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. 
 
EFFECTS OF PHYSICAL AND BIOCHEMICAL 
CHARACTERISTICS OF CONTRASTING 
LEGUME SWARDS ON SELECTIVE BEHAVIOUR 
OF GRAZING CATTLE 
A thesis presented in partial fulfilment of the 
requirements for the degree of 
Doctor of Philosophy (Ph.D.) 
Institute of Natural Resources 
College of Sciences 
Massey University 
Palmerston North, New Zealand 
CESAR H. E. C. POLl 
1998 
ABSTRACT 
In order to assess the effects of sward physical characteristics and secondary compound 
concentration on cattle ingestive behaviour and diet selection, two sets of experiments 
were carried out using two legumes [birdsfoot trefoil (Lotus corniculatus L.); red clover 
(Trifolium pratense L.)] with approximately neutral partial preference. The fIrst set 
investigated the ingestive behaviour and diet selection response to manipulation of sward 
area, maturity and height using alternating sward strips. The second set tested the effects of 
plant morphology and secondary compound concentration on preference using sequences 
of spaced plants. 
The fIrst set of four experiments was conducted at Agresearch Flock House, near Bulls. 
Yearting heifers in groups of three grazed a sward formed by alternate 2.4 m wide strips 
of a mixture of birds foot trefoil cv. Goldie and white clover (Trifolium repens L.) cv. Pitau, 
and strips of red clover cv. Colenso. The experiments were formed by combinations of 
four treatments and fIve groups of animals over four successive three-day periods in a 
Row-Column Design balanced for previous treatment. This design was used to estimate 
the difference between periods, the difference between groups of heifers and the effect of 
previous treatments. Observations of the distribution of grazing activity and biting rate 
were made over 3-hour periods each evening. The distribution of grazing activity 
assessed the changes during three days of grazing (55 hours). The effects of contrasting 
areas, maturity and height of the alternate swards were examined in Experiments 1 ,  2 
and 3, respectively. In the fIrst experiment four treatments were imposed, the area ratio 
in percentage of each sward per treatment being: 20:80; 33:67; 67:33; 80:20. In 
Experiment 2 the treatments provided four combinations of maturity (immature/mature) of 
the two swards. In Experiment 3 the treatments were arranged in order to compare 
contrasts in height at the same vegetative stage of growth for the two swards. Experiment 4 
was a small trial that showed that the proximity of a particular sward to the perimeter 
fence did not influence the proportion of grazing time spent on that sward. 
Abstract iii 
The results of Experiments 1, 2 and 3 demonstrated that the physical contrasts between 
swards imposed by the treatments, and the variation in herbage mass and sward surface 
height between the fIrst and third day of grazing, had important effects on selective 
behaviour. The effect of relative sward area was demonstrated to be important mainly 
when herbage mass and sward height were high, when the animals showed preferential 
selection for the sward of smaller area irrespective of which species was present with 
smaller relative area. The sward maturity effect was closely related to the preference for 
leaves and rejection of stems, though as the herbage mass and height decreased, the 
selection for leaf was offset by a selection of greater sward height and bulk density. The 
animals showed selection for taller and greater herbage mass swards, however, at high 
levels of herbage mass and height selectivity was reduced by the preference for a mixed 
diet. An overall analysis of the three experiments showed that there was a general partial 
preference for the two swards close to 50:50, though preference for birdsfoot trefoil was 
lower in Experiment 2 (40:60) than in either Experiment 1 or 3 (close to 50:50). This effect 
was mainly related to sward maturity and also indicated a need for further research on the 
effect of secondary compounds on animal preference. 
The second set of experiments, Experiments 5 and 6, were conducted at Massey University 
and Agresearch, Palmerston North. In these experiments the response of grazing animals 
to contrasts in plant morphology and specifIc plant secondary compounds were 
examined in trials in which trained dairy cows grazed spaced plants of two "genotypes" 
(one accession and one cultivar) of birdsfoot trefoil with high or low concentration of 
extractable condensed tannins (BeT) (PI273938 and Goldie, respectively) and two 
"genotypes" (cultivars) of red clover with high or low formononetin concentration 
(Pawera and G-27, respectively). Plants were established in 4 linear sequences of 26, 
each providing three blocks (replicates) of balanced sets of 2 plant species, 2 genotypes 
(within each species), and plants either not trimmed or trimmed to minimise physical 
differences between genotypes within species. The plant sequences in Experiment 5 
were grazed by four lactating cows and in Experiment 6 by two rumen-fIstulated dry 
Friesian cows. In Experiment 6 the effects of rumen manipulation on preference were 
also tested by inserting minced material into the cow's rumen through the fistula to 
provide contrasts of low [birdsfoot trefoil (Lotus comiculatus L.) cv. Goldie] and high 
Abstract iv 
[lotus maku (Lotus pedunculatus L.)] concentrations of condensed tannin, and low [red 
clover (Trifolium pratense) cv. Astred] and high [red clover cv. Pawera] concentrations of 
formononetin. 
The results of Experiments 5 and 6 demonstrated that the animals showed an immediate 
preference for large, dense and leafy plants. High concentrations of ECT also had an 
important negative effect on preference for birdsfoot trefoil, but this effect was 
confounded with a positive effect of plant morphology, mainly proportion of leaf. 
Formononetin did not have an important effect on preference of cattle. 
The overall analysis of the six experiments showed that there was a relatively stable 
partial preference between birdsfoot trefoil and red clover demonstrating neutrality in 
preference between these two species. However this stability was sensitive to changes in 
sward area, plant morphology, sward structure (height and herbage mass) and secondary 
compound concentration. Observations showed that the animals did not graze randomly, 
but with the objective of obtaining a mixed diet. In tall, high mass and similar stage of 
maturity swards, the animals grazed preferentially the sward offered in smaller area or 
lower mass offered. In this context, the importance of leaf/stem ratio and high EeT 
concentration in affecting selection showed scope for manipulation of preferential 
behaviour through manipulation of the plant attributes. Improvement of leaf/stem ratio 
of birdsfoot trefoil and red clover, and reduction of ECT concentration in birds foot 
trefoil could therefore have a practical effect on animal preference. The preference for a 
mixed diet and the adjustment of this behaviour as sward conditions changed can be 
explained by interactions between three possible hypotheses: (i) animals tried to obtain 
a balanced diet; (ii) animals selected swards that provided the potentially higher rate of 
intake; (iii) animals sampled to constantly reinforce awareness of sward conditions. 
This thesis is dedicated to my wife Beatriz 
ACKNOWLEDGEMENTS 
I would like to express my deepest gratitude to my chief supervisor, Professor John 
Hodgson for his excellent help and guidance in organizing and arranging a Ph.D. 
program for me to pursue, directing my research, commenting on manuscripts, and 
making me realize the importance of plant-animal interactive research. 
My grateful thanks to my co-supervisor, Dr. Gerald Cosgrove for his encouragement 
and enthusiastic supervision and helpful advice, and Mr. Greg Arnold for the valuable 
statistical advice to analyze the data collected during the experiment. 
Thanks to CAPES (Brazilian Ministry of Education) for the scholarship that allowed me 
to have an excellent experience in New Zealand. 
Sincere thanks to AgResearch and Massey University for providing area, animal 
laboratory and financial support that allowed me to run my research and do my 
chemical analyses. 
I would also wish to express my grateful thanks to AgResearch staff Dr. Gary Waghorn, 
Dr. Warren McNabb, Dr. Reg Keogh, Mr. Jason Peters and Dr. Abdul Molan for their 
co-operative manner, willingness to help, friendly attitude and prompt dealing with 
problems. 
I am greatly indebted to Mr. Craig Anderson for his technical assistance on the field and 
for his friendship. Thanks to Mr Terry Lynch (in memory), Mr. Mark Osborne and Mr 
Roger Levy for providing field technical assistance in my research. 
Thanks for the support and friendship provided by the staff of the Pastoral Science 
Group and Animal Science Department, Massey University, specially to Dr. Cory 
Matthew, Mr. Parry Matthews, Dr. Colin Holmes, Dr. Peter Kemp, Dr. Ian Valentine 
Acknowledgements vii 
Mrs Kathy Hamilton, Mrs Hera Kennedy, Mr. Matt Alexander and Mr. Ruwan 
Dissanayake. 
I wish to thank Dr. Aino A. Jacques (Professor of Pastoral Science) from the Faculty of 
Agronomy, UFRGS, Brazil for his guidance and encouragement in the field of pasture 
science. 
I am grateful to many friends for the help and support during my studies. Your help 
made this thesis possible. Special thanks must be due to all postgraduate students of the 
pastoral and animal science group, particularly Alberto Torres, Aurelio Guevara, 
Carolina Realini, Claudio Machado, Daniel Laborde, Daniel Real, Fabio Montossi, 
Fulton Hughes, Greg Bishop-Hurley, Ignacio Lopez, Nigel Johnston, Fuyuan Liu, Luis 
Barioni, Mauricio Padilla, Mark Hyslop, Naba Devkota, Silvia A ssuero, Stephanie 
Bluett, Suzanne Hodgkinson, Walter Ayala, Wagner Beskow, Wendy Griffiths and 
others. Thank you for your friendship. 
I wish to thank Mrs Midge and Mr Henk Janssen, part of my family in New Zealand. 
People like them are making the world a better place to live in. 
My Dad, example of friend, father, agronomist and researcher, he taught me that 
nothing is impossible, but it is important to work hard. 
A big thanks to my Mum and sisters for their great love, encouragement and education 
throughout my life. To my parents in-law for their love and support. 
An immense appreciation is to my wife Beatriz for her encouragement to come to New 
Zealand. This thesis was only possible because of her love, dedication, physical and 
mental support during my studies, being always by my side. She is an example of wife, 
mother and friend that went beyond physical possibilities. Thanks to my little daughter 
for the happiness she brought home. 
Thanks also to all people who offered help to me but I failed to name individually here. 
TABLE OF CONTENTS 
Abstract. ......................................................................................................................... ii 
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  vi 
Table of Contents . . . . . . . . . . . . . . . . . . .. . . . . . . .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  viii 
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . .... . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  xviii 
L· f F' 
. . .  1st 0 19ures ......................................................... ...................... .......................... XXVlll 
List of Plates . . . . . . . . . . . . . . . . . . . . . . . . .. .. .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  xxxii 
L· f A  d' 
. . .  1st 0 ppen Ices .......... ................................................. .... . ............................ ...... XXXlll 
CHAPTER 1: INTRODUCTION AND OBJECTIVES .......................................... 1 
CHAPTER 2: LITERATURE REVIEW .................................................................. 4 
2. 1 .  IN'fRODUCTION ............... ....................... ....... ................................ .. .............. . 4 
2.2. DIET SELECTION BY GRAZING ANIMALS . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  5 
2.2. 1 .  Models of diet selection ....................................................................... , . . . . . .  5 
2 .2.2.  Role of senses in diet selection ..................... . ............................................ 10  
2.2.2. 1 .  Sight. ......................................................................... ................. ........ 1 1  
2.2.2.2. Taste . .............................................. ......................... ........................... 1 2  
2.2.2.3.  Smell .......................................... ..... . ...... ............................................. 1 3  
2.2 .2.4. Touch ..........
............................. ........................................................... 1 5  
2.3 .  EFFECT OF SWARD CHARACTERISTICS ON GRAZING BEHAVIOUR, 
DIET SELECTION AND HERBAGE INTAKE .......................... .......... ......... 1 5  
2 .3 . 1 .  Spatial variability affecting grazing behaviour and diet selection ........... 16  
2.3.1.1. Plant part level ............................. ......... . ............................................. 1 7  
Table of Contents lX 
2.3 .1.2. Plant level .................................................... ......... .............................. 18 
2.3 .1.3 .  Patch level .......................................................................................... 18 
2.3 .1.4. Landscape level ......................... ....... .................................................. 19 
2.3 .2. Spatial heterogeneity and the process of diet selection ............................. 19 
2.3 .3 .  Temporal variability affecting grazing behaviour and diet selection ........ 22 
2.3 .4. Effect of physical sward characteristics on herbage intake and diet 
selection 
............................................. . ......................... ........................ ...... 24 
2.3 .5 .  Effect of sward nutritional characteristics on herbage intake .. ................. 25 
2.3 .6.  Effect of biochemical characteristics on grazing behaviour, diet selection 
and herbage intake ............................ ........................................................
. 29 
2.3 .6.1. Effect of plant secondary compounds on grazing behaviour and diet 
selection ............................................. ................. '" . . . . . . . . . .. .  '" . . . . . . . . . . . . . .  31 
2.3 .6.2. Effect of condensed tannins on grazing behaviour and diet selection 32 
2.3 .6.3 .  Effect of formononetin on grazing behaviour and diet selection ....... 36 
2.4. CONCLUSIONS ................................................ .............................................. 39 
CHAFfER 3: EXPERIMENTS 1, 2, 3 AND 4 ....................................................... 41 
3.1. INTRODUCTION .................................. . ..... ................................................. ... 41 
3 .2 .  MATERIAL AND METHODS .................................................. ..................... 42 
3 .2.1. Experimental site ... . . . . . . ................................................ .............. ................ 43 
3 .2.2. Swards ................................................... ... ..... . ... . ....................................... 43 
3 .2 .3 .  Design ........................................................................................................ 45 
3 .2.4. Animals ................... ........ ................ . ....... ......... . ................. ....................... 46 
3.2 .5 .  Measurements .............................................................................. . ....... . . . ... 46 
Table of Contents 
3.2.5.1. Sward measurements .......................................................................... 46 
Herbage mass ................................................ ............................................... 46 
Botanical composition 
................................................................................. 46 
Pasture height and bulk density ................................................................... 47 
Pasture structure 
.......................................................................................... 47 
3.2.5.2. Grazing Behaviour .............................................................................. 47 
3.2.6. Chemical analysis ...................................................................................... 48 
3.2.7. Statistical analysis ... .................................................................................. 49 
3.2.8. Experimental layouts 
................................................................................ 50 
3.2.8.1. Experiment 1 ...................................................................................... 50 
3.2.8.2. Experiment 2 ...................................................................................... 50 
3.2.8.3. Experiment 3 ...................................................................................... 51 
3.2.8.4. Experiment 4 .............................. ........................................................ 51 
3.3. RESULTS ......................................................................................................... 56 
3.3.1. Experiment 1: Effect of the proportion of area of birds foot trefoil (Lotus 
comiculatus L.) and white clover (Trifolium repens L.) sward in relation to 
red clover (Trifolium pratense L.) sward on grazing behaviour, diet selection 
and herbage intake . .................................................................................... 56 
3.3.1.1. Sward measurements .......................................................................... 56 
Herbage mass, sward surface height and sward bulk density . .................... 56 
Sward composition ...................................................................................... 57 
3.3.1.2. Canopy structure within the sward ......... ............................................ 59 
3.3 .1.3. Sward chemical composition ........... ................................................... 59 
Extractable condensed tannin concentration 
............................................... 59 
Forrnononetin concentration ........................................................................ 62 
x 
Table of Contents 
General chemical composition ................................................................ ... , 63 
3.1.1.2. Animal measurements ........................................................................ 64 
Grazing time and intake ............................................................. ................ 64 
Rate of biting ............................................................................................... 71 
xi 
3.3.2. Experiment 2: Effect of the maturity of birds foot trefoil (Lotus corniculatus 
L.) and white clover (Trifolium repens L.) in relation to red clover (Trifolium 
pratense L.) on grazing behaviour, diet selection and herbage intake . ...... 72 
3.3.2.1. Sward measurements .......................................................................... 72 
Herbage mass, sward surface height and sward bulk density ..................... 72 
Sward composition ...................................................................................... 74 
3.3.2.2. Canopy structure within the sward ..................................................... 76 
3.3.2.3. Sward chemical composition .............................................................. 79 
Extractable condensed tannin concentration ............................................... 79 
Formononetin concentration ........................................................................ 79 
General chemical composition ............. , ...................................................... 80 
3.3.2.4. Animal measurements 
........................................................................ 82 
Total grazing time and intake ...................................................................... 82 
Grazing time per kg of DM offered . ............................................................ 85 
Rate of Biting .................................................. . ........................................ '" 86 
3.3.3. Experiment 3: Effect of height of birds foot trefoil (Lotus corniculatus L.) 
and white clover (Trifolium repens L.) in relation to red clover (Trifolium 
pratense L.) on grazing behaviour, diet selection and herbage intake . ...... 89 
3.3.3.1. Sward measurements .......................................................................... 89 
Herbage mass, sward surface height and sward bulk density ..................... 89 
Sward composition ...................................................................................... 91 
Table of Contents XlI 
3.3.3.2. Canopy structure within the sward ..................................................... 93 
3.3.3.3. Sward chemical composition .............................................................. 93 
Extractable condensed tannin concentration ............................. '" ............... 93 
Formononetin concentration ........................................................................ 96 
General chemical composition .................................................................... 96 
3.3.3.4. Animal measurements ........................................................................ 98 
Total grazing time and intake ...................................................................... 98 
Grazing time per kg of DM offered .......................................... . ................ 102 
Rate of biting .......................... ................................................................... 103 
3.3.4. Experiment 4: Effect of position in the plot of birdsfoot trefoil (Lotus 
corniculatus L.) and white clover (Trifolium repens L.) sward in relation 
to red clover (Trifolium pratense L.) sward on grazing activity 
distribution . ........................................ ...................................................... 105 
3.3.3.1. Sward measurements ........................................................................ 105 
Herbage mass, surface sward height and sward bulk density ................... 105 
3.3.3.1. Animal measurements .......................................................................... 106 
Grazing time .................................. ............................................................ 106 
3.4. DISCUSSION ................................................................................................. 107 
3.4.1. Evaluation of experimental procedures ................................................... 107 
3.4.2. Definition: Selection and Preference ....................................................... 111 
3.4.3. Components of ingestive behaviour and selection .................................. 111 
3.4.4. Pattern of selection across experiments ........... ........................................ 114 
3.4.5. Comparison of diet selection in Experiments 1,2 and 3 ......................... 117 
Effect of alternative sward ......................................................................... 119 
Table of Contents xiii 
3.4.6. Diet selection changes over time (from Day 1 to Day 3) ........................ 119 
3.4.7. Diet selection in Day 1 (high herbage mass and height) ......................... 120 
3.4.8. Diet selection within each sward .............................................................  124 
3.5. CONCLUSIONS ............................................................................................ 125 
CHAPTER 4: EXPERIMENTS 5 AND 6 ............................................................. 127 
4.1. INTRODUCTION .......................................................................................... 127 
4.2. MATERIAL AND METHODS ..................................................................... 128 
4.2.1. Experimental site ..................................................................................... 128 
4.2.1.1. Experiment 5 .................................................................................... 128 
4.2.1.2. Experiment 6 .................................................................................... 129 
4.2.2. Glasshouse sowing and management ...................
................................... 129 
4.2.3. Site preparation and management ........................................................... 130 
4.2.4. Design ...................................................................................................... 131 
4.2.5. Animals and sequence allocations ........................................................... 134 
4.2.5.1. Experiment 5 .................................................................................... 134 
Animals ..................................................................................................... 134 
Sequences allocation ................................................................................. 134 
4.2.5.2. Experiment 6 .................................................................................... 135 
Animals 
........
............................................................................................. 135 
RUMEN MODIFICATION .................................................................. 135 
Sequence allocation ........... ........................................................................ 136 
4.2.6. Grazing behaviour assessment ................................................................ 137 
Table of Contents xiv 
4.2.7. Plant assessment ...................................................................................... 139 
Plant height ................................................................................................ 139 
Plant diameter ............................................................................................ 139 
Plant density .............................................................................................. 139 
Leafiness .................................................................................................... 139 
Habit .......................................................................................................... 140 
Herbage mass ............................................................................................. 140 
4.2.8. Morphological and chemical analysis ..................................................... 140 
4.2.8.1. Morphlogical analysis ...................................................................... 141 
4.2.8.2. Chemical analysis ............................................................................. 141 
4.2.9. Statistical analysis ................................................................................... 141 
4.2.9.1. Experiment 5 .................................................................................... 142 
4.2.9.2. Experiment 6 
.................................................................................... 142 
4.2.9.3. Correlation Analyses ........................................................................ 142 
4.2.9.4. Covariance analyses ......................................................................... 143 
4.3. RESULTS ............................................................................ ........................... 144 
4.3.1. Experiment 5: Effect of condensed tannin in birds foot trefoil (Lotus 
comiculatus L.) and formononetin in red clover (Trifolium pratense L.) on 
preference and grazing behaviour of dairy cows . .................................... 144 
4.3.1.1. Plant characteristics .......................................................................... 144 
Comparisons between species ................................................................... 144 
Comparisons between cultivars of the same species ................................. 145 
BIR.DSFOOT TREFOJI., ....................................................................... 145 
RED CLOVER .............................................................. ........................ 148 
Table of Contents xv 
4.3.1.2. Sward chemical composition ............................................................ 150 
Extractable condensed tannin concentration ............................................. 150 
Formononetin concentration ...................................................................... 151 
General chemical composition .................................................................. 151 
4.3.1.3. Number of bites per plant ................................................................. 152 
4.3.1.4. Rate of Biting ................................................................................... 154 
4.3.1.5. Correlation Analyses ........................................................................ 155 
Number of bites vs General plant chemical characteristic ...................... 156 
ECT concentration vs Plant morphology and Number of bites in birdsfoot 
trefoil plants ............................................................................................... 157 
Formononetin concentration vs Pant morphology and Number of bites in red 
clover plants ................................................. .............................................. 157 
Number of bites vs Plant morphology ..................................................... 158 
BIR.DSFOOT TREFOIL ....................................................................... 158 
RED CLOVER ...................................................................................... 159 
4.3.1.6. Use of Covariates ............ ................................................................. 160 
Genotype effect (class variable) vs Plant morphological characteristics 
(covariate) .................................................................................................. 160 
TRIMMED AND UNTRIMMED PLANTS ......................................... 160 
UNTRIMMED PLANTS ...................................................................... 164 
EeT concentration (covariate) vs Plant morphological characteristics 
(covariate) .................................................................................................. 167 
4.3.2. Experiment 6: Effect of condensed tannin in birds foot trefoil (Lotus 
comiculatus L.) and formononetin in red clover (Trifolium pratense L.) on 
diet selection and grazing behaviour of dairy cows: rumen content 
modification approach ............................................................................. 170 
4.3.2.1. Plant characteristics ....... ................................................................... 170 
Table of Contents XVI 
Comparisons between species ................................................ , .................. 170 
Comparisons between genotypes within each species .............................. 172 
BIRDSFOOT TREFOIL ....................................................................... 172 
RED CLOVER ...................................................................................... 176 
4.3.2.2. Sward chemical composition ............................................................ 179 
Extractable condensed tannin (BCT) ......................................................... 179 
Formononetin concentration ...................................................................... 179 
General chemical composition .................................................................. 180 
4.3.2.3. Number of bites per plant. ................................................................ 181 
4.3.2.4. Rate of biting .................................................................................... 185 
4.3.2.5. Correlation Analyses ........................................................................ 186 
Number of bites vs General plant chemical characteristics ..................... 186 
ECT concentration vs Plant morphology and Number of bites in birdsfoot 
trefoil plants ............................................................................................... 187 
Formononetin concentration vs Plant morphology and Number of bites in 
red clover plants ........................................................................................ 188 
Number of Bites vs Plant Morphology .................................................... 188 
BIRDSFOOT TREFOIL ....................................................................... 188 
RED CLOVER ...................................................................................... 189 
4.3.2.6. Use of covariates .............................................................................. 189 
Genotype effect (class variable) vs Plant morphological characteristics 
(covariate) 
.................................................................................................. 190 
TRIMMED AND UNTRIMMED PLANTS ......................................... 190 
UNTRIMMED PLANTS .................... .................................................. 193 
ECT concentration (covariate2 vs Plant morphological characteristics 
(covariate) 
.................................................................................................. 196 
Table o/Contents XVll 
4.4. DISCUSSION ................................................................................................. 199 
4.4.1. Evaluation of experimental procedures ................................................... 199 
4.4.2. Plant morphological characteristics ......................................................... 204 
4.4.3. Plant chemical composition ..................................................................... 205 
4.4.4. Grazing behaviour ................................ ................................................... 209 
4.5. CONCLUSIONS ............................................................................................ 217 
CHAPTER 5: 
GENERAL DISCUSSION ...................................................................................... 219 
GENERAL CONCLUSIONS ................................................................................. 225 
REFEREN CES ...............................................................................................•........ 228 
APPEND I CES ......................................................................................................... 268 
LIST OF TABLES 
Table 3.1. Distribution of four treatments with five groups of three heifers over 
four periods ................................................................................................... 45 
Table 3.2. Herbage mass, sward height and bulk density before and after 
grazing, and estimation of the herbage mass removed for birds foot 
trefoil and white clover (BW) and red clover (RC) swards in 
Experiment 1 . ................................................................ '" ............................ 57 
Table 3.3. Botanical composition of birdsfoot trefoil and white clover (BW) and 
red clover (RC) swards before and after grazing (DM basis): (a) 
percentage of components in live fraction, (b) percentage of live 
matter in total DM of each sward and ( c) ratio of the total live matter 
of birdsfoot trefoil and white clover (BIW) in the BW sward, 
Experiment 1 . ............................................................................................... 58 
Table 3.4. Extractable condensed tannin (ECT) concentration (%) of birdsfoot 
trefoil leaf and stem in Experiment 1 (DM basis) . ....................................... 59 
Table 3.5. Formononetin concentration (%) of leaf, petiole, stem and flower of 
birdsfoot trefoil (BT), white clover rvvC) in birdsfoot trefoil and 
white clover sward (BW), and red clover (RC) in red clover sward 
(RC) in Experiment 1 . ....................................................................... ........... 62 
Table 3.6. Crude protein (ep), lipid, acid and neutral detergent fibre (ADF, 
NDF), carbohydrates (soluble sugars plus starch)(CHO), ash and in 
vitro dry matter digestibility (IVDMD) determined by Near Infrared 
Reflectance Spectroscopy (NJRS) of the main components of 
birdsfoot trefoil and white clover, and red clover swards in 
Experiment 1 (percentage of OM basis) ....................................................... 63 
Table 3.7. The effect of treatments (area ratios) on grazing time (minutes) in the 
first, second and third days of grazing (Days 1, 2 and 3), and average 
DM intake per animal per day (kg drnlhdJday) during 55 hours of 
grazing in Experiment 1 
. ....................................... ....................................... 65 
Table 3.8. Treatment (20, 33, 67 and 80 % of the total area offered) effects on the 
proportion of grazing time (in relation to the total grazing time spent 
in plot) devoted to birdsfoot trefoil plus white clover swards (BW) in 
Experiment 1 . ....................................................................................... ........ 67 
List of Tables 
Table 3.9. Regression slopes of the proportion of grazing time (propngt) in 
relation to the proportion of area (propnarea) and dry matter (propndm) 
offered in the first, second and third days of grazing observation in 
Experiment 1 (slope significance in relation to neutrality value of 
xix 
1.0) ................................................................................................................ 67 
Table 3.10. The effect of swards of birds foot trefoil and white clover (BW), and 
red clover (RC) on rate of biting (bites/minute) in the first, second 
and third days (total 55 hours) of grazing assessment in Experiment 
1 . ................................................................................................................... 71 
Table 3.11. Herbage mass (kg DMlha), sward height (cm) and bulk density (mg 
DM/cm3) before and after grazing, and estimation of the herbage 
mass removed (kg DMlha) of birdsfoot trefoil and white clover (BW) 
and red clover (RC) swards according to treatment in Experiment 2 . ......... 73 
Table 3.12. Botanical composition of birdsfoot trefoil plus white clover (BW) 
and red clover (RC) swards before and after grazing, according to the 
treatments (plant maturity: Immature and Mature) (DM basis): (a) 
percentage of components in live fraction, (b) percentage of live 
matter in total DM of each sward and (c) ratio of the total live matter 
of birdsfoot trefoil and white clover (BIW) in the BW sward, 
Experiment 2 . ............................................................................................... 75 
Table 3.13. The effect of extractable condensed tannin (ECT) concentrations of 
birds foot trefoil leaf and stem in Experiment 2 (%DM basis) 
according to the treatment. ........................................................................... 79 
Table 3.14. Formononetin concentration (%) of leaf, petiole, stem and flower of 
birdsfoot trefoil (BT), white clover (WC), in birdsfoot trefoil and 
white clover sward (BW), and red clover (RC), in red clover sward 
(RC) in Experiment 2 . .................................................................................. 80 
Table 3.15. Crude protein (CP), lipid, acid and neutral detergent fiber (ADF, 
NDF), carbohydrates (soluble sugars plus starch)(CHO), ash and in 
vitro dry matter digestibility (IVDMD) determined by Near Infrared 
Reflectance Spectroscopy (NIRS) of the main components of 
birds foot trefoil plus white clover, and red clover swards in 
Experiment 2 according to the period of re growth (immature and 
mature) (percentage of D M basis) . ..... ..................... . .................................... 81 
Table 3.16. The effect of swards of birdsfoot trefoil and white clover (BW), and 
red clover (RC) on grazing time (minute) during the three hours of 
grazing assessment in Days 1, 2 and 3 and on average intake per 
animal per day (kg DMlhdlday) in Experiment 2 ......................................... 82 
List of Tables 
Table 3.17. The effect of treatment on total grazing time (minute) during the 
three hours of grazing assessment in Days 1, 2 and 3 and on average 
intake per animal per day (kg DMlhdlday) on birdsfoot trefoil plus 
white clover (BW), and red clover (RC) sward in Experiment 2 
(treatment A= BW and RC immature; treatment B=BW immature 
and RC mature; treatment C= BW mature and RC immature; D= BW 
xx 
and RC mature) ............................................................................................. 83 
Table 3.18. Sward maturity contrast effect on total grazing time and 
intake/animal/day contrasting the effect within each sward type (BW 
or RC) and the effect of alternative sward in the days 1, 2 and 3 of 
grazing assessment in Experiment 2 ............................................................. 84 
Table 3.19. Treatment effects on the proportion of grazing time (in relation to the 
total grazing time spent in the plot) devoted to birdsfoot trefoil and 
white clover swards (BW) in Experiment 2 Treatment A= BW and 
RC immature; Treatment B=BW immature and RC mature; 
Treatment C= BW mature and RC immature; Treatment D= BW and 
RC mature) . .................................................................................................. 85 
Table 3.20. The effect of interaction between treatment (maturity: immature 
(Imm) and mature (Mat» and sward type (birdsfoot trefoil and white 
clover (BW), and red clover (RC) on grazing time per unit of DM 
(minlkg DM), in the first and third days of grazing in Experiment 2 ........... 86 
Table 3.21. The effect of swards of birdsfoot trefoil and white clover (BW), and 
red clover (RC) on rate of biting (bites/minute) in the first, second 
and third days of grazing assessment in Experiment 2 according to 
the treatment. ................................................................................................ 87 
Table 3.22. Sward maturity effect [contrast: immature - mature (negative 
numbers represent greater effect of mature than immature sward)] on 
rate of biting contrasting the effect within each sward type [either 
birdsfoot trefoil and white clover (BW) or red clover (RC)] and the 
effect of adjacent sward in the first, second and third days of grazing 
assessment in Experiment 2 .......................................................................... 88 
Table 3.23. Herbage mass (kg DMlha), sward height (cm) and bulk density (mg 
DMlcm3) before and after grazing, and estimation of the herbage 
mass removed (kg DMlha) of birdsfoot trefoil and white clover (BW) 
and red clover (RC) swards according to the treatment in Experiment 
3 
. ................................................................................................................... 90 
List of Tables 
Table 3.24. Botanical composition of birdsfoot trefoil plus white clover (BW) 
and red clover (RC) swards before and after grazing, according to the 
treatments (plant height: Short and Tall) (DM basis): (a) percentage 
of components in live fraction, (b) percentage of live matter in total 
DM of each sward and (c) ratio of the total1ive matter of birds foot 
xxi 
trefoil and white clover (BfW) in the BW sward, Experiment 3 . . . . . . . . . . . . . . . . . .  92 
Table 3.25. Extractable condensed tannin (ECT) concentration (%) of birds foot 
trefoil leaf and stem in Experiment 3, according to the sward 
characteristic (tall or short) imposed by the treatments (DM basis) . . . . . . ... . . .  93 
Table 3.26. Formononetin concentration (%) of leaf, petiole and stem of 
birdsfoot trefoil (BT), white clover (WC) in birds foot trefoil and 
white clover sward (BW), and red clover (RC) in red clover sward 
(RC) in Experiment 3 . . . . . . . . . . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  96 
Table 3.27. Crude protein (CP), lipid, acid and neutral detergent fibre (ADF, 
NDF), carbohydrates (soluble sugars plus starch) (CHO), ash and in 
vitro dry matter digestibility (IVDMD) determined by Near Infrared 
Reflectance Spectroscopy (NIRS) of the main components of 
birdsfoot trefoil and white clover, and red clover swards in 
Experiment 3 according to height treatment (percentage of DM 
basis) ................................................................................................... .......... 97 
Table 3.28. The effect of treatments (sward height) on grazing time (minutes) in 
Days 1 ,  2 and 3, and average DM intake (kg) per animal per day 
during 55 hours of grazing in Experiment 3, according to the 
treatments . ........................................................... .......... ............................... 99 
Table 3.29. Sward height contrast effect on total grazing time and 
intake/animal/day contrasting the effect within each sward type (BW 
or RC) and the effect of alternative sward in Days 1, 2 and 3 of 
grazing assessment in Experiment 3 .. . . . . . . . . ..... . ........... . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . .  101 
Table 3.30. Treatment effects on the proportion of grazing time devoted to 
birdsfoot trefoil and white clover swards (BW) in Experiment 3 
Treatment A= BW and RC short; Treatment B=BW short and RC 
tall; Treatment C= BW tall and RC short; Treatment D= BW and RC 
tall) .............................................................................................................. 102 
Table 3.31. The effect of treatment on grazing time per kg of dry matter in the 
plot (minlkg DM) in the first and third day of Experiment 3 
(Treatment A= BW and RC short; Treatment B=BW short and RC 
tall; Treatment C= BW tall and RC short; Treatment D= BW and RC 
tall) . .. . . . . .. . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  102 
list of Tables 
Table 3.32. The effect of swards of birdsfoot trefoil and white clover (BW), and 
red clover (RC) on rate of biting (bites/minute) in the first, second 
and third days of grazing assessment in Experiment 3 according to 
xxii 
the treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 03 
Table 3.33. Sward height effect [contrast: tall - short (negative numbers 
represent greater effect of short than tall sward)] on rate of biting 
contrasting the effect within each sward type [either birds foot trefoil 
and white clover (BW) or red clover (RC)] and the effect of adjacent 
sward in the first, second and third days of grazing assessment in 
Experiment 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . .. . . . .. . . . . . . . .  104 
Table 3.34. Herbage mass (kg DMlha), and bulk density (mg DMlcm3) before 
grazing and sward surface height (cm) before and after grazing 
birdsfoot trefoil and white clover (BW) and red clover (RC) swards 
in Experiment 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  105 
Table 3.35. Proportion of time grazing (in relation to the total grazing time spent 
in plot) birdsfoot trefoil plus white clover (BW) and red clover (RC) 
swards according to the strip position (side: close to the fence; 
central: in the middle of the plot) in the plot in Experiment 4 . . . . . . . . . . . . . . . . . . .  106 
Table 3.36. Overall averages of the proportion of grazing time on BW swards 
(proPGT), physical sward characteristics, extractable condensed tannin 
concentration (ECT - leaves of birdsfoot trefoil) and formononetin 
concentration (leaves of red clover) in Experiments 1, 2 and 3 (El ,  
E2 and E3) carried out in November (Nov), February (Feb) and 
April-May (Apr-May) 1995/1996 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  115 
Table 3.37. Amount of time (dwell time - minutes) the animal spent grazing in 
swards of BW or RC before moving to the alternative sward in 
Experiments 1, 2 and 3 (El ,  E2 and E3, respectively), Day 1. . . . . . . . . . . . . . . . . . .  122 
Table 4.1. Management schedule, agrochemical and fertiliser applied in 
Experiment 5 and 6 . . . . . . . . . . . . . . . . . . .  '" . . . . . . . . . . . . . . . . . . . . . . . . . .  '" . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  131 
Table 4.2. Descriptions of the 8 treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... . . . . . . . . .  132 
Table 4.3. Schedule of rumen modification in the first period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  137 
Table 4.4. Schedule of rumen modification in the second period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  137 
Table 4.5.Characteristics of untrimmed (Ntrim) and trimmed (Trim) plants of 
birdsfoot trefoil cultivar Goldie (Low Tannin) and accession 
PI273938 (High Tannin), and red clover cultivars G27 [Low Form 
(formononetin)] and Pawera (High Form (formononetin)] in Periods 
2 and 3 of Experiment 5 . . . . . . . . . . . . ..... . ...... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  146 
List of Tables 
Table 4.6. Characteristics of birds foot trefoil plants before grazing, and 
difference between before and after grazing (removed from height, 
mass, area and leafiness) according to trimming (Ntrim ::::: untrimmed; 
Trim ::::: trimmed) and genotype [secondary compound concentration 
(sec. comp. conc.):  high (accession PI273938), low (cultivar Goldie)] 
xxiii 
effects in Period 2 and 3 of Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  147 
Table 4.7. Characteristics of red clover plants before grazing and difference 
between before and after grazing (removed from height, mass, area 
and leafiness) according to trimming (Ntrim ::::: untrimmed; Trim == 
trimmed) and genotype [secondary compound concentration (Sec. 
