in the stubble s ignificantly compared with l ight defoliation ( C 1 ) and continued through to the end of the drought ( day 8 4 ) . At thi s stage ( day 8 4 ) the starch fraction in the stubble was also depressed viz . the more severe the cutting intens ity the lower the starch concentration . In the inf lorescence , both starch and sugar , - and hence TNC were also depressed by the severe and moderate cutting ( C 1 and C 2 ) . During the recovery period , a l l regrowth components showed a cons iderable decrease in both sugar and starch . There was no ef fect of previous water stress on sugar and starch in a l l plant components , except the stubble where sugar concentration showed a greater increase in previously severely stressed plants compared with mildly s tressed plants . Cutting intensity had no e f fect on sugar and starch in a l l plant components ( Figure 5 . 9 ) , except the s tubble where starch concentration was depre ssed by the severe and moderate cutting under previously severe water stress ( Appendix 6 ) o/o 1 1 8a A : WATER STRESS 20�--�==========�--------�----------------� * I'll * ns na na na na ** I'll ** 0 10 5 0 20 1 0 SUGAR STARCH TNC SUGAR STARCH TNC SUGAR TNC SUGAR STAR<>i 1NC ns ns ns ns ** •• I'll I'll na B: LEAF na na na na na I'll na na ns C: ST E M ns ns ns ** ** ** na na ns D : INFLOR E S E N C 1 ... DROU G H T P E RIOD A: S T U B BLE B:CUTT/NG INTENSITY L SD ::11: I na ns - na % sx na na na na na ns na na B: LEAF % na na na nt na ns C : ST�M D: INFLORESCENC 1 0 ns ns nt na ril na na 5 oL-----�----�--���----L---����--���2�1�L-� DAYS DAYS AFTER RECOVERY F . 5 9 ua; n e ffect of water stress ( A ) and defoliation on �gure • : l"'l .... carbohydrates concentrations in the plant components ( % of dry matter ) 1 1 9 In terms of the carbohydrate concentration in the root fraction , sugar was found in both the tap and fibrous roots but was not significantly af fected by cutting or water stress throughout the entire drought period . Nevertheless the sugar concentration was much higher in the tap root than in the fibrous roots during the f irst half of the drought ( Figure 5 . 1 0 ) . Thereafter , both tap and fibrou� roots had similar sugar level s . In contrast , starch was predominant in the tap root but negligible in the fibrous roots . Thus , the di f ference between treatments in TNC levels was due mainly to the starch fraction . There was no ef fect of cutting on TNC concentration in the fibrous roots . In contrast , severe cutting depressed TNC , and particularly starch , in the tap root s ignificantly . The ef fect of water stres s was also apparent throughout the entire drought period but only in the tap root where severe moisture l imitation increased TNC concentration significantly over mild water stres s . However , on re-watering , TNC level s decreased significantly , particularly the starch fraction , and were negl igible in the tap and fibrous roots . Sugar was also negl igible in the fibrous roots and at only low levels in the tap root . When converted to TNC yields , the differences between treatments during the drought period were highly s igni f icant and followed the plant weight responses presented earlier , i . e . TNC yields were lower under severe than under mi ld water stress and were increasingly depressed with increas ing level s o f cutting ( Figure 5 . 1 1 ) . These responses were s imilar in both the sugar and starch fractions ( Figure 5 . 1 2 and 5 . 1 3 ) . Also in terms of the plant components both the sugar and s�arch , and hence the total TNC yields of all plant components , were s imilarly depressed by water stress . I n contrast , the main effect of cutting on TNC yields in the leaves and including both sugar and starch fraction were not a f fected . However , TNC yields in the stem and including both sugar and starch were increasingly depressed with increasing level s of cutting intensity . The main ef fect of cutting intensity on TNC , sugar and starch yields in the inf lor��cence was evident only during the first 4 2 days of ;/. 10 5 0 10 ;/. 5 0 1 0 x 5 X 5 1 19a WAT E R STRESS A: TAP ROOT * * ** AR TNC STARCH T NC 8: FI BROUS ROOT ns ns ns SUGAR STARCH TNC D E FOLI AT I ON A : TAP ROOT 8 : F I BR O U S ROOT ns ns ns 0 ns *"' .IUt ns ns ns ns :I: I ns hs ns 42 ns ** ** ns ns ns ns = ::c ns ns ns 84 DAYS A F T E R D EFOLIATION DROUGHT PERIOD ns ns ns NEGLIGIBLE ns - ns NEGL I G IBLE 105 RECOVERY Figure 5 . 10 : Main e f fect of water stress ( A ) and defoliation ( B ) non-structural carbohydrates concentration in the tap and fibrous roots . ( % of dry matter ) ..... z <[ .....J n. .....J <( 0: � ..... 0 � 0: ..... (/) I z 0 z .....J � � 1 19b r L I s ::z:; I I T ( TOTAL l I I % 11 I S I T l L I I I I I I I I I I I I 42 DAYS 84 DAYS I s 2 1 DAYS RECOVERY T DROUGHT PERIOD �--------------------------------------���----------------�--� Figure 5 . 1 1 : Ef fect of water stress and defoliation on tota l non-structural carbohydrates ( TNC ) yields ( mg/plant ) t­z <( ..J a.. (/) 0 ..J lJJ ): 0:: <( (!) � (/) 1 19c :r L INFLORESCENCE C I ) J I I I I :1: I I s I TOTAL I 42 I I I I DAYS AFTER DEFOLIATION I s 84 DROUGHT PERIOD I I I L s ::c I I TOTAL 2 1 DAYS AFTER RECOVERY RECOVERY Figure 5 . 1 2 : Effect of water stress and defoliation on sugar yields of the plant components ( mg/plant ) 1 19d 1400 1200 � 1 0 00 z Cl ..J Q.. ... G L S T DROUGHT PERIOD--...,..,_ __ RECOVERY __.. L S KEY : • � T ( TOTAL ) INFLORESCENCE ( I ) LEAF ( L ) a. 800 WICI 0 E D STEM ( S ) 0 ..J LLI >- I u Q: � (f) 42 DAYS 84 DAYS 21 DAYS Figure 5 . 1 3 : Ef fect of water stress and defoliation on starch yields of the plant components ( mg/plant ) 1 2 0 regrowth . This e ffect was dependent on the s everity of cutting , i . e . the more severe the cutting the lower the TNC , sugar and starch yields in the inf lorescence . The greater ef fect of hard cutting on total starch and TNC yield of the s tem and inf lorescence components under mild water stress compared with that under severe water stress caused a significant interaction between water and cutting but only on day 4 2 . This effect had disappeared by the end of the drought ( day 8 4 ) by which time the total starch and TNC yield under both mild and severe water stress were similarly depressed by hard cutting relative to the moderate and lax cutting ( Figure 5 . 1 1 and 5 . 1 3 ) . This interaction was not apparent in terms of sugar yields . During the recovery period , although the concentrations of TNC were low , there was a substantial increase in sugar yields in all treatments ( Figure 5 . 1 2 ) . This was greater in those plants previously under mild water stress ( W 1 ) than under severe water stress ( W2 ) , and particularly in the leaf and inflorescence components . By contrast , starch yields dec lined appreciably over this period and continued to show the depress ing ef fect of cutting but only under previous severe water stress ( Figure 5 . 1 3 ) . In terms of the yields of TNC , sugar and s tarch in the root fraction of the various treatments they were relatively low in the first hal f of the drought period but increased noticeably by the end of the period ( Figure 5 . 1 4 ) . By this time yields of both sugar and starch were general ly higher under mild water stress than under severe water stress and in terms of starch signif icantly depressed under hard cutting . During the recovery period , TNC level s dropped appreciably with the complete absence of starch , only hal f the previous quantity of sugar and with no dif ference between previous treatments ( Figure 5 . 1 4 ) . 1 20a � 400 <( _J � WATER STRESS � o. 300 E I (/) b � 200 (/) ::::> 0 a: m LL (/) 100 ::::> ...J Q_ (/) 1- 0 0 0: z 200 0 _J ·w r u z 1- 100 SUGAR : * STARCH : n s TNC : * WI W2 SUGAR ' n s STARCH : n s TNC n s CI C2C3 n s n s n s n s LSD SY• * * * * n s I I n s n s n s n s n s oL-�0�------��L_--���------llll __ 42 84 105 DAYS AFTER DEFOL IAT ION �-�------------�D�R�O�U�G�H�T�PoE�R�I�O�D�--------�•- �ECOVERY• Figure 5 . 1 4 : E f fect of water stress and defol iation on tota l non-structural carbohydrate yield in the roots ( tap + f ibrous ) ( mg/plant ) D . RELATIONSHIP BETWEEN TOTAL PLANT DRY WEIGHT AND THE MAIN GROWTH PARAMETERS ( BRANCH NUMBER , LEAF NUMBER AND LEAF AREA 1 2 1 Highly significant and positive correlations were found between plant dry weight and the main growth parameters ( branch number , leaf number and leaf area ) for all harvests ( Table 5 . 1 1 ) suggesting that these parameters are important for determining yield , as found in the non-water stress experiments ( Experiments 2 , 3 and 4 ) . IV DISCUSSION Soi l moi sture is the most frequent and ma jor limitation to plant growth ( Whiteman , 1 9 8 0 ) and can reduce dry matter accumulation and the expans ion of plant parts s ignificantly ( Waikakul , 1 9 8 3 ) . However , the severity of thi s effect is dependent upon the degree of this limitation as shown in the present experiment . Severe soil moisture l imitation resulted in a cons iderable reduction in plant si ze and plant dry weight compared with mild water stress ( Table 5 . 3 ) . This was also ref lected in the lower absolute and relative growth rate recorded ( Table 5 . 4 ) . As a result , the accumulation of total plant dry weight showed little increase under severe water stress throughout the drought period . The e f fects of water stress on plant dry weight were experessed mainly through the stem fraction and to a lesser extent the leaf and inf lorescence components ( Figure 5 . 