Comp. Cone.): High (accession PI273938), Low (cultivar Goldie)] 
effects in Period 1, 2 and 3 of Experiment 5 . . . . . . . . . . . . . . . . ... .... ..... . . . .. . . . . . . . . . . . . .  149 
Table 4.8. Extractable condensed tannin (ECT) concentration (%DM) of 
birds foot trefoil genotypes considering the interactions with period 
and trimming effects (untrimmed: Ntrim; trimmed: Trim) of 
Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  150 
Table 4.9. Formononetin concentration (%DM) of red clover genotypes 
considering the interactions with period and trimming (untrimmed: 
Ntrim; Trimmed: Trim) effects of Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  151 
Table 4.10. Percentage of crude protein (CP), lipid, acid and neutral detergent 
fibre (ADF, NDF), carbohydrates (soluble sugars plus starch)(CHO), 
ash and in vitro dry matter digestibility (IVDMD) determined by 
Near Infrared Reflectance Spectroscopy (NIRS) of plants of birdsfoot 
trefoil and red clover of Experiment 5 (percentage of DM basis ) . . . . . . . . . . . . . .  152 
Table 4.11. Average of number of bites per plant in birds foot trefoil (BT) and red 
clover (RC), in relation to plant secondary compound concentration 
(Sec.Comp.Conc.) and trimming (Trim ::::: trimmed; Ntrim ::::: 
untrimmed) characteristics in Period 1 and Period 2, Experiment 5 . . . . . . . . . .  153 
Table 4.12. Average of number of bites per minute (biting rate) in birdsfoot 
trefoil in relation to secondary compound concentration 
(Sec.Comp.Conc.) and trimming (Ntrim ::::: untrimmed; Trim ::::: 
trimmed) characteristics in Period 2 and Period 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  155 
Table 4.13. Average of number of bites per minute in red clover, in relation to 
secondary compound concentration (Sec.Comp.Conc.) and trimming 
(Trim ::::: trimmed; Ntrim ::::: untrimmed) characteristics in Periods 1, 2 
and 3, Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  155 
List of Tables 
Table 4.14. Pearson Correlation coefficients (r) from correlation analysis 
between number of bites per plant and percentage of protein, lipid, 
acid and neutral detergent fibre (ADF, NDF), carbohydrates (soluble 
sugars plus starch)(CHO), ash and in vitro dry matter digestibility 
(NDMD) of plants of birdsfoot trefoil and red clover of Experiment 
xxiv 
5 (percentage of DM basis) . . . . . . . . . . . . . . . . . . .  , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  156 
Table 4.15. Pearson Correlation coefficients (r) from correlation analysis 
between extractable condensed tannin concentration and plant area, 
height, volume, leafiness and number of bites per plant (N. Bites) of 
birdsfoot trefoil in Periods 2 and 3 of Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  157 
Table 4.16. Pearson Correlation coefficients (r) from correlation analysis 
between formononetin concentration and plant area, height, volume, 
leafiness and number of bites per plant (N. Bites) of red clover in 
Periods 1 and 2 of Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  158 
Table 4.17. Pearson Correlation coefficients (r) from correlation analysis 
between number of bites per plant and area, height, volume and 
leafiness of birds foot trefoil plants in Periods 2 and 3 of Experiment 
5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  158 
Table 4.18. Pearson Correlation coefficients (r) from correlation analysis 
between number of bites per plant and area, height, volume and 
leafiness of red clover plants in Periods 1, 2 and 3 of Experiment 5 . . . . . . . . .  159 
Table 4.19. Characteristics of untrimmed (Ntrim) and trimmed (Trim) plants of 
birdsfoot trefoil Goldie (Low Tannin) and accession PI273938 (High 
Tannin), and red clover cultivars G27 (Low Form) and Pawera (High 
Form) in Period 1, Runs 1 and 2 (rumen content modified with 
formononetin and tannin, respectively) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  171 
Table 4.20. Characteristics of untrimmed (Ntrim) and trimmed (Trim) plants of 
birdsfoot trefoil Goldie (Low Tannin) and accession PI273938 (High 
Tannin), and red clover cultivars G27 (Low Form) and Pawera (High 
Form) in Experiment 6 Period 2, Runs 1 and 2 (rumen content 
modified with tannin and formononetin, respectively) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  172 
Table 4.21. Characteristics of birdsfoot trefoil plants before grazing and 
difference between before and after grazing (removed from height, 
mass, area and leafiness) according to trimming (Ntrim == untrimmed; 
Trim == trimmed) and genotype [secondary compound concentration 
(Sec. Comp. Conc.) :  high (accession PI273938), low (cultivar 
Goldie)] effects in Period 1, Runs 1 and 2 of Experiment 6 . . . . . . . . . . . . . . . . . . . . . .  174 
List of Tables 
Table 4.22. Characteristics of birdsfoot trefoil plants before grazing and 
difference between before and after grazing estimates (removed) of 
height, mass, area and leafiness according to trimming (Ntrim = 
untrimmed; Trim ::: trimmed) and genotype [secondary compound 
concentration (Sec. Comp. Conc.): High (accession PI273938), Low 
xxv 
(cultivar Goldie)] effects in Period 2, Runs 1 and 2 of Experiment 6 . . . . . . . .  175 
Table 4.23. Characteristics of red clover plants before grazing and difference 
between before and after grazing estimates (removed from height, 
mass, area and leafiness) according to trimming (Ntrim = untrimmed; 
Trim = trimmed) and genotype [secondary compound concentration 
(Sec. Comp. Cone.): High (cultivar Pawera) and Low formononetin 
(cultivar G-27)] effects in Period 1, Runs 1 and 2 of Experiment 6 . . . . . . . . . .  177 
Table 4.24. Characteristics of red clover plants before grazing and difference 
between before and after grazing (removed from height, mass, area 
and leafiness) according to trimming (Ntrim ::: untrimmed; Trim ::: 
trimmed) and genotype [secondary compound concentration (Sec. 
Comp. Cone.): High (cultivar Pawera), Low formononetin (cultivar 
G-27)] effects in Period 2, Runs 1 and 2 of Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . .  178 
Table 4.25. Extractable condensed tannin (ECT) concentration (%DM) of 
birdsfoot trefoil genotype main effect and interactions with period 
and trimming effects (untrimmed: Ntrim; Trimmed: Trim) of 
Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  179 
Table 4.26. Formononetin concentration (%DM) of red clover genotype main 
effect and interactions with period and trimming effects (untrimmed: 
Ntrim; Trimmed: Trim) of Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 80 
Table 4.27. Percentage of dry matter (DM), crude protein (CP), lipid, acid and 
neutral detergent fibre (ADF, NDF) , carbohydrates (soluble sugars 
plus starch)(CHO), ash and in vitro dry matter digestibility (IVDMD) 
determined by Near Infrared Reflectance Spectroscopy (NIRS) of 
plants of birds foot trefoil and red clover of Experiment 6 (percentage 
of DM basis) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  181 
Table 4.28. Average of number of bites in birdsfoot trefoil (BT) and red clover 
(RC), in relation to secondary compound concentration (High and 
Low) and trimming (Ntrim ::: untrimmed plants Trim ::: trimmed 
plants) effect in Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  182 
Table 4.29. Average of number of bites in birdsfoot trefoil (BT) and red clover 
(RC) plants in relation to rumen chemical modification (Tannin and 
Formononetin) effect in Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  '" . . . . . . . . . . . . .  182 
List of Tables 
Table 4.30. Average of number of bites in birdsfoot trefoil (BT) in relation to 
plant secondary compound concentration (Plant Sec. Comp. Conc. : 
High and Low tannin concentration), plant trimming characteristic 
(Ntrim = untrimmed; Trim = trimmed plants), type of rumen 
chemical modification [Tannin and Formononetin (Form.)] and 
rumen concentration effect (within each type of rumen chemical 
xxvi 
modification) in Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . ... . . . . . ... . . . . ... . .. . .. 183 
Table 4.31. Average of number of bites in red clover in relation to type of rumen 
chemical modification [Tannin and Formononetin (Form.)] and plant 
trimming characteristic (Ntrim = untrimmed plants; Trim = trimmed 
plants) and, in relation to rumen concentration effect (within each 
type of rumen chemical modification), plant secondary compound 
concentration (Plant Sec. Comp. Conc.) and plant trimming 
characteristic, in Experiment 6 . . . . . . . . . . . . . . . . ... . . . ... . . . . . . . . . . .... . . .... . . . . . . . . . . . . . . . . . . . . . .  184 
Table 4.32. Average of number of bites per minute (rate of biting) in birdsfoot 
trefoil and red clover in relation to plant trimming characteristic 
(Ntrim = untrimmed plants; Trim = trimmed plants) and plant 
secondary compound concentration (Sec. Comp. Conc.: High and 
Low) in Experiment 6 . . . ... ..... . . . ... . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . .  185 
Table 4.33. Pearson Correlation coefficients (r) from correlation analysis 
between number of bites per plant and percentage of protein, lipid, 
acid and neutral detergent fibre (ADF, NDF), carbohydrates (soluble 
sugars plus starch)(CHO), ash and in vitro dry matter digestibility 
(IVDMD) of plants of birds foot trefoil and red clover of Experiment 
6 (percentage of DM basis) . . . . . . ... . . . . . . ... . . . . . . ... . . . . . . . ... . . . . . . . . . . . . ... . . ... . . . . . . . . ... . . . .  187 
Table 4.34. Pears on Correlation coefficients (r) from correlation analysis 
between extractable condensed tannin concentration and plant area, 
height, volume, leafiness and number of bites per plant (N. Bites) of 
birdsfoot trefoil in Periods 1 and 2 of Experiment 6 . . . . . . . . ... ... . . . . . . . . . . . . . . . . . . . .  187 
Table 4.35. Pears on Correlation coefficients (r) from correlation analysis 
between formononetin concentration and plant area, height, volume, 
leafiness and number of bites per plant (N. Bites) of red clover in 
Periods 1 and 2 of Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . 188 
Table 4.36. Pearson Correlation coefficients (r) from correlation analysis 
between number of bites per plant and area, height, volume and 
leafiness of birdsfoot trefoil plants in Periods 1 and 2 of Experiment 
6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  188 
Table 4.37. Pears on Correlation coefficients (r) from correlation analysis 
between number of bites per plant and area, height, volume and 
leafiness of red clover plants in Periods 1 and 2 of Experiment 6 . . . . . . . . . . . . .  189 
List of Tables 
Table 4.38. Range of values (minimum and maximum) of individual plant 
morphological and biochemical characteristics observed in 
Experiments 5 and 6. These values were extracted from averages 
presented in Tables 4.5, 4.6, 4.7, 4.19, 4.20, 4.21, 4.22, 4.23 and 
xxvii 
4.24 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  204 
Table 4.39. Extractable condensed tannin (ECT) concentration (%) in birdsfoot 
trefoil (cultivar Goldie and accession PI273938) between November 
and April-May of Experiments 1, 2 and 3 (only leaves) and 
Experiments 5 and 6 (intact stems: leaf and stems) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  206 
Table 4.40. Formononetin concentration (%) in leaves of red clover cultivar 
Colenso and intact stems (leaf, petiole and stems) of cultivar G-27 
and Pawera between November and April-May of Experiments 1, 2 
and 3 and Experiments 5 and 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . .  208 
Table 4.41. Summary of the R-square improvement when the covariates plant 
leafiness, volume or ECT concentration were added before and after 
fitting the alternative covariate to the model in analyses 1, 2 and 3 
for Periods 2 and 3 of Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  213 
Table 4.42. Summary of the R-square improvement when the covariates plant 
leafiness, volume or ECT concentration were added before and after 
fitting the alternative covariate to the model in analyses 1, 2 and 3 
for Periods 1 and 2 of Experiment 6 . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  213 
LIST OF FIGURES 
Figure 2. 1 .  Schematic presentation of interactions between effects of sward 
physical and biochemical characteristic on intake and diet 
selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 6  
Figure 3. 1 .  Experiment 1 - treatment layout : area ratio birds foot trefoil and white 
clover (BW) : red clover (RC) (not to scale - strips of 2.4 cm width). 
Total area of each plot = 405 m2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  52 
Figure 3.2. Experiment 2 - treatment layout: plant maturity contrast [ 3 weeks 
( ) and 9 weeks (filiI) of regrowth] of birds foot trefoil and white 
clover (BW) : red clover (RC) (not to scale - strips of 2.4 cm width) . . . . . . . . . .  53 
Figure 3.3. Experiment 3 - treatment layout: plant height contrast [short( ) and 
tall(f///I)] of birds foot trefoil and white clover (BW) : red clover (RC) 
(not to scale - strips of 2.4 cm width) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 
Figure 3.4. Experiment 4 - treatment layout : spatial distribution contrast of 
birdsfoot trefoil and white clover (BW) : red clover (RC) (not to scale 
- strips of 2.4 cm width). Total area of each plot = 159.4 m2 . . . . . . . . . . . . . . . . . . . . . . .  55 
Figure 3.5. The stratum structure of plant parts within the sward canopy of 
birds foot trefoil and white clover before and after grazing in the four 
treatments (area ratio) in Experiment 1 .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 
Figure 3.5. The stratum structure of plant parts within the sward canopy of red 
clover, before and after grazing in the four treatments (area ratio) in 
Experiment 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  , . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  6 1  
Figure 3.7. Proportion of grazing time in relation to the proportion of area 
offered of birdsfoot trefoil plus white clover (BW) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  68 
Figure 3.8. Proportion of grazing time in relation to the proportion of herbage 
mass offered of birdsfoot trefoil plus white clover (BW) . . . . . . . . . . . . . . . . . . . . . . . . . . . .  69 
Figure 3.9. Proportion of intake in relation to the proportion of area offered of 
birdsfoot trefoil plus white clover (BW) ., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  70 
Figure 3. 1 0. The stratum structure of plant parts within the sward canopy of 
birdsfoot trefoil and white clover (BW), before and after grazing in 
the four treatments (RC = red clover sward) in Experiment 2 . . . . . . . . . . . . . . . . . . . . .  77 
list of Figures 
Figure 3.11. The stratum structure of plant parts within the sward canopy of red 
clover (RC), before and after grazing in the four treatments (BW = 
XXiX 
birdsfoot trefoil and white clover sward) in Experiment 2 . .. . . . . . . . . . . . . . . . . . . . . . . . .  78 
Figure 3.12. The stratum structure of plant parts within the sward canopy of 
birds foot trefoil and white clover (BW), before and after grazing in 
the four treatments in Experiment 3 (RC == red clover sward) . . . . . . . . . . . . . . . . . . . . .  94 
Figure 3. 13. The stratum structure of plant parts within the sward canopy of red 
clover (RC), before and after grazing in the four treatments in 
Experiment 3 (BW = birdsfoot trefoil and white clover sward) . . . . . . . . . . . . . . . . . . .  95 
Figure 4.1. Distribution of the set of four sequences with the eight treatments 
(treatments: 1 to 8 - see the description of the 8 treatments in Table 
4.2) arranged in 3 blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .  133 
Figure 4.2. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering all experimental plants of birdsfoot trefoil in Period 2 of 
Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  162 
Figure 4.3. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering all experimental plants of birds foot trefoil in Period 2 of 
Experiment 5 . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . .. . . . . .  163 
Figure 4.4. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering only untrimmed birdsfoot trefoil plants in Period 2 of 
Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . .. . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 
Figure 4.5. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering only untrimmed birdsfoot trefoil plants in Period 3 of 
Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . .  166 
List of Figures 
Figure 4.6. Effect of extractable condensed tannin concentration (covariate) 
(ECT conc.) and plant morphology (covariate) (Leafiness or Volume) 
on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant 
morphology or ECT cone. had been added. Considering plants of 
xxx 
birdsfoot trefoil in Period 2 of Experiment 5 . . . . . . . . . . .  , . . . . . . . . . . . . . . . .  , . . . . . . . . . . . . . . . .  1 68 
Figure 4.7. Effect of extractable condensed tannin concentration (covariate) 
(ECT cone.) and plant morphology (covariate) (Leafiness or Volume) 
on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant 
morphology or ECT cone. had been added. Considering plants of 
birdsfoot trefoil in Period 3 of Experiment 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1 69 
Figure 4.8. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering all experimental plants of birdsfoot trefoil in Period 1 of 
Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  19 1  
Figure 4.9. Effect of  genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering all experimental plants of birdsfoot trefoil in Period 2 of 
Experiment 6 . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .. . . . . .. . . .  192 
Figure 4. 10. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering only untrimmed birdsfoot trefoil plants in Period 1 of 
Experiment 6 . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  194 
Figure 4. 1 1. Effect of genotype (class variable) and plant morphology (covariate) 
(Leafiness or Volume) on R-squares (from the analysis of variance of 
number of bites), when added to the model before (independent) or 
after plant morphology or genotype effect had been added. 
Considering only untrimmed birds foot trefoil plants in Period 2 of 
Experiment 6 . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  195 
List of Figures 
Figure 4.12. Effect of extractable condensed tannin concentration (covariate) 
(ECT cone.) and plant morphology (covariate) (Leafiness or Volume) 
on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant 
morphology or ECT cone. had been added. Considering plants of 
xxxi 
birdsfoot trefoil in Period 1 of Experiment 6. ' "  . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .  197 
Figure 4.13. Effect of extractable condensed tannin concentration (covariate) 
(ECT conc.) and plant morphology (covariate) (Leafiness or Volume) 
on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant 
morphology or ECT cone. had been added. Considering plants of 
birdsfoot trefoil in Period 2 of Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . ... 198 
LIST OF PLATES 
Plate 3. 1 .  General view of experimental swards formed by alternate 2.4 m wide 
strips of a mixture of birds foot trefoil (Lotus comiculatus L.) cv. Goldie 
and white clover (Trifolium repens L.) cv. Pitau, and strips of red clover 
(Trifolium pratense L.) cv. Colenso . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  44 
Plate 3.2. General view of experimental swards with grazing animals . . . . . . . . . . . . . . . . . . . . . . . . . . .  44 
Plate 4. 1 .  General view of the sequence of plants, before grazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 38 
Plate 4.2. Animal grazing observation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  138 
LIST OF APPENDICES 
APPENDIX 3. 1 .  DMACA-HCL Protocol -using plate reader. .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . .  268 
APPENDIX 3 .2. Rainfall and soil temperature - Experiments 1 , 2, 3  (Chapter 3) . . . . . .  272 
APPENDIX 3.3 .  Experiment 1 (Chapter 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  275 
APPENDIX 3.4. Experiment 2 (Chapter 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . .. . . . . . . . . . . .  277 
APPENDIX 3.5. Experiment 4 (Chapter 3) . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  278 
APPENDIX 3.6. Grazing time - morning observation (Chapter 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  279 
APPENDIX 4. 1 .  Rainfall and temperature - Experiments 5 and 6 (Chapter 4) . . . . . . . . . . . .  282 
APPENDIX 4.2. Equations to estimate herbage mass per plants of birdsfoot 
trefoil and red clover, using probe GrassMaster reading (X) . . . . . . . . . . . . .  283 
APPENDIX 4.3.  Correlation matrices with Pearson Correlation Coefficients and 
P-values of Experiment 5 - plant nutritional characteristics . . . . . . . . . . . . . . 284 
APPENDIX 4.4. Correlation matrices with Pears on Correlation Coefficients and 
P-values of Experiment 5 Period 1 - characteristics of red 
clover plants . . . . . . . . . . . . . . . . . . . . . . . . . . . .  , . . . .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . .. . . . . . . . . . . . . . . .  285 
APPENDIX 4.5. Correlation matrices with Pears on Correlation Coefficients and 
P-values of Experiment 5 Period 2 - characteristics of birdsfoot 
trefoil and red clover plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  286 
APPENDIX 4.6. Correlation matrices with Pearson Correlation Coefficients and 
P-values of Experiment 5 Period 3 - characteristics of birdsfoot 
trefoil and red clover plants . . . ... . . . . . . . . ... . . . . . . . . . . '" . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  288 
APPENDIX 4.7. Correlation matrices with Pearson Correlation Coefficients and 
P-values of Experiment 6 - plant nutritional characteristics . . . . . . . . . . . . . .  290 
APPENDIX 4.8.  Correlation matrices with Pearson Correlation Coefficients and 
P-values of Experiment 6 Period 1 - characteristics of birdsfoot 
trefoil and red clover plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  291 
APPENDIX 4.9. Correlation matrices with Pearson Correlation Coefficients and 
P-values of Experiment 6 Period 2 - characteristics of birdsfoot 
trefoil and red clover plants . . . . . . . . . . . . . . . . . . . . . . . .
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  293 
List of Appendices xxxiv 
APPENDIX 4.10. Rumen manipulation - Experiment 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  295 
APPENDIX 5 . 1 .  Preliminary report of Experiment 1 published in the 
proceedings of the xvm International Grassland Congress, 
Canada, 1997 . . .. ... . . . . . . . .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  298 
APPENDIX 5.2. Preliminary report of Experiment 5 published in the 
proceedings of the New Zealand Society of Animal Production, 
1998 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  300 
CHAPTER 1 
INTRODUCTION AND OBJECTIVES 
To meet their requirements for maintenance, growth and reproduction, ruminants are 
faced with a complex of decisions that reflect the heterogeneity of the environment in 
which they are foraging (Gordon and Lascano, 1993). In contrast to carnivorous 
predators, herbivores usually need to deal with an environment where the food 
components are more evenly distributed in space and time and at low density (lllius and 
Gordon, 1990). In this environment the animal could take its entire daily intake of dry 
matter grazing unselectively from a few square metres of pasture. However, ruminants 
very seldom graze in a completely non-selective manner (Parsons et al., 1994a). It has 
been widely recognised that the diet of grazing animals is affected by the physical (e.g. 
mius and Gordon, 1990; Gordon and Lascano, 1993; Hodgson et al. , 1994; O'Reagain 
and Schwartz, 1995; Laca and Demment, 1996; Hodgson et al., 1997) and biochemical 
(e.g. Barry, 1989; Provenza and Balph, 1990; Van Soest, 1994; Provenza, 1995; 
Launchbaugh, 1996) characteristics of the sward, and understanding of these effects 
requires knowledge of the concomitant changes in ingestive behaviour (Hodgson, 
1985). 
Pasture characteristics vary spatially and temporally (O'Reagain and Schwartz, 1995) 
forming heterogeneous swards that need to be explored by the animals (Gordon and 
Lascano, 1993). In this heterogeneity, the effects (as demonstrated in several studies) of 
sward structural characteristics such as height, bulk density and herbage mass, and plant 
morphology, interact with the effect of sward biochemical characteristics to determine 
diet selection and intake (Hodgson et al. 1994). Several studies have been carried out to 
determine the effects of either physical sward characteristics (e.g. Milne et al, 1982; 
Burlison et al., 1991; Mitchell et al., 1991, 1993a; Ungar et al.,1991; lllius et al., 1992; 
Laca et al., 1992; Demment et al., 1993) or biochemical characteristics (e.g. Francis 
Introduction and Objectives 2 
1973, Whittaker and Feeny, 1971; Rhoades and Cates, 1976; Barry and Manley, 1984; 
Provenza and Malachek, 1984; Provenza et al. 1990, 1994; Kyriazakis et al. 1998) 
separately on ingestive behaviour. However, due to the complexity involved at the 
plant-animal interface, many of the interactions between grazing animals and sward 
characteristics are not well understood (Gong, 1993). As a consequence there is limited 
information on the influence of spatial and temporal heterogeneity in swards upon 
animal behaviour (Taylor, 1993). There is a major need for studies (lllius and Hodgson, 
1996; Hodgson et al. ,  1997) where the effect of interactions between sward physical and 
biochemical characteristic are assessed. 
Studies have shown that when animals are offered two different species, they do not 
demonstrate an absolute preference for one species, but a partial preference, resulting in 
a mixed diet (Newman et al., 1992; Parsons et al. ,  1994a; Cosgrove et al. ,  1996). So far, 
studies on partial preference have been restricted largely to either pen feeding or large 
blocks of monocultures. Newman et al. (1992) and Parsons et al. ( l994a) demonstrated 
that this preferential behaviour has an important influence on dietary balance, and was 
itself influenced by the relative area of the alternative species, by the animals' previous 
experience, and by the time of the day. These findings indicate a new dimension for 
further research (Hodgson et al., 1997). 