3 ) . The same response was reflected in the number of branches ( Figure 5 . 5 ) , number of leaves and leaf area ( Figure 5 . 6 ) and the highly s ignificant and positive correlations between these parameters and plant dry weight ( Table 5 . 1 1 ) . Thus , it i s not surprising that the poorer response of these parameters recorded under severe water stress resulted in lower dry matter yield compared with those plants under mild moisture l imitation . 1 2la Table 5 . 1 1 Linear correlation coeff icients between plant dry weight ( DM ) and main growth parameters ( branch number ( B ) , leaf number ( LNO ) and leaf area ( LA ) ) Parameters of growth Weeks from defoliation B LNO LA 6 ( during drought period ) 0 . 9 1 3 * * 0 . 7 5 4 * * 0 . 8 6 3 * * 9 ( during drought period ) 0 . 9 0 9 * * 0 . 8 4 6 * * 0 . 8 4 3 * * 1 2 ( during drought period ) 0 . 8 6 3 * * 0 . 9 3 9 * * 0 . 8 8 8 * * 1 5 ( during recovery period ) 0 . 9 4 5 * * 0 . 8 8 5 * * 0 . 8 0 3 * * 1 2 2 All growth parameters mentioned above ( plant dry weight , branch number , l eaf number and leaf area ) were increased substantially during the recovery period in both the previously severe and mild water stress treatments but the e f fect of the previous intens ity of water l imitation was still evident i . e . plants previously under severe moisture l imitation produced significantly lower yields than those previously under mild water stres s ( Table 5 . 3 ) . One of the more interesting responses on re-watering was the remarkable recovery in al l growth parameters , especial ly in leaf number and leaf area , of those plants previously under severe water stress ( Figure 5 . 6 ) . S latyer ( 1 9 7 3 ) stated that developing tis sues appear to enter a re juvenating phase , and when the water stress is el iminated , the relative growth rates of such plants may be more rapid than in the control ( non-stres sed ) plants . F isher and Campbell ( 1 9 7 7 ) showed that only at early vegetative growth rate growth of stages did the water stress reduce the Townsville stylo but when the stress was el iminated , a period of rapid growth ensued which compensated for the loss . Thus , the rapid increase in the growth of all plant parameters can possibly be attributed to the younger physiological state of the previously severely stressed plants compared with those under previous mild moisture limitation . Ludlow and Ng ( 1 9 7 7 ) also found rapid extension growth on re-watering after drought in Panicum maximum and they concluded that the stimulated rates resulted from the rapid expansion of cells accumulated during the stress period , as cel l divi sion i s less sensitive to water stress than cel l expansion . I t i s also of interest to note that in the one parameter that- was increased during the recovery period ( at 7 days ) vi z . number of branches , there was an immediate rapid increase in the previous mild water stress treatment but a delay in the previous severe s tress treatment ( Figure 5 . 5 ) . However , once over this initial delay , by the end of the first week , branch number on the previously severely water stressed plants increased at a similar rate to that of the previously mi ldly water stres sed plants . Presumably the early advantage shown by the previously mildly water stressed 1 2 3 plants was due to their greater reserve of carbohydrates ( Figure 5 . 1 1 ) , greater leaf area and leaf number ( Figure 5 . 6 ) and possibly even greater root weight for better nutrient uptake ( Figure 5 . 3 ) . During the drought period crude protein concentrations in the leaf , stem and inf lorescence were higher under severe than under mild water stress ( Table 5 . 9 ) , as also reported by Carvalho ( 1 9 7 8 ) . However , this increase in crude protein concentration was not suff icient to compensate for the reduction in the amount of protein when converted to a yield bas i s . Mi ldly water stressed plants still produced a signi ficantly higher crude protein yield than did severely water stressed plants ( Table 5 . 1 0 ) - reflecting the greater dry weight of the mildly stres sed plants , as discussed ear l ier . The ma jor contributor to this response was the crude protein yield in the stem and to a lesser extent in the leaf and inflorescence components ( Figure 5 . 8 ) . S imi lar results have been recorded by Bennett et al ( 1 9 6 4 ) for sudan grass , mil let and sorghum . On re-watering , the crude protein concentrations markedly increased in the stem and especially in the leaf and inf lorescence components due possibly to the translocation of crude protein from the stubble and roots under both previous mild and severe water stres s . The differences in concentrations between treatments were generally sma l l but differed markedly in the crude protein yields and fol lowed the plant dry weight response - lower under previous severe than under mild water stress . In turning now to the carbohydrate levels , there are several reports which state that water stress increases the carbohydrate concentration in forage e . g . of cocks foot and tall fescue by Brown and Blaser ( 1 9 7 0 ) , of rye grass by Norris and Thomas ( 1 9 8 2a , 1 9 8 2b ) , o f green panic , buf fel and spear grass by Ford and Wilson ( 1 9 8 1 ) but , in contrast , this trend was not observed in Siratro by Ford and Wil son ( 1 9 8 1 ) . The result of the present work indicated that a mi ld intensity of water stress a l so increased carbohydrate concentration , particularly that of starch , leaf and inf lorescence ( Figure 5 . 9 ) . 1 2 4 but only in the However , severe moisture l imitation depressed the carbohydrate level in the leaf and inf lorescence s ignificantly but continued to increase it in the tap root . Of the total carbohydrate fraction , starch maintained a much higher concentration than sugar in all treatments throughout the drought period , particularly in the stem and stubble ( Figure 5 . 9 ) . In terms of the amounts of carbohydrate reserves , there were large differences due to water stress with the severely water stressed plants accumulating only hal f the reserves of the mildly water stressed plants during the drought period . This reduction was particularly noticeable in the leaf and inflorescence components , relative to the stem ( Figure 5 . 1 1 ) and roots ( Figure 5 . 1 4 ) . Under both mild and severe water stress the stem was the ma jor accumulator of these reserves , particularly the starch fraction ( Figure 5 . 1 2 and 5 . 1 3 ) . During the recovery period , although carbohydrate yields were high due to high dry matter production , concentrations were low , particularly of starch , and probably reflected their use in the production of new leaf and new branches ( Alberda , 1 9 57 ; May , 1 9 6 0 ; Davidson and Milthorpe , 1 9 6 6 ; Norris and Thomas , 1 9 8 2a , 1 9 8 2b ) . It is interesting to note the dif ferences between sugar and starch in terms of both concentrations and amounts immediately prior to the re­ watering period and after 3 weeks of adequate water ing . Starch concentration and yield , particularly in the stem and stubble , showed a substantial drop in the " recovery " period , presumably being used for early regrowth , while the sugar fraction showed much less change . However , owing to the lack of intensive measurements during the re-watering period , i t is not possible to detect the dynamics o f these reserves , particularly of the sugar fraction . Turning now to the ef fects of cutting , the results clearly showed that total plant dry weight during the water stress period was reduced under hard defoliation , especially when plants were sub jected to mild moisture l imitation 1 2 5 ( Figure 5 . 2 ) . This ef fect was evident throughout the drought period . The ma jor contributors in this regard were the stern and to a lesser extent the leaf fractions . Branch production was also depressed by cutting , but again only under mild moisture l imitation , as indicated by the s ignificant interaction between water stress and cutting . When under severe moisture stress - which in itself depressed plant weight and branch number s ignif icantly - plants showed no further depression from light , moderate or hard cutting . These results tend to conf l ict with the statement of Jantii and Heinonen ( 1 9 5 7 ) who recommended light or moderate defoliation when plants are under severe moisture stress . Perhaps the dif ference between their work and the current study is due to the relative levels of water stress imposed i . e . the current findings could support their recommendation if their eo-called severe water stress equated with our mild water stress . Cutting had no s ignificant ef fect on l eaf number throughout the drought period , except immediately a fter cutting , but did depress leaf area . However , this depres sion only occurred during the f irst 6 weeks when cutting was hard and under mild water stre s s ( Figure 5 . 6 ) . Thereafter , this treatment showed a substantial increase resulting in a s igni f i cantly higher leaf area than light and moderate cutting by the end of the drought period ( day 8 4 ) . In contras t , there was no e f fect due to cutting under severe moi sture limitation , except immediately after cutting , throughout the drought period . On re-watering , -the effects of previous cutting on a l l growth parameters tended to be relatively minor compared with the e f fects of previous water stress and differences were often non-signi ficant . The exception was the previous hard defoli ation treatment under mild water stress whi ch showed a sma l l but s ignificant increase in leaf area and leaf number ( Figure 5 . 6 ) . However , one of the more interesting responses on re-watering was the remarkable recovery in leaf area and especially leaf number in all the cutting treatments particu- 1 2 6 larly under previous severe water stress - as a l so mentioned earlier following water stres s . Apparently even hard defoliation followed by an extended period of water s tress did not prevent the plant from responding dramatically to the application of adequate water , particularly in terms of leaf area and leaf number , over the subsequent three weeks . In terms of the crude protein concentration in the various p lant components , the effects of cutting were generally sma l l although sometimes signi ficant . Hard cutting tended to increase crude protein concentration under both mild and severe water stress particularly in the leaf and inf lorescence ( Figure 5 . 7 ) . However , this increase in crude protein concentrations was not sufficient to compensate for the reduction in the amount of protein when converted to a yield bas i s . The greater the severity of cutting especially under mi ld water stress , the lower the total crude protein yields . Crude protein concentrations and amounts in the l eaf , stem and inf lorescence were a l l markedly increased during the recovery period , particularly in the leaf and inflorescence fractions . In terms of carbohydrate levels in the plant components , the ef fects of cutting during the drought period were only evident in the stubble , inf lorescence and tap root fractions - the l evels declining with increasing intensity of defoliation , particularly of the starch fraction ( Figure 5 . 9 ) . However , the concentration of TNC ( sugar plus starch ) was much higher than that found in Experiments 3 and 4 which were conducted with adequate water supply . Thi s suggests that under conditions where soil moisture is l imiting and growth rate reduced , carbohydrates may be able to accumulate to a greater extent than under conditions of rapid continuous growth , as a lso reported by Brown and Blaser ( 1 9 7 0 ) , particularly in the stem , stubble and tap root . Although total TNC concentration in the leaf was not a s high a s i n the apparent storage components of stubble , stem 1 2 7 and tap root , i t did , nevertheless , contain a higher proportion of sugar relative to starch - whereas starch was by far the ma jor fraction in the former components . It was this starch fraction in the stubble , stern and especially in the tap root that a lmost totally disappeared during the rapid recovery phase following re-watering - presumably used in regowth - as discus sed earlier . This supports the f indings of Fisher and Ludlow ( 1 9 8 4 ) who showed that the sugar concentrations in the leaves of Verano stylo tend to increase under water stress - and presumably enabled , in this experiment , the appreciable build-up of starch in the stubble , stern and tap root during the water stress period which was subsequently used during the recovery phase . Returning to the ef fect of cutting on carbohydrates , it was apparent that during the drought period hard cutting signif icantly depressed the concentration and accumulation of sugar and starch mainly in the stubbl e and in terms of yield , especially under mild water stress . In the roots only the starch fraction was so af fected . Simi larly during the recovery period , previous hard cutting again depressed both carbohydrate concentration and yield , particularly in those plants under previous severe water stress - but only in terms of the starch fraction . CHAPTER 6 EXPERIMENT 6 : EFFECT OF GRAZ ING MANAGEMENT ON STYLOSANTHES HAMATA PRODUCTION AND SURVIVAL I . INTRODUCTION 1 2 8 Although the number of dairy cattle in Thailand has substantially increased over the pas t ten year s , the main feeding systems for these cattle have shown l ittle change . Large amounts of money are still being spent on concentrate feeding during the dry season and even during the rainy season for milk production . ( Thai Dairy Promotion and Organization , 1 9 8 4 ) . The recognised alternative of us ing cultivated grasses with selected legumes , such a s Siratro ( Macroptil ium atropurpureum ) , Centrosema ( Centrosema pubescens ) and Stylo ( Stylosanthes guianens is ) as a means of reducing this dependency on concentrates sti l l lacks acceptance and adoption due to inadequate information on good graz ing or cutting management practices . The succes s ful establishment and persistence of Verano stylo ( Stylosanthes hamata cv Verano ) under grazing has been reported from the northeastern part of Thailand ( Wilaipon and Humphreys , 1 9 8 1 ; Wi laipon , 1 9 8 5 ) . As such the potential for using this legume to improve feeding qual i ty of improved pastures for dairy production is of real practical significance . Under intens ive studies in the Control led Cl imate Rooms at the Plant Physiology Divi sion , DSIR , Palmerston North , New Zealand , it has been clearly shown that Verano stylo has a­ real potential to produce acceptable yield and qual ity under an appropriate defoliation system . Early c utting , for example , when 5 0 % of the plant population reaches the f lowering stage , results in rapid recovery a lthough subsequent yield is depressed , while late cutting results in slow recovery of growth and may lead to a reducti on in yield . The extent of thi s reduction depends on the severity of cutting and in particular the number and s i z e of the residual 1 2 9 primary branche s l e f t after cutting . I n the s ubsequent s tudy compar ing frequency and intens ity of cutting , i t was a l s o shown that graz ing management should a im a t a l l owing the plants to retain at least 7 nodes on the main stern ( approxi­ mately 7 - 8 primary branches ) and at least 4 nod e s a long the primary branches a fter defoliation to achieve high yield and s urviva l . However , the s e experiments were conducted under artif ic ial growth room and cutting conditions and therefore it was cons idered e s s enti al to tes t this hypothes i s under reali stic f ie ld condition s . Hence a f ield expe riment was conducted in Thai land under graz ing conditions , t o s tudy the effect of selected gra z i ng management pract ice s on the productivity and per s i stence of a Verano stylo sward . I I MATERIAL AND METHODS A . ENVIRONMENTAL CONDITIONS AND PLANTING PRODUCURES The experiment was conducted at the Tha i Dai r y Promot ion and Organisation of Thai l and , located at Muaklek , 1 8 0 Km Northeas t of Bangkok . I n the area s e lected , Guinea gras s ( Panicum maximum ) and other native gras se s wer e initi a l l y G IYPho�ate , s prayed witn f Roundup ,and left for 7 days to ach a good kil l , and then ploughed , cultivated and subs hamata cv Verano ( Verano s tylo ) at y s own to 2 0 kg/ha . A basal ferti l i zer of super s ( 2 5 0 kg/ha ) s h ( 1 2 5 kg/ha was l ied and mur 1ate s owi ng ; nitrogen fert i l i z e r was added . The so i l was de s cr o f no by the Land Deve ( 1 9 7 9 ) as a f ine textured s i lty c l ay loam with poor s tr ucture in B hori zons and wi th a pH o f 6 . 5 . C l imatic conditions at the experimental s A are monsoonal with the ra iny s ea son extending from with peak precipitation in September and to October 1 , 0 1 2 mm annual l y . From November to April the weather i s relat l y dry with l i ttle rain occurring i n t h i s period . Mean and minimum temperature s a re 3 4 . 1 8 and 1 8 . 7°C r e s with a relative humidity averaging 7 7 % . Deta i l s o f I the 1 29a P late 6 . 1 : A genera l v iew of the experimenta l s ite , showing the we l l prepared seed-bed be fore sowing . P late 6 . 2 : The area was irrigated a fter sowing to ensure good germination . 1 3 0 c l imatic cond i tions for the p a s t ten years a r e given i n Figure 6 . 1 . However , during t h e experimental period , s o i l moi sture a t s owing , on 2 9 th Apri l 1 9 8 3 , was low . A s there was l i ttle rain fol lowing sowing , i rrigation was appl ied to ensure good seed germination and e stabli shment . Only l ight s howers continued throughout May and much of June and it was not until late June that good heavy rains occurre d , with some minor f l oodi ng in August . Good rainfa l l continued through September , October and into earl y November when pa sture growth began to be restricted by l ower temperature s . re was a l so s ome s l ight evidence o f a n Anthracno s e infection eh might a l s o have affected growth towards the end o f the growing season . Plant e s tabl ishment was exc e l l ent and the re latively few weeds appearing were removed by hand . B . TREATMENTS The pas ture was first gra z ed at an early s tage of growth v i z . when 5 0 % of plants began f l owering . Thereafter , two graz ing management treatments wer e impos ed vi z . at f i r s t f lower ( approximately 4 weekly i nterval s ) a nd at f u l l f lowering ( approximately 8 weekl y i nterva l s ) ( F i gure 6 . 2 ) . The pasture was gra z ed down to the 6th - 7 th node ( from the ground ) on the ma stem for both treatments which was equ iva lent to under the prev ng above node 7 of the main s tem ( E- 7 - 4 ) control led room exper Dry cows were used for gra z ing which extended ove r a 4 8 - 7 2 hour per number of ls us was on amount of pre sent . Mowing with an Auto s cythe was carri out after each gra z ing to ach a f orm res idual ant l { 6 th - 7th node ) , approximately 1 2 cm above ground l eve l . The d e s i gn of thi s exper iment was a randomi s ed ete block with 4 repl ications . i s total led 8 paddocks with each paddock be ing 2 0 x 20 m and separate l y . 