Torres-Rodriguez (1997) showed little contrast in the preference of grazing cattle 
between relatively large blocks of monocultures of birdsfoot trefoil and red clover. Thus, 
these species provide the opportunity to study the effects of manipulation of plant physical 
and biochemical contrasts on dietary preference and selective behaviour by grazing cattle 
against the background of relative neutrality in partial preference. The present study was 
designed to investigate these effects using experimental structures that minimise the 
requirement for animal movement to demonstrate partial preference. In addition, birdsfoot 
trefoil and red clover contain important secondary compounds (condensed tannin and 
formononetin). Very little information is available in relation to how their concentrations 
might influence animal appraisal and how their effects interact with sward physical 
characteristics in influencing selective behaviour. 
Introduction and Objectives 3 
The fIrst studies involved alternate strips of two simple swards where the effects of sward 
area, maturity and height were investigated. The variation in discriminatory behaviour 
between and within experiments led to two further experiments with spaced plants where 
the effects of secondary compound concentration and morphological characteristics on 
partial preference were investigated. 
The main objectives of the six experiments reported in this thesis were: 
1) Evaluate the impact of the relative areas, maturity and height of alternate 
simple swards, formed basically by either birdsfoot trefoil and white clover or 
red clover, on diet selection by cattle in circumstances where the physical 
distribution of the alternative swards provided maximum opportunity for 
selection. 
2) Investigate the specific effects of secondary compounds (condensed tannin in 
birdsfoot trefoil and formononetin in red clover) and plant morphology on 
cattle preference. 
3) Assess seasonal variation in selective behaviour in relation to variations in 
sward physical characteristics and in concentrations of secondary compounds. 
CHAPTER 2 
LITERATURE REVIEW 
2.1. INTRODUCTION 
Cattle production is a function of the ability of the animal to harvest nutrients in an 
effective and efficient manner (Stuth, 1991). Selection exerts an important influence on 
the ability of the animal to harvest its daily nutrient requirement where the variation in 
sward structure and the distribution of components within the sward allow the opportunity 
for selection (Hodgson et al., 1994). 
Several pasture factors affect animal preference. This preference allows the grazing animal 
to select a diet that differs from the overall composition of the feed available (Leigh and 
Mulham, 1966ab; Dudzinski and Amold, 1973; L'Huillier et al 1986). In this complex 
relationship Provenza and Balph (1990) suggest that forage environments present at least 
five problems or challenges to ruminants selecting dietary items: (1) variation among 
dietary items in kind and nutrient content (2) variation among potential dietary items in 
kind and amount of chemical defence, (3) variation in plant morphological defences, (4) 
temporal and spatial variation in the quantity and quality of forage, and (5) exposure of 
ruminants to unfamiliar foraging environments. 
Literature Review 5 
In this thesis attention is concentrated on a study of the preferences exhibited by grazing 
cattle between swards of red clover (Trifolium pratense L.) and birdsfoot trefoil (Lotus 
comiculatus L.)/white clover (Trifolium repens L.), and the influence of some physical 
and biochemical sward characteristics on these preferences. These species provide a useful 
model for investigation since Torres-Rodriguez (1997) has shown only limited 
discrimination between them by cattle when offered monocultures of these species in 
adjacent areas, in contrast to stronger preference shown by animals offered monocultures 
of these legumes in relation to perennial ryegrass. 
The main objectives in this literature review are: (i) to describe and discuss the main 
models of animal grazing decisions in relation to the effects of physical and biochemical 
sward characteristics on diet selection; (ii) to investigate the role of animal senses in diet 
selection; (iil) to discuss the effects of sward spatial and temporal variability, and plant 
biochemical characteristics on grazing behaviour and diet selection. Particular emphasis 
is concentrated on tannins and formononetin as potential biochemical factors 
influencing selection. 
2.2. DIET SELECTION BY GRAZING ANIMALS 
2.2.1. Models of diet selection 
Modelling diet selection is to a certain extent frustrating. Several attempts have been 
made to explain and predict diet selection. However the complexity of the subject has so 
far prevented the development of comprehensive models. Many different constraints 
affect the diet selected by free-ranging animals (Parsons et al., 1994a). The variation in 
diet selection happens not only due to differences among animal species (eg. Hofmann, 
1989; Van Soest, 1994) but also due to differences in pasture composition and structure 
(e.g. Stephens and Krebs, 1986; Senft et al., 1987; Gordon and Lascano, 1993; 
Demment and Laca, 1995),  environmental conditions (eg. Gordon 1989abc; 
Coughenour, 1991), animal physiological state and experience (eg. Newman et al. ,  
Literature Review 6 
1992, 1994b; Parson et al., 1994a; Provenza, 1995), time of the day (Parsons et al. 
1994a), period of the year (Arnold 1987), plant biochemical characteristics (e.g. Barry 
and Blaney, 1987;  Provenza and Malechek, 1984; Launchbaugh, 1996), and 
relationships with other animals (e.g. Penning et al., 1993). 
Not only are there several factors that affect diet selection but also there are several 
possible approaches for modelling their effects. The approaches can be described as 
empirical, teleonomic or mechanistic (Thornley et al., 1994), but none are perfect in 
describing diet selection. Several models describe intake in functional terms, or 
empirically (Spalinger and Hobbs, 1992). However, this approach seems to be 
inappropriate for describing choice of food on most occasions since animals can modify 
intake according to changes in the relative availability of alternatives (Parsons et al. ,  
1994a). A purely mechanistic approach is  very difficult for describing diet selection 
since this can not consider aspects of behaviour that are not well enough understood to 
describe mechanistically (Parsons et al., 1994b). On the other hand, Thornley et al. 
(1994), taking a purely teleonomic approach, provided only a simple view of what is in 
fact a very complex problem. 
Because of the complexity, no experimental program could expect to cover all aspects 
of the system at one time (Gordon and Hutchings, 1993). It seems that most of the 
models are additive rather than exclusive. The more models are created the more 
different situations can be understood. A group of researchers in the UK, for example, 
developed different models to explain diet selection. Thornley et al. (1994) using a cost­
benefit model found that the animal maximises the benefits minus the costs when 
grazing. This agrees with what was empirically found by Kenney and Black (1986) and 
Black and Kenney (1984), who showed that animals prefer the forage that provides 
higher intake rate. This knowledge is expanded in the model of Parsons et al. (1994b), 
which demonstrates that the way plant species are distributed in the sward can have 
different effects on diet selection. This point is clarified by Newman et al. (1995), who 
developed a stochastic dynamic programming model of grazing behaviour. This 
Literature Review 7 
demonstrates that diet preference may depend on the relative intake rates of the 
alternative plant species. 
Diet selection models are summarised by Provenza & Balph (1990), who describe five 
main group of models to explain how ruminants select their diet. In the first case it is 
assumed that animals have the innate ability to sense, through taste and smell, specific 
nutrients and toxins in plants (euphagia). In the second, animals select vegetation that is 
immediately "pleasing" to olfactory, gustatory, and tactile senses and avoid that which is 
not (hedyphagia). The third assumes that diet selection is a function of animals' 
morphophysiology and size. The fourth assumes that the animals learn through foraging 
consequences, including pre- and post-ingestive consequences of foraging experience 
and social relationship. The last argues that animals forage to maximise nutritional gain 
per unit cost. 
The first and second models were created to explain how animals could obtain a more 
nutritious diet than the forage on offer. These models, in fact, were not supported by 
research (Amold and Hill 1972; Marten and Andersen, 1975). They are also criticised 
for not including the learning process (post-ingestive consequences) in selecting a diet 
(Provenza and Balph, 1990). However there are cases where animals show "nutritional 
wisdom". Bell and Sly (1983), for example, demonstrate that sodium deficient cattle can 
detect a few millimoles of salt by smell, and retain a "memory" of the locations. 
Although the third model does not include differences among different individuals 
within the same species, it complements the learning models. Provenza and Balph 
(1990) state that morphophysiology models provide a broad explanation for diet 
selection, while learning models fine tune to the level of individual animals.  The last 
model, optimal foraging theory (Stephens and Krebs, 1986) gives a different explanation 
for animal diet selection. This theory states that evolutionary pressure selectively formed 
animals that hunt or graze for their food efficiently. Laca and Demment (1996) argue 
that this theory offers the strongest theoretical basis and framework to study foraging 
strategy of grazing animals. In fact, researchers (eg. Laca et al., 1993; Kenney and 
Black, 1984; and Black and Kenney, 1984) have shown that a choice between 
Literature Review 8 
alternative forage or patches is strongly affected by the potential intake rate. However 
Griffiths et al. (1997) found that, contrary to expectations from optimal foraging 
strategy, grazing behaviour of cows at a current sward patch was unaffected by sward 
conditions at adjacent patches. Provenza and Balph (1990) criticised the optimal 
foraging theory because of the need for grazing animals to explore a varied diet where 
several other characteristics (eg. biochemical plant characteristics, plant morphology 
etc.) can influence the grazing decision. Crawley (1983) states that this model is more 
suitable for carnivores than herbivores, because of the limited variability of nutritional 
quality of prey, relative to the diet of herbivores. 
The models outlined by Provenza and Balph (1990) do not accommodate more complex 
phenomena of selection that have been more recently observed, such as the effects on 
preference of recent dietary experience (Newman et al., 1992; Parsons et al., 1994a), 
animal state (Newman et al. 1994; Parsons et al.,  1994a), time of the day (Parsons et al. 
1994a), and vigilance requirements (mius and Fitzgibbon, 1994). Newman et al. (1992), 
offering turfs of perennial ryegrass and turfs of white clover, observed that sheep 
preferred the opposite species to the one they had previously grazed. Parsons et al. 
(1994a), testing this hypothesis in swards, found that although sheep showed a 
preference for a high proportion of clover (the sheep did not graze at random), they 
sustain a mixed diet even in situations where a monospecific diet was readily possible. 
They suggest that sheep have a partial and changing preference for white clover in 
contrast to perennial ryegrass. Forbes (1995) argued that animals show a sensory­
specific satiety. Animals choose to eat a variety of foods when none of them is aversive. 
These results advise caution when short term results are extrapolated to a long-term test 
of preference (Parsons et al. , 1994a). In fact, it is wrong to suggest that sheep always 
show preference for white clover in relation to perennial ryegrass. Before the publication 
of Parsons et al. (1994a), only few studies (see mius and Gordon, 1990) on animal 
preference comment on whether the preference for clover is total or is partial. Many 
studies have come to the conclusion that sheep prefer clover (van Dyne and Head, 1965; 
Hodge and Doyle, 1967; Bedell, 1973; Leigh and Holgate, 1978; CurU and Wilkins, 
1980; Frame and Newbould, 1986; Curll and Gleeson, 1987; Lascano and Thomas, 
Literature Review 9 
1988; Vallentine, 1990; Ridout and Robson, 1991), although some suggest a lack of 
selectivity (eg. Clark et al. , 1982; L'Huillier et al., 1984; 1986; Amold, 1987). On the 
other hand other studies reported that selection is a response to variation in vertical 
distribution between species (Hodgson, 1981b; Milne et al., 1982; lllius et al., 1992) 
(see section 2.3.1) .  
More recently, Provenza ( l996b) offers a new explanation for diet preference, based on 
avoidance of toxins and acquisition of nutrients. A key concept in this theory is food 
aversion. The decrease in preference for a food is a function of the sensory and post­
ingestive feedback of that food. In this way animals also prefer the familiar to the novel 
and regard anything new with caution (Provenza, 1996a). However, they can become 
averse to what is too familiar, eaten too frequently or in excess, and search for a novel 
food and varied diets (Provenza, 1996a) 
Nevertheless, diet selection also varies according to intrinsic differences among 
different animals. Although some studies (eg. Hoffman, 1989; Provenza and Balph, 
1990; Van Soest, 1994) agree in modelling diet selection as a function of animal 
morphophysiology and size, they do not consider the differences between different 
animals of the same species. Hodgson (1985) mentions that increased nutrient demand 
will usually increase forage intake, but the effects on the individual components of 
intake appear to be variable and difficult to predict. Good examples of contrasting 
nutrient demand are lactating and non-lactating animals, shorn and unshorn sheep, 
fasted and non-fasted animals. Parsons et al. (1994a) comparing the diet preference of 
dry vs. lactating ewes found that, despite major differences in energy requirement and 
intake behaviour, no significant effects of physiological state on preference were 
detected. On the other hand, Newman et al. (1994) and Edwards et al. (1994), using 
different methods, have observed evidence of state dependent changes in preference 
when comparing fasted vs. unfasted sheep. Despite what was expected, fasted animals 
spent a significantly lower proportion of their time grazing clover (the higher intake rate 
component) . These studies demonstrate that animal state should be 
Literature Review 10 
considered in modelling diet preference. However more studies need to be done in this 
area (Parsons et al. 1994a). 
Diet selection modelling also needs to consider the vigilant behaviour of ruminants. In 
evolutionary terms the best defence of ruminants against predators is to be vigilant. 
Ruminants are constantly vigilant, with a great capacity of the rumen to accumulate food 
to be processed later. Evidence of a trade-off between energy gain and scanning for 
predators while foraging suggests that vigilance is costly (Barnard, 1980; Underwood, 
1 982; Lendrem, 1983; Metcalfe and Fumess, 1984). lllius and Fitzgibbon ( 1994), 
calculating the costs of vigilance in foraging ungulates, pointed out that an animal 
feeding selectively from the vegetation voluntarily accepts a reduced density of bites. 
This behaviour requires a trade-off between the advantage of being selective and the 
reduced opportunity to scan for predators. Since small animals incur lower vigilance 
costs than large animals, they can afford to be more selective (lllius and Fitzgibbon, 
1 994). This fact can be connected with the social behaviour of the animals. Penning et 
al. ( 1993) showed that sheep kept individually, at pasture, may not behave in the same 
way as when they graze as members of a flock. Although there is no clear evidence to 
explain this, it could be related to the fact that the animals do not feel so threatened in 
large groups, when they are better protected by more vigilant companions (Pulliarn and 
Caraco, 1984). 
2.2.2. Role of senses in diet selection 
Several studies have been carried out on the function of the senses in dietary preference 
(e.g. Arnold, 1966a, 1966b; lackson et al., 1968; Goatcher and Church, 1970ab; Bell 
and Sly, 1 983; Warden and Dyk, 197 1 ;  Arave et al., 1 989; Bazeley and Ensor, 1 989). 
All the senses, in some way, influence diet selection. Preference, in fact is the result of a 
complex of sensing behaviours where the final choice is determined by the responses 
elicited to stimuli from the food (Arnold, 1966a). Edwards ( 1994) found that sheep have 
the ability to form associations between cues and rewards to direct their foraging, and 
increase their encounter rate with preferred patches in the environment. In addition, 
Literature Review 11 
Edwards (1994) demonstrated that sheep are able to distinguish between ryegrass and 
white clover without sampling the two species, and can remember and associate 
different food types with their cues. More recently Provenza (1996a) argued that 
animals acquire preferences for the flavours of familiar foods that have been associated 
with the positive post-ingestive effects of nutrients. Provenza (1996b) explained that the 
reason why animals have varied diets is due to the decrease in preference for food as a 
result of sensory (taste, smell, texture) and post-ingestive feedback unique to each food. 
How the senses (sight, taste, smell and touch) affect animal preference and diet selection 
is discussed below. 
2.2.2.1. Sight 
Although sheep eyes possess cones (C.V. Ensor, in Forbes, 1996), they are thought to be 
colour-blind (Tribe and Gordon, 1949). Bazeley (1988) suggests that sheep may use 
sight to distinguish different shades of green. However Bazely and Ensor (1989) found 
that while sheep learned to discriminate between visual cues which varied in brightness, 
they failed to do so when the cues varied only in hue. This does not negate the 
possibility of sheep having a colour vision, but brightness might be important to 
distinguish grass, for example, with high and low protein content (Bazeley and Ensor, 
1989). 
Sight is seldom thought to be the primary sense acting in grazing preference. However it 
is known that cattle, like sheep and goats, can make quite complex discrimination 
between shapes (Forbes, 1996), Electrical activity of single neurones in regions of the 
brain was thought by Kendrick (1992) to be involved in feeding control. Mainly when 
food was moved towards the mouth, cells in the lateral hypothalamus and zona incerta 
respond to the sight but not the smell of food. Arnold (1966a) working with blinkered 
sheep under pasture conditions found that diet composition differed for control and 
blinkered sheep. These differences were attributed to the use of sight by sheep to 
orientate themselves while grazing. The bUnkered sheep tended to graze all the strains 
of a species to the same height. Edwards (1994) also demonstrated the importance of 
Literature Review 12 
sight on diet selection. Using identical bowls whose contents could not be seen, except 
from directly above, he removed the possibility of sheep identifying patch type. In fact, 
the sense of sight is very important for orientating animals in space. 
Sight also affects fine selection. Arnold ( l966a) found that species preference was 
unaltered by sight impairment, but the habit of grazing was modified. On short swards, 
sheep with blinkers took more taller components, those that they could feel first as the 
head was bent to graze (Arnold, 1966a). In fact, sheep can distinguish patterns of the 
pasture at quite a fine level. Cahn and Harper (1976) suggested that sheep select clover 
on the basis of leaf mark polymorphism. They found that unmarked leaves were initially 
preferred to marked ones. Edwards (1994) observed, in fact that sheep used visual cues 
and/or olfactory cues to determine patch type when directly above a patch, as patches 
were often rejected without sampling the patch. 
2.2.2.2. Taste 
Taste is one of the most powerful effects on animal preference. The ability of calves, 
for example, to discriminate among various sugar solutions and to show a preference for 
a specific sugar has been amply demonstrated (Stubbs and Kare, 1958; Bernard, 1964; 
Waldern and vanDy, 1971). Taste is believed to be involved in at least two 
psychological processes: food seeking behaviour and reinforcing value of a food 
(Goatcher and Church, 1970a). Many mammals readily accept sweet and reject bitter 
tastes (Provenza and Balph, 1990). This led several authors (Bate-Smith, 1972; Garcia 
and Hankins, 1975, 1977; Grill et al. 1984; Beauchamp and Cow art, 1987) to relate 
sweet taste with nutrients and bitter with toxins. In fact, Provenza et al. (1990) argued 
that odour and taste may be more important for enabling ruminants to identify and 
discriminate among subtle differences in diet items, than for innately recognising plants 
that are nutritious or toxic items based on pleasing or adverse gustatory and olfactory 
sensations. 
Literature Review 13 
Ruminant species differ among themselves in relation to preference (Arnold and Hill, 
1972; Church, 1979; Hofmann, 1989). Goatcher and Church (1970b) comparing the 
taste responses of goats, sheep and cattle, using different concentrations of acetic acid 
and quinine hydrochloride, found that cattle were usually the first to make a 
discrimination, goats were generally second and sheep were normally last. These authors 
confirm what was found previously only with sheep (Goatcher and Church, 1970a) that 
stimulating effectiveness was greatest for bitter, followed by sour, salty and sweet tastes. 
Taste (as well as smell and sight) helps the animal to discriminate according to what is 
pleasing. The taste-feedback interaction is noncognitive and is not related to a feedback 
event memory (Provenza, 1996b). Animals can change preference despite the 
knowledge of the cause of the feedback event. According to Provenza (1995a, 1996a, 
1996b), animals acquire preferences for the flavours of familiar foods that have been 
associated with the positive post-ingestive effects of nutrients. The preference of 
animals increases if the food is adequate in nutrients (Provenza, 1996a) . Preferences for 
flavours paired with energy (starch), for example, persist for at least 2 months following 
conditioning, which suggests animals acquire a liking for flavours paired with energy 
(Provenza, 1996a). 
2.2.2.3. Smell 
Together with taste, smell forms the flavour of a food. Although Tribe (1949) thought 
that odour had little effect on selection of plant species by grazing animals, Arnold 
(1981,1970) argued that the chemical signals, which mainly influence food selection, are 
those received at receptor sites for taste and smell. Arnold (1966) comments that marked 
changes in the relative acceptability of species, or strains of a species, occur when the 
sense of smell is impaired. McLaughlin et al. (1974) observed that sheep that had 
olfactory bulbectomy had less intense feeding, though with more re-entry into the feeder 
during meals, although they did not find any difference in daily feed intake. 
Literature Review 14  
The sense of smell i s  thought to be  not very precise in ruminants (Arnold, 1981) .  
However it i s  enough to  make precise decisions about where to  graze. Arave et  al. 
(1989) demonstrates that flavour agents significantly increased preference for 
concentrates in dairy cows. 
In addition, the importance of sense of smell in grazing animals is very clear when they 
avoid grazing close to patches of dung. Arnold (1981)  observed in a sward with many 
dung patches that animals avoided the patches but ate very close to them. Norman (in 
Garner, 1963), comparing dung with urine, demonstrated that dung had the greater and 
more lasting influence, and the effect of urine disappeared in a relatively short period. 
The negative effect of the undesirable smell of dung on pasture palatability is enhanced 
by the fact that the ungrazed patches will become fibrous and coarse (Garner, 1963). 
The olfactory effect in ruminant grazing decisions is also clear in relation to salt 
preference. Bell and Sly (1983) show that olfactory and gustatory receptors are able to 
detect very small amounts of sodium salt. Ruminants do not ingest salt as such unless 
they are in the metabolic state of sodium deficiency. Salt appetite increases in proportion 
to the bodily depletion of sodium (Bell and Sly, 1976, 1977). With the development of 
sodium depletion in cattle the ability to detect a very low concentration of salt increases. 
Olfactory senses play an important role in the process of animal learning. Animals learn 
initially about the flavour of foods in utero and from mother's  milk (Provenza and 
Balph, 1990). When young, learning from other animals is very important. Animals can 
smell where others have passed and how long ago others were there (Hart, 1985), 
influencing diet selection directly. In this process the animal learns the ability to 
discriminate between subtle chemical differences in different plants through odour and 
taste, and they associate the food flavour with post-ingestive consequences 
(Launchbaugh and Provenza, 1993). 
Literature Review 15 
2.2.2.4. Touch 
Hyde and Witherly (1993) and Garner (1963) claim that changes in texture have a great 
impact in the palatability of food. Animals usually select against rough, harsh, and spiny 
material (Vallentine, 1990). Van Niekerk et al. (1973) argue that the greater quantities 
eaten of pelleted forages are in part attributed to textural effect. According to Garner 
(1963), leaves that are very harsh to the human touch are harsh in the animal's  mouth. 
This author mentions that the unpalatability of Yorkshire fog and cocksfoot is due to the 
hairiness and the silicious teeth on the leaves, respectively. In fact, leaf toughness is 
regarded as the most important mechanical attribute influencing grazing (Theron and 
Booysen, 1966; Coley, 1983). However Briske (1996) points out that there is not enough 
experimental evidence that mechanical attributes of plants are deterrents to grazing by 
vertebrate herbivores. 
2.3. EFFECT OF SWARD 
GRAZING BEHAVIOUR, 
HERBAGE INTAKE 
CHARACTERISTICS ON 
DIET SELECTION AND 
Diet selection and intake are a function of plant and animal characteristics. To ingest an 
adequate level of nutrients animals need to deal with plant physical variability and 
biochemical defence. The way they explore the physical and biochemical variability 
depends also on their own characteristics. The interactions between effects of sward 
physical and biochemical characteristics on intake and diet selection are illustrated in 
Figure 2.1., and are discussed in the following sections. 
Although herbivores spend little time searching for food and face a relatively abundant 
and conspicuous food resource compared to carnivores (Stephens and Krebs, 1986), 
searching time is an important component of grazing time (Laca and Demment, 1996) 
that varies with grazing conditions. To graze a diet, which has an adequate level of 
nutrients to meet requirements for maintenance, growth and reproduction, a herbivore is 
faced with a series of short-term tactical decision about what diet to select, how long to 
Literature Review 16  
search between bites and the resulting rate of  food intake. In the longer term, strategic 
decisions concern the length of time to spend feeding and where to feed (Gordon and 
Lascano, 1993). The complexity of the decision depends on the heterogeneity of the 
environment (Gordon and Lascano, 1993), One of the most important decisions the 
animals need to make is the trade-off between quality and quantity (Senft et al. ,  1987), 
Figure 2 . 1 .  Schematic presentation of interactions between effects of sward physical 
and biochemical characteristic on intake and diet selection. 
ANIMAL 
+ 
physiological state 
+ 
pre and post 
ingestive experienc 
+ 
Animal 
=> 
Intake 
Diet 
<=> 
Selection 
Plant 
{:= 
Defence 
/ SWARD ........... 
PHYSICAL BIOCHEMICAL 
vertical 
variability 
+ 
spatial 
distribution 
+ plant structural 
compounds 
+ 
metabolites 
Although studies of how spatial variation of vegetation influences diet selection are very 
useful, they are relatively recent. The influence of spatial and temporal variation in food 
availability on diet selection is discussed below. 
2.3.1. Spatial variability affecting grazing behaviour and diet selection 
Diet selection is sensitive to variation in the horizontal and vertical distribution of 
dietary components (Edwards, 1994), and in sward component distribution (Arditi and 
Literature Review 1 7  
D '  Acorogna, 1 984; Parsons et al. ,  1994a; Thorney et al. ,  1 994; Newman et al. ,  1 995). 
Edwards ( 1 994) suggests that the diet selected by sheep from two different swards, with 
the same proportions of the component plant species, may be quite different if spatial 
distributions of the plant species in the two swards are different. In fact, diet selection 
happens at different levels of scale: plant part level, plant level, patch level and 
landscape level (O'Reagain and Schwartz, 1995). 
2.3.1.1. Plant part level 
Selection for different parts of the plant is demonstrated in several studies. Animals are 
faced with differences in the nutrient content of different plant parts (Arnold, 1 960), in 
potential bite dimensions due to size and or specific mass (O'Reagain and Schwartz, 
1 995), potential ingestion rate due to tensile strength or location and spatial arrangement 
within the canopy (O'Reagain and Schwartz, 1995). Animals usually respond to this 
variation by selecting leaf from stem (Arnold, 1960; Juko and Bredo, 1 96 1 ;  Arnold, 
1 964; van Dyne and Heady, 1965; Guy et al . ,  1 98 1 ;  Arnold, 1 98 1 ;  L'Huiller et al. 1984; 
L'Hiuller, 1986; Edwards, 1994), young components in relation to old (Arnold, 1960), 
and green in preference to dead material (Arnold, 1 960; Juko and Bredon, 1961 ) .  
The factors that determine selection of plant parts are not well understood (Hodgson, 
1990). Selection between plant components at a fine spatial scale is in several cases only 
the reflection of the easy prehension of one component in relation to another 
(Hendricksen and Minson 198 1 ;  Hodgson, 1990) . However selection is also an animal 
ability. Edwards ( 1 994), using homogeneously mixed pellets, found that sheep have the 
ability to select on a small scale. In reality, sheep tend to be more selective than cattle, 
and goats shows a greater preference for fibrous components than sheep and cattle 
(Hodgson, 1 990). 
Literature Review 1 8  
2.3.1.2. Plant level 
Herbivores select one plant in relation to others using several strategies. Plants affect 
animal selection according to their nutritive content, mainly leaf quality, and intake rate 
due to plant structure (O'Reagain, 1993) . Considering the same availability and height, 
animals select different species according to the amount of stem and accessibility of 
leaves of high nutrient content and low tensile strength (O'Reagain and Schwartz, 
1995). Plants with a high proportion of stems and leaves of high tensile strength are 
avoided (Theron and Booysen, 1966, Field, 1 976; O'Reagain and Mentis, 1989, 
O'Reagain, 1 993). The importance of each factor depends on the animal and 
environmental conditions. Nutritive content and tensile strength are also important 
factors in terms of selection of different plants of the same species. Gammon and 
Roberts ( 1978), for example, showed that animals tended to select plants that were 
defoliated previously mainly because grazed tufts remain green and leafy. Animals 
avoid tufts which are rank and contain senescent material (Mo tt, 1985; Ganskopp et al, 
1993). In addition to selection for plants of high nutritive value and intake rate, animals 
also avoid some secondary compounds (O'Reagain and Schwartz, 1 995 -see section 
2.3.6. 1 ) .  
2.3.1.3. Patch level 
The variation in a sward is to some extent related to the grouping of plant species. 
Patches can vary from a single plant to landscape scale. However, beside plant species 
variability, the variability can be caused by several biotic processes, with the most 
important being grazing (O'Reagain and Schwartz, 1995). Grazing forms a mosaic of 
patches of varying size in a sward (Mott, 1985; Willms et al. ,  1 988). The process of 
grazing increases the difference between the preferred and non-preferred species. 
Because animals tend to select previously defoliated plants (Gammon and Roberts, 
1 978), once the mosaic is formed, it tends to be maintained. The preferred patches are 
usually short and the ungrazed, rank and stemmy (Mott, 1985). This difference is also 
enhanced through the variability promoted by urination, defecation and trampling 
Literature Review 19  
(Thorhallsdotir, 1 990; Ledgard et al. ,  1982; laramillo and Detling, 1992) . Animals, 
particularly cattle, avoid grazing close to their dung (Hodgson, 1990) 
2.3.1.4. Landscape level 
The variation explored by free-ranging animals can also be described in relation to the 
landscape. A landscape is described by O'Reagain and Schwartz ( 1 995) as an 
association of areas which differ markedly in species composition, vegetation structure 
and/or some physical characteristic such as slope, rockiness or soil fertility. Selection 
within the landscape is a complex process, involving several factors which are beyond 
the scope of this review. The reader is referred to recent reviews in this topic by Stuth 
( 199 1 )  and O'Reagain and Schwartz ( 1995). 
2.3.2. Spatial heterogeneity and the process of diet selection 
The animal may select in a horizontal dimension, from which patch to graze, and 
vertically from which plant part to graze and for how long (Gordon and Lascano, 1 993). 
Animals graze certain patches where the density of a particular species exceeds some 
threshold (Arditi and Dacoragna, 1 988). In this process, when the density is below the 
threshold, it is energetically more profitable for the animal to move on and continue 
searching for other patches (Kacelnik and Bemstein, 1988). Intake rate increases in 
areas where forage items are dense or most abundant (Dudzinski and Amold, 1 973; 
Trudell and White, 198 1 ;  Wickstrom et aL 1 984). Kenney and Black(1984) and Black 
and Kenney ( 1 984) found in intensive studies that the animal was strongly influenced by 
the potential intake rate. These results were then confirmed in large scale studies with 
sheep and cattle in swards with a range of variation in canopy height and bulk density 
(mius et al. ,  1 992; Demment et al.,  1 993). Bazely ( 1990) illustrated this point, showing 
that sheep graze tall patches more intensively than shorter ones, and Amold ( 1987) 
showed that sheep consume more from high biomass than the low biomass areas. Cl ark 
et al. (in Gordon and Lascano, 1990) observed that sheep and goats moved from patch to 
Literature Review 20 
patch but never stayed long on the shorter patch, and this contributed to the incomplete 
selection of the taller sward. 