1 30a 250 E E 200 _J _J it z � 15 0:: w a.. � w 1- 10 5 MAY I JUNE 14 TREATMENT I ( FREQUENT ) JUNE 12 MAY I JUNE 14 TREATMENT 2 ( INFREQUENT J JULY 18 ·�- . ·- SEPT. 19 OCT. 20 NOV. 23 -·-• - '" 1 CYCLE 4 A CYCLE 5 I I 3 28/ 3 AUGUST I9 SEPT. 23 OCT. 24 AUGUST 14 CYCLE 2 CYCLE 3 3/ AUGUST 19 OCT. 24 GRAZIN6 PERIODS IN DAYS REGROWTH PERIODS NOV. 23 F igure 6 . 2 : Timing of grazing and regrowth period throughout the experiment 1-' w 0 tr 1 3 1 C . MEASUREMENTS Seed l ing e stab l ishment and surviva l o f the s own plants ( Verano stylo ) was recorded on 1 2th June , 1 st July , 1 4 th Jul y , 3 rd August , 1 4 th August , 1 9th September , 2 0 th October , 9 th November and 2 3 rd November 1 9 8 3 , approximatel y 0 , 2 , 4 , 6 , 8 , 1 2 , 1 6 , 1 8 and 2 0 weeks a fter first graz ing , by taking twel ve quadrats ( 5 0 x 5 0 cm ) per p lot on each occa s i on . Dry matter y ields were measured immediate l y before and a f ter each gra z ing and on some occas ions during the regrowth period ( approximately 2 , 4 , 6 , 8 , 1 2 , 1 6 , 1 8 and 2 0 week s after f irst gra z ing ) . Ten quadrats { 5 0 x 5 0 cm were taken at random in each paddock and cut to ground l evel , care be ing taken to avoid sampl ing the s ame quadrat area twice . The cut samples were then bulked , weighed and subsampled to determine dry matter percentage . Drying was for 7 2 hours at 9 0°C . A further subsample , compri s ing at l east 1 0 plants , was a l s o taken for detailed measurements o f plant component s ( s tem , leaf and inflorescence ) , leaf area and branch number . Additional measurements of branch numbers were made between dry matter yield determinations , to g ive weekly counts , from norma l l y 2 0 random plants per plot . 1 3 2 I I I RESULTS A . SEEDLING ESTABLISHMENT Seed germination and seedl ing e stabl i shmen t , a lthough s l ow initially , was excel lent and by 1 2t h June a count of 1 7 8 plants/m2 was recorded . F irst f l owers appeared 3 2 days a fter s owing and the f irst graz ing commenced on 1 2th June 1 9 8 3 ( 4 2 days a f ter sowing ) . As shown in Figure 6 . 3 , plant dens ity f or both treatments decl ined with time but to a much greater extent under infrequent graz ing . O n ly 1 2 % of t h e plants s urvived under infrequent gra z i ng compared with 3 9 % under f requent ref l e cted graz ing ( Table 6 . 1 ) . Thes e dif ferences wer e a l s o i n the dry matter yie lds ( kg/ha ) p r esented i n Figure 6 . 4 . Phenological observation a l s o showed that p lants i n the infrequently gra z ed plots grew tal l er in height , were l e s s branched and had a l eaf canopy concentrated on t h e upper 4 0 cm o f the sward . This was apparent by the end o f the f i r s t eight weeks fo l lowing the f i r s t gra z ing . Many p lants d i ed after the second graz ing and the survivors regrew s lowl y . On the other hand , fewer plants d ied in on ly the f requently graz ed plot and thos e r emaining regrew vigorous ly . P lants in the f requently grazed plot a l so produced many f i ne branches nearBround leve l , prostrate stems and leave s . ng a dense mat o f relat ly B . PLANT B . 1 Plant Dry Weight and Production Total net regrowth y lds for both treatments over ful l experimental periods are presented in Figure 6 . 4 on an area bas i s ( Kg/ha ) . There was a marked d i f ference between treatments with the frequently g ra z ed treatment produci ng a s igni f icantly h i gher d ry matter yield ( P = 0 . 0 1 ) than the frequently gra zed treatment . I n terms of dry matter production dur ing the exper imental period , the the f requently gra z ed treatment was c l early ior i ty of ( Figure l 32a Table 6 . 1 ant den sity on the day before f ir st graz i ng and Trea tments Frequent I n frequent at the end of the ( plant s/m2 ) At end o f At f i r s t graz i ng exper iment % ( 1 2 /6/8 3 ) ( 2 3 / 1 1 /8 3 ) 1 5 9 6 2 1 7 4 2 0 Signi f icance n s * * survival 3 8 . 9 9 1 1 • 4 9 ns ns • ; ,J " w CD :E :::> z �· ·� 0 2 DATES: 1-HJ·83 1·1'83 ns �- ** ' � - 4 G 8 1 0 WEEKS AFTER GRAZING 14·1'83 3·883 19-883 .... � 12 19·9-83 F i gure 6 . 3 : Changes i n l egume ant dens i ty wi th t i me ( number/m 2 ) 14 ** *)11- "** • GRAZED � - FREQUENT INFREQUENT IS 18 20 20·10-83 !H 183 231/·83 1-' w N t! 1 3 3 6 . 5 ) . The i nfrequently graz ed treatment showed a marked peak in product i on to the f i r st e ight-week defo l i at i on in mid­ Augus t but thereafter showed poor recovery and regrowth . I n contrast the frequently graz ed treatment maintained re latively h igh production up to the 4th defo l iation and then dec l ined s omewhat over the remaining two regrowth period s . Thi s dec l ine may wel l have been a reflection o f the drop i n plant numbers over thi s per iod , a s shown in Figure 6 . 3 , together with the onset o f the dry and cooler s ea son . The regrowth pattern on a per plant bas i s { g/pl ant ) unde r the two gra z ing managements i s shown in F i gure 6 . 6 . A changing pattern of plant regrowth in the infrequently gra zed treatment wa s again most evident , with a marked peak in plant dry weight on the 1 5th o f August , after which the y ield dec l i ned sharply . I n contra s t , plant dry wei ght achieved in the f requently gra zed treatment was relativel y constant over the experimental peri od . The components of p l ant dry weight are s hown in Figure 6 . 4 and 6 . 7 on a total unit area bas i s , and on a per p l ant bas i s , respectively . S tem was the ma jor component in both treatments but with a cons iderably higher proport ion and amount of l eaf per p lant being produced under frequent gra z ing than under in frequent gra z ing . As shown in Figure 6 . 8 , rapid branch deve ( no /plant ) was recorded in the frequent graz ing treatment a fter ever y graz ing . I n contra s t , branch deve in frequent gra z ing was marked ly depre s s ed fol l owing the f t eight week de fol iation in mid August . However , branch pro­ duc tion was encouraged dur ing the last four of the experiment in both treatments . Thi s was pos s y due to an increas e in reproductive branches as the plant res l ower temperature and shorter day-length . to (/) LLJ � DATE : 6 5 0 14·6-83 + GRAZED 2 4 1·7·83 14·7·83 + 6 8 10 WEEKS AFTER GRAZING 3-8·83 I&-8·A3 12 14 16 19·9·83 20-10·83 Figure 6 . 5 : Ef fect o f graz i ng management on dry matter producti on ( ton /ha ) 1 8 9-1 1·83 20 23·1 1 ·83 ..... w w (lJ 10 9 - 1-z 8 <{ ...J 0.. ... 7 ID Q. e. 6 1-:X: !:2 5 w � >- 4 er 0 � 3 <{ ...J 0.. 2 ..J � 0 1- 0 1 · 5-83 Q INFREQUENT • FREQUENT � l +GRAZING .L.__ _ _ L_ _ _ _L I I DATE : 14·6·83 2 4 6 8 10 12 14 20 WEEKS AFTER GRAZ ING I ·NI3 1<4·7· 83 3·8·83 le-&·83 19·9·83 20· 10 · 83 !J- 1 1 · 83 21•1 1 83 Figure 6 .6 : Tota l plant dry wei ght ( g/pl ant ) ,._, w w t) -1-z <( � l 1-0: � 1-z <( ..J 0.. ... • 0. a. - Vi 1-0: <( Q. 1-z <( ..J Q. � (f) 1-z LtJ z 0 Q. � 0 u 9 8 A : INFREOUENT 8: FREOUENT 0 DATE : 14·6·83 2 1-7·83 4 14'7·83 8 D sTEM � INFLORESCENCE - LEAF 10 WEEKS AFTER GRAZING 15·8·83 19 9·83 16 20·10·83 : Ef fect of grazi ng management on the plant components dry we i ght ( g =-.;::,.AL::;..;.;_:;___:...::......, 20 9·11 ·83 23· 1 1 ·83 r-> w w () 100 90 80- 1-z z :X: 30 u z 1-z � c[ ! lOO Ol z ...... a:: LiJ CO � 50 ::J z 1..1... � LiJ • FREQUENT 0 I N FREQUENT t GRAZED � ...J 0 K: / > 6 8 10 12 14 16 18 20 0 2 DATE: Hs-133 14·6·� 1·7·83 14·7· 83 WEEKS AFTER 3·8·83 15-8·83 19·9·83 Figure 6 . 9 : Ef fect of grazi ng management on l ea f area ( cm 2 ) and lea f number per oLant 20·10·83 9 1 1 83 23-11·83 1-' '-" ,):> OJ ns P£RC£NTAG£ OF W££0$ /. 0 2 4 DATE : 14·6·83 1·7-83 14-7·83 n s 6 8 10 WEEKS AFTER GRAZING 3·8-83 1!5-8·83 ..... 1 2 19·9·83 Figure 6 . 1 0 : Percentage of weeds dur i ng the exper imenta l per i od ...... ... ... ...... FREQUE�T G��ING -o- -a 14 16 18 20 20·10·83 9·1 1 ·83 23·11 ·81 1-' w J:>. IJ Plate 6 . 5 : Dry cows on Verano stylo at first grazing . ( Note : Ready acceptance by grazing anima l s ) . Plate 6 . 6 : Immediately after first grazing - plant height approximately 1 2-15 ems . 1 3 5 1 3 5 a Plate 6 . 3 : A general v iew showing good Verano stylo establishment . Plate 6 . 4 : Immediately before f irst grazing when approximately 50% of plants begin to f lower - approximately 30-35 ems in height . 1 3 6 Plate 6 . 9 : Dur i ng second gra zi ng a t e i ght weekly i nterva l a nd th i rd graz i ng a t four wee k l y i nterva l . ( Note : Ea se of prehens i on of four weekl y gra z i ng and severe wa stage of ei ght week l y grazi ng ) . 1 36a Plate 6 . 7 : Immed i ate l y be fore thi rd g raz i ng of four weekly interva l s . ( Note : Large number of green leaves and branches c lose to ground ) . P l ate 6 . 8 : Immediate l y be fore second grazing a t e ight weekly interva l s . ( Note : Neg l i g i ble number of green l eaves and branches c lose to the ground ) . 