Preferential grazing between the different species in a sward is also influenced by spatial 
availability. Ridout and Robson ( 1 99 1 ), for example, re-analysing the data of Clark and 
Harris ( 1 985) who worked with different horizontal distributions of white clover and 
perennial ryegrass, showed that the percentage of clover in a diet was, in general, 
higher when white clover and perennial ryegrass were in strips of monocultures than 
intermixed. On the other hand, Armstrong et al. ( 1 993) found that the size and distance 
between clover patches had no specific effect on the proportion of clover in the diet of 
weaned lambs in addition to that attributable to the overall proportion of grass and 
clover in the sward. 
Animals also respond to the vertical variation between species. Hodgson ( 198 1 b), Milne 
et al. ( 1 982) and mius et al. ( 1992) found that for temperate pastures there is little 
difference between the composition of the diet and that of the upper strata of the canopy 
within which animals are known to be grazing. The variations in the frequency or 
severity of defoliation are likely to be directly related to the size of individual plants and 
their proximity to the surface of the vegetation canopy (Gammon and Roberts, 1 978;  
Bircham and Hodgson, 1983; Briske, 1 986). However this variation can be influenced 
by several factors such as plant biochemical characteristics (e.g. Malechek and Balph, 
1987;  Launchbaugh, 1996) and plant maturity (eg. Gardener, 1980; L'Huillier et al., 
1 984). In mature pasture, for example, animals prefer to graze shorter and younger 
plants (Hodgson and Ollerenshaw, 1969; Gibb and Ridout, 1988). 
Spatial availability has different effects according to the animal species. Cattle are 
relatively indiscriminate surface grazers compared to sheep. Sheep tend to be more 
selective and to penetrate to a greater depth within the vegetation canopy (Grant et 
al. , 1 985;  Collins, 1989). Goats, on the other hand, concentrate their attention on the 
vegetation at intermediate levels, shallower than sheep (Collins, 1989) .  In a tropical 
environment, goats consume a diet with a higher legume content than sheep (Norton et 
Literature Review 2 1  
al. ,  1990), reflecting the vertical distribution of  the plants . Combinations of animal 
species in range swards can allow higher biomass of herbivores per hectare by 
complementary grazing (Cumming, 1982; Gordon, 1988). However this does not 
happen on sown pastures (Nolan and Connolly, 1977; Brelin, 1979; Nicol et al. 1 993) 
because of the lack of vertical and horizontal variability for the different species to 
explore. In this case, cattle, sheep and goats have similar preference for green leaf 
components. Although Hughes et al. ( 1984) found more clover in sheep diet than in 
cattle or goat diets, this difference was negligible. In fact, diet selection is not only due 
to the animals' preference but also due to the combination of the distribution of the plant 
in the sward (vertical and horizontal) and the grazing depth. 
In this complex of sward spatial variability, animal "memory" plays an important role. 
The ability to remember spatial distribution helps the animals to increase its encounter 
rate with preferred species (Olton et al. ,  198 1 ;  Bell, 1 99 1 ). Certainly, animals can learn 
about spatial distribution (Brown and Gass, 1 993). Bailey et al. ( 1 989), using food 
patches in parallel and radial arm mazes, showed that cattle can remember the location 
of food. However lllius and Gordon ( 1 990) and mius et al. ( 1 987) argue that animals 
need to sample to learn about the alternative foods on offer. In fact, it is very difficult 
for the animals to learn about plants and the location of different patches in a complex 
community, particularly if it is changeable (Edwards, 1 994). However, Edwards ( 1994) 
demonstrated that when food patches remained in the same location, sheep learned to 
visit them. The ability of animals to return to the same preferred and known patches can 
lead to plants being changed or even grazed to extinction (Edwards, 1994), making the 
memory of limited use. However this behaviour shows the advantage to the animal of 
the partial preference described by Parsons et al ( 1994a) (see section 2.2. 1 ) .  Edwards 
( 1 994) also demonstrates that sheep rapidly abandon discredited information and 
quickly learn the new distributions of food that they experience, making it easy for 
sheep to explore new environments. Sheep remember the spatial location of food for as 
long as three days (Edwards, 1 994). Although cattle can remember for as long as 15  
days (Laca et al. (unpublished data - in Edwards, 1994), memory declines after 8 hours 
Literature Review 22 
(Bailey et al. ,  1 989) . In fact, animals use a combination of skills (eg. Spatial memory, 
sensory cues - see section 2.2.2) to choose their diet. 
2.3.3. Temporal variability affecting grazing behaviour and diet selection 
Animals have to cope not only with variability in spatial distribution in a sward, but also 
with variation over time. The temporal variation in a sward is a function of the normal 
changes in plant physiology, phenology and growth associated with seasonal or diurnal 
changes in environmental conditions (O'Reagain and Schwartz, 1995). The grazing 
action also induces temporal variability. An animal, for example, can deplete in seconds 
one plant and leave other plants intact. Grazing is a continuous process that alters sward 
component availability in time. Animals graze different parts of plants, and different 
patches, at different times. This asynchrony of defoliation affects the community 
structure (Edwards, 1 994). Some plants may be disadvantaged in terms of plant 
competition (Lubcheno, 1978;  Crawley and Pacala, 199 1 ). Selective grazing, rather than 
random grazing, increases the disadvantage of the preferred species. 
Plants at various developmental stages may possess various degrees of grazing 
protection (grazing avoidance) resulting from what Briske ( 1996) called developmental 
resistance. These mechanisms can vary throughout the growing season and with time 
following plant defoliation. Briske reviewed mechanical (production of thorns, silica, 
etc.) and biochemical plant defences that affect animal preference. Cyanide is a good 
example of short term modification. This increases within 1 8  hours of defoliation in 
potted plants of C. plectostachyus (Georgiadis and MacNaughton, 1988) .  Cyanide in 
this case acted as a deterrent to grazers. In the same way, Furstenburg and Hoven ( 1994) 
monitoring giraffe feeding behaviour realised that condensed tannin negatively affected 
acceptability and nutritional value of dietary browse. Tannin content increased due to 
browse disturbance and its level also changed daily due to temperature, light intensity 
and phenological status of the foliage (see section 2.3 .6.2).  
Literature Review 23 
The temporal modifications that affect grazing behaviour and diet selection vary from a 
very short term (occurring over a few seconds to a few hours) to the very long term 
(years) (O'Reagain and Schwartz, 1 995). Short term variability is influenced by the 
diurnal variation in plant chemical composition due to the normal plant physiological 
processes like photosynthesis, transpiration etc. This variation influences the grazing 
behaviour of ruminants. Sheep, on perennial ryegrass and on white clover swards, 
concentrated their grazing in the 4 hour before sunset when there is higher concentration 
of starch and sucrose in clover leaves and sucrose in grass (Penning et al. ,  199 1) .  In this 
way the animals probably try to maximise their net intake of energy (O'Reagain and 
Schwartz, 1 995). Animals are also affected, in the short term, by reduction in herbage 
quantity and qUality. The grazing process, for example, promotes the reduction of plant 
size, reducing the bite size and intake rate, which in turn may reduce the preference for 
the grazed plant (Laca et al. ,  1 994). 
In the medium and long term, animal grazing behaviour and diet selection respond to the 
environment (eg. soil nitrogen, soil moisture, seasonal environment variations) and 
plant (eg. seasonal cycles of vegetation growth) variation (O'Reagain and Schwartz, 
1995). Phenological development is known as one of the processes plants use to avoid 
grazing (L'Huiller et al. ,  1984; Briske, 1996). Reproductive culm development and 
accumulation of dead leaves, for example, are known to reduce animal preference for 
grasses (Willms et al. ,  1 988;  Ganskopp et al. ,  1 992). The avoidance of mature seed 
heads and stem is explained by Hodgson et al. ( 1994) as reflection of differences in 
structural strength and shear strength rather than any direct perception by the animal of 
nutritional difference. In fact, Gammon and Roberts ( 1978) found that chances of 
defoliation of individual tillers in range grassland increased with height until culm 
development, after which defoliation declined sharply. 
Herbivores demonstrate a range of strategies to cope with plant temporal variation. As 
the plant matures, pasture structure and quality change, and leaf accessibility may also 
be modified (O'Reagain, 1993). The first strategy that animals use is often to simply 
move to a different place in search for better grazing (O'Reagain and Schwartz, 1 995). 
Literature Review 24 
However, if animals are restricted to a specific area they modify their diet selection and 
grazing behaviour according to the circumstance. One of the strategies used is to 
increase dietary breadth by including other less palatable species (Owen-Smith, 1 994) or 
utilising other less palatable components (e.g. seed pods, leaf litter) of the same 
preferred species (Skinner et al., 1984 in O'Reagain and Schwartz, 1 995). Herbivores 
also increase grazing time to compensate for the decrease in intake rate (Joblin, 1 960) 
and modify their grazing time to increase selectivity. In order to increase the encounter 
rate with a preferred diet animals can increase travel speed (Collins et al. ,  1 978) or look 
more thoroughly for the preferred species (Owen-Smith, 1994). These strategies may be 
followed by metabolic changes. Sheep, for example, accumulate body fat when sward 
conditions are good, for using in time of deficit (O'Reagain and Schwartz, 1995). On the 
other hand, some animals modify their digestive strategy to cope with the poor quality of 
the forage, increasing their digestive capacity or particle retention time (Holland, 1994; 
Lechner-Doll et al. , 1 990). 
2.3.4. Effect of physical sward characteristics on herbage intake and diet selection 
Several authors (Holmes, 1 987; Nicol and Nicoll, 1987; Poppi et al, 1987;  Rattray et al, 
1 987) agree that pasture allowance has the major effect on the quantity of feed 
consumed by a ruminant. However pasture allowance can not be used by itself to 
represent the effect of sward characteristics on intake. In some cases, for example, 
pasture allowance can be high due to low stocking rate, but intake is restricted by sward 
height or density. Pasture allowance has to be accompanied by the pasture characteristic 
like height or density to define the animal effect on herbage intake. 
Sward height is one of the most important pasture characteristics that affect intake of 
grazing animals. Grazing cattle, for example, uses their tongues to pull a bunch of grass .  
It is  very difficult for cattle to graze shorter than about 10 mm (Forbes, 1995). As 
herbage declines in height, bite mass decreases, and grazing time and number of bites 
increase (Chacon and Stobbs, 1976). However as herbage becomes further defoliated (for 
cattle around 1000 kg ha- I), there is reduction in grazing time, number of bites and 
Literature Review 25 
biomass due to low leaf density (Forbes, 1995). Biting rate and grazing time are often 
regarded as the primary compensating responses of the animal to limitations in intake 
per bite (Hodgson, 198 1b). Nevertheless, increases in grazing time are seldom great 
enough to compensate for reductions in intake rate (Hodgson, 198 1 b ;  Penning, 1 986). 
Intake per bite is probably the primary animal response to variations in sward conditions 
(Gordon and Lascano, 1 993). 
Like sward height, herbage bulk density is also an important sward component that 
influences intake. Burlison et al ( 199 1 )  and Mitchell et al ( 199 1 ,  1993) show a 
continuous pattern of response in intake per bite over a wide range of sward height and 
density. According to Burlison et al ( 199 1 )  and Laca et al ( 1992) height and density 
effects are independent and additive. Estimates of both sward density and height are 
necessary to predict bite weight. Laca et al ( 1992) found that animals obtain heavier 
bites in tall sparse swards than on short dense ones of equal mass. However in short 
swards, less than 100 mm, bulk density does not seems to have a great effect (Ungar et 
al, 1 99 1 ). Bulk density, in fact, seems to be more important for tropical (tall and sparse) 
than for temperate (short and dense) pasture (Stobbs,  1973a) .  
The general rate of jaw movement (prehension, biting and chewing) in grazing animals 
is remarkably constant (Penning, 1986). Variations in biting rate (bites min-I) reflect 
variations in the relative proportions of the three jaw activities, and they are therefore 
largely influenced by the manipulation necessary to graze effectively in swards of 
different structure (Penning, 1986; Laca et al, 1993). Animals taking small bites in a 
short sward, for example, graze uninterruptedly. They swallow faster than they can eat. 
As the bite weight increases, the animals are forced to perform exclusive chewing jaw 
movements between set of bites. When animals obtain a mouthful, they lift their heads 
during the longer period of chewing which is necessary before more bites can be taken 
(Laca et al 1 993). 
A better understanding of how sward characteristics influence intake and their 
interactions with animal variables is given by Burlison et al ( 1 991 ). They defined intake 
Literature Review 26 
per bite as a product of bite volume (BV) and bulk density of herbage in grazed strata 
(BD). They also defined bite volume as a product of bite depth and bite area. In fact, 
animal intake represented by bite weight varies less than bite dimensions because of 
compensatory effects between bite depth, bite area and sward bulk density (Black and 
Kenney, 1984; Laca et al, 1992). Laca et al ( 1992), for example, found little variation in 
bite weight amongst their cattle. Animals with small bite areas took deep bites. 
Nevertheless, because of the distribution of biomass between strata within the sward, 
bite depth has a major influence on variation in bite weight (Mursan et al, 1989). 
Because of this variation, herbage mass, by itself, is not a good predictor of intake (Laca 
et al, 1 992). 
Burlison et al ( 1 99 1) ,  Mitchell et al ( 199 1 ,  1993), Laca et al ( 1992) and Gong et al 
( 1 993) found that bite depth is much more responsive than bite area to variation in 
sward conditions. In most circumstances, it is the major determinant of both bite volume 
and intake per bite. In temperate swards bite depth increases as sward height increases. 
Recent research has shown that the relationship, rather than being linear (Milne et al, 
1 982;  Burlison et al, 199 1 ), is asymptotic (Mitchell et al, 199 1 ;  Laca et al, 1992). In this 
case, the effort required to detach plant material near the ground may restrict bite depth. 
Barthram and Grant ( 1 994), working with perennial ryegrass, and Dougherty et al 
( 1 992) with tall fescue, found that pseudo stem may act as a deterrent to deep grazing 
penetration within the sward canopy. In this case, plant maturity would be an important 
factor. However some studies show that this is not necessarily the case. Burlison et al 
( 1 99 1 )  and Gong et al ( 1 993) did not find any marked difference in the relationship 
between height and bite depth for vegetative and reproductive swards. 
Several researchers (Burlison, 1 99 1 ;  Hughes et al. ,  199 1 ,  Mitchell et al. ,  199 1 ,  1 993) 
agree that bite area is less sensitive to sward change than bite depth. However, in more 
controlled conditions,  with hand constructed swards, bite area decreased linearly with 
bulk density and increased quadratic ally with sward height (Black and Kenney, 1 984; 
Laca et al. ,  1 992). In this case a positive relationship between sward height and bite area 
occurs mainly on sparse rather than on dense swards (Mitchell et al. ,  199 1 ) .  This 
Literature Review 27 
relationship is better explained by mius and Gordon ( 1987), Hughes et al. ( 199 1)  and 
Laca et al. ( 1 993) where on short swards the bite area is limited by the difficulty of 
clamping plants between incisors and dental pad. 
2.3.5. Effect of sward nutritional characteristics on herbage intake 
The biochemical characteristics of a sward have a great influence in detennining 
herbage intake. However factors that regulate dry matter intake (DMI) by ruminants are 
complex and not understood fully (NRC, 1996). 
One of the main sward nutritional characteristics that affects intake is related to energy 
and fibre content. Animals compensate for changes in the concentration of available 
energy in the food, unless the physical capacity of the rumen restricts intake (Forbes, 
1995). This effect is very well described by Conrad et al. ( 1964). They worked with 
dairy cows and concluded that intake of forage is controlled primarily by physical 
means, and the intake of more concentrated diets is controlled mainly by the cows' 
energy requirement. Later on, Bines ( 1979) working with lactating cows summarised the 
effects of the proportion of forage in the diet on the voluntary intake. The general trend 
is that there is a reduction in intake both above and below approximately 50% of forage 
in the diet. Below 50%, the reduction in intake is probably due to metabolic control, 
whereas above 50% it is due to physical limitation (Forbes, 1 995). 
Nevertheless this trend is not always true for ruminants. Infusions of glucose into the 
blood (Manning et al. ,  1959) and more recently duodenal infusions in dairy cows 
(Farvedin et al. ,  1992) failed to decrease dry matter intake (DMI). This shows that 
metabolic control of ruminants is different from monogastrics. Hodgson ( 1 982a), 
observing a simple rectilinear relationship between DMI and forage digestibility (up to 
80% of digestibility) from several trials, concluded that under grazing conditions the 
herbage intake of productive animals is seldom, if ever, likely to be affected by 
metabolic limits. Farvedin et al. ( 1 995) explain that the metabolic limit observed in 
some studies can be more related to rumen activity than to metabolic action. 
Literature Review 28 
On the other hand, the reduction in intake due to physical limitation is determined by the 
rumen capacity. The faster the rate of disappearance of food from the digestive tract the 
less the physical limit of intake. Hovell et al. (1986) shows a very close linear 
relationship between the potential degradability of the DM and voluntary intake in 
sheep. Although intake and digestibility are somewhat interdependent, they are separate 
parameters of forage qUality. Intake depends on the structural volume, and therefore the 
cell wall content, and its availability to digestion is determined by lignification and other 
factors (Van Soest, 1994). 
The relationship between various forage constituents and animal intake depends on their 
association with plant structure. Cellulose, for example is more closely related with 
intake than digestibility. On the other hand, lignin is more closely associated with 
digestibility than with intake (Van Soest, 1994). The total structural matter - the plant 
cell wall, represented by NDF (Neutral-Detergent Fibre) - is the most consistent fraction 
related to intake because the cell wall contains the entire structural substance of the 
plant within which all other components are contained (Van Soest, 1994). 
Intake is also controlled by the energy consumed. Donefer et al (1963) showed that 
intake in ruminants was not only controlled by physical limitation, but also by metabolic 
factors. Using pellets of alfalfa hay and concentrates, they found that sheep controlled 
their intake to a constant intake of digestible energy. At the same time, several 
experiments showed that infusion of short chain fatty acids depressed intake, confirming 
the possible mechanisms of metabolic control (Forbes, 1995). 
Beside energy, there are other specific nutrients that affect intake in ruminants. As in 
other animals, low protein content of the food depresses voluntary intake, but the critical 
level in ruminants is lower than in monogastric species. In ruminants, the 
microorganisms can be supplied with urea from the saliva (Forbes, 1995). In part the 
lower voluntary intake due to the deficiency of protein can be explained because protein 
deficiency reduces the activity of the rumen microflora and thus the rate of digestion of 
Literature Review 29 
cellulose (Forbes,1 995).  Minerals are also important in relation to intake. A deficiency 
of essential minerals results in reduced food intake, and an excess of many of the 
minerals causes toxic effects (Forbes, 1995) . Forbes ( 1995) comments that depression of 
voluntary intake in ruminants can be caused by: excess of arsenic, fluorine, 
molybdenum and selenium; deficiency of cobalt, magnesium, manganese and 
potassium; and excess and deficiency of calcium, copper, sodium, and zinc. He also 
mentions that deficiencies of vitamins A or D cause inappetence, and Riboflavin 
deficiency causes depressed intake in calves . 
2.3.6. Effect of biochemical characteristics on grazing behaviour, diet selection and 
herbage intake 
Plants produce a relatively distinct set of defensive chemicals and these chemical 
defences affect different animals in different ways (Freeland and Janzen, 1 974). The 
chemical defence originates from the differentiation of the cells. The cellular level of 
development can be classified as growth (cell division and enlargement) or 
differentiation (chemical and morphological changes leading to cell maturation and 
specialisation) (Herms and Mattson, 1992) . There is then a trade-off between growth 
and differentiation. A plant needs to grow fast enough to compete with other plants and 
at the same time to differentiate some cells for defence against pathogens and herbivores 
(Herms and Mattson, 1 992) . 
There are several ways that ruminants can protect themselves against the toxic effect of 
plant chemicals. One way is through grazing a variety of species (Freeland and Janzen, 
1 974; Laycock et al., 1 988). In this way, animals avoid toxic effects by eating plants or 
plant parts that do not contain large amounts of these chemicals and use several different 
detoxification pathways (Freeland and Janzen, 1974). To a certain extent this behaviour 
agrees with the partial preference postulated by Newman et al. ( 1 992) and Parsons et al. 
( 1994a) (see section 2.2. 1) .  The preference for mixed diet might indicate an 
evolutionary adaptation to reduce the possibility of toxicity. 
Literature Review 30 
In fact, free-grazing animals are faced with complex decisions, including where and for 
how long to graze (Gordon and Lascano, 1 993). Ruminants seem to avoid plants with 
strong odour or taste (Provenza et al. ,  1988), and acquire aversion to the food that 
causes illness (Burrit and Provenza, 1989a, 199 1 ;  Provenza, 1993). Animals that 
become ill after a meal of novel foods, avoid the foods whose flavours are most novel 
(Kalat, 1974; Burritt and Provenza, 1989a; Launchbaugh et al. ,  1 993; Provenza et al., 
1 994). According to Provenza ( 1995) there is increasing evidence that neurally mediated 
interactions between the senses (taste and smell) and the viscera enable ruminants to 
sense the consequences of food ingestion, and these interactions operate in subtle, but 
profound ways to affect food selection and intake, as well as the hedonic value of food. 
Animals are also able to identify plant toxins by associating food flavour with post­
ingestive feedback (Garcia, 1 989; Provenza et al. ,  1990). According to Provenza and 
Balph ( 1990) any physiochemical agent that causes nausea can cause aversion. However 
ruminants have difficulties in learning response to toxic compounds that do not affect 
the emetic system of the midbrain and brainstem (Provenza et al. 1988,  1990). Bloating, 
allergies, lower intestinal discomfort and drugs that do not affect the emetic system of 
the midbrain and brainstem are examples that animals do not learn to avoid (Garcia, 
1 989). 
On the other hand, animals show preference for the flavours of familiar foods that have 
been associated with the positive post-ingestive effects of nutrients (Provenza, 1 995a, 
1996b). Ruminants, for example, acquire preference that is paired with energy 
(Provenza, 1 996a). However, flavour may not always be a good indicator of toxicity 
(Launchbaugh et al. ,  1 993) . Often chemical changes can not be detected by animals 
through taste and smell (Bryant et al. ,  1992; Provenza et al., 1 992a). This makes it more 
difficult for mammalian herbivores to avoid phytotoxic plants. 
Animals prefer the familiar food and regard anything novel with caution. However, 
preference decreases when familiar foods are eaten too frequently or in excess, which 
encourages the consumption of novel foods and varied diets (Provenza, 1996a) . In fact, 
Literature Review 3 1  
mammalian herbivores must sample food because the nutrient content and toxicity of 
the familiar food change frequently (Freeland and Janzen, 1974; Westoby 1974, 1978).  
In this grazing process, ruminants demonstrate ability to remember the food location 
and discriminate among subtle chemical differences within plant species based on 
flavour and quickly associate with the post-ingestive consequences (Provenza and 
Balph, 1990). Provenza ( 1 996a) offered an explanation of how ruminants select diets 
from an array of plant species that vary in nutrients and toxins. Provenza argued that 
animals show aversion to the food (rather than preference) as a result of sensory 
(flavour) and post-ingestive feedback unique to each food. In this way aversions cause 
animals to sample novel foods and eat varied diets (Provenza, 1 996a). 
Another adaptation enabling ruminants to cope with plant chemicals is through the 
presence in the rumen of a diversity of bacterial and protozoal flora that can degrade a 
wide variety of secondary compounds (Freeland and Janzen, 1 974, Launchbaugh, 
1 996) . The presence of micro-organisms helps ruminants generally suffer fewer 
negative effects from poisonous plants than non-ruminants (Smith, 1992). However 
there are interactions between micro-organisms and plant chemicals that are not 
beneficial. Examples are nitrate (AlIison, 1978); cyanogenic glycosides (Conn, 1 979) 
and formononetin (Keogh et al., 1996). Formononetin (an oestrogenic compound in 
some leguminous species - see section 2.3.6.3), for example, has only a very weak 
oestrogenicity effect, but it is converted to equol in the rumen,  which is oestrogenically 
active (Shutt and Braden, 1968), and readily absorbed (Shutt et al. ,  1970) . 
2.3.6.1. Effect of plant secondary compounds on grazing behaviour and diet 
selection 
Plants produce chemicals that initially were thought not to be involved in metabolic 
processes supporting growth, development or reproduction. They were named as 
secondary compounds (Launchbaugh, 1996) . However nowadays these chemicals are 
known to be involved as regulators of plant growth or biosynthetic activities, transport 
Literature Review 32 
facilitators and nutrient or waste storage compounds (Rosenthal and Bell, 1 979). 
Ecologically, secondary compounds are very important, acting as defence substances 
against herbivory (Whittaker and Feeny, 1 97 1 ;  Rhoades and Cates, 1 976; Lindroth, 
1 989, Van Soest, 1 994) and they also enhance the ability of plants to survive stress 
conditions (Harborne, 1993). This review will deal with two specific secondary 
compounds - tannin and formononetin - of importance in this study. 
2.3.6.2. Effect of condensed tannins on grazing behaviour and diet selection 
Tannin historically was classified as the substance that converted hide into leather (Van 
Soest, 1 994 ; Bernans et al. ,  1 989; McLeod, 1 974). Nowadays it is known that tannin is 
any phenolic compound that contains enough phenolic hydroxyls to form strong 
complexes with protein and other macromolecules (Van Soest, 1 994). The classification 
of tannins into two groups, hydrolysable and condensed, by Frendenberg ( 1920) (quoted 
in McLeod, 1 974), was until now accepted. However Van Soest ( 1994) argued that this 
division is an oversimplification because some tannins contain functional properties 
characteristic of both groups, and other polyphenolics (with tannin-like properties) do 
not fit into either category. 
Condensed tannins (CT) are the most widely distributed tannins in vascular plants 
(McLeod, 1 974; Swain, 1 979), while hydrolysable tannins are restricted only to 
angiosperms (Swain, 1 979). Tannin is of little importance in the lower orders of plants 
and most of the monocotyledons, like grasses (McLeod, 1 974). In fact, the importance is 
greater in dicotyledonous plants, like the legurninosae (McLeod, 1974). Information on 
herbage and browse plants are included in this review, recognising differences in pattern 
of distribution and diet selection of these two kind of plants. 
The tannins are a group of soluble phenolic compounds distributed in several plants 
which provide defence against pathogens (bacteria and fungi) (Swain, 1 979) and 
herbivores (Rhoades and Cates, 1 976; Zucker, 1 983; Freeland et al. 1 985), and delay 
decomposition when plant tissue becomes litter (Zucker, 1 983).  Tannin does not seem to 
Literature Review 33  
have any role in  plant physiological processes (McKey, 1 979). In addition it is 
energetically expensive to the plant to produce CT. This probably explains why some 
plants only produce CT in reproductive tissue (Barry, 1 989). 
CT production fluctuates in relation to genetic and environmental variables. Roberts et 
al. ( 1 993), quantifying the amount of CT in ninety-seven accessions of birdsfoot trefoil, 
concluded that tannin concentration decreased from summer to autumn, but it was also 
related to geographic origin. Seasonal tannin changes were also verified with other 
species (Donnelly, 1 959;  Cope and Burns, 197 1 ;  Cope et al. 1 977 ; Windham et al, 1988 ;  
Furstenburg and Hoven, 1 994; lason et al. ,  1 995). Several of these studies (Clark et  al. ,  
1 939; Stitt and Clarke, 1 94 1 ;  Donnelly, 1959 ;  Cope e t  al. ,  1 97 1 ;  Windham e t  al., 1 988 ;  
lason et  al. ,  1 995) showed an increase in plant tannin concentration in summer, and 
some (Clark et al. ,  1939; Stit and Clark, 1 94 1 ;  Cope et al. ,  1 97 1 ;  Windham et al. ,  1 988), 
like Roberts et al. ( 1 993), reported a decline from summer to autumn. It is  not very clear 
how temperature and moisture affect tannin concentration. Donnelly ( 1 959) explained 
that tannin content increased with increase in temperature and decreased with rainfall. 
However, Cope et al. ( 1 97 1 )  were not able to associate rainfall and temperature with 
tannin content. On the contrary, Furstenburg and Hoven ( 1994), working with 25 tree 
and shrub species verified that tannin levels decreased with increasing temperature 
during the day and increased with descending temperature through the night, and that 
tannin content was found to be higher in the shade than in direct sunlight. They also 
noticed that tannin content increased due to browse disturbance. 
Tannin concentration is also variable according to the proportion and age of leaves. 
Leaves have higher concentration of tannin than stem (lason et al. , 1 995 and Douglas et 
al. ,  1 993). Douglas et al. ( 1 993), comparing 12 herbaceous species found that CT in 
lamina was 2-5 times that of stem. On the other hand, Donnelly ( 1 959) showed that 
tannin content increased with maturity. lason et al. ( 1 995), working with Yorkshire fog, 
reported significantly higher concentration of CT in dead versus living leaf. However, 
Coley ( 1 983) and Furstenburg and Hoven ( 1994) working with tropical trees and shrubs 
reported the opposite behaviour. Furstenburg and Hoven ( 1994) found in African 
Literature Review 34 
species that young leaves contained twice as much condensed tannin as old and mature 
leaves. 
CT levels are also related to forage quality and soil fertility. High concentrations of CT 
are in general related to high levels of lignin (Barry, 1 989). Low fertility soils are 
associated with increases in both lignin and tannin content (Barry, 1 989). CT and lignin 
are both produced in plants from the shikimic acid biochemical pathway (Swain, 1 979). 
Barry and Manley ( 1 986) argue that the most probable explanation for an increase in CT 
and lignin concentration when nutrient stress increases, is that environmental stress 
stimulates the shikimic acid biochemical pathway. 
Tannins protect plants, acting in the process of diet selection and voluntary intake. There 
are different processes that make large herbivores avoid or reduce the intake of high 
tannin content plant species. Traditionally, condensed tannins have been thought to 
decrease plant preference by digestion inhibition (Fenny, 1 969; Rhoades and Cates, 
1 976; Swain, 1 979). It is known that tannins defend plants against grazing by reducing 
protein, cell wall, and sodium digestion and retention (Rhodes and Cates, 1 973; Zucker, 
1 983 ;  Robbins et al. ,  1 987ab). Condensed tannins can form complexes by bonding with 
both carbohydrates and proteins (Barry, 1 989). At neutral pH, CT form a stronger bond 
with protein (McLeod, 1 974) . CT can then complex and render digestive enzymes 
inactive (Swain, 1979), and precipitate dietary proteins (Feeny, 1 969) making them less 
easily degraded. This may suppress microbial activity in the rumen, decreasing fibre 
digestibility and consequently intake (Barry and Blaney, 1 987; Barry, 1 989). Barry and 
Manley ( 1 984) found that binding tannin with PEG 3350 increased forage intake and 
digestibility by sheep. Barry and Blaney ( 1987) explain that this fact might be the result 
of blocking the effects of tannins (by PEG 3350) on the rumen. 