1 3 7 IV D ISCUSSION The reduction in plant density for both treatments from the beginning of the wet s eason to the end of the growing season , agrees with the f indings of Gil lard et a l ( 1 9 8 0 } and Wilaipon and Humphreys ( 1 9 8 1 } . However , mortal ity was much greater under the longer spe l l ing interval . The ma jority of plant deaths also occurred much earlier under infrequent graz ing , leading to a ma j or drop in production . Obviously plant competition was more intense under the longer spell ing interval leading to greater death of plants , compared with the shorter spe l l between gra z ings . The importance of res idual leaf area remaining after graz ing was a l so evident in achieving rapid r ecovery of leaves and branches for regrowth . Under the controlled climate room studies , it was shown that Verano s tylo had low level s of carbohydrate reserves even under adequate water supply , so that the primary branches were needed to compensate for the lack of the residual leaf areas and thereby to provide the f lush of new buds . In the present experiment , the plants were grazed and mown to a s imilar intensity to treatment E- 7 - 4 in Experiment 2 ( Chapter 4 } . However , under the controlled cl imate room conditions , the plant was better able to withstand cutting several times , probably due to the greater number of growing points remain­ ing below cutting height . In the field conditions , under long spelling intervals , the elevation of growing points above the graz ing and cutting height was probably a factor contributing to subsequent plant mortal ity . Similar ef fects have been reported in Townsvil l e stylo ( Stylosanthes humi l i s } ( Robertson et a l , 1 9 7 6 ) , where they recorded a 5 3 % mortal ity of Townsville stylo plants under an 8 week cutting interval compared with 2 % under a 2 week cutting regime . This they considered to be due , mainly , to the elevation of the growing points above cutting height . Trampl ing by cattl e may also have been a factor in causing plant mortality in the present study , as discussed by Watkin and Clements ( 1 9 7 8 } . I n contrast , under the shorter spelling interval , Verano stylo develop a more prostrate habit with many branches and 1 3 7 a P l ate 6 . 10 : Four weekly grazing treatment a t end of exper imental per iod . ( Note : Ma intenance of relative ly dense and c l ean s tand of Verano style ) . P late 6 . 1 1 : Eight weekly graz ing treatment a t end of experimental per iod . ( Note : R e lativel y l ow density of Verano style and subsequent ingras s o f weeds ) . 1 3 8 leaves growing close to the ground and hence les s suscepti­ bil i ty to prehension and removal by animal or machine . With other tropical legumes , Fi sher ( 1 9 7 3 ) a lso noticed the dif ference in plant and sward morphology under f requent and infrequent cutting in � humilis . The death of plants under frequent graz ing occurred mainly a fter the fourth graz ing . It is dif ficult to explain whether this was due to the limited longevity of the plant or to the treatments imposed . Verano stylo is a short-lived perennial plant ( Humphreys , 1 9 8 0a ) and under graz ing conditions , Gardener ( 1 9 8 1 ) found that the ma jor i ty of Verano stylo plants died in their seeding year . He recorded only 0 . 0 3 % survival to the end of the third year . Wilaipon and Humphreys ( 1 9 8 1 ) also reported a s imilar result under graz ing conditions in Thailand . They also noticed that graz ing l ate in the wet season increased the number of perennating plants . Although this decline in plant numbers did occur in the present experiment , it is encouraging to note the relatively high number and percentage of plants present in the frequently graz ed treatment , at the end of the season . The longer spelling interval also reduced plant s i ze and branch recovery , particularly after the second graz ing . The infrequently defoliated plants took a longer time to build up their branch numbers compared with the more frequently defol iated plants . Weeds , mainly Guinea gras s ( Panicum maximum ) , were therefore able to invade and increased from 20% to 8 0 % by the second and third graz ing respectively . The results of this f ield experiment , a lthough only covering one year ' s production , c learly support and substan­ tiate the results obtained from the controlled c l imate room studies . Although the carbohydrate reserves were not determined in this field work , the slower recovery of the infrequently grazed plants was considered to be due to the lack of residual leaf area , growing points a nd branches capable of regrowth , as demonstrated in the controlled room studie s . Primary branches on the main stem are a l so es sen­ tial to achieve high yield and survival as stated earl ier . 1 3 9 CHAPTER 7 GENERAL DISCUSSION A . GROWTH AND DEVELOPMENT Verano style has become increasingly important in recent years because it has shown a potential for high yields of good quality feed ( Gardener , 1 9 8 0 ; Wilaipon , 1 9 8 5 ) in con j unction with rapid or moderately rapid recovery a fter cutting or graz ing throughout a wide range of conditions ( Edye et al , 1 9 7 5a ; Bishop et al , 1 9 8 0 ; Gil lard et al , 1 9 8 0 ; Gardener , 1 9 8 1 ) . I t has a reputation for being a drought res i s tant forage legume ( Wi lliam and Gardener , 1 9 8 4 ; F i sher and Ludlow , 1 9 8 4 ) which , although highly productive under high rainfall ( Bishop et a l , 1 9 8 0 ) , is also wel l suited to areas of low ( 5 0 0 mm ) or erratic rainfall ( Edye et a l , 1 9 7 5b ) . These characteristics have been clearly demonstrated in the control led environment studies ( Experiments - 5 ) under a wide range of defol iation and water stre s s regimes . Growth and development of Verano style in the controlled environment clearly showed the highly product ive capacity of the species over the experimental period of 1 3 1 days . However , early growth i n terms of dry matter y ield , leaf area , plant height and number of branches , was relatively slow up to the onset of f lowering . This trend was also observed under field conditions in Thailand ( Exper iment 6 ) . Nevertheless , growth of the p�ant increased rapidly after the onset of f lowering ( Figure 1 . 2 ) and reached a maximum absolute growth rate around day 8 0 which closely approximated the attainment of maximum leaf area . Maximum dry matter production of 1 0 5g/plant based on the predicted growth model , occurred on day 1 0 8 after s eedl ing emergence . Thi s increase in plant dry weight was mai nly through the stem and inf lores­ cence fractions and to a l esser extent the leaves - as also ref lected in an increase in the number of branches and leaves � Beyond the 1 0 8 day period production noticeably decl ined due to ageing processes with remaining the dominant component , as Gardener et al ( 1 9 8 2 ) . the s tem fraction , a lso reported by 1 4 0 B . DEFOLIATION AND PLANT GROWTH High yields of good qual ity feed in con j unction with the adaptability to low soil fertility has lead to a rapid and growing use of Verano stylo throughout the tropi c s . However , the widespread adoption of this legume in farm practice has been constrained by its failure to persist . Thi s i s in part due to a lack of understanding not only of the management whi ch must be applied to ensure the persistence of the legume but also of the appropriate method of utilization to achieve satisf actory legume growth as well as maintain plant vigour and stand longevity , and of the appropriate pathway of plant replacement . Work in the controlled environment study has clearly shown the possibil ity of achieving at least some of these goal s through correct cutting and graz ing management , as i llustrated in Experiment 2 - 6 . As one might expect , the results from the controlled environment studies clearly showed that the greater the intens ity of defol iation the greater the depress ion of regrowth ( Experiment 4 ) . However , this e f f ect was very dif ferent depending on whether it was defoliation of the main stem or the primary branches ( Experiment 3 ) . Reducing the s i z e of the primary branches had a greater detrimental impact on plant regrowth than defoliating the main stem and hence reducing the number of primary branches . For example , a reduction in the number of primary branches f rom node 7 to node 3 on the main stem ( i . e . from 8 branches to 4 branches ) reduced total yield by approximately 3 0 % whi le a reduction in the size of primary branches from node 4 to node 0 ( along the branch ) reduced yield by over 6 0 % ( Figure 3 . 2 ) . The most severe cutting treatment of both the main s tem and the primary branches ( E-3-0 ) resulted in the maximum reduction in yield and on several occasions resulted in s ignificant plant death . Thi s highlights the importance of controlled graz ing to ensure a residual of an adequate number and particularly s i z e of primary branches and hopefully a greater residual leaf area for regrowth . In view of the strong relationship between number of branches and plant weight it i s perhaps not surprising that defoliation practices which lead to a marked 1 4 1 depression or delay in new branch development results in slow recovery and often low dry matter production . The effect of reducing the s i ze of the primary branches was even more apparent when defoliation was delayed to the later stage of growth ( Experiment 2 ) . In thi s s ituation , virtual ly complete removal of the primary branches ( i . e . to node 0 ) resulted in extensive plant death after only a s ingle cut whereas retaining some l ength of primary branches ( up to node 4 ) assisted plants to survive and regrow at a s low but steady rate and subsequently achieve a total dry matter yield not too dissimilar from that of the uncut control six weeks after defoliation ( Figure 2 . 