Tannin is also known as a defence compound due to its astringent properties (Bate­
smith, 1 973). According to Van Soest ( 1994), astringency is caused by the precipitation 
of salivary mucoproteins. He describes the tannin astringency flavour through the taste 
of beer, wine, tea and some fruit juice. In relation to ruminants Provenza and Malechek 
Literature Review 35 
( 1 984) argue that plants that contain high levels of tannin may also contain high levels 
of energy and nutrients, but the astringent sensation animals probably experience when 
consuming these plants may lead to their rejection. In this case, they worked with goats 
grazing two kinds of shrubs and concluded that goat nutrition was affected more by the 
adverse effects that tannins apparently had on palatability, than by the negative effects 
they had on digestibility. 
Mammalian herbivores eat nutritive plants that contains toxins, but they generally limit 
intake in accordance with the concentration of the toxin (Provenza, 1995). Animals, 
therefore, must either instinctively recognise or learn to avoid the biochemical 
compound (Chapman and Blaney, 1979; Provenza and Balph, 1 990). Instinctively 
animals would associate the flavour of the plant tissue with aversive post-ingestive 
consequences (Provenza et al., 1 990). However, there are plants high in CT that are 
highly palatable (e.g. plants of blackbrush twig studied by Provenza and Malechek, 
1984). In addition, the large variation of plants in a sward make it very difficult for 
animals to recognise and avoid plants that contain tannin (Provenza et al. ,  1 990). 
Provenza et al. ( 1990) argue that animals learn to avoid plants high in eT because of the 
internal malaise promoted by CT and not because of its flavour. This post-ingestive 
feedback of malaise is a quick process (Provenza, 1995) . Goats for example learn to 
limit intake of twigs containing tannin within one hour (Provenza et al. ,  1994). 
The kind of animal also affects the degree of selection for tannin. The greater amount of 
tannin consumed by browser ruminants and wild animals is probably associated with 
active defences against plant tannins (Robbins et al. ,  1 987a) .  Browsers, for example, 
show increased secretion of salivary proteins that bind and neutralise tannins (Mehansho 
et al. , 1987). This binding factor seems to be absent or reduced in sheep and cattle saliva 
(Austin et al. ,  1989). Long-term ingestion of tannins induces enlargement of the salivary 
glands, although it is restricted to species capable of the adaptation (Van Soest, 1 994). 
Wild animals also may have increased detoxification capabilities (Harborne, 1 993). The 
faecal losses, for example, of metabolic nitrogen (probably represented by the 
indigestible tannin-mucoprotein complex) are higher in white-tailed deer than sheep and 
Literature Review 36 
cattle (Austin et al. ,  1 989). However wild animals must still balance the rate of intake 
with the rate of detoxification (Robbin et al. ,  1 987a) . 
2.3.6.3. Effect offormononetin on grazing behaviour and diet selection 
In the 1 930' s reports related severe abnormalities in conception by ewes with the 
consumption of subterranean clover. These abnormalities was then called "clover 
disease" (Marshall, 1 973). In fact, Clover disease is promoted by isoflavonoid 
compounds, known as phytoestrogens, infabacea (Bush and Burton 1 994). High intake 
of phytoestrogens by ewes can promote oestrogenic activities, causing several 
reproductive effects: reduced fertility, dystocia, prolapse of the reproductive tract, high 
tail, increased death rate, lactation in virgin ewes and wethers, enlarged bulbo-urethral 
glands and urinary obstructions in wethers (Marshal1, 1 973). 
The most important oestrogenic compounds known to be present in pasture include the 
Isoflavones and coumestans, produced by legume species; and zearalenone, produced by 
Fusaruim species and wide spread over New Zealand pasture (Keogh, 1 995). 
Coumestans and Isoflavones are acetate-derived fragments and phenylpropanoids (see 
Wong, 1 973;  Bush and Burton, 1 994). Coumestan (main compounds coumestrol, and 
methyl coumestrol) are important phytoestrogens in lucerne and other medics 
(McDonald, 1 995). However they are generally not present in sufficient quantity, except 
in foliage that has been affected by pests and/or diseases, to cause reproductive 
problems in livestock (Keogh, 1 995). The Isoflavones are the main compounds 
responsible for the "clover disease". The best known Isoflavone phytoestrogens are 
genistein, biochain A, daidzein and formononetin (Bush and Burton 1 994). They are 
present in many Trifolium species including white clover (Trifolium repens) red clover 
(Trifolium pratense) and subterranean clover (Trifolium subterraneum) (Keogh, 1 995). 
However subterranean and red clovers are the most commonly reported species with 
moderate to high oestrogenic potency (e.g. Keogh, 1 995; Keogh et al. ,  1 996; Marshall, 
1 973; Kelly et al. ,  1 979). This is mainly due to their high concentration of 
formononetin. 
Literature Review 37 
Fonnononetin is the main Isoflavone phytoestrogen responsible for reproductive 
problems in sheep (Davies et al. ,  1 970; Keogh, 1 995; Keogh et al. ,  1 996). In fact, 
Millington et al. ( 1964) reported that the oestrogenic effects in sheep were linked to the 
fonnononetin level of different subterranean clover strains, but not biochain A or 
genisten. This was later explained by Shutt and Braden ( 1 968) who showed that 
biochain A and genistein are degraded to non-oestrogenic phenols in the rumen while 
fonnononetin is converted to equol. Equol is oestrogenically active and readily absorbed 
in the rumen (Shutt et al. ,  1 970). The presence of free equol in the blood indicates the 
oestrogenic effect in sheep. 
The concentration of fonnononetin varies within a plant. According to Keogh ( 1995) in 
both red clover and subterranean clover the highest concentrations occur in the youngest 
leaves, declining progressively as the leaves get older. Sterns show the lowest 
concentration, with increasing concentrations in petioles, expanded laminae and 
expanding laminae (Bush and Burton, 1994). However Francis and Millington ( 1965) 
found for subterranean clovers that fonnononetin concentrations are usually higher in 
leaf laminae and in sterns than in petioles. In the same way as leaves, the younger the 
stern, the higher the fonnononetin concentration (Keogh, 1 995). The distribution of 
fonnononetin within the plant shows that a high proportion of the fonnononetin is 
situated in a readily accessible position for animals to graze (Keogh, 1 995). However 
the fonnononetin concentration varies also according to the season. Kelly et al. ( 1979) 
working with the red clover cultivar Pawera in New Zealand, showed the highest 
concentration ( 1 .38%) and oestrogenic activity (32.5 /lg equivalents of oestradiol - 1 7  {J) 
in March and found the lowest concentration (0.64%) and oestrogenic activity ( 1 5.5  /lg) 
in January. Plant fonnononetin concentration can also increase with marked phosphate 
deficiency (Marshall, 1 973), waterlogging, low temperature, and defoliation (Neil and 
Marshall, 1 970; Rossiter, 1 970). 
Genetic variation has been an important tool to overcome the problem of reproductive 
problems associated with "clover disease" (Nicollier and Thompson, 1 982; Anwar, 
Literature Review 38  
1 994). In the same way as subterranean clover (Smith e t  al. ,  1 986), new cultivars of  red 
clover have been successfully selected for low fonnononetin concentration. In New 
Zealand the low fonnononetin red clover cuitivar, G27, was selected from the late­
flowering, tetraploid cultivar Pawera (New Zealand, 1 995). 
The effect of fonnononetin on diet selection is still not very clear. Rossiter and Ozanne 
( 1970) observed that when a choice is given to sheep, they show preference for 
particular cultivars of subterranean clover. However most of these cultivars have a large 
concentration of Isoflavone glycosides. Frands ( 1 973) clarified this difference in 
palatability using chemically induced mutations of the Geraldton variety of subterranean 
clover. He showed that the mutant that lacks P.glucosidase enzyme was significantly 
less palatable than the other clovers. This means that the flavonoid glycosides will 
remain intact during mastication whereas in the other cultivars they are almost 
instantaneously hydrolysed. Francis ( 1 973), therefore, concluded that in strains of clover 
high in Isoflavones, it is likely that larger amounts of unhydrolyzed glycosides will 
remain after the initial mastication and these could contribute to unpalatability. 
However, Harborne ( 1993) argues that there is no evidence of preferences for clover 
lines deficient in Isoflavone, reporting that tests show that sheep can not discriminate 
between a high Isoflavone strain of Trifolium subterraneum and a strain essentially 
lacking these compounds. Harborne ( 1993) concluded that presumably Isoflavones are 
not sufficiently repellent in taste to deter feeding. In fact, some secondary compounds 
like fonnononetin, rather than affecting diet selection or appetite, reduce the fitness of 
herbivores to avoid or limit future grazing (Rhodes and Cates, 1 976; Launchbaugh, 
1 996). Phytotoxins that affect the reproductive system of the ingesting herbivores could 
selectively remove traits from the gene pool that allow animals to detoxify or tolerate a 
particular phytotoxin (Launchbaugh, 1 996). 
Literature Review 39 
2.4. CONCLUSIONS 
This literature review covered the progress over the last three decades on the influence 
of sward characteristics on animal ingestive behaviour. Much research carried out 
elucidated and quantified the effect of sward physical and biochemical characteristics on 
animal ingestive behaviour. In the effects of physical sward characteristics, herbage 
accessibility, sward surface height, bulk density and plant botanical composition are the 
most important sward characteristics studied. The empirical observations of their effects 
on animal ingestive behaviour generated important models,  but because of the 
complexity involving the plant-animal interface, they can not be generalised for all 
grazing situations. Some of the models, for example, postulated that the animal prefers 
plants or components which maximise the intake rate, agreeing with optimal foraging 
theory. However several studies show that this is a rather simplistic way to explain a 
complex relationship. More recently research has demonstrated, for example, that 
animals constantly move between alternative swards, in most of the cases they do not 
have a unique preference, but they show a partial preference for dietary components. In 
fact, a unique explanation for the effect of physical sward characteristics on ingestive 
behaviour is apparently not practical because of the complex interactions among the 
factors concerned. Although the effects of sward physical characteristics on animal 
ingestive behaviour are becoming better understood nowadays, there is still need for 
research, involving more complex conditions. 
Together with physical sward characteristics, ruminants also respond to sward 
biochemical characteristics. More recent studies have shown that ruminants respond not 
only to a direct sensory perception but also to a post-ingestive feedback where they 
avoid grazing the component which caused malaise. In this plant-animal interface, 
secondary compounds are important plant products where some have the function of 
protecting the plants against grazing. However their effect on animal ingestive 
behaviour is variable mainly because their concentration has been shown to vary 
Literature Review 40 
according to several factors such as plant species, plant morphology, environmental 
condition, and season of the year. 
The literature shows that condensed tannin is one of the most important secondary 
compounds that apparently protect the plants against grazing. The information about 
condensed tannin has been growing mainly in relation to its effect on nutritional value, 
providing an understanding of the mechanisms of ruminant digestion and absorption 
affected by condensed tannin in forages. Studies also have shown that avoidance of 
condensed tannin may reflect either astringent taste or post-ingestive feedback. However 
knowledge of the effect of condensed tannins is still limited in most of the cases to 
either penned animals or controlled conditions where physical sward characteristics are 
not relevant. There is scope for studies that involve variation in both physical and 
biochemical sward characteristics in grazing situation. 
On the other hand, very little is known about the effects of several other secondary 
compounds on ingestive behaviour. In this context formononetin has been related to 
reproduction problems since 1 930' s, but little research has been carried out on the effect 
of formononetin on ingestive behaviour. In addition, most of the research carried out in 
this area involved sheep grazing subterranean clover. No report was found in the 
literature of a study of the influence of formononetin concentration in red clover on 
cattle diet selection. 
Although several studies in the literature have demonstrated that in nature the effects of 
sward physical and biochemical characteristics on diet selection have spatial (e.g. plant 
part level, plant level, patch level and landscape level) and temporal (e.g. seasonal 
variation, plant maturity, secondary compounds concentration) variation, few studies 
have incorporated both variations. Because of the complexity of the animal-plant 
interface, most of the studies have restricted to either sward physical or biochemical 
characteristics, using either penned animals or uniform swards. There is a need for 
research combining these effects, involving both spatial and temporal variation, on 
ingestive behaviour of grazing animals.  
CHAPTER 3 
EXPERIMENTS 1, 2, 3 AND 4 
3.1. INTRODUCTION 
Selective grazing behaviour has an important effect on influencing the dietary balance of 
animals (Parsons et al., 1994a). Recent studies on partial preference (Newman et al. ,  1992; 
Parsons et al. ,  1 994a; Cosgrove et al. ,  1996) are restricted to either pen feeding, or large 
adjacent blocks of monocultures. The literature also demonstrates that there is a lack of 
studies assessing grazing behaviour involving more complex conditions (Taylor, 1993; 
mius and Hodgson, 1996; Hodgson et al., 1997). These experiments were therefore set up 
to investigate the effects of physical sward characteristics on diet selection where the 
physical distribution of the alternative swards minimised the requirement for animal 
movement, and where the physical and nutritional characteristics of the alternative sward 
were manipulated independently. In this context, previous studies have shown little 
contrast in preference between birdsfoot trefoil and red clover (Torres-Rodriguez, 1 997). 
The small contrast observed provided an opportunity to use preferentially neutral species 
for determining the influence of sward physical characteristics on diet selection by grazing 
animals. This project also had the objective to assess the seasonal variation in diet 
selection in relation to variation in sward physical characteristics and in concentrations of 
secondary compounds. 
To achieve these objectives four experiments were carried out using swards formed by 
alternate strips of birds foot trefoil mixed with white clover, and strips of red clover, with 
individual objectives as follows: 
Experiments 1, 2, 3 and 4 42 
Experiment 1 :  to evaluate the impact of the relative areas of alternative simple swards 
on the demonstration of grazing behaviour and diet selection by cattle. 
Experiment 2:  to assess the effect of sward maturity on grazing behaviour and diet 
selection. 
Experiment 3: to assess the effect of sward height at the similar vegetative stage of growth 
on grazing behaviour and diet selection. 
Experiment 4: to give support to previous experiments in relation to the effect of the 
perimeter electric fence on animal grazing distribution on the plot. 
3.2. MATERIAL AND METHODS 
Four experiments were carried out in order to understand how the contrasting 
characteristics of two swards (birdsfoot trefoil with white clover, and red clover) affect 
diet selection, grazing behaviour and intake: 
Experiment 1 :  contrasting area ratio of the two swards offered, to assess the effects 
of horizontal sward structure on selective grazing. 
Experiment 2: contrasting periods of plant regrowth imposed, to assess the 
influence of plant maturity on selective grazing . 
Experiment 3 :  contrasting sward height offered at similar stages of vegetative 
development, to assess the effect of vertical sward structure on 
selective grazing. 
Experiment 4: assessment of the effects of spatial distribution of each sward in the 
plots upon selective grazing. This was a limited trial, the main 
objective being evaluation of proximity to fences on animal 
behaviour. 
General details of site and experimental techniques are outlined here. Specific procedures 
applicable to individual experiments are considered later in the chapter. 
Experiments 1, 2, 3 and 4 43 
3.2.1. Experimental site 
The experiments were carried out between 30 October 1 995 and 20 May 1 996 at 
AgResearch Flock House in the ManawatulRangitikei region (40° 1 6'S, 1 75° 17'E). The 
site is a sandy soil classified as Rangitikei loamy sand (Cowie et al. ,  1 972) on low-lying 
alluvial flats bordering the Rangitikei River, about 9 m above sea level. Average annual 
precipitation in this area is 875 mm with a dry period from January to March and strong 
westerly winds during October to November (spring). The weather conditions, 
determined approximately 300m from the experiment site, show an average monthly 
temperature ranging from 9°C (July) to 20°C (January). The daily rainfall and mean soil 
temperature ( 10  cm depth) during Experiments 1 ,  2 and 3 are presented in Appendix 
3.2.  
3.2.2. Swards 
The trials were set up on a sward formed by alternate 2.4 m wide strips (see Plate 3 . 1 .) of a 
mixture of birds foot trefoil (Lotus corniculatus L.) cv. Goldie and white clover (Trifolium 
repens L.) cv. Pitau, and strips of red clover (Trifolium pratense L.) cv. Colenso. The area 
was sown in November 1 993 with 8 kg of coated seedlha of birdsfoot trefoil and red 
clover. The white clover originated from the seed bank formed by the previous sward of a 
mixture of white clover, perennial ryegrass (Lolium perenne L.) and cocksfoot (Dactylis 
glomerata L.). In contrast to a substantial content of white clover in birdsfoot trefoil, very 
small amounts of volunteer white clover were found with red clover. 
In the two years preceding the trial, the sward formed from alternate strips was rotationally 
grazed by steers. The previous sward had been grazed by cattle, ewes with lambs at foot, 
and weaned lambs. 
Prior to sowing birdsfoot trefoil and red clover, 250 kglha of superphophate was 
applied. In addition, 200 kglha of DAP 13S was applied to the area annUally. 
Experiments 1, 2, 3 and 4 44 
Plate 3 .  1 .  General view of experimental swards formed by alternate 2.4 m wide strips of 
a mixture of birdsfoot trefoil (Lotus corniculatus L.) cv. Goldie and white 
clover (Trifolium repens L.) cv. Pitau, and strips of red clover (Trifolium 
pratense L.) cv. Colenso 
Plate 3 .2. General view of experimental swards with grazing animals 
Experiments 1, 2, 3 and 4 45 
3.2.3. Design 
The experiments were set out in a Row-Column Design balanced for previous treatment, 
using four treatments and five replicate groups of three heifers in each experiment 
(Table 3. 1 .) .  In each experiment, four treatments were randomised in each period and 
replicate groups were allocated to treatments over time. The fifth replication was 
allocated as one extra treatment in each period. This design was used to control the 
difference between periods, the difference between groups of heifers and the effect of 
previous treatments. In this case, five replications were used in order to provide enough 
degrees of freedom for the residual. 
Table 3 . 1 .  Distribution of four treatments with five groups of three heifers over four 
periods. 
Period 1 2 3 4 
Group of 
Heifers 
1 B D A C 
2 D B C A 
3 C A D B 
4 A B C D 
5 D C B A 
In order to obtain similar groups, the animals were separated into three main classes 
according to weight: heavy, medium and light. One animal from each class was randomly 
chosen to form each experimental group of three animals. The animals grazed for fifty five 
hours on each replication from 1 .00 p.m. on Day 1 to 8.00 p.m. on Day 3 .  Between 
replications there were four-day intervals for pasture assessment and animal 
acclimatisation for the next treatment. Each group of three animals stayed together, 
allocated to treatments over time in each experiment. 
Experiments 1, 2, 3 and 4 46 
3.2.4. Animals 
Three different yearling dairy-cross heifer groups were used in Experiments 1 ,  2, and 3 ,  
respectively. The same heifers were used in Experiments 3 and 4 .  The overall average 
weight before starting each trial was 264 kg (ranging from 237 to 290 kg), 288 kg (ranging 
from 222 to 365 kg), and 17 1  kg (ranging from 145 to 208 kg) for Experiments 1 , 2 and 3,  
respectively. All heifers were drenched with anthelmintic to remove any internal parasites, 
and bloat capsules were administered prior to each experiment. 
For at least two weeks prior to the trial and during four days between replications, all 
heifers grazed an area adjacent to the plots composed of the same species and strips as the 
experimental swards. These adjacent areas were used for acclimatisation and reduction of 
previous treatment effects. 
3.2.5. Measurements 
3.2.5.1. Sward measurements 
Herbage mass 
Six samples were cut to ground level before and after grazing in each sward type of each 
treatment ( 12 samples per plot), using an electric shearing handpiece and a square 
sampling frame 0. 1 m2 in area. After cutting, the samples were washed, dried to constant 
weight in a forced-draught oven at a temperature of 70-80oC, and weighed. 
Botanical composition 
Six samples of each sward in each treatment, at least 50g each, were also cut to ground 
level (excluding litter), before and after grazing, from the area beside the sampling frame 
used in the dry matter assessment. They were sealed in polythene bags and taken to the 
laboratory in an icebox. The six samples were bulked within plots and from this two sub-
Experiments 1, 2, 3 and 4 47 
samples were taken. The fIrst was used for assessment of botanical composition and 
separated into species, and then into leaf, petiole (only in clovers), stem (or stolon in white 
clover), flower and dead material. All the samples were dried and weighed individually. 
The second sub-sample was separated, as for the botanical composition determination, and 
freeze-dried for chemical analysis. These samples were ground prior to analysis using a 
hammer mill fItted with a 1 mm screen. 
Pasture height and bulk density, 
Forty random readings were recorded from each sward in each plot using a sward stick 
(Bircham, 198 1 ;  Barthram, 1986). Readings were taken during the pre-grazing assessment, 
before the second day of grazing observation and during the post -grazing assessment. 
Sward bulk density was calculated by dividing herbage mass (g DMlcm2) by sward height 
(cm) for each plot. 
Pasture structure 
An inclined point quadrat (Rhodes and Collins 1993; Montossi et al., 1 994) was used to 
assess the vertical distribution of plant tissue within the sward canopy. Forty contacts per 
sward per treatment were recorded. The point quadrat observation was taken randomly 
before and after grazing. 
3.2.5.2. Grazing Behaviour 
The animals were observed during each the three days of grazing in each treatment for 3 
hours/day, using the method of Jarnieson and Hodgson ( 1 979). The distribution of grazing 
by individual animals between swards was recorded each evening from 4.00 to 7.00 pm at 
intervals of 10 minutes. ill addition, morning observations (from 6.40 am to 9.40 am.) 
were carried out in the morning of the second day in order to assess the importance of 
diurnal variation in selective behaviour. 
Experiments 1, 2, 3 and 4 48 
Between each 10  minutes of recording, rates of biting were measured using the 20 bite 
method of Forbes and Hodgson ( 1 985). The seconds spent for the animal to take 20 bites 
were recorded. However, if the animal lifted its head the watch was stopped until the 
animal started grazing again. If the animal did not resume grazing in less than a minute, 
the recording was not considered. At least two assessments for each animal on each sward 
were recorded in each observation period. 
3.2.6. Chemical analysis 
Separated samples (see section 3.2.5. 1) were stored at -20°C. The samples were then 
freeze dried and ground to pass through a 1 mm diameter screen. 
Each plant part of birdsfoot trefoil, white clover and red clover was analysed for 
formononetin concentration. Analyses were carried out on pre-grazing samples. One 
sample of leaf, petiole, stem and flower (when available) of birds foot trefoil, white 
clover and red clover was randomly chosen from each replication for the analysis of 
formononetin. In Experiment 2, separate analyses were carried out on samples of 
immature and mature plants. The formononetin content was determined by a fluorimetric 
assay described by Gosden and Jones ( 1978) and modified by Anwar ( 1 994). 
The analysis of extractable condensed tannin concentration was done only on birdsfoot 
trefoil (leaves and stems) since preliminary analysis showed insignificant concentration of 
extractable condensed tannin in leaves, petioles and stems of red and white clover. In 
Experiment 1 three samples, of three different replications, were randomly chosen; in 
Experiments 2 and 3 one sample, of each replication, from immature and mature and from 
short and tall swards, respectively, were randomly chosen and analysed. The analysis of 
extractable condensed tannin was carried out by a modification of the DMACA-HCl 
Protocol described by Li et al. ( 1 996). The modified methodology is described in 
Appendix 3 . 1 .  
Experiments 1, 2, 3 and 4 49 
Similar amounts of samples were bulked to perform the herbage quality analysis. In 
Experiment 1 the samples were bulked across treatments and replicates to get one 
sample of each plant part for birdsfoot trefoil, white clover and red clover. In 
Experiments 2 and 3 the samples were bulked as in Experiment 1 ,  but contrasts in plant 
maturity, and plant height characteristics, respectively, were bulked separately. 
Conventional indices of forage quality (protein, neutral detergent fibre (NDP), acid 
detergent fibre (ADP), carbohydrates (soluble sugars plus starch), ash and lipid) were 
determined by Near Infrared Reflectance Spectroscopy (NlRS) (Shenk and Westerhaus, 
1994). In vitro dry matter digestibility (IVDMD) were calculated using the following 
equation (Roberts and Packman, 1983): 
IVDMD = (-0.896 x ADP) + 95. 85 
3.2.7. Statistical analysis 
The sward and animal data were analysed using the statistical package SAS (SAS Institute 
Inc.,  1985 and 1990) and S-plus (Math Soft, 1 995). Analyses of variance were carried out 
to obtain information about the differences between legume species in relation to grazing 
time and rate of biting, balanced for previous grazing experience. Analyses of variance 
were also performed for comparison of the sward characteristics. Least square means and 
standard error of the difference was used to quantify the contrasts in grazing behaviour and 
sward characteristics. Regression analyses were carried out in Experiment 1 to clarify the 
relationships between the proportion of area or herbage mass allocated to the birdsfoot 
trefoil/white clover sward and the proportion of either grazing time or intake on that 
sward. Contrasts of sward maturity (in Experiment 2) and sward surface height (in 
Experiment 3) effects were also performed to determine the effects of the treatments 
(within each sward type and of the alternative sward) on grazing behaviour and animal 
intake. The point quadrat data were summarised using a point quadrat package (Butler, 
1 99 1) .  
Experiments 1, 2, 3 and 4 50 
3.2.8. Experimental layouts 
3.2.8.1. Experiment 1 
The fIrst experiment was carried-out from 30 October until 27 November 1 995 . Each 
replication was composed of four 405 m2 plots fenced to provide distinct area ratios 
(treatments) of birdsfoot trefoil and white clover (BW) in relation to red clover (RC). Four 
treatments were imposed, the area ratio in percentage of each sward per treatment being: 
20:80; 33 :67;  67:33; 80:20 (see Figure 3 . 1 .). 
3.2.8.2. Experiment 2 
Experiment 2 started on 5 February, and fInished on 1 March 1 996. The treatments were 
arranged in order to compare differences in maturity of BW in relation to RC. 
Four treatments were imposed. Each treatment was formed by different regrowth periods 
of BW and RC to provide maturity differences. The swards with a short period of 
regrowth (immature) were mowed fIrst to 5 cm, and then topped after 4 weeks of 
regrowth to remove the flowers, leaving an residual of approximately 12 cm. After 
topping, this area had 3 more weeks of regrowth. Swards with a long period of regrowth 
(mature) had 9 weeks of regrowth after been mowed to 5 cm. Treatments provided all four 
combinations of maturity (immature/mature) and sward type (see Figure 3 .2.). 
The plot sizes were calculated in order to provide similar total quantities of herbage per 
group of heifers. Therefore, the plot size was determined according to the treatment. The 
treatment with more mature plants, for example, was estimated to provide greater herbage 
mass and, consequently, less area was allocated than for treatments with less mature 
plants. 
Experiments 1, 2, 3 and 4 5 1 
3.2.8.3. Experiment 3 
Experiment 3 ran from 1 5 April to 10  May 1 996. The treatments were arranged in order 
to compare contrasts in height at the same vegetative stage of growth for birdsfoot trefoil 
and white clover in relation to red clover. The short swards were formed by mowing to 
approximately 7 cm height, then allowing four weeks of regrowth until the experiment 
started. The tall swards were left without being mowed (approximately 8 weeks of 
regrowth). 
As in Experiment 2, plot sizes were calculated in order to provide a similar amount of 
herbage per group. Therefore, the plot size was determined according to the treatment. The 
treatment with taller plants, for example, was estimated to provide greater herbage mass 
and, consequently, less area than treatments with shorter plants (see Figure 3 .3.) .  
3.2.8.4. Experiment 4 
Experiment 4 was a small trial with the objective of assessing the effect of position in the 
plot on sward use. Each plot was fenced to incorporate four strips, one of which was 
mown to ground level to provide all combinations of one or two grazeble strips, adjacent 
to or separated from a fence, for the two sward types (see Figure 3 .4.). 
Experiments 1, 2, 3 and 4 52 
Figure 3 .  1 .  Experiment 1 - treatment layout : area ratio birdsfoot trefoil and white 
clover (BW) : red clover (RC) (not to scale - strips of 2.4 cm width) . Total 
area of each plot = 405 m2 
Treatment: 20% : 80% 
1---------------------------------------------67.5 m-----------------------------------------------I 
RC 
BW 
RC 
1------------------33.9 m-------------------I 
Treatment: 33% : 67% 
1----------------------------------56.2 m---------------------------------------I 
RC 
BW 
RC 
Treatment: 67% : 33% 
1---------------------------------56.2 m----------------------------------------I 
BW 
RC 
BW 
Treatment: 80% : 20% 
1------------------------------------------67.5 m---------------------------------------------------I 
BW 
RC 
BW 
1------------------3 3 .9 m-------------------I 
7.2 m 
Experiments 1, 2, 3 and 4 53 
Figure 3 .  2. Experiment 2 - treatment layout: plant maturity contrast [3 weeks ( ) and 9 
weeks (//Ill) of re growth] of birdsfoot trefoil and white clover (BW) : red 
clover (RC) (not to scale - strips of 2 .4 cm width). 
Treatment: 3 weeks : 3 weeks 
1------------------------------------------3 1 .25 m------------------------------------------I 
BW 
RC 
BW 
RC 
BW 
RC 
Treatment: 3 weeks : 9 weeks 
1-------------------------------22.5 m-----------------------------I 
BW 
I I I I I I I I I I I I I I I I I I 1//1 I I I II I I I I I I I IRCI I I I I I I I I I I I I I //I I I I I I I I I I I I I I I I I I I 
BW Total area = 324 m2 
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I IRCI //I I I I I I I I I I //I I I I I I I I I I I I I I I I I I I I 
BW 
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIRCIIIIIIIIIIIIIIIIII//111111111111111 
Treatment: 9 weeks : 3 weeks 
1------------------------------22.5 m------------------------------I 
II//IIIIIIIIIIIII//IIIIIIIIIIIIIIIIBWIIIIIIIIIIIIIIIIII//IIIIIIIIIIIIII 
RC 
I I //I II // I I / I Ill/fill I Ill/If/If I I I lBW 1/////////// I I I III //I // I III I //I // I Total area = 324 m2 
RC 
1//11/11/111111/111///////// I///I/IBW //I 11//1//1///11////11/1//1////// I 
RC 
Treatment: 9 weeks : 9 weeks 
1---------------------- 17.57 m ---------------------1 . 