3 ) . This indicates that severe graz ing of Verano stylo at this late stage may lead to a severe reduction in subsequent yield through an insuf ficiency of s ites for regrowth as well as through plant mortal ity . Such e f fects on plant mortality after both late and intense cutting has also been reported in other tropical l egumes such as Stylosanthes humil i s ( Fi sher , 1 9 7 3 ) and Crotalaria juncea ( Kess ler and Shelton , 1 9 8 0 ) . When total dry matter yield was examined in terms of its components , it was found that the differences in yield were largely due to changes i n the stem , and to a les ser extent the inflorescence and lea f fractions . The importance of the stem as a ma jor component of yield was also apparent under infrequent graz ing - as the increase in yield from six weekly versus three weekly cutting was largely through an increase in the stem and t o a l esser extent the inf lorescence components . The beneficial ef fect of retaining an adequate number and especially size of the primary branches after defoliation i s obviously associated with the greater number of s ites i n terms o f active " growing points " ( potential new branches ) and branches , the amount of carbohydrate reserves and the residual leaf area remaining for initial regrowth a fter cutting , as revealed in the high and signi ficant correlation of these parameters with subsequent regrowth . Removal of the primary branches ( i . e . to node 0 ) greatly reduced a l l these 1 4 2 parameters . Thus , growth of these plants in terms of dry matter , the number of branches , leaf number and leaf area was slow to recover as reflected in the slow rate of growth ( Figure 3 . 4 ) , resulting in low dry matter production . I n contrast , the presence of primary branches resulted i n a rapid i ncrease in all these parameters especially in those plants having the greatest res idual a fter defoliation even under repeated cutting . It was also interesting to note that when cutting was delayed to the later stage of growth and then subj ected to only a lax intensity - hence more residual sites for new branch development ( Table 2 . 4 ) - initial plant recovery was slow due to lack of res idual leaf area but subsequently produced a real " f lush" of branches and leaves during the later period of growth . This slow initial recovery growth particularly under hard cutting warrants emphasis as it was a common reaction in all the experiments and reflected the time required for adequate leaf and especially branch development to occur - often a period of 3 to 4 weeks . However , once achieved then subsequent growth was general ly rapid over the following 2 to 3 weeks . Root weights were also s ignif icantly depressed by the removal of the primary branches especial ly when the main stem was a l so severely defoliated ( Figure 3 . 6 ) . Such an e f fect must a l so limit the plants ' abi lity to obtain soi l nutrients and moisture for regrowth . Humphreys ( 1 9 8 1 ) also highlighted the importance of new shoot development by stating that a high rate of shoot replacement after cutting or gra z ing is necessary to ensure high yield and persistence under grazing or cutting . Graz i ng management that leads to complete loss of the primary branches may not only reduce growth in terms of dry matter , green leaf number and number of branches but also in terms of the legume longevity . The frequency determining yield . of cutting was also important In fact frequency of defol iation had in a 1 4 3 much greater impact on yield than i ntensity of defoliation . Many workers have clearly demonstrated the advantage of cutting infrequently rather than frequently ( e . g . Topark­ Ngarm and Akkasaeng , 1 9 7 8 ) . Experiment 4 showed that growth in terms of dry matter , number of branches , leaf number and leaf area was s ignificantly depres sed under frequent cutting in spite of a greater number of growing points present . Presumably the 3 -weekly cutting was too short an interval to permit the full exploitation and development of leaf f rom the greater number of sites . Brown and Blaser ( 1 9 6 8 ) have shown the importance of high levels of canopy light interception for maximum growth rate . Thus , the growth rate of plants under infrequent cutting was greater than under frequent cutting ( Table 4 . 2 ) , resulting in higher dry matter production . However , the results from the field experiment in Thai land were quite di f ferent from the growth room study - this wil l be discussed later . C . WATER STRESS AND PLANT GROWTH The results from the control led environment study confirmed that Verano stylo is a dehydration tolerator ( Fisher and Ludlow , 1 9 8 4 ) and showed that the plant is able to survive a prolonged moisture l imitation without serious ly impairing its ability to recover . This dehydration tolerance is associated with a reduction in leaf area through leaf senescence and a reduction in leaf expansion . An increase in specific leaf weight due to leaf thickening , and an increase in chlorophyll concentration as observed in the dark green colour of the leaves during the stress period , were a l so noted ( Plate 5 . 7 and 5 . 8 ) . These phenomena are considered a s morphological mechanisms o f adaptation which a s s ist plant survival ( Peak et a l , 1 9 7 5 ; Fisher , 1 9 8 3 ; Waikakul , 1 9 8 3 ; Fisher and Ludlow , 1 9 8 4 ) . In addition , an increase in the carbohydrate concentrations in the leaves is al so regarded as a physiological mechanism of the plant ( Wil son et a l , 1 9 8 0 ) . This mechani sm is known as osmotic ad justment , which i s the process by which leaf osmotic potential decreases and off sets the lowered water potential , and hence ful ly or partially 1 4 4 maintains turgor . Maintenance of turgor is necessary for shoot and root growth , stomatal opening , and many metaboli c processes ( Ludlow , 1 9 80a , 1 9 8 0 b ; Fi sher and Ludlow , 1 9 8 4 ) . These phenomena are found in many species of Stylosanthes and Centrosema but not in Macroptilium atropurpureum cv . S iratro ( Ludlow et a l , 1 9 8 3 ; Fisher and Ludlow , 1 9 8 4 ) . Under field conditions , it was also found that Verano stylo di splays deep-rootedness ( Gutteridge , 1 9 8 2 ) parahel iotropic leaf movements which reduce absorption and of solar ( Fisher adapted radiation , and hence leaf temperature and water loss and Ludlow , 1 9 8 4 ) . Verano stylo is therefore well to a dry monsoonal environment where drought during growth is a common occurrence . The results f rom the control led environment ( Experiment 5 ) indicated that mild and severe water resulted in a reduction in plant growth in terms study deficit of dry matter , compared number of green leaves and number of branches with wel l watered plants ( Experiment 3 ) . when The severity of this e f fect was related to the degree of water def icit i . e . the more severe the water deficit the greater the reduction in these plant structures . For example , at the end of the drought ( day 8 4 ) , total dry matter yield , green leaf number and branch number under severe water stress were only 4 4 % , 4 5 % , and 4 0 % of these respective parameters under mild water stres s . However , enhanced growth and yield were obtained fol lowing re-watering in both previous ly mild and severe water stres sed treatments - especially in the former treatment . Although it was not poss ible to establish whether there was any compensatory growth following rewatering of the previous ly stressed plants , they nevertheless displayed a remarkable abil ity to recover from a signif icant period of water stress - thi s is of considerable practical importance . The most striking ef fect of water def icit was on total leaf area ( Figure 5 . 6 ) . For any crop , the leaf area i s determined by the number o f leaves produced and the s i z e and rate of development of these new leaves . In thi s study ( Experiment 5 ) the total number of leaves per plant ( Figure 1 4 5 5 . 6 ) and the rate of new leaf appearance ( Table 5 . 6 ) were the ma j or contributors to the dif ferences in total leaf area recorded and showed marked reductions under s evere water stre s s . By compari son individual leaf s i ze appeared less sensitive to stres s and in fact showed both a positive and negative response on occasions ( Table 5 . 7 ) . The paramount importance of leaf area on the productivity of plants has been repeatedly emphasi zed ( Watson , 1 94 7 ) . During the recovery period , total leaf area was markedly increased through a rapid increase in leaf number and at a s imilar rate in both the previous ly mild and severe stress treatments although the total leaf area achieved over the three weeks of re-watering was greater in the former treatment ( Figure 5 . 6 ) . In this respect , it is important to note that mild stressed plants had higher amounts of total TNC at the end of the drought than did the severely stressed plants ( F igure 5 . 1 1 ) . It i s suggested that these higher reserves of carbohydrate may well have a s sisted the early recovery growth of the previously mild stressed plants ( Davidson and Milthorpe , 1 9 6 6 ) leading to their greater production o f leaf area ( Figure 5 . 6 ) and dry matter ( Table 5 . 3 ) . Dry matter y ields throughout the drought and recovery periods were greater in the mild stres s treatment than in the severe stress treatment largely due to an increase in the stem fraction and to a lesser extent in the leaf and inf lorescence f ractions ( Figure 5 . 