////I/I///III/I/IIIII////IBWI////I/I///II//////I/IIII/ 
1///III/I/IIIIIIII//II///IRCIII/II/IIIIIIIIIIIIIIIIII11 
//1/1/11/1 IIIII//II/II//IBW 1111///111111111//11/11111 I Total area = 253 m2 
IIIIIIII/IIIIII/III/IIIIIRCIIIIIIIIIIIIIIIIIIIIIII//III 
IIIII IIII 1111111//1 I III lBW IIIIII 1//1 I I I I 111//1// //I I I I 
I I I I I I I I I I I I I // I I I I I I I /IRCI 11// I / I / I //I I I 11///11// I I I I / 
14 .4 m 
Total. area 
= 450 m2 
I 
Experiments 1, 2, 3 and 4 54 
Figure 3. 3 .  Experiment 3 - treatment layout : plant height contrast [short( ) and tall (Ill)] 
of birdsfoot trefoil and white clover (BW) : red clover (RC) (not to scale - strips of 
2.4 cm width). 
Treatment: short : short 
12.2 m 
BW 
RC 
BW 
RC 
Treatment: short : tall 
1---------------------------------------35.2 m----------------------------------------I 
BW 
-I 
9.6 m 
Total area 
= 405 m2 
1111111111111/111111/111111111111111/11111/111 IRCIIIII/III 111111111111111111111111/11/1111111 I11 Total area = 338 m2 
BW 
I I I I I I I I I I I I I I I I I I 11 I I I I I I I I I I I I 1111/111 I I I I I I IRCI I I I I I I I I I I I I I I I I I 11/1 I I I I I I I I 1111/1 I I / I I I Ill/I 
Treatment: tall : short 
1---------------------------------------35.2 m----------------------------------------I 
1111/1/11 I I I I 11/ I 11/ / I I I I I I I I I I I I I I I I 11/11111 I lBW 11/11/1/11/1 I I / I / I I I I / I / I I I I / I I I 11// I / I I I I / I I I 
RC Total area = 338 m2 
11/111/1/1/1111/11/1/11/1/1 I I I / I I /1111// Ill/I /IBW 1/11/1/11/111/1 I I I I I / 11/ I 11// I I /111/1/11/1/11/ 
RC 
Treatment: tall : tall 
1-----------------------2 1 . 1 m---------------------I 
1111/1/111/111 III/IIIIII/IBW 111/11/11/1111/11/11/1111/1 
III/II/II/IIII/IIIIII/IIIIRCII/I/III/IIIIIII/IIIIIIIII/ Total area = 202.6 m2 
I I I 11// Ill/ 1/ 11111111 1/11 lBW 111/111/11/111/11/11/11/1/1 
II/III! I I / I I / I I I / I I I 11/ I /IRC/II/III! I / I I I I 111/11/11/11/1 
Experiments 1, 2, 3 and 4 55 
Figure 3 .  4. Experiment 4 - treatment layout : spatial distribution contrast of birdsfoot 
trefoil and white clover (BW) : red clover (RC) (not to scale - strips of 2.4 
cm width). Total area of each plot = 1 59.4 m2 
Treatment: BW beside 2 strips of RC 
1-------------------- 16.6 m--------------------------I 
RC 
////////////////////I////mowedl/l////I/////////I//////I/ 9.6 m 
RC 
BW 
Treatment: BW between 2 strips of RC 
RC 
BW 
RC 
/ / / / / //I / / / / / / / / / / / / / / / / /mowedl / / / //1// / / / / / / / / /1/ / //I / / 
Treatment: RC beside 2 strips of BW 
BW 
/I//////I// //////II//////mowedl////////I/////II//////// 
BW 
RC 
Treatment: RC between 2 strips of BW 
BW 
RC 
BW 
//1// //I I I I I I I I I I I / / I I I I I Imowedl ///11/1 I I I I I I / /1 I I I I I I I I 
Experiments 1, 2, 3 and 4 56 
3.3. RESULTS 
The results of the first four experiments are presented in this section. The results of each 
experiment are shown separately. 
3.3.1. Experiment 1: Effect of the proportion of area of birdsfoot trefoil (Lotus 
corniculatus L.) and white clover (Trifolium repens L.) sward 
in relation to red clover (Trifolium pratense L.) sward on 
grazing behaviour, diet selection and herbage intake. 
3.3.1.1. Sward measurements 
Herbage mass, sward surface height and sward bulk densib!. 
Most of the interactions between treatments and sward type were not significant, so results 
are presented as main effects only. The results for herbage mass, surface sward height and 
sward bulk density are given in Table 3.2. The interaction means between treatments and 
sward type are presented in Appendix 3.3. 
Both before and after grazing, red clover swards had significantly greater herbage mass 
and height, but significantly lower bulk density. This resulted in a similar amount of 
herbage removed per sward. 
After grazing, sward height and bulk density showed a significant interaction between 
sward type and treatments (Appendix 3.3). In this case the difference in height between 
RC and BW swards within each treatment increased as the BW area decreased. 
Experiments 1, 2, 3 and 4 57 
Table 3 .2. Herbage mass, sward height and bulk density before and after grazing, and 
estimation of the herbage mass removed for birdsfoot trefoil and white 
clover (BW) and red clover (RC) swards in Experiment 1 .  
Herbage mass (kg DMIha) 
Pre-grazing 
Post grazing 
DM removed 
Sward height (cm) 
Pre-grazing 
After 1 day grazing 
Post-grazing 
Bulk density (mg DMkm3) 
BW RC SEDi P-value2 
3940 
2380 
1560 
19. 1 
12 .3 
7.5 
4570 
3340 
1230 
27.7 
2 1 .9 
1 3 .3 
1 64 
145 
206 
1 .33  
0.87 
0.35 
0.0014  
0.0000 
0. 1 329 
0.0000 
0.0000 
0.0000 
Pre-grazing 2.06 1 .67 0. 172 0.0027 
Post-grazing 3 . 17 2.5 1  0.2 16  0.0074 
1 SED - Standard error for differences of means when comparing BW with RC swards. 
2 P-value of the sward main effect. 
There was also a significant interaction between sward type and treatments for bulk 
density post-grazing. The BW sward became denser, from pre to post grazing, more than 
the RC sward, except when 80 % of the area offered was BW. 
Sward composition 
The botanical composition of each sward before and after grazing is shown in Table 3 .3 .  
The interactions between treatments and sward types were not significant for most of 
plant components, so results are presented as main effects only. The interactions 
between treatments and sward types are given in Appendix 3 .4. 
Experiments 1, 2, 3 and 4 58 
Table 3. 3. Botanical composition of birdsfoot trefoil and white clover (BW) and red 
clover (RC) swards before and after grazing (DM basis): (a) percentage of 
components in live fraction, (b) percentage of live matter in total DM of 
each sward and (c) ratio of the total live matter of birds foot trefoil and white 
clover (BIW) in the BW sward, EX12eriment 1 .  
P-value'i BW RC SED' 
Pre-grazing 
(a) Leaf 44.0 4 1 . 1  2.89 0.3293 
Petiole 19.5 15 .8  2.99 0.3264 
Stem 1 5.5 34.3 2.58 0.0000 
Flower 0.8 0.9 0.30 0.7000 
Grass 8.3 1 .8 2.86 0.0350 
Other species 12 . 1  6.2 2.98 0.0680 
(b) Total live matter 95.7 9 1 .5 1 .97 0.0003 
(c) BIW ratio 0.68 
Post-grazing 
(a) Leaf 27.8  1 7.3 2.56 0.0008 
Petiole 29. 1 1 6.2 1 .53 0.0001 
Stem 27.7 48. 1 3 .36 0.0000 
Flower 0.7 0.8 0.52 0.8229 
Grass 5 .3 5.5 3 .8 1 0.9464 
Other species 9. 1 1 1 .9 2.90 0.3362 
(b) Total live matter 9 1 .3 84.3 1 .74 0.0003 
(c) BIW ratio 0.40 
I SED - Standard error for differences of means when comparing BW with RC swards. 
2 P-value of the sward type main effect. 
Before grazing, the RC sward had significantly more stem material, more dead matter 
and less other broad leaf species than BW. Flowers were a minor component in both 
swards. 
There were modifications in botanical composition as a consequence of grazing, though 
no statistical analyses were performed to compare pre with post grazing. The percentage 
of dead matter and stem increased from pre to post grazing in both swards. The decline 
in leaf content was greater in RC than in BW swards. After grazing, there were 
significantly more leaves and petioles in BW than in RC swards. White clover was the 
predominant species in BW before and after grazing, but there was greater difference 
after grazing. 
Experiments 1, 2, 3 and 4 59 
3.3.1.2. Canopy structure within the sward 
The point quadrat data from the two swards are summarised in Figures 3 .5 and 3.6.  
Comparing the four treatments within each sward type, either before or after grazing, the 
structures and distributions of plant parts were similar. In all cases the dead material was 
distributed mainly in the bottom strata. In the BW sward before grazing, white clover 
was the main contributor, distributed mainly in the bottom and medium strata. Birdsfoot 
trefoil was more evenly distributed vertically in the canopy. In the RC sward before 
grazing, there was a major contribution of leaves and petioles mainly in the medium and 
upper strata. The proportion of stems was larger in the mid and bottom strata of the 
canopy. The relative contribution of red clover leaves, petioles and stems in the bottom 
strata increased after grazing. 
3.3.1.3. Sward chemical composition 
Extractable condensed tannin concentration 
The concentration of extractable condensed tannin in leaves and stems of birdsfoot 
trefoil is presented in Table 3 .4. 
Table 3. 4. Extractable condensed tannin (ECT) concentration (%) of birdsfoot trefoil 
leaf and stem in Experiment 1 (DM basis). 
Components ECT 
Leaf 1 .69 
Stem 0.05 
1 SEM - Standard error of the means. 
Number of observations contributing for each mean (n=3) 
0.090 
0.003 
The data in Table 3 .4. are the average of three replications for leaves and stems. The 
concentration of ECT was higher in leaves than in stems. The stems showed negligible 
amount of ECT. 
60 
BIRDSFOOT TREFOIL & WHITE CLOVER 
Before grazing After grazing 
54-57 54-57 � o BT leaf Treatment 20% I Treatment 20% 1 BT stem 45-48 45-48 � • BT flower 
36-39 36-39 J WC leaf o WC petiole 
27-30 27-30 WC stolon o WC llower 
1 8-2 1 1 8-21  o Grass o Other broad leaf 
9- 1 2  9- 12  o Dead rraterial 
0-3 0-3 
0 10  20 30 40 50 60 70 80 90 0 1 0  20 30 40 50 60 70 80 90 
54-57 54-57 ] 
Treatment 33% Treatment 33% 
45-48 45-48 j 
36-39 36-39 
27-30 27-30 .j 
1 8-21  1 8-2 1 
E 
9- 1 2  9- 12  
� 0-3 0-3 
� g 0 1 0  20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 
Cl) 54-57 1 54-57 
0 Treatment 67% Treatment 67% a:: 45-48 45-48 � 
Cl) 36-39 36-39 
27-30 27-30 
1 8-2 1  1 8-2 1  
9- 12  9- 12  
0-3 0-3 
0 1 0  20 30 40 50 60 70 80 90 0 1 0  20 30 40 50 60 70 80 90 
54-57 54-57 = 
Treatment 80% Treatment 80% 
45-48 45-48 � 
I 
36-39 36-39 J 
27-30 27-30 
1 8-2 1  1 8-21  
9- 12  9- 12  
0-3 0-3 
0 1 0  20 30 40 50 60 70 80 90 0 1 0  20 30 40 50 60 70 80 90 
NUMBER OF CONTACTS FOR PLANT PARTS 
Figure 3 . 5 .  The stratum structure of plant parts within the sward canopy of birdsfoot 
trefoil and white clover before and after grazing in the four treatments (area 
ratio) in Experiment 1 .  
6 1  
RED CLOVER 
Before After 
90-93 90-93 .RC leaf Treatment 80% Treatment 80% 
75-78 75-73 ORC petiole RC stem 
60-63 60-63 .RC flower .WC leaf 
45-48 45-48 O WC petiole WC stolon 
30-33 30-33 OWC flower o Grass 
1 5 - 1 8  1 5- 18  
o Other broad leaf 
o Dead material 
0-3 0-3 
0 1 0  20 30 40 50 60 0 1 0  20 30 40 50 60 
84-87 Treatment 67% 84-87 Treatment 67% 
72-75 72-75 
60-63 60-63 
48-5 1 48-5 1 
36-39 36-39 
24-27 24-27 
1 2- 1 5  1 2- 1 5  
E .!::- 0-3 0-3 
< � 0 1 0  20 30 40 50 60 0 1 0  20 30 40 50 60 < � � CIl 84-87 Treatment 33% 84-87 Treatment 33% 
� 72-75 72-75 < 
� 
CIl 60-63 60-63 
48-5 1 48-5 1 
36-39 36-39 
24-27 24-27 
1 2- 1 5  12- 15  
0-3 0-3 
0 1 0  20 30 40 50 60 0 1 0  20 30 40 50 60 
90-93 90-93 
Treatment 20% Treatment 20% 
75-78 75-78 
60-63 60-63 
45-48 45-48 
30-33 30-33 
1 5- 1 8  1 5- 1 8  
0-3 0-3 
0 1 0  20 30 40 50 60 0 1 0  20 30 40 50 60 
NUMBER OF CONTACTS FOR PLANT PARTS 
Figure 3 .6 .  The stratum structure of plant parts within the sward canopy of red clover, 
before and after grazing in the four treatments (area ratio) in Experiments 1 .  
Experiments 1, 2, 3 and 4 62 
Formononetin concentration 
The fonnononetin concentrations in BW and RC sward main species are presented in 
Table 3 .5 .  
Red clover had significantly higher concentration of  formononetin than birdsfoot trefoil 
and white clover in leaves, petioles and stems. Red clover had higher concentration of 
fonnononetin in leaves than in petioles or stems. The formononetin contents of red and 
white clover flowers were similar. 
Table 3 .5 .  Formononetin concentration (%) of leaf, petiole, stem and flower of 
birdsfoot trefoil (BT), white clover (WC) in birdsfoot trefoil and white 
clover sward (BW), and red clover (RC) in red clover sward (RC) III 
Experiment 1 .  
Plant BW sward R C  SED1 P-value2 
components 
Leaf 
Petiole 
Stem 
Flower 
BT 
0.08 
6 
0. 12  
WC 
0. 19  
0. 1 6  
0. 1 6  
I SED - Standard error for differences of means 
2 P-vaIue for the comparison of species. 
sward 
RC 
0.0103 0.00213 
0.6 1 0.0804 0.00604 
0.0805 0.01435 
0.50 0.056 0.0089 
0.30 0.029 0.0 1 16 
0. 1 7  0.045 0.9296 
3 Standard error for differences of means and P-value for comparison between BT and Wc. 
4 Standard error for differences of means and P-vaIue for comparison between BT and RC. 
5 Standard error for differences of means and P-vaIue for comparison between WC and RC. 
6 _ no sample 
Number of observations contributing to each mean (n=4) 
Experiments 1, 2, 3 and 4 63 
General chemical composition 
The general chemical composition of each plant part is presented in Table 3 .6. Because 
the values represent a single sample derived by bulking across treatments and 
replications, statistical analysis was not possible. In general, the three species, white 
clover (WC), birdsfoot trefoil (BT) and red clover (RC) were similar in the percentage 
of crude protein (CP), acid and neutral detergent fibre (ADF, NDF) and in vitro dry 
matter digestibility (IVDMD). Although BT had a slightly lower percentage of CP, 
ADF and NDF in leaves than did red clover and white clover, the three legumes had low 
percentages of ADF and NDF and high percentages of all other chemical components, 
showing the high nutritive value of these species. BT had higher CHO in leaves, but 
lower percentage in stems, than WC and RC. 
Table 3. 6. Crude protein (CP), lipid, acid and neutral detergent fibre (ADF, NDF), 
carbohydrates (soluble sugars plus starch)(CHO), ash and in vitro dry matter 
digestibility (IVDMD) determined by Near Infrared Reflectance 
Spectroscopy (NIRS) of the main components of birdsfoot trefoil and white 
clover, and red clover swards in EXEeriment 1 (Eercenta�e of DM basis). 
CP LIPID ADF NDF CHO ASH IVDMD 
White clover 
leaf 34.0 2.9 1 9.2 20.6 5.3 1 2. 1  78.6 
petiole 14.9 1 . 1  3 1 .7 33.6 8 . 1 10.7 67.4 
stolon 1 7.9 0.9 25.3  2 1 .5 1 2.4 9.0 73.2 
flower 20.6 3 .3 26.9 17 .0 7 . 1  1 2.6 71 .7 
Birdsfoot trefoil 
leaf 27.9 3 .9 17.0 15.7 1 3 .6 10.4 80.6 
stem 13 .8  1 .6 33 .5 40.5 6.4 7.9 65.8 
Red clover 
leaf 33.3 3.0 20.5 2 1 .5 1 1 . 1  1 1 .3  77.5 
petiole 14.6 1 .5 32.5 34.6 10.4 10.0 66.7 
stem 12.4 0.8 32.6 34.5 1 3 .0 8.0 66.6 
flower 1 9.7 3.4 27. 3  2 1 .6 9.7 1 1 .6 71 .4 
Experiments 1, 2, 3 and 4 64 
3.1.1.2. Animal measurements 
Grazing time and intake 
Measurements of grazing time and estimates of herbage intake per day are summarised 
in Table 3 .7 .  
There was a significant interaction between treatment and sward type for grazing time in 
each of the three days of grazing. The magnitude of this interaction increased from the 
first to the last day. In absolute terms, as time passed the animals spent more time 
grazing the larger areas and consequently the difference in time spent grazing between 
different swards increased from Day 1 to Day 3. The total number of minutes spent 
grazing per animal also increased from Day 1 to Day 3 .  
Results of the morning observation for Day 2 was consistently intermediate between 
evening observations of Day 1 and Day 2 in this and the following experiments 
(Appendix.3 .6). Further evaluation of behaviour results is confined to evening 
observations. 
Table 3 .  7. The effect of treatments (area ratios) on grazing time (minutes) in the first, second and third days of grazing (Days 1 ,  2 and 3), 
and average DM intake per animal per day (kg DMlhd/day) during 55 hours of grazing in Experiment 1 .  
Treatment A Treatment B Treatment C Treatment D 
BW RC BW RC BW RC BW RC 
20% 
Grazing time (min) 
Day 1 26.3 
Day 2 14.5 
Day 3 1 1 .7 
80% 
47.6 
82.5 
100.4 
33% 
4 1 .9 
25.0 
19. 1 
67% 
49.3 
65.0 
76.5 
67% 
5 1 .9 
57.5 
72.4 
33% 
3 1 .2 
25.5 
39. 1 
80% 
58.5 
72.3 
82.7 
20% 
28.5 
1 6. 3  
1 9.4 
10.75 
5.46 
10.25 
Intake 2. 1 7  6.34 3.04 4.98 5.75 2 .24 5 .88 1 .62 1 .425 
(kgDMlhd/day) 
I SED - Standard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
P­
value2 
0.0129 
0.0000 
0.0000 
0.0004 
Experiments 1, 2, 3 and 4 66 
There were significant effects of treatment on the proportion of grazing time (Propngt) 
activity on BW strips for all three days (Table 3 .S),  with no indication of a significant 
effect of the previous treatment between periods and no significant differences between 
groups of animals (P>0.05). Within days, variation in the distribution of grazing activity 
was linearly related to the proportional area of BW available (Propnarea, P<O.OOl ) ,  with 
no significant residual treatment effects (P>0.05) and no significant differences between 
groups of animals (P>0.05). On Dayl the animals allocated grazing activity 
preferentially to the minor sward component on each treatment, and the regression slope 
of Propngt on Propnarea was correspondingly low (Table 3.9, Figure 3 .7). However, in 
Days 2 and 3 the regression did not differ significantly from a 1 :  1 relationship, implying 
neutrality of choice (Figure 3.9, Figure 3.7). 
Regression based on the proportions of herbage DM on BW strips (PrOPIldm) also 
showed similar linear trends between the proportion of DM offered and proportion of 
grazing activity with no significant residual treatment effects (P>0.05) (Table 3.9).  This 
regression also differed significantly from 1 :  1 relationship (neutrality) on the first day 
but not on the last day of grazing (Figure 3.S) .  In these linear trends, there was no 
significant previous treatment and group of animals effect. In both regressions, propngt 
on propnarea and propngt on Propndm, the first day grazing activity was more variable.  
There was a significant interaction between treatment and sward type effects on 
estimated forage intake per animal per day. On average an animal consumed between 
7.5-S.5 kg DMlday. There was a linear relationship between the proportion of total 
intake from BW and proportion of area offered to the animals. The slope of the 
regression was not significantly different from 1 :  1 (slope == 0.76 ± 0.25), but the mean 
was significantly higher than neutrality (Figure 3 .9). There was also no indication of a 
significant effect of previous treatment between periods, and no significant residual 
treatment effect and significant group of animals effect (P>0.05) .  
Table 3 .  8 .  Treatment (20, 33,  67 and 80 % of the total area offered) effects on the proportion of grazing time (in relation to the total 
grazing time spent in plot) devoted to birds foot trefoil plus white clover swards (BW) in Experiment 1 .  
Proportion of area of BW 
20% 33% 67% 80% SED1 P-value2 
Day 1 0.35 0.46 0.60 0.69 
Day 2 0. 15  0.26 0.70 0.80 
Day 3 0.09 0.20 0.66 0.79 
1 SED - Standard error for differences of means when comparing different treatments. 
2 P-value of the treatment main effect. 
0.069 
0.035 
0.074 
0.0040 
0.0001 
0.0001 
Table 3 .  9. Regression slopes of the proportion of grazing time (propngrJ in relation to the proportion of area (Propnarea) and dry matter 
(propndm) offered in the first, second and third days of grazing observation in Experiment 1 (slope significance in relation to 
neutrality value of 1 .0). 
Regression of Day ] Day 2 Day 3 
Propngt on  propnarea 0.57 ± 0. 1 154 **  1 .09 ± 0.0583 NS 1 . 1 7  ± 0 . 1235 NS 
Propngt on  propndm 0.55 ± 0.0929 *** 1 .20 ± 0. 1093 NS 
Significance of  difference from 1 .0: **  P�O.Ol and NS (not significant). 
The slopes are originated from the analysis of variance, after period and group of animals had been added to the model. 
Experiments 1, 2, 3 and 4 
0.9 
0.8 
:> 
co 0.7 
c 
0 
Cl 
c 0.6 ·N � 
Cl 
<l) 0.5 ,§ 
-0 
c 0.4 0 :e 
0 
e 0. 3 0-
0.2 
0.1 
0 
0 0 .1  0.2 0.3 0.4 0.5 0.6 0.7 0.8 
proportion of area offered of BW 
• Day 1 
• Day 2 
A Day 3 
- - Day 1 
- - - Day 2 
Day 3 
---Neutrality 
0.9 
68 
Figure 3. 7 .  Proportion of grazing time in relation to the proportion of area offered of 
birdsfoot trefoil plus white clover (BW) 
Experiments 1, 2, 3 and 4 
1 � OBl 
c o 
Q) � 0.6 
0> c 
·N 
� 
0>0.4 .... o 
c o 
t 8. 0.2 A 
• 
• • Day 1 
A Day 3 
-N:t..traity 
• Day 1  
- -Day3  
§. A 
O��
A�i 
____ � ____ � ____ � __ __ 
o 0.2 0.4 0.6 0.8 
r:rop:nim ri t-e1::Ega rrass (0'111) dfered ri eJV 
69 
Figure 3 .  8. Proportion of grazing time in relation to the proportion of herbage mass 
offered of birdsfoot trefoil plus white clover (BW) 
Experiments 1, 2, 3 and 4 70 
1 .6 
• 
1 .4 
• 
� 1 .2 • 
CD 
c: 
0 
Q) • .:L 
� 0.8 /. '0 A c: 0 • / • :e 0.6 0 a. 
e 
a. 
0.4 • 
• Intake 
0.2 -Neutrality 
• - Intake 
0 
0 0.2 0.4 0.6 0.8 
proportion of area offered of SW 
Figure 3. 9. Proportion of intake in relation to the proportion of area offered of birdsfoot 
trefoil plus white clover (BW) 
Experiments 1, 2, 3 and 4 7 1  
Rate orbiting 
There was no significant interaction between treatment and sward type on the rate of 
biting on any of the three days of grazing. There was a significant main sward type 
effect (Table 3 . 1 0) with no significant difference between treatments. Bite rates were 
consistently higher on BW than on RC swards over all three days of measurements (52.3 
vs 46.3 ± 0.59 bites/minute P=::;;O.OOO l ) . The rate of biting increased in both species and 
the statistical significance of the difference between BW and RC decreased from the 
first to the last day of grazing. The interaction means between treatments and sward type 
are presented in Appendix 3 .3 .  
Table 3 .  1 0. The effect of  swards of birds foot trefoil and white clover (BW), and red 
clover (RC) on rate of biting (bites/minute) in the first, second and third 
days (total 55 hours) of grazing assessment in Experiment 1 .  
BW RC SED] P-value2 
Rate of biting (biteslmin) 
Day 1 49.5 45. 1 0.72 0.0000 
Day 2 52.7 46.2 1 .36 0.0003 
Day 3 54.3 47.5 1 .63 0.0238 
1 SED - Standard error for differences of means when comparing BW with RC swards. 
2 P-value of the sward main effect. 
Experiments 1, 2, 3 and 4 7 2  
3.3.2. Experiment 2: Effect of the maturity of birdsfoot trefoil (Lotus corniculatus L.) 
and white clover (Trifolium repens L.) in relation to red clover 
(Trifolium pratense L.) on grazing behaviour, diet selection and 
herbage intake. 
3.3.2.1. Sward measurements 
Herbage mass. sward surface height and sward bulk density 
The results of herbage mass, sward surface height and sward bulk density are given in 
Table 3 . 1 1 . 
There was a significant interaction between treatment and sward type in relation to 
herbage mass. Before grazing, the sward with the longer period of regrowth (mature) 
had higher herbage mass. Considering only mature swards, herbage mass was greater in 
RC than BW. In contrast, for immature swards, herbage mass was greater in BW. 
Although there was more herbage mass removed from mature swards, there was no 
significant interaction between treatment and sward type in relation to the difference 
between pre and post grazing herbage mass. However there was also significant 
treatment main effect. The two treatments that had mature RC had the greatest amount 
of mass removed and the treatment with both swards (BW and RC) immature the 
smallest. 
There was a significant interaction between treatment and sward type in relation to pre 
grazing height. Within each treatment, RC was always the tallest sward and mature 
plants were taller than immature. After the first day of grazing, there was no significant 
interaction between treatment and sward type in relation to height. RC swards were 
significantly taller ( 1 1 .8 vs 1 5 .2 cm, SED 0.700, P<O.OOOO4) after one day of grazing, 
but not after three days ( 7 . 8  vs 8 .4, SED 0.90, P<0.5409). 
Table 3 .  1 1 . Herbage mass (kg DMlha), sward height (cm) and bulk density (mg DMfcm3) before and after grazing, and estimation of the 
herbage mass removed (kg DMlha) of birds foot trefoil and white clover (BW) and red clover (RC) swards according to 
treatment in EXEeriment 2 .  
Treatment A Treatment B Treatment C Treatment D 
BW RC BW RC BW RC BW RC SEDi p-
Immature Immature Immature Mature Mature Immature Mature Mature value2 
Herbage mass 
(kgdmlha) 
Pre-grazing 4270 3550 4240 6470 5450 3830 5660 5760 7 1 0  0.0044 
Post grazing 2860 2730 2050 4080 3320 2550 3090 3290 393 0.0002 
DM removed 1 420 820 2 190 2400 2 1 30 1 290 2570 2470 672 0.6859 
Sward height (cm) 
Pre-grazing 1 8.5 27.6 19.0 39.3 24.3 26.9 22.2 38.6 1 .92 0.0000 
After 1 day grazing 1 2.3 1 5.4 10.8 1 5.4 1 2.7 1 5. 5  1 1 .5 14.5 1 .40 0.7840 
Post-grazing 8 .5  9 .0  7.6 8.4 7.7 8.3 7.6 7.9 1 .80 0.9960 
Bulk 
density( mgdm/cm3) 
Pre-grazing 2.33 1 .28 2.20 1 .65  2.28 1 .50 2.55 1 .50 0.266 0.5 1 90 
Post-�razin� 3 .77 3 .94 2.73 5.30 4.75 3 .4 1  4.08 4.74 0.973 0.0608 
1 SED - Standard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
Experiments 1, 2, 3 and 4 74 
There was also no significant interaction between treatment and sward type for bulk 
density. BW was consistently denser before grazing (main effect 2.34 vs 1 .47 
mgDMlcm3, SED 0. 1 33 ,  P<O.OOOO). After the third day of grazing the swards did not 
differ significantly (main effect 3 .83 vs 4.35 mgDMlcm3, SED 0.486 ,  P<0.2966). Both 
swards had much higher density after grazing than before. 
Sward composition 
Before grazing, there was a significant interaction between sward type and treatment 
effect for percentage of leaves and a marginal interaction for stems (Table 3 . 1 2) .  The 
difference in percentage of leaves was smaller when both swards were in the same stage 
of regrowth. Swards with long period of regrowth (mature) had a smaller proportion of 
leaves than with short period of regrowth (immature). Comparing within each treatment 
before grazing, immature swards had proportionally fewer stems and more leaves than 
mature. Although the interaction was not significant, there was much higher amount of 
flowers in mature than in immature swards. There was also a marginally higher amount 
of broad leaf weeds on RC than on BW (5. 1 vs 9.7, SED 2 .30, P<0.0579). 
The sward type x treatment interaction for stem content was stronger after than before 
grazing, but that for leaf content declined. The percentage of stems increased from pre 
to post grazing and the difference between immature and mature within each treatment 
also increased. The percentage of flowers decreased but there was still a large contrast 
between immature and mature within each treatment. The percentage of dead matter 
increased in the sward from pre to post grazing. 
Table 3. 12 .  Botanical composition of birdsfoot trefoil plus white clover (BW) and red clover (RC) swards before and after grazing, according to 
the treatments (plant maturity: Immature and Mature) (DM basis):  (a) percentage of components in live fraction, (b) percentage of 
live matter in total DM of each sward and (c) ratio of the total live matter of birdsfoot trefoil and white clover (BIW) in the BW 
sward, EXEeriment 2. 
Treatment A Treatment B Treatment C Treatment D 
BW RC BW RC BW RC BW RC SEDl P-value2 
Immature Immature Immature Mature Mature Immature Mature Mature 
Pre-grazing 
(a) Leaf 30.2 3 1 . 1  34.7 24.9 22.3 3 1 .8 23 .7 25.5 3 .70 0.0094 
Petiole 13 . 8  14 .7 1 3 .4 8 .6 14 . 1 14.5 1 0.5  9 .8  4.25 0.7825 
Stem 47.7 35 .2 36.5 44.6 46.2 36.7 48.4 39.3 5.55 0.0535 
Flower 5. 1 6 .3  4.9 1 2.6 1 2.4 6.4 1 3 .5  1 9.2 3.75 0.0734 
Grass 0.0 0.0 1 .7 0.0 0.0 0.0 0.3 0.2 0.68 0.2279 
Other species 3.2 1 2.8  8.7 9.3 5 .0 10.5  3 .5 6.0 4.60 0.54 16  
(b) Total live matter 90.3 88 .7 90.3 9 1 .3 90.2 9 1 .0 93.6 93.0 2.92 0.9 147 
(c) BIW ratio 1 .79 2.76 4.76 7 .07 2.606 0.6467 
Post-grazing 
(a) Leaf 1 5 .2  1 3 .6 1 5 .0 8 .6  8 .6 9 .8 9 .2 5 .4 2.74 0.2530 
Petiole 1 3 . 1 1 4.5 1 5 .0 8 . 3  8 .6  1 0.6 6.2 6.8 2.46 0.0627 
Stem 65.5 47.7 53.5 64. 1 62. 1 37.5 63.2 66. 1 6.50 0.00 1 5  
Flower 2.3 3 .3 1 . 1  9 .8  8 .6  2.9 10 . 1 5 .4 3 . 1 0  0.0102 
Grass 1 . 1  0.0 2.6 0.0 3 .7 0.4 0.4 1 .2 2.23 0.6003 
Other species 2.7 20.8 1 2.8  9 . 1 8 .5  38 .6 10 .9 1 5 .2 7 . 1 0  0.0 109 
(b) Total l ive matter 79. 1  7 1 .3 79.2 8 1 . 1  8 1 .3 68.8 89.3 69.2 5 .99 0.0920 
(c) BIW ratio 2.04 0.49 2.77 4.99 0.698 0.01 23 
I SED Standard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
Experiments 1, 2, 3 and 4 76 
Considering only the sward type main effect, BW had, on average, significantly higher 
percentage of live matter than RC (82.2 vs 72.6 %, SED 3.00, P<0.0032). 
The amount of birds foot trefoil was higher than the amount of white clover in all 
treatments, except in the treatment where BW was immature in post-grazing. There 
were only significant differences between treatments after grazing. The treatment where 
both swards were mature had significantly higher amount of birdsfoot trefoil in relation 
to white clover. 
3.3.2.2. Canopy structure within the sward 
The vertical distributions of the plant parts of BW and RC swards within the sward 
canopy are shown in Figure 3 . 1 0  and 3 . 1 1 . In BW swards birds foot trefoil had a 
proportionally larger contribution in upper strata than white clover. However the 
differences between treatments decreased after grazing. In RC swards there was a 
similar distribution of plant parts, comparing treatments within each period (before of 
after grazing), but mature swards had components in higher strata. In both swards the 
dead matter was distributed in the bottom strata and the leaf and petiole components 
were proportionally greater in the medium and high strata. The number of contacts in 
the lower strata increased after grazing. 
54-57 
45-48 
36-39 
27-30 
1 8-2 1 
9- 1 2  
0-3 
0 10  
54-57 
45-48 
36-39 
27-30 
1 8-2 1 
9- 1 2  
E 
� 0-3 
� 
0 1 0  g 
Cl) 54-57 
Cl 
� 45-48 ::: 
Cl) 
36-39 
27-30 
1 8-2 1 
9- 1 2  
0-3 
0 1 0  
54-57 
45-48 
36-39 
27-30 
1 8-2 1 
9- 1 2  
0-3 
0 1 0  
77 
BIRDSFOOT TREFOIL & WHITE CLOVER 
Before grazing After grazing 
54-57 DBT  leaf Treatment SW lrnrmture:RC lrnrmture Treatment SW lrnrmture: 
45-48 RC lmmature BT stern 
• BT flower 
36-39 • WC leaf 
tlI WC petiole 
27-30 WC stolon 
D WC flower 
1 8-21  DGass 
D Other broad leaf 
9- 12  
0-3 
20 30 40 50 60 70 80 
54-57 
Treatment SW lrnrmture:RC mature 
45-48 
36-39 
27-30 
1 8-2 1  
9- 1 2  
0-3 
20 30 40 50 60 70 80 
Treatment SW Mature:RC lrnmature 
54-57 
45-48 
36-39 
1 8-2 1  
9- 1 2  
0-3 
20 30 40 50 60 70 80 
Treat:rn::nt SW Mature:RC Mature 
54-57 
45-48 
36-39 
27-30 
1 8-2 1 
9- 1 2  
0-3 
20 30 40 50 60 70 80 
D Dead material 
0 10  20 30 40 50 60 70 80 
Treatment SW Immature:RC mature 
0 10  20 30 40 50 60 70 80 
Treatment SW Mature:RC Immature 
I 
r r 
.� .' - 1 
• I •• 1 - 11." 
· - , • 
0 10  20 30 40 50 60 70 80 
Treat:rn::nt SW Mature:RC Mature 
0 10  20 30 40 50 60 70 80 
NUMBER OF CONTACTS FOR PLANT PARTS 
F igure 3 . l 0.The stratum structure of plant parts within the sward canopy of birdsfoot 
trefoil and white clover (BW), before and after grazmg in the four 
treatments (RC = red clover sward) in Experiment 2 .  
90-93 
75-78 
60-63 
45-48 
30-33 
90-93 
75-78 
60-63 
45-48 
30-33 
1 5- 18  
78 
RED CLOVER 
Before 
90-93 
Treatment BW Immature : RC Immature 
75-78 
60-63 
45-48 
30-33 
[ 5- 18  
0-3 
o 10 20 30 40 50 60 70 80 90 1 00 
90-93 
Treatment BW Immature : RC Mature 
75-78 
60-63 
45-48 
30-33 
15- 1 8  
0-3 
After 
Treatment BW Immature: 
RC Immature 
.RC leaf 
ORC petiole 
RC stem 
.RC flower 
.WC leaf 
[J WC petiole 
WC stolon 
OWC flower 
o Grass 
o Other broad leaf 
[J Dead material 
o 10 20 30 40 50 60 70 80 90 1 00 
Treatment B W Immature : RC Mature 
o 1 0  20 30 40 50 60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 
90-93 90-93 
Treatment BW Mature : RC Immature Treatment BW Mature:RC Immature 
75-78 75-78 
60-63 
45-48 
30-33 
1 5- 1 8  
90-93 
75-78 
60-63 
45-48 
30-33 
1 5- 1 8  
o 1 0  20  30 40  50  60  70  80  90  1 00 
Treatment BW Mature : RC Mature 
o 1 0  20 30 40 50 60 70 80 90 1 00 
60-63 
45-48 
30-33 
1 5- 18  
0-3 
90-93 
75-78 
60-63 
45-48 
30-33 
1 5-1 8 
o 10 20 30 40 50 60 70 80 90 l OO 
Treatment BW Mature : RC Mature 
o [ 0  20 30 40 50 60 70 80 90 l OO 
NUMBER OF CONTACTS FOR PLANT PARTS 
Figure 3 . 1 1 .The stratum structure of plant parts within the sward canopy of red clover 
(RC), before and after grazing in the four treatments (BW = birdsfoot trefoil 
and white clover sward) in Experiment 2. 
Experiments 1, 2, 3 and 4 79 
3.3.2.3. Sward chemical composition 
Extractable condensed tannin concentration 
The concentrations of extractable condensed tannin (ECT) in leaves and stems of 
birds foot trefoil plants for swards with long (Mature) and short (Immature) period of 
regrowth are given in Table 3 . 1 3 . The concentration of ECT in leaves of immature 
plants was twice that in mature plants, a difference approaching significance at the 5 % 
level of probability over four replicates. The concentration of ECT in stems was 
insignificant at both maturities. The concentration of ECT in white clover and red clover 
was also negligible. 
Table 3 . 1 3 . The effect of extractable condensed tannin (ECT) concentrations of 
birdsfoot trefoil leaf and stem in Experiment 2 (%DM basis) according to 
the treatment. 
Leaf 
Stem 
Mature 
0.37 
0.005 
Immature 
0.79 
0.005 
0. 1478 
0.0200 
P-valui 
0.0665 
1 .0000 
I SED - Standard error for differences of means when comparing Mature 
with Immature swards. 
2 P-value of the treatment main effect. 
Number of observations contributing for each mean (n:::::4) 
Formononetin concentration 
There was no significant difference in formononetin concentration between immature 
and mature swards (Appendix 3 .4). The formononetin concentrations in the main 
species of BW and RC swards are presented in Table 3 . 14. Compared with birdsfoot 
trefoil and white clover, red clover had a higher percentage of formononetin in all plant 
components, except flowers. The main concentrations of formononetin in red clover 
were in leaves and petioles. Birdsfoot trefoil did not differ from white clover in 
formononetin concentration in leaves. Although flowers were significantly different, 
they all had relatively low concentrations of formononetin. 
Experiments 1, 2, 3 and 4 80 
Table 3 . 1 4. Fonnononetin concentration (%) of leaf, petiole, stem and flower of 
birdsfoot trefoil (BT), white clover (WC), in birdsfoot trefoil and white 
clover sward (BW), and red clover (RC), in red clover sward (RC) in 
Experiment 2. 
Leaf 
Petiole 
Stem 
Flower 
BWsward 
BT WC 
0. 1 2  
4 
0. 10  
0.07 
0. 17  
0.2 1 
0. 1 3  
R C  sward 
RC 
0.50 
0.42 
0.33 
0. 1 0  
0.04003 
0.0458 
0.0358 
0.0 12 1 (BT-RC) 
0.01 1 3  (WC-RC) 
1 SED - Standard error for differences of means when comparing means of each specie. 
2 P-value for the comparison of species. 
3 SED and P-value for the comparison of RC vs BW and RC vs Wc. 
4_ no sample 
P-value2 
0.0001 3 
0.0034 
0.0007 
0.0014 
Number of  observations contributing to the mean of each specie (n=8), but n=5 for birdsfoot trefoil, n=6 
for white clover and n=8 for red clover in flower assessment. 
General chemical composition 
The general chemical composition of each plant part is presented in Table 3 . 15 .  There was 
a similar chemical composition between the three main species, and only a small reduction 
in quality, comparing immature and mature swards. Birdsfoot trefoil had lower percentage 
of crude protein in the stems, compared with the other two species, and, as in Experiment 
1 ,  had higher proportion of soluble sugars plus starch in the leaves than in the stems. 
Experiments 1, 2, 3 and 4 8 1  
Table 3 .  1 5 .  Crude protein (CP), lipid, acid and neutral detergent fiber (ADF, NDF), 
carbohydrates (soluble sugars plus starch)(CHO), ash and in vitro dry matter 
digestibility (IVDMD) determined by Near Infrared Reflectance 
Spectroscopy (NlRS) of the main components of birdsfoot trefoil plus white 
� 
§ F-. � 
� 
� 
§ F-. 
� 
clover, and red clover swards in Experiment 2 according to the period of 
regrowt h C d ) ( f DM b 
. ) Immature an mature lpercentage 0 asIS . 
CP LIPID ADF NDF CHO ASH 
White clover 
Leaf 35.4 2.3 1 7 . 1 1 8.3 5 . 3  1 2.5 
Petiole 1 6.9 1 .0 30.6 3 1 .9 7 .0 1 1 .5 
Stolon 17. 9  0.7 22.2 1 9.5 1 8.5 8 . 1  
Flower 2 1 .0 3 . 3  30.5 30.2 5 . 1 1 3 . 8  
Birdsfoot trefoil 
Leaf 32.8 3 .4 1 6.2 1 7.7 9.9 1 1 .7 
Stem 9.9 1 .5 40.5 54.7 1 .7 7 . 6  
Flower 28.0 7.6 1 8 .0 6.5 8 . 2  1 1 . 9  
Red clover 
Leaf 36.2 2.7 20. 8 25.0 9.8 1 2.4 
Petiole 1 5 . 8  1 .5 3 3 .4 35.6 8 .6 1 0.6 
Stem 1 1 .4 0.9 36.7 43. 1 9.0 8 . 3  
Flower 20.4 3 . 1 30. 1 28.2 7 . 1  1 3 .3 
White clover 
Leaf 33.6 2.6 1 8 .5 23.0 5.3 1 1 .9 
Petiole 1 6 .4 0.9 3 1 .4 33.4 6.2 10.9 
Stolon 1 8. 8  0.6 22.0 1 6.9 1 8 .2 8 . 2  
Flower 20.5 3.3 30.4 30.2 5 . 5  1 3. 6  
Birdsfoot trefoil 
Leaf 3 1 . 1  3 .2 1 6.7 1 6.5 9.9 1 1 .8 
Stem 7.8 1 .4 4 1 .3 56.9 2.3 7 . 3  
Flower 1 1 .7 4.3 24.2 22.9 1 1 .9 8.2 
Red clover 
Leaf 34.9 2.6 20.9 24.9 1 0.0 1 2.2 
Petiole 14.5 1 .6 34.3 3 8.5 8 .4 1 0.9 
Stem 8 . 3  0.8 39.3 49.7 9.8 7 . 3  
Flower 20.0 3 . 1  30.6 29.2 6.7 1 3 . 6  
NDMD 
80.5 
68.4 
76.0 
68.5 
8 1 .3 
59.6 
79.7 
77.2 
65.9 
63.0 
68.9 
79.3 
67.7 
76. 1  
68. 6  
80.9 
58.8 
74.2 
77. 1 
65. 1  
60.6 
68.4 
Experiments 1, 2, 3 and 4 82 
3.3.2.4. Animal measurements 
Total grazing time and intake 
Interactions between sward type and treatment were not significant for grazing time and 
intake variables, so results are presented as main effects only in Table 3 . 1 6  and Table 
3 . 1 7. The interactions are given in Appendix 3.4. 
Although estimates of intake from each sward did not differ, the animals consistently 
spent more time on red clover than on birdsfoot trefoil, independent of the sward 
maturity (Table 3 . 1 6) .  The total time spent grazing each treatment increased from the 
first to the third day, but there was no significant difference between treatments (Table 
3 . 1 7.) .  Except when both swards where mature, the animals seemed to increase more the 
total grazing time per treatment from Day I to Day 2 than from Day 2 to Day 3 .  
Table 3 .  1 6. The effect of swards of  birds foot trefoil and white clover (BW), and red 
clover (RC) on grazing time (minute) during the three hours of grazing 
assessment in Days 1 ,  2 and 3 and on average intake per animal per day (kg 
DM/hd/day) in Experiment 2. 
BW RC SED} P-value' 
Grazing time (min) 
Day 1 
Day 2 
Day 3 
3 1 .8 
33.6 
32.3 
44.9 
52.3 
55.6 
4.97 
4.40 
5 . 1 1  
0.0 1 83 
0.0006 
0.0003 
Intake (kg DM/hd/day) 4.86 3 .95 0.7 1 8  0.2228 
1 SED _ Standard error for differences of means when comparing BW with RC swards .  
2 P-value of the sward main effect. 
Experiments 1, 2, 3 and 4 83 
Table 3. 17. The effect of treatment on total grazing time (minute) during the three hours 
of grazing assessment in Days 1 ,  2 and 3 and on average intake per animal 
per day (kg DMlhd/day) on birdsfoot trefoil plus white clover (BW), and red 
clover (RC) sward in Experiment 2 (treatment A= BW and RC immature; 
treatment B=BW immature and RC mature; treatment C= BW mature and 
RC immature; D= BW and RC mature) .  
Treatments 
A B D 
Grazing time(min) 
Day 1 36.6 38.2 40.3 38.3 
Day 2 44. 8  4 1 .3 48.5 37.2 
Day 3 45.9 42.7 46.5 40.6 
Intake (kg DM/hd/day) 3.57 5 .45 3 .98 4.62 
1 SED - Standard error for differences of means when comparing treatments. 
2 P-value of the treatment main effect 
SEDl P-valui 
5 .48 0.92 1 4  
5 . 63 0. 1 998 
6.74 0.7670 
1 . 1 00 0. 3 1 05 
The effects of maturity contrast (immature - mature) within each sward type and within 
the alternative sward type on grazing time and intake within each type are shown in 
Table 3 . 1 8 . The contrasts within each sward type were calculated as the differences in 
grazing time or intake/animal/day between immature and mature swards, at the same 
maturity for the alternative swards. The contrasts (where the alternative swards were 
either mature or immature) were then averaged. Contrasts within the alternative sward 
type was calculated as the differences in grazing time or intake/animal/day, within sward 
type, when the alternative sward was either immature or mature, and averaging to the 
results. Negative numbers represented greater effect of mature than immature sward. 
The Table 3 . 1 8  shows that there was no significant effect of BW or RC maturity on the 
time spent grazing in BW or RC, respectively_ However in the third day the animals 
spent more time grazing BW when RC was immature than when it was mature. 
Experiments 1, 2, 3 and 4 84 
Although there were no significant differences in intake/animal/day between swards or 
treatments (Tables 3 . 1 6 and 3 . 1 7), mature RC swards had on average higher intake than 
immature (Table 3 . 1 8) .  There was no apparent effect of BW maturity on 
intake/animal/day from BW, or of alternative sward maturity on intake from either 
sward. 
The effect of treatment on the proportion of time spent grazing BW is shown in Table 
3 . 1 9. There were no significant effects of treatment on the proportion of total grazing 
time on BW swards in Days 1 and 2 of grazing. However there was a significant 
treatment effect on Day 3 (Table 3 . 1 9.)  when the animals spent proportionally less time 
grazing immature BW sward and more time grazing mature RC. 
Table 3. 1 8 .  Sward maturity contrast effect on total grazing time and intake/animal/day 
contrasting the effect within each sward type (BW or RC) and the effect of 
alternative sward in the days 1 ,  2 and 3 of grazing assessment in Experiment 
2. 
Grazing time 
Day 1 Day 2 Da>: 3 
Effect on behaviour on BW 
Maturity effect of B W  sward 4.542 - 1 .49 -3 .34 
Maturity effect of RC sward -5.59 1 .7 1  1 5 .35* 
Effect on behaviour on RC 
Maturity effect of RC sward 2. 1 5 
Maturity effect of BW sward 5.37 
3.54 
-8.99 
-6.73 
0.90 
5. 730 5.426 5.902 
Intake 
-0.23 
-0. 14 
-2.27* 
0. 1 2 
0.991 
I SED - Standard error when comparing means with same levels of specie. 
Contrast significance: * P::;O.05 
2 Contrast: immature - mature (negative numbers represent greater 
effect of mature than immature sward) 
Experiments 1, 2, 3 and 4 85 
Table 3 .  1 9.Treatment effects on the proportion of grazing time (in relation to the total 
grazing time spent in the plot) devoted to birdsfoot trefoil and white clover 
swards (BW) in Experiment 2 Treatment A= BW and RC immature; 
Treatment B=BW immature and RC mature; Treatment C= BW mature and 
RC immature; Treatment D= BW and RC mature) .  
Treatments 
SED1 P-valui 
A B C D 
Day 1 0.39 0.5 1  0.39 0.4 1 0.0567 0.4334 
Day 2 0.42 0.39 0.38 0.38 0.06 14  0.97 1 4  
Day 3 0.42 0.27 0.46 0.34 0.042 1 0.0464 
I Standard error for differences of treatment means 
2 P-value for the comparison of treatrnent main effect. 
Grazing time per kg ofDM offered 
The total time the animals spent grazing during the 3 hours of observation in Day 1 and 
Day 3 was divided by the amount of herbage mass available pre and post-grazing, 
respectively, in order to verify the effect of herbage mass on the distribution of grazing 
activity. The interaction between treatments and species for the number of minutes the 
animals spent grazing per kg of DM in each sward is given in Table 3 .20. There was a 
significant difference between species (and marginal significant interaction between 
treatment and species) in Day 1 (0.41 vs 0.64 minJkgDM, SED: 0.079, P=0.0 105, in BW 
and RC, respectively) and a significant interaction between treatment and species in Day 
3 .  In both days, the animals spent more time, per kg of DM, grazing red clover swards, 
except where BW was immature and RC mature. The difference between swards (within 
each treatment) was greater when BW was mature. The largest difference, in both days, 
happened when RC was immature and BW was mature. The time the animals spent 
grazing per kg of DM increased from Day 1 to Day 3 .  
Experiments 1, 2, 3 and 4 86 
Table 3 .  20. The effect of interaction between treatment (maturity: immature (Imm) and 
mature (Mat» and sward type (birdsfoot trefoil and white clover (BW), and 
red clover (RC» on grazing time per unit of DM (minlkg DM), in the first 
and third days of grazing in Experiment 2. 
Treatments 
A B C D 
B W  RC BW RC B W  RC BW RC 
Imm Imm Imm Mat Mat Imm Mat Mat 
Day 1 0.34 0.61 0.57 OA2 0.32 0.80 OAl 0.73 0. 158 
Day 3 0.64 0.80 0.88 0.79 0.66 1 .5 1  0.87 1 .35 0.216  
1 SED- Standard error for differences o f  means when comparing swards within each treatment. 
2 p_ value of the interaction between treatment and sward type. 
Rate orBiting 
p-
0.0701 
0.0403 
The results of rate of biting are presented in Table 3 .2 1 .  There were significant interactions 
between sward type and treatment effects in the three days of grazing. Bite rate was 
consistently higher on BW than on RC in all four treatments on Day 1 (average 46.7 vs 
40.8 bites/min, SED 1 .57, P=O.OOO l ) ,  and the difference between BW and RC was larger 
when RC was mature and BW immature. In Days 2 and 3 there was also a higher rate of 
biting in BW swards, except for the treatment where RC was immature and BW mature 
(Treatment C). In these two days the largest difference in rate of biting between swards 
also happened when BW was immature and RC mature (Treatment B) .  
Maturity effect contrasts (immature - mature) of rate of biting within each sward type 
and effects of the adjacent sward are shown in Table 3 . 22. There were highly significant 
effects of maturity on rate of biting for both species in the three days of grazing 
assessment, but there was no significant effect of the adjacent sward on rate of biting. 
Immature swards in both sward types had significantly higher rate of biting then mature 
swards, and the contrasts were consistently greater in BW than in RC swards. This 
difference appeared to increase from Day 1 to Day 2 and decrease from Day 2 to Day 3 
in both species. 
Table 3 .  2 1 .  The effect of swards of birdsfoot trefoil and white clover (BW), and red clover (RC) on rate of biting (bites/minute) in the first, 
second and third days of grazing assessment in Experiment 2 according to the treatment. 
Treatment A treatment B treatment C treatment D 
Rate of biting BW RC BW RC BW R C  BW RC SED1 P-value2 
(biteslmin) Immature Immature Immature Mature Mature Immature Mature Mature 
Day 1 48.9 46.0 5 1 .5 38.3 44. 0  4 1 .4 42.6 37.3 2.35 0.01 90 
Day 2 52.6 47 .6 55.4 40.6 4 1 .5 45.7 43.8 37.6 2.82 0.0022 
Da� 3 55.4 44.4 5 1 .0 38.8 42.2 47. 1  45.2 43.0 2.27 0.0002 
I SED - S tandard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
Experiments 1, 2, 3 and 4 88 
Table 3 .  22. Sward maturity effect [contrast: immature - mature (negative numbers 
represent greater effect of mature than immature sward)] on rate of biting 
contrasting the effect within each sward type [either birdsfoot trefoil and 
white clover (BW) or red clover (RC)] and the effect of adjacent sward in 
the first, second and third days of grazing assessment in Experiment 2 .  
Rate of Biting 
Day 1 Day 2 Day 3 
Effect on behaviour on BW 
Maturity effect of BW sward 6.87* 
Maturity effect of RC sward -0.63 
Effect on behaviour on RC 
Maturity effect of RC sward 5.89* 
Maturity effect of BW sward 2.82 
1.664 
1 1 .36* 
-2.54 
7.6 1 * 
2.46 
1 .737 
9.5 1 *  
0.7 1  
4.8 1 *  
-3. 4 1  
1. 717  
I SED - Standard error when comparing means with same 
levels of specie 
Contrast significance: • P$O.05 
Experiments 1, 2, 3 and 4 89 
3.3.3. Experiment 3: Effect of height of birdsfoot trefoil (Lotus corniculatus L.) and 
white clover (Trifolium repens L.) in relation to red clover 
(Trifolium pratense L.) on grazing behaviour, diet selection and 
herbage intake. 
3.3.3.1. Sward measurements 
Herbage mass, sward surface height and sward bulk density 
Interactions between treatments and sward type were significant, so results are presented 
as interaction effects of sward type and treatment in Table 3.23. 
Before grazing, there was a significant interaction between treatments and sward type 
for herbage mass. Taller swards had greater herbage mass. However in the treatment 
where both swards were short, BW had higher herbage mass than RC, and in the 
treatment where both swards were tall, RC sward had higher herbage mass than BW. 
These differences disappeared after grazing. This modification can be visualised as a 
significant interaction of DM removed. Taller plants had higher DM reduction. 
Short swards were on average 8 to 1 0  cm high, while the taller were between 1 4  to 1 6  
cm. There was a significant interaction between sward type and treatment effects for the 
sward surface height before and after the first day of grazing. After 5 5  hours of grazing 
there was no residual significance, either in the interaction or in the sward main effect, 
but there were significant differences between treatments (main effect). After grazing, 
the treatment where both swards were initially tall (Treatment D) was on average the 
shortest, and the treatment where RC was tall and BW was short (Treatment B) was the 
tallest. Treatments where BW was tall and RC short (Treatment C), and where both 
swards were short (Treatment A) was intermediate in height (5.9, 6.6, 5 . 6  and 5.0 cm for 
Treatments A, B,  C and D, respectively, SED 0.46, P=0.0407). 
Table 3. 23. Herbage mass (kg DMlha), sward height (cm) and bulk density (mg DMlcm3) before and after grazing, and estimation of the 
herbage mass removed (kg DMlha) of birdsfoot trefoil and white clover (BW) and red clover (RC) swards according to the 
treatment in EXEeriment 3 .  
Treatment A Treatment B Treatment C Treatment D 
BW RC BW RC BW RC BW RC SEDl P_value2 
Short Short Short Tall Tall Short Tall Tall 
Herbage mass (kg DMlha) 
Pre-grazing 2 1 1 0  2040 201 0  3060 2870 2 120 2530 2920 234 0.00008 
Post grazing 1 290 1 200 1 6 10  1 5 80 1 4 1 0  1 540 1 670 1 570 1 85 0.8062 
DM removed 820 850 4 10  1480 1 460 580 870 1 350 224 0.00001 
Sward height (cm) 
Pre-grazing 8 .7 9.7 8.6 1 5 .5 1 5 . 1  9.3 14.7 1 4.4 0.89 0.0000 
After 1 day grazing 7.3 7.7 7.5 10.6 9.7 7.3 7.8 9.0 0.76 0.0002 
Post-grazing 5 .6 6.2 6 . 1  7 . 1  5 .9 5 .4 5 .2 4 .8  0.43 0.0629 
Bulk density (mg DM/cm3) 
Pre-grazing 2.4 1 2.20 2.36 2.00 1 .89 2.27 1 .72 2.03 0 . 1768 0.0 1 20 
Post-�azin� 2.45 2. 1 9  2.9 1 2.46 2.68 3.37 3.24 3 .4 1  0.554 1 0.49 1 1 
1 SED - S tandard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
Experiments 1, 2, 3 and 4 9 1  
Bulk density also showed a significant interaction between sward types and treatment 
effects before grazing. Tall swards were on average less dense. After grazing this 
interaction disappeared. Bulk density of herbage remaining after grazing was greater 
than the herbage offered. This difference in bulk density was greater for taller swards. 
Sward composition 
Before grazing, BW had a significantly higher proportion of live matter than RC (9 1 .6 
vs 8 1 .5 %, SED 1 .63, P<O.OOOO). However there were no other significant differences 
between BW and RC swards, or between heights before grazing. The botanical 
composition differences were affected by the interaction between species and treatment 
(Table 3 . 24). There was a marginally significant interaction effect for the percentage of 
leaf and stem (Table 3 .24), short swards had a higher proportion of leaves. Within each 
treatment, short swards also had less stem than tall swards. 
After grazing, there was no significant interaction between treatments and sward type. 
There were significant main effects for the percentage of stems, grasses, other broad leaf 
species, and total live matter. BW had significantly more stems (43 .2 vs 26 %, SED 
8 .42, P=0.0498), grasses (6.4 vs 1 .7 %, SED 0.04, P<0.0400) and live matter (87 vs 73 
%, SED 3 .0 1 ,  P=0.OOOO7) than RC. On the other hand, RC swards had significantly 
more other broad leaf species ( 1 4.0 vs 26.8 %, SED 5.60, P=0.0289). 
Before grazing, the taller swards had significantly more birdsfoot trefoil than the short 
ones (birdsfoot trefoil/white clover (B/w) ratio: 1 . 1 0 vs 0.34 , SED 0.32 1 3 ,  P=0.0446). 
After grazing, there was only a marginally significant contrast between short and tall .  In 
both swards (tall and short) there was more white clover than birds foot trefoil (BIW 
ratio 0.7646 vs 0. 3290, SED 0.2 1 1 0, P=0.07 1 4, for tall and short swards, respectively) . 
Table 3 . 24. Botanical composition of birdsfoot trefoil plus white clover (BW) and red clover (RC) swards before and after grazing, 
according to the treatments (plant height: Short and Tall) (OM basis): (a) percentage of components in live fraction, (b) 
percentage of live matter in total OM of each sward and (c) ratio of the total live matter of birds foot trefoil and white clover 
(BIW) in the BW sward, EXEeriment 3 .  