3 ) . This sensitivity of stem elongation to water stress has also been reported by Wil l iams and Gardener ( 1 9 8 4 ) . As in the previous experiments , the number of branches per plant. was strongly correlated to plant dry weight , with the severely stressed plant producing s igni ficantly fewer branches than mildly stressed plants . This d i f ference in branch number and hence plant dry weight establ i shed during the drought period was maintained through into the recovery period in spite of a greater increase in the rate of branching of those plants previously under s evere water stress ( Table 5 . 5 ) . 1 4 6 As cell divis ion i s reported to be less sensitive to water stress than cel l expansion ( Hsiao , 1 9 7 3 ) , it i s poss ible that under mild stress cel l division and hence branch initiation occurred but cel l expansion was restricted . With the removal of water stress there was a rapid expansion of these cells resulting in massive branch development - as recorded during the first 7 days following re-watering . In contrast , those plants that were previous ly under severe water stress showed slower branch development suggesting that both cell divi sion and cell expansion . were s everely restricted during the drought period and hence s lower in recovery fol lowing re-watering ( Table 5 . 7 ) . As di scussed in the experimental section ( page 1 2 1 ) , this response i s probably closely l inked to the di f ferences in carbohydrate reserves , leaf area , leaf number and root weights of the respective treatments . Defoliation shortly after the onset of drought stre s s caused a s igni f icant depression i n subsequent yield particularly when severe and i f the plant was under mild water stress . Thi s ef fect was evident throughout the drought per iod but was not s ignificant following re-watering . It suggests therefore that hard grazing during the dry s eason should be avoided and preference given to merely " taking the top " of f the l egume and thereby ensur ing an adequate number and size of res idual primary branches . Although there were no significant dif ferences in root treatments made at the onset of the nevertheless suggested that lax defoliation prior to the dry season weights between drought per iod , rather than ( Experiment 3 ) cutting it i s severe may be important in encouraging deeper root development as an aid to plant survival through the drought . D . HERBAGE QUALITY Although the results from the cutting experiment have limited application to grazed pasture where defoliation i s usually incomplete and may be of higher frequency , it can sti l l be used as a guide for quality prediction . 1 4 7 As shown in Tables 2 . 5 , 3 . 5 , 4 . 6 and 5 . 9 , Verano stylo generally contained sufficient crude protein to meet animal requirements . These data also show that the content of crude protein in these studies varied ( 1 ) with stage of growth , ( 2 ) with cutting management ( 3 ) in different organs and ( 4 ) with the watering regime . Regarding the f irst aspect the results of a number of studies ( Fisher , 1 9 6 9 , 1 9 7 0 ; Winter et al , 1 9 7 7 ; Gardener et al , 1 9 8 2 ) have shown that the crude protein i s highest in young tissue at the early part of the growing season , and decl ines with advancing plant maturity . These changes were also observed with Verano stylo in the uncut control treat- ment ( Tables 2 . 5 and 3 . 5 ) . However the rate of decl ine with maturity varies between plant parts with the leaf and inflorescence having a higher crude protein concentration than the stem at a l l stages of maturity . The changes in leaf and stem values with time may be a s sociated with tran s loca­ tion of nitrogen to the seeds as demonstrated by Robinson and Jones ( 1 9 7 2 ) with Townsvil le stylo . The crude protein concentration in al l plant parts was increased by defoliation especial ly in the stem fraction ( Tables 2 . 5 , 3 . 5 and 4 . 6 ) a s a lso reported in Townsvil le sty le ( Hendy , 1 9 7 1 ; Ive , 1 9 7 4 ) . However , the dif ferences between the treatments as a result of varying the cutting intensity and frequency were general ly smal l . was a l so s l ightly higher in the stem of the plants ( every three weeks ) compared with Crude protein frequently cut the stem of infrequently cut plants ( every six weeks ) . S imi l ar e f f ects were reported by Mufandaedza ( 1 9 7 6 ) who found that the crude protein level in several strains of � guianensi s increased with more frequent cutting . Robertson et al ( 1 9 7 6 ) working in Thailand , a l so reported an increase in crude protein level s of S . humilis under more frequent cutting . Leaf and inf lorescence contained higher concentrations of crude protein than stem at a l l cutting intensities and f requencies ( Tables 2 . 5 , 3 . 5 and 4 . 6 ) . In contrast , stubble had the lowest crude protein level in a l l cutting treatments . 1 4 8 The concentrations of these plant components mentioned above respond dif ferently under water stress . Lea f and inflorescence were depressed in crude protein level relative to wel l watered plants in Experiments 2 , 3 and 4 , and appeared to be due to the redistribution of nitrogen f rom leaves and inflorescence to the root f raction . Thi s in part supports the results obtained by Fi sher ( 1 9 8 0 ) working with s. hurnilis , who found that the plant redistributed its nitrogen and phosphorus to the roots primarily in response to water stress rather than maturity . This transfer of nitrogen appeared to be less apparent under severe than under mild water stress ( Table 5 . 9 ) . Although the crude protein percentage in di f ferent plant parts and for dif ferent cutting intensity and frequency and for different watering regimes was small , the amounts per plant ( Figures 2 . 9 , 3 .9 , 4 . 8 and 5 . 8 ) were largel y due to the large and s ignif icant di f ference obtained in dry weights between the di f ferent treatments . The effect of defoliating the main stern on crude protein yields fol lowed a similar pattern to that of plant dry weight vi z . defol iating to node 5 or node 3 significantly depressed yield compared with defol iating to node 7 , with l ittle di f ference between medium ( to node 5 ) and severe ( to nod� 3 ) defol iation ( Figure 4 . 8 ) . However , reduction in the s i ze of the primary branches again had a greater impact on crude protein yield than reduction in the number of primary branches - with the greatest depression resulting from the de foliation of both the main stern and the primary branches . The importance of s i ze of primary branches on crude protein yield was also apparent even . when cutting was delayed to the later stage of growth . Increas ing the frequency of cutting from 6 to 3 weekly intervals decreased the crude protein yields of a l l above ground components in spite of the s l ight increase in nitrogen concentration in the stern under frequent cutting . The e f fect of water stres s on crude protein yield again ref lected the dry matter responses . At 42 days a fter cutting 1 4 9 the depression from severe cutting was only signif icant under mild water stress , but by day 8 4 the e f fect of severe cutting resulted in a reduction in crude protein yield under both mild and severe stress . Of perhaps greater significance was the fact that cutting , even when s evere , had less e f fect on crude protein yields than water def i c iency ( Figure 5 . 8 ) , and was reflected in all plant components . Also , mi ldly stressed plants produced substantially greater crude protein yields than severely stressed plants . The results of these experiments clearly show that the complete removal of the primary branches especially in con j unction with the hard cutting of the main stem i s detrimental to both quantity and qual ity of product ion . Whi l e lax cutting of the primary branches may not be too damaging to dry matter yield and qual ity under adequate soil moisture , this may not apply during periods of soil moisture limitation . Figure 5 . 8 clearly shows that even under mi ld water stress crude protein yields were greatly reduced but were nevertheless capable o f rapid recovery following adequate rainfall . It is interesting to note in Experiment 3 that a lthough the stern was the ma jor contributor to total dry matter under lax cutting , the crude protein yield arose largely from the inflorescence . In fact the crude protein yield of leaf plus inflorescence represented more than 5 0 % of the total and indicates the h igh quality of this species even under infrequent cutting . This was a l so i l lustrated in the uncut control treatment of Experiment 2 ( Figure 2 . 9 ) and suggests that cutting for hay at a later stage of growth can achieve high dry matter of high qual ity . . However , in Experiment 5 , where plants were sub jected to water stress , results showed that the increase in crude protein in mildly stres sed plants arose mainly from the stem and to a les ser extent the leaf , whi le the inflorescence which was severely reduced in development contributed only a small amount to the total crude protein yield . However , on re-watering the crude protein yield of the leaf and the inflorescence showed a marked increase during the recovery period . 