Treatment A Treatment B Treatment C Treatment D 
BW RC BW RC BW RC BW RC SED
! P-value2 
Short Short Short Tall Tall Short Tall Tall 
Pre-grazing 
(a) Leaf 3 8 .4 42.0 45 .8 3 6.4 35.5 42.4 39.6 3 1 .7 5. 1 6  0.0794 
Petiole 1 8 .0 24. 1 22.3 26.3 1 9.5 22.3 1 8 . 2  2 1 .2 3 .77 0.9269 
Stem 32.6 1 4. 6  1 9.7 24.5 20.4 1 9.6 20.3 1 8 .6 5.88 0.0577 
Flower 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.08 0.3288 
Grass 1 .5 2.4 0.6 0.7 4.9 0.8 1 .8 0.7 2.76 0.5993 
Other species 8 . 1 1 7 .0 8.0 1 3 .0 1 8 .6 1 3 .9 7 . 9  27 .8 8 .72 0.2763 
(b) Total live matter 90.2 77.8 93.4 84. 1 90.0 79.7 92.9 84.3 3 .26 0.855 1 
(c) BIW ratio 0.40 0.29 1 . 1 9 1 .02 0.457 0.9 1 03 
Post-grazing 
(a) Leaf 28.6 25.2 26.0 24.2 24.9 26.7 2 5 . 3  2 2 . 3  5 .49 0.9078 
Petiole 1 8 . 1 1 8 .5 1 5 .9 22.2 1 9 .0 1 9.9 24.5 22.3 5 . 1 2  0.6974 
Stem 37.7 2 1 . 5 35.8 2 8 . 8  67.5 27 .9 3 1 .9 25.8 1 6.85 0.4745 
Flower 0.0 0.0 0.0 0.3 0.0 0.0 0.6 0.0 0.27 0. 1 8 8 1 
Grass 6.6 2 . 3  6 . 6  1 .3 5.4 2 .4 7 .0 0.6 4.35 0.9360 
Other species 1 3 .6 32.3 1 6. 3  24.0 9 . 8  2 1 .9 1 6. 1 29. 1 1 1 . 2 1  0.9 1 83 
(b) Total live matter 82. 1 75.4 8 8 .4 75.4 87.8 7 1 .7 8 9. 6  69.4 6.02 0.4546 
(c) BIW ratio 0.28 0.38 1 .06 0.46 0.300 0. 1 877 
I SED - Standard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
Experiments 1, 2, 3 and 4 93 
3.3.3.2. Canopy structure within the sward 
The canopy structure determined from point quadrat data is summarised in Figures 3 . 1 2  
and 3 . 1 3. The tall swards had greater variation in height. The leaf and petiole 
components were relatively more important in the medium and higher strata. In both 
swards the number of contacts of dead matter was greater in the bottom strata, and there 
was an increase of dead matter after grazing. In RC swards, the proportion of red clover 
leaves close to the ground increased from before to after grazing. In BW swards the 
proportion of birds foot trefoil in the bottom strata was larger in tall than in short swards. 
In contrast, the number of white clover contacts after grazing in the bottom strata 
increased more in short than in tall swards. 
3.3.3.3. Sward chemical composition 
Extractable condensed tannin concentration 
There were no significant differences between short and tall birds foot trefoil in relation 
to concentration of extractable condensed tannin (EeT) (Table 3 . 25).  Leaves had 
consistently higher concentration of EeT than stems. The concentrations of ECT in 
stems were negligible. 
Table 3. 25. Extractable condensed tannin (EeT) concentration (%) of birdsfoot trefoil 
leaf and stem in Experiment 3, according to the sward characteristic (tall or 
short) imposed by the treatments (DM basis). 
Components Short Tall SED} 
Leaf 0.66 0.4 1 0. 1 257 
Stem 0.03 0.02 0. 0 1 3 1  
P-valui 
0. 1 257 
0.3548 
I SED _ S tandard error for differences of means when comparing S hort with Tall 
swards. 
2 P-value of the treatment main effect. 
Number of observations contributing to each mean (n::::4). 
94 
BIRDSFOOT TREFOIL & WHITE CLOVER 
Before grazing After grazing 
42-4S 
Treatment BW SHORT: O BT leaf Treatment BW SHORT: RC SHORT 
36-39 36-39 RC SHORT BT stem 
_ BT flower 
30-33 .WC leaf 27-30 
24-27 OWC petiole 
1 8-21 WC stolon 18-21 OWC flower 
12-IS o Grass 9-1 2  o Other broad leaf 
6-9 o Dead material 
0-3 
0-3 
0 1 0  20 30 40 50 60 70 80 90 1 00 10 20 30 40 50 60 70 80 90 100 
42-45 42-45 
Treatment SW SHORT: RC TALL Treatment BW SHORT: RC SHORT 
36-39 36-39 
30-33 30-33 
24-27 24-27 
1 8-2 1 1 8-21 
1 2- 1 5  1 2- 1 5  
6-9 6-9 
E � 0-3 0-3 
..: 
f- 0 1 0  20 30 40 50 60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 ..: � f- 42-45 42-45 r;n 
0 Treatment SW TALL: RC SHORT Treatment SW TALL: RC SHORT � 36-39 36-39 ..: 
� 
r;n 30-33 30-33 
24-27 24-27 
1 8-21 1 8-2 1 
1 2- 15  1 2- 1 5  
6-9 6-9 
0-3 0-3 
0 1 0  20 30 40 50 60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 
42-45 42-45 
Treatment BW TALL: RC TALL Treatment SW TALL: RC TALL 
36-39 36-39 
30-33 30-33 
24-27 24-27 
1 8-2 1 1 8-2 1 
12- 15  1 2- 15  
6-9 6-9 
0-3 0-3 
0 1 0  20 30 40 50 60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 
NUMBER OF CONT ACTS FOR PLANT PARTS 
FigW"e 3 . 1 2 .The stratum structure of plant parts within the sward canopy of birdsfoot 
trefoil and white clover (BW), before and after grazing in the foW" 
treatments in Experiment 3 (RC = red clover sward). 
95 
RED CLOVER 
Before After 
Treatment SW SHORT: Treatment SW SHORT: .RC leaf 
36-39 RC SHORT 36-39 RC SHORT O RC petiole 
RC stern 
27-30 27-30 .WC leaf 
O WC petiole 
1 8-2 1 1 8-2 1 WC stolon 
o Grass 
9- 1 2  9- 1 2  o Other broad leaf 
o Dead material 
0-3 0-3 
0 1 0  20 30 40 50  60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 
Treatment SW SHORT : RC TALL Treatment SW SHORT: RC TALL 
36-39 36-39 
27-30 27-30 
1 8-2 1 1 8-2 1 
9-1 2  9- 1 2  
a 0-3 0-3 � 
-< 0 1 0  20 30 40 50 60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 r,... -< � Treatment SW TALL : RC SHORT r,... Treatment SW TALL : RC SHORT Cl) 36-39 36-39 
� 
-< 
� 27-30 27-30 Cl) 
1 8-2 1 1 8-2 1 
9-12 9-1 2  
0-3 0-3 
0 10  20  30 40 50 60 70 80 90 100 0 1 0  20 30 40 50 60 70 80 90 1 00 
Treatment SW TALL : RC TALL Treatment SW TALL : RC TALL 
36-39 36-39 
27-30 27-30 
1 8-21 1 8-21 
9-12 9-12 
0-3 0-3 
0 1 0  20 30 40 50 60 70 80 90 1 00 0 1 0  20 30 40 50 60 70 80 90 1 00 
NUMBER OF CONTACTS FOR PLANT PARTS 
Figure 3 . l 3 .The stratum structure of plant parts within the sward canopy of red clover 
(RC), before and after grazing in the four treatments in Experiment 3 (BW = 
birdsfoot trefoil and white clover sward) . 
Experiments 1, 2, 3 and 4 96 
F ormononetin concentration 
Table 3 .26 shows the concentration of forrnononetin in leaves, petioles and/or stems of 
birds foot trefoil, white clover and red clover. There was a significantly higher 
concentration of forrnononetin in red clover than in birdsfoot trefoil or white clover, and 
red clover leaves had the highest concentration. In the leaf component, white clover had 
a higher concentration than birds foot trefoil. In contrast to red clover, birdsfoot trefoil 
had a higher concentration in stems than in leaves. 
Table 3 .26. Forrnononetin concentration (%) of leaf, petiole and stem of birdsfoot trefoil 
(BT), white clover (WC) in birds foot trefoil and white clover sward (BW), 
and red clover (RC) in red clover sward (RC) in Experiment 3 .  
Leaf 
Petiole 
S tem 
BW sward RC sward 
BT WC RC 
0.07 0. 1 6  0.70 
6 0. 1 6  0.54 
0. 1 4  0.56 
I SED _ Standard error for differences of means 
2 P-value for the comparison of species. 
P-value2 
0.0203 0.0203 
0.0404 0.0004 
0.0405 0.0005 
0.0204 0.0003 
0.0850 0.0 1 54 
3 Standard error for differences of means and P-value for comparison between BT and Wc. 
4 Standard error for differences of means and P-value for comparison between BT and RC. 
5 Standard error for differences of means and P-value for comparison between WC and RC. 
6 - no sample 
Number of observations contributing to the mean of each specie (n:::::4). 
General chemical composition 
The general chemical composition of each species, when part of the short and tall swards, 
is given in Table 3 .27. Although some general comparisons are reported, statistical 
analysis was not possible. The values represent single samples bulked across treatments 
and replications. 
The three species had similar chemical composition. Within each species, leaves had the 
highest concentration of crude protein (CP), lipid, ash and digestibility (IVDMD), and the 
Experiments 1, 2, 3 and 4 97 
least of ADF and NDF. The distribution of sugar plus starch (CHO) varied according to 
the species. For birds foot trefoil, leaves had highest concentration of CHO, while for white 
clover, concentration was higher in stolons. Red clover had an even distribution between 
leaves, petioles and stems. There were no important differences between tall and short 
swards. 
Table 3 .  27.Crude protein (CP), lipid, acid and neutral detergent fibre (ADF, NDF), 
carbohydrates (soluble sugars plus starch) (CHO), ash and in vitro dry 
matter digestibility (IVDMD) determined by Near Infrared Reflectance 
Spectroscopy (NIRS) of the main components of birds foot trefoil and white 
clover, and red clover swards in Experiment 3 according to height treatment 
(Eercenta�e of DM basis). 
CP LIPID ADF NDF CHO ASH IVDMD 
White clover 
leaf 34.8 3.0 1 6.3 1 7.6 5.8 1 2. 1  81 .2 
petiole 1 8.2 1 .3 28. 1 26. 1  6.0 1 1 .6 70.7 
stolon 1 6.8 2.2 22.3 26.2 2 1 .2 8.3 75.9 
1:: Birdsfoot trefoil 
� leaf 34.5 3.3 1 6.3 1 7.4 8.0 1 2. 1  81 .2 
CI::I stem 1 5.0 1 .4 34.6 38.4 1 .6 9.8 64.8 
Red clover 
leaf 36.2 3.4 1 7.5 22.5 9.8 1 1 .8 80.2 
petiole 1 8.0 1 .7 29.3 26.6 8.6 1 1 .5 64.9 
stem 1 3.7  0.8 34.5 36.1 8.9 9 . 1  69.6 
White clover 
leaf 35.0 2.4 1 6.3 1 2.5 5.7 1 2.5 81 .2 
petiole 1 8.4 0.8 28.7 24.7 6.6 1 1 .4 70. 1 
stolon 1 9.7 2.4 2 1 .0 25.8 23.7 8.0 77.0 
_ Birdsfoot trefoil � leaf 31 . 1  4.0 1 5.5 1 8.4 1 0.4 1 1 . 1  82.0 
stem 1 4.9  1 .9 32.6 37.9 4.0 9.3 66.6 
Red clover 
leaf 35.6 3 . 1  1 9.2 2 1 .7 1 0.2 1 2.0 78.6 
petiole 1 8. 1  1 .2 30.4 26. 1 9.2 1 1 .3 68.6 
stem 1 6. 1  0.8 32. 7  30.3 9.6 9.3 66.6 
Experiments 1, 2, 3 and 4 9 8  
3.3.3.4. Animal measurements 
Total grazing time and intake 
There were significant interactions between sward type and treatment for grazing time 
and intake (Table 3 .28) during the three days of grazing measurement, though in the 
second day the significance for grazing time was marginal. The animals spent more time 
grazing taller swards on each of the three days, and also ate more herbage from tall 
swards. When both swards were offered at similar height, the animals spent similar time 
grazing each sward. The only exception was for Day 3 in the treatment with short BW 
and short R C ,  where animals spent more time grazing RC. 
Although no statistical analyses were performed to compare days, the total grazing time 
per animal in each treatment increased from Day 1 to Day 3, except in treatment D 
(where both swards were tall). From the second to the third day the increase in grazing 
time happened mainly on red clover, except in the treatment where RC was short and 
BW was tall (Treatment C). 
Animals ate more from tall swards when contrasts of tall and short swards were offered. 
However, considering the total intake/animal/day in each treatment, the lowest intake 
was found on plots where both species were talL Within each plot, more was consumed 
from RC swards than from BW swards, except when BW was tall. However when both 
swards were short the difference in intake from each species was smalL 
Table 3 .2 8 .  The effect of treatments (sward height) on grazing time (minutes) in Days 1 ,  2 and 3 ,  and average DM intake (kg) per animal 
per day during 55 hours of grazing in Experiment 3, according to the treatments. 
Grazing time (min) 
Day 1 
Day 2 
Day 3 
Intake 
Treatment A Treatment B Treatment C 
BW RC BW RC BW R C  
Short Short Short Tall Tall Short 
46. 1 49.0 26.6 65.9 56.9 3 3 . 7  
49. 1 49.8 47.0 67.9 68.4 44. 1 
49.4 6 1 .4 39.2 78.3 70.3 6 3 . 6  
(kgDM/hd/day) 2.34 2.42 0.94 3 . 67 3 .68 1 .40 
I SED - Standard error for differences of means when c omparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect 
Treatment D 
BW RC 
Tall Tall 
52.3 5 8 . 7  
5 3 . 6  49. 1 
54.4 53.9 
1 .3 1  2 . 04 
P-value2 
1 2 .08 0.0236 
1 0.23 0.0600 
1 0 .40 0.0397 
0.576 0.0003 
Experiments 1, 2, 3 and 4 1 00 
The effect of height contrasts (tall - short) within each sward type and of the alternative 
sward on grazing time and intake is presented in Table 3.29, As described in 
Experiment 2 results (section 3 3 . 2.2) the contrasts within each sward type were 
calculated subtracting the total grazing time (or intake/animal/day) the animals spent 
grazing on tall from short swards, considering that the alternative swards were the same 
height. The contrasts (where the alternative swards were either tall or short) were then 
averaged. On the other hand, the contrasts within the alternative sward type was 
calculated subtracting the averages of total grazing time (or intake/animal/day) the 
animals spent in a sward type that had as alternative sward either tall or short sward. In 
these cases negative numbers represented greater effect of short than tall sward, There 
was a significant BW height effect on time spent grazing BW in the first and last days of 
assessment. In these days the animals spent significantly more time grazing the tall than 
the short sward, In Day 3 the animals were also affected by the sward height of the 
alternative area. The animals spent significantly more time grazing BW when the RC 
alternative area was short. The grazing time in RC was only significantly affected by 
height on the first day. In this day the animals spent more time grazing taller than short 
swards. Behaviour on RC swards did not seem to be affected by the BW sward height. 
The average intake/animal/day on BW and RC appeared to be significantly affected by 
sward height. Taller swards had greater intake than shorter. However the intake either in 
BW or RC was also affected by the alternative sward. The animals had greater intake 
from the swards when the alternative sward was short. 
Experiments 1, 2, 3 and 4 1 0 1  
Table 3 .29. Sward height contrast effect on total grazing time and intake/animal/day 
contrasting the effect within each sward type (BW or RC) and the effect of 
alternative sward in Daxs 1 , 2 and 3 of grazing assessment in Experiment 3 .  
Grazing time Intake 
Day 1 Day 2 Day 3 
Effect on behaviour on BW 
Height effect of BW sward 1 8 .25*2 1 2.95 1 8.05 *  0.85* 
Height effect of RC sward 1 2.06 -8.44 - 1 3.06*t - 1 .88* 
Effect on behaviour on RC 
Height effect of RC sward 20.9 3 *  
Height effect of BW sward - 1 1 .26 
6.572 
1 1 .52 
1 2 .22 
6.491 
3 .60 
- 1 1 . 1 1 
5.901 
0.94* 
- 1 .32* 
0.349 
1 SED - Standard error when comparing means with same levels of specie. 
Contrast significance: * P:;;O.05 ; *t P=O.054 16  (marginal significance) 
2 Contrast: tall - short (negative numbers represent greater effect of short than 
tall sward) 
The treatment effect on the proportion of grazing time on BW swards is shown in Table 
3 . 30 .  There were significant effects of treatment on the proportion of grazing time on 
BW swards for Day 1 ,  but there were no significant treatment effects for Day 2 and 3 .  In 
Day 1 the animals spent proportionally less time grazing short BW swards that had tall 
RC as alternative sward. They also grazed proportionally more time in BW swards that 
had short RC sward as alternative sward. 
Experiments 1, 2, 3 and 4 1 02 
Table 3 . 30. Treatment effects on the proportion of grazing time devoted to birdsfoot 
trefoil and white clover swards (BW) in Experiment 3 Treatment A= BW 
and RC short; Treatment B=BW short and RC tall; Treatment C= BW tall 
and RC short; Treatment D= BW and RC tall) . 
Treatments 
A B 
Day 1 0.46 0.29 0.63 
Day 2 0.52 0.4 1 0.59 
Day 3 0.44 0.34 0.5 1 
IStandard error for differences of treatment means 
2 P-value for the comparison of treatment main effect. 
Grazing time per kg ofDM offered 
D 
SEDi P-value2 
0.46 0.070 1 0.0493 
0.54 0.0555 0.2303 
0.50 0.0545 0 . 1 667 
The total time the animals spent grazing during the 3 hours of observation in Day 1 and 
Day 3 was divided by the amount of herbage mass available before and after grazing, 
respectively, in order to verify the effect of herbage mass on the distribution of grazing 
activity. The number of minutes per kg of DM the animals spent grazing each treatment 
is shown in Table 3 .3 1 .  There were no significant differences between swards for Day 1 
and for Day 3 of grazing. There was only a significant difference between treatments 
(main effect) in Days 1 and 3 of grazing. In both cases the animals grazed significantly 
longer per kg of DM when both swards were tall (Treatment D). The treatments with tall 
RC (C and D) had greater increase from Day 1 to Day 3 in grazing time per kg of dry 
matter in the plot than treatments with short RC (A and B), though no statistical 
analyses were performed to compare days. 
Table 3 . 3 1 .  The effect of treatment on grazing time per kg of dry matter in the plot 
(minJkg DM) in the first and third day of Experiment 3 (Treatment A= BW 
and RC short; Treatment B=BW short and RC tall; Treatment C= BW tall 
and RC short; Treatment D= BW and RC tall). 
Treatments 
A B 
Day 1 1 . 1 8 1 .06 
Day 3 2.32 2.23 
I Standard error for differences of treatment means 
C 
1 .08 
2.77 
2 P-value for the comparison of treatment main effect. 
D 
2.07 
3 . 62 
0. 1 52 
0.329 
P-valui 
0.0002 
0.0082 
Experiments 1, 2, 3 and 4 1 03 
Rate orbiting 
There was a significant interaction between sward type and treatment for rate of biting 
in Day 1 and a marginal interaction in Day 3 (Table 3 .32). In Day 2, there was a 
significant species main effect where the rate of biting was significantly higher on BW 
than on RC (55 . 8  vs 50.0 ± 2.014 bites/min, P= 0.0 1 1 0). As in Day 2, the rate of biting 
was greater on BW than on RC for most of the treatments in Days 1 and 3 .  However in 
the treatment where BW was tall and RC was short there was higher rate of biting on 
RC than on BW. Comparing Days 1 and 3, there was a decrease in number of bites per 
minute in all treatments and both species. 
Table 3 .32. The effect of swards of birdsfoot trefoil and white clover (BW), and red 
clover (RC) on rate of biting (bites/minute) in the first, second and third 
days of grazinS assessment in Experiment 3 accordins to the treatment. 
Treatments 
A B C D SED1 P-valui 
BW RC BW RC BW RC BW RC 
short short short tall tall short tall tall 
Day 1 58.2 54. 1 59.6 47 .7 56.6 5 8.2 54.2 48.0 1 .882 0.00 1 6  
Day 2 57.5 47.7 55.6 48.5 57.5 53.2 52.6 50.7 4.028 0 . 5892 
Dal 3 56.5 52.0 5 1 .9 44.2 5 1 .5 53.6 46.5 44.5 2.39 1 0.0800 
I SED- Standard error for differences of means when comparing swards within each treatment. 
2 P-value of the interaction between treatment and sward type. 
Height effects (contrasts: tall - short) on rate of biting within each sward type and of the 
adjacent sward are shown in Table 3.33 .  The rate of biting in BW swards was only 
marginally significantly affected by height in the third day of grazing. In this day, short 
BW swards had greater rate of biting than tall BW swards. In RC swards height had 
significant effects on rate of biting in the first and third days. In both days, short RC 
swards had significantly higher rate of biting. The rate of biting of both sward types was 
not affected by the adjacent sward treatments. 
Experiments 1, 2, 3 and 4 1 04 
Table 3 . 3 3 .  Sward height effect [contrast: tall - short (negative numbers represent 
greater effect of short than tall sward)] on rate of biting contrasting the effect 
within each sward type [either birdsfoot trefoil and white clover (BW) or red 
clover (RC)] and the effect of adjacent sward in the first, second and third 
days of grazing assessment in Experiment 3. 
Rate of Biting 
Day 1 Day 2 Day 3 
Effect on behaviour on BW 
Height effect of BW sward -3 .492 - 1 .52 -5. 1 9*t 
Height effect of RC sward -0.52 -3 . 3 6  -4.8 3  
Effect on behaviour on RC 
Height effect of RC sward -8.32* 
Height effect of BW sward 2.22 
-0. 89 
3.83 
- 8.44 *  
0.90 
SED} 1.905 2.430 2.372 
1 SED - Standard error when comparing means with same 
levels of specie 
Contrast significance: * PS;O.05; 
*t P::::O.05643 (marginal significance) 
2 Contrast: tall - short (negative numbers represent greater 
effect of short than tall sward) 
Experiments 1, 2, 3 and 4 1 05 
3.3.4. Experiment 4: Effect of position in the plot of birdsfoot trefoil (Lotus 
corniculatus L.) and white clover (Trifolium repens L.) sward 
in relation to red clover (Trifolium pratense L.) sward on 
grazing activity distribution. 
3.3.3.1. Sward measurements 
Herbage mass, sudace sward height and sward bulk density 
The interactions between treatments and sward type were not significant, so results are 
presented as main effects only. The results of measurements of herbage mass, surface 
sward height and sward bulk density are given in Table 3 . 34. The interaction between 
treatments and sward type is presented in Appendix 3.5. 
Table 3.34 .  Herbage mass (kg DMJha), and bulk density (mg DMlcm3) before grazing 
and sward surface height (cm) before and after grazing birdsfoot trefoil and 
white clover (BW) and red clover (RC) swards in Experiment 4. 
Herbage mass (kg DMJha) 
pre-grazing 
Sward height (cm) 
pre-grazing 
post-grazing 
Bulk density(mg DMlcm3) 
BW RC SED] P-value2 
2050 
7.0 
4.3 
2200 
7.5 
4.5 
1 53 
0.37 
0.27 
0. 3237 
0 . 1 68 3  
0.5560 
pre-grazing 2 .93 2.98 0.2 1 62 0.80 1 9  
I Standard error of the difference comparing means of different sward main effect 
2 P-value for sward main effect differences 
There was no significant difference between BW and RC swards, which had similar 
herbage mass, height and density. They were relatively shorter and denser than the 
previous experiment (Experiment 3 ) .  
Experiments 1, 2, 3 and 4 1 06 
3.3.3.1. Animal measurements 
Grazing time 
There were no significant differences between treatments, comparing the proportion of 
grazing time the animals spent grazing the smaller sward in area of each plot. There was 
no influence of the minor sward location (close to the fence or in the middle of the plot) 
in the distribution of grazing activities. The proportion of grazing time the animal spent 
in the minor swards according to the position in the plot is given in Table 3.35.  The 
average grazing time (in minutes) the animals spent grazing each sward in each plot is 
shown in Appendix 3 . 5 .  
Table 3 . 3 5 .  Proportion of time grazing (in relation to the total grazing time spent in plot) 
birdsfoot trefoil plus white clover (BW) and red clover (RC) swards 
according to the strip position (side: close to the fence; central: in the middle 
of the plot) in the plot in Experiment 4. 
Proportion o/grazing time 
BW BW RC RC 
side central side central 
P­
value2 
first day 0.32 0.29 0.34 0.33 0.0246 0.5277 
1 Standard error of the difference comparing means in the same species in different treatments 
2 P-value for treatment differences. 
Experiments 1, 2, 3 and 4 1 07 
3.4. DISCUSSION 
3.4.1. Evaluation of experimental procedures 
The use of a pair of swards sown in strips provided a flexible and reliable condition to 
test animal preference. The use of two separate swards, rather than intermixed species 
avoided the inter-species competition and its effect on species-balance. The use of 
separate swards also made it easier to determine the animal selection, and for animals to 
find their preferred species (Parsons et al., 1994a), without needing to search the 
location (Newman et al., 1 995). Swards formed by strips also provided a flexible 
approach for modification of each strip according to the treatment required in each 
experiment. 
Before starting the experiment, and between each replicate run, the animals were 
continuously grazed on an adjacent area with the same species and design (alternate strips 
swards) of the experimental area to establish and maintain experience of experimental 
conditions. In this period all animals grazed together to avoid group effects. Previous 
studies with different animal species have shown that previous experience has a marked 
effect on grazing preference (Arnold and Maller, 1977; Provenza, 1996a). The previous 
experience in these experiments was clearly effective since the previous treatment effect 
and group effect were not significant in any of the three experiments. 
Because of the relatively small size of the plots there was a concern about possible bias 
caused by the electric fences on animal preferential behaviour. In order to clarify this 
effect, Experiment 4 was carried out. This experiment showed that the proximity of a 
sward to the fence did not influence the proportion of grazing time spent on that sward. 
The comparisons in this experiment were very clear since both swards had similar physical 
characteristics (height, herbage mass and bulk density) (Tables 3.34, 3.35) 
Experiments 1, 2, 3 and 4 1 08 
Five replications provided an accurate statistical assessment of the animal behaviour. 
Five replications, rather than four, were important to efficiently investigate the carry 
over effect, and showed clearly that selective behaviour was not affected by previous 
treatment. 
There were several advantages for having white clover intermixed with birdsfoot trefoiL 
Although the use of pure birdsfoot trefoil swards would provide a clearer evaluation of 
the animal preferences between birdsfoot trefoil and red clover, the mixture with white 
clover was more realistic in practical farm terms. White clover also provided a better 
soil cover, not allowing weeds to invade the BW sward. In addition, white clover did not 
influence the secondary compound concentration effect, showing negligible 
concentration of ECT and low concentration of formononetin. Jones and Lyttleton 
( 1 97 1 ), Jones et al. ( 1 973) and Li et al. ( 1 996) reported that white clover is either free 
of, or has very low concentrations of condensed tannin. However there is evidence that 
there is tannin in flowers (Jones et al., 1 976; Stockdale and Dellow, 1 995). Small 
concentrations of formononetin were also found in other studies in white clover plants 
(Francis et al. ,  1 967) .  
The decline in herbage available to the animals from the first to the second day and from 
the second to the third day was illustrated by the decline in surface height. The 
assessment of herbage mass on the second day by cutting quadrats was not possible 
because of the disturbance to animal behaviour. Although the use of a rising plate meter 
and sward surface height for predicting mass was tested before starting the experiment, 
they did not provide a reliable basis for prediction. 
Point quadrat data provided a good illustration of the vertical distribution of sward 
components in the canopy, and provided estimates of botanical composition similar to 
hand separation. However there were variations between the two techniques.  In 
Experiment 2, for example, the high proportion of stems in BW swards before grazing 
(Table 3 . 1 2) was not demonstrated in the point quadrat illustration (Figure 3 . 1 0) .  This 
effect may be related to the small stem surface area in relation to leaves to be touched by 
Experiments 1, 2, 3 and 4 109 
the point quadrat needle, relative to dry matter content in the hand separated samples. In 
addition, the percentage of dead matter was often underestimated by point quadrat. This 
agrees with results reported by Hodgson ( 198 1 a) and Grant ( 1985), who explain that the 
concentration of dead matter in the bottom strata is too dense to be recorded accurately. 
Although there were variations between techniques, the two techniques tended to 
complement one another. 
Because concentrations of ECT in samples of white clover and red clover was 
negligible, the analysis of BCT was done mainly on samples of birdsfoot trefoil plants. 
Other studies also reported low concentration of BCT in white clover (Jones and 
Lyttleton, 1 97 1 ;  Jones et al. ,  1 973; Li et al. ,  1996) and red clover (Jackson et al. ,  1 996). 
BCT corresponds to the fraction of the total condensed tannin in the forage that is not 
bound to protein or fibre (Terrill et al. ,  1992). Douglas et al. ( 1 993), Jackson et al. 
( 1 996) and Wang et al. ( 1 995) found that the extractable fraction varied from 62.8 to 
70.9% of the total condensed tannin in birdsfoot trefoil plants. The analysis of BCT 
concentration followed the DMACA-HCL protocol described by Li et al. ( 1 996), Li et 
al. ( 1 996) compared the DMACA-HCL with vanillin-acid procedure, and they found 
that DMACA-HCL was more sensitive in condensed tannin detection, particularly for 
material with low concentration of tannin. The DMACA-HCL procedure is also a more 
specific methodology to determine the condensed tannin concentration than butanol­
HCL (W. C. McNabb, pers. comm.) .  
The BCT concentrations observed in these experiments for cultivar Goldie (Tables 3 .4, 
3 . 1 3, 3 .25) were lower than values reported elsewhere (Douglas et al. ,  1995, and 
Douglas et al. ,  1 993), though the concentration of BCT in leaves was within the limits 
found by previous studies for birds foot trefoil in New Zealand (Jackson et al. ,  1 996; 
John and Lancashine, 1 98 1 ;  Li et al., 1 996; Lowther et al. ,  1987; Terrill et al., 1 992; 
Wang et al. ,  1 995; Waghom et al. 1 987ab) (see comments in section 4.4.3 - Chapter 4). 
Formononetin was determined by a modification (Anwar, 1994) of the fluorimetric 
assay described by Gosden and Jones ( 1 978). According to Anwar ( 1 994), the 
Experiments 1, 2, 3 and 4 110 
modification was necessary to ensure complete hydrolysis of formononetin glycoside, 
especially in plant components such as stems which contain relatively small amounts of 
B-glucosidase. In Experiments 1, 2 and 3 the formononetin concentration of red clover 
was in agreement with published results of several cultivars of red clover (Francis et al. ,  
1966; Kelly e t  al. ,  1979; McMurray e t  al. ,  1986 Anwar 1994) 
Although variation of preference between morning and afternoon grazing periods has been 
found in grazing (Parsons et al. ,  1994a) and indoor feeding (Fisher et al. ,  1997) studies, the 
grazing behaviour observation in this study was restricted to late evening (4.00 to 7.00 
p.m. in winter and 5 .00 to 8 .00 p.m. in summer). Thus, grazing preference measured in 
these experiments may not entirely reflect the preferences of animals during a full day. 
However grazing behaviour in the evening reflects the preference in one of the main 
grazing periods of the day (Amold, 1981 and Hodgson, 1990) and consequently for a 
substantial portion of the total grazing activity. Morning observations were carried out 
between Day 1 and Day 2 in Experiments 1, 2 and 3 to clarify the difference between 
morning and afternoon animal