1 5 0 E . MANAGEMENT RECOMMENDATIONS The f indings from this study suggest that Verano style is capable of producing acceptable dry matter yields of high qual ity under cutting or graz ing in tropical monsoonal cl imates such as that o f Thailand , provided that an adequate number and s i z e of primary branches is retained . The results also indicate a high tolerance to water stress and an ability to recover f rom drought . In order to achieve good establ i shment of the legume it is important to ensure good seedbed preparation , appropriate fertili zer appl ication , weed and insect control and correct inoculum . Since Verano style has a relatively high percentage of hard seed , scarification using sand paper or soaking in hot water at 80°C for 1 0 minutes is necessary and effective . Verano stylo should also be sown early in the wet season when soil moi sture is adequate for germination and emergence . Although it was not a consideration in this study , Verano style is normal l y sown in a mixture with appropriate tropical gra sses and subsequently cut for feeding directly or for storage as hay or grazed in s itu . E . 1 Cutting In a " cut and carry " system , Verano style can achieve relatively high production of high quality provided it is cut laxly at approximately 6 weekly intervals . From the current experiments , it is possible to recommend a cutting intens ity down to the 7 th node on the main stem to achieve this level of production . However , in the f ield under practical conditions , such a requirement is somewhat academic . Never­ theless in view of the evenness of the stand , as found in the f ield experiment at Muaklek ( Plate 6 . 3 and 6 . 4 ) , it may well be possible for the farmer to cut within an acceptable range of the optimum - between the 6th and the 8th node - and still achieve high yields . 1 5 1 Results also highlighted the importance o f retaining primary branches of adequate s i z e for regrowth . From a practical stand point the recommendation of cutting to a mean level approximating the 7th node on the main stem should a l so ensure this residual of adequate branch size . These primary branches pro j ecting into the upper regions of the canopy at later cutting wil l certainly be defoliated but should also retain the residual s i ze recommendation { i . e . at least 4 nodes remaining ) . Under a cut and carry system it also appears poss ible to achieve even higher dry matter production under relatively lax cutting when the plant is defoliated at a much later stage of growth - approximating maximum yield . Although qual ity { crude protein ) is still relatively high the extra is largely due to greater stem production which palatable to stock . The ma j or l imitation to such production may be less late and a l so - of recovery cutting is yields and to infrequent cutting i s the 3 to 4 weeks . Nevertheless s l ow initial provided the lax and adequate time is allowed for recovery high quality can be achieved . In marked contrast to the plant responses recorded under adequate soil moisture , the reaction of plants to s imi lar defoliation under soil moisture stress was almost nil . This suggests that under such climatic conditions , the farmer i s able to adopt a much more flexible cutting practice without a f fecting production - a f lexibility that he can even maintain up to 3 weeks of recovery growth e . g . the first cut , fol lowing good rains . E . 2 Graz ing Possibly the most important f inding in thi s series of experiments is the comple�ely opposite plant response obtained depending on the method of defoliation . Under cutting , 6 weekly defoliation produced significantly higher yields than 3 weekly defoliation , whereas under graz ing , 4 weekly defol iation was markedly superior in yield to 8 weekly defoliation . 1 5 2 It i s wel l known that grazing is a very dif ferent form of defoliation from cutting - as i t can be highly selective between plants and plant parts , it includes the animal impacts of treading and excreting and it is more di f f icult to impose preci se intensities and frequencies of defoliation compared with cutting . A further factor associated with grazing i s the ability of the plant to modi fy its growth form in response to animal prehension . This reaction was strikingly evident in the f ield study , where the more frequently grazed sward ( approxi­ mately 4 weekly ) developed a more prostrate habit of growth whi le the infrequently grazed sward ( approximately 8 weekly ) maintained an erect and semi-erect habit . With more frequent removal of the main stem under 4 weekly grazing , the primary branches became more abundant and productive as they grew at a more acute angle closer to the ground and hence better able to avoid tota l prehension by the graz ing animal . In contrast the more erect growth under 8 weekly grazing suf fered more complete defoliation at grazing l eaving a rather " nude " main stem with fewer residual branches and buds for regrowth . A further factor contributing to the relatively poor productivity of the infrequently grazed sward was the much greater decl ine in Verano stylo plant density with time . From an initial and very s imi lar plant density o f approxi­ mately 1 7 0 p lants/m2 at the first graz ing , the f requently grazed sward ended the season 20 weeks l ater with 6 2 plants/m2 while the infrequently grazed ended with only 2 0 plants/m2 ( Table 6 . 1 ) . It i s suggested that the various f actors of plant competition were more damaging to plant density under infrequent graz in9 than under frequent _ graz ing . This "opening up " of the sward , particularly under infrequent grazing , obviously accounted for the significant ingress of weeds recorded in the treatment by the end of the season approximately 8 0 % compared with approximately 4 0 % under frequent graz ing ( Figure 6 . 1 0 ) . In terms of the practical recommendation to farmers , it i s clear that frequent rather than infrequent graz ing should be encouraged on the grounds of both quantity and quality of 1 5 3 dry matter and the probable improvement in forage accepta­ bility and intake due to the greater amount of leaf produced . Again , the di f ficulty and very practical que stion of grazing intensity must be faced - and it is recommended that graz ing intensity should be ad justed to leave approximately 7 residual nodes on the main stern after grazing . As stated earl ier , thi s may wel l be feasible in view of the relative evenness of sward that can be established - and i f nece ssary , maintained by mechanical trimming . Obviously farmers must not a llow the sward to become rank a s it wil l lead to lower production and more rapid sward deterioration through legume death and ingress of weeds . In terms of the recommendations relating to the graz ing management of Verano stylo at the early stage of the drought period , it appears that farmers may be able to graze these stands to a lower level ( e . g . down to the 5th node rather than the 7th node ) without causing a s ignificant reduction in regrowth during the remainder of the drought and without restricting recovery following the onset of the early wet season rains . On the other hand s evere grazing ( down to the 3 rd node ) during the early drought period wil l depress regrowth s ignificantly dur ing the remaining drought period at a time when even small additions of quality feed may be highly important . However , such a practice doe s not appear to ef fect recovery growth detrimentally with the subsequent onset of rains . This s eries of experiments did not attempt to address the question of management of mixed legume/gras s pastures generally found in practice . Hence caution must be shown in attempting to extrapolate these current findings from pure Verano swards to Verano/gras s pastures . Nevertheless it is argued that in view of the paramount importance of legumes in tropical pastures , the productivity and pers istency of the legume component should be favoured . Such an emphas i s could wel l lead to tropical Verano/grass pastures of high animal performance through the production and maintenance of feed of high quality , high acceptabil ity and high digestibil ity , even though thi s may be at the expense of some dry matter production . 1 5 4 Appendix 1 Control led environment conditions . Temperature : Humidity : Lighting : day/night 7 0/ 9 0 + 5 % RH 1 3 /2 mb VPD day/night 1 2 hour photoperiod W m-2 PAR* pre 1 5 9 post 1 3 5 mean 1 4 7 - 2 - 1 uE . m . sec 7 4 2 6 4 0 6 9 1 * 4 0 0 - 7 0 0 nm photosynthetical ly active radiation . 1 5 5 Appendix 2 Climate laboratory - N .C .S .U . Phytotron nutrient 2 ml A + 2 ml B/ litre water r-blecular weight ( g ) Stock solution A Ammonium nitrate NH4oofr Calcium ni ate Ca(00�2 X 4 H20 Sequestr e 330 1 0% DTPA Na Fe Stock solution B Potassium phosphate KH2P04 Potassium phosphate K2HP04 Potass1um nitrate KNO .Magnesi� sulphate MgS04 X 7H20 Sodium sulphate NafS04 Zinc su phate Znso4 x 7H 0 ' Manganese chforide MnCl Copper s�phate �so4 � 5H2o Bor1c ac1d H3BOf Sodium rro ybdate NaMc04 . x 2H2o Nutrient N p K s ea Fe Mg PPM 1 1 1 . 55 7 . 65 6 1 . 54 24 . 06 54 . 06 5 . 96 6 . 08 80 . 2 236 . 1 5 468 . 20 1 36 . 08 1 74 . 1 7 1 0 1 . 1 1 246 . 50 1 42 . 05 287 . 55 1 97 . 92 249 . 68 6 1 . 82 24 1 . 93 urn NH4 2200 003 6 1 30 P04 6 . 8 80 804 750 1 280 1 25 250 pH of final solution = 6 . 5 - 7 . 5 grams/litre Cone . 80 . 05 1 59 . 25 29 . 8 1 2 . 5 5 . 5 63 . 9 30 . 8 1 35 . 5 0 . 025 0 . 26 0 . 0 1 0 . 35 0 . 0027 Nutrient B Mn Cu Zn r-b Cl Na Final PPM soln . • 1 60 1 N ea . 3 1 85 N Fe . 0596 Na K . 025 p K • 0 1 1 p K • 1 278 N Mg . 06 1 62 s Na . 071 s Zn . 00005 s Mn . 00052 Cl Cu . 00002 s . 0007 B Na . 0000054 r-b PPM urn 0 . 1 23 803 0 . 1 44 0 . 005 0 . 0 1 1 0 . 002 0 . 1 86 25 . 9 1 1 56 . 05 54 . 06 37 . 79 5 . 96 2 . 93 7 . 1 8 5 . 69 4 . 94 1 . 96 49 . 42 1 7 . 7 1 6 . 08 8 . 02 22 . 98 1 6 . 03 0 . 0 1 1 0 . 006 0 . 1 44 0 . 1 86 0 . 005 0 . 003 0 . 1 23 0 . 00 1 0 . 002 1 1 . 39 2 . 55 0 . 08 0 . 1 7 0 . 02 5 . 36 1 1 25 t- z - er 0 ..J er 50 0 1-z C( 40 _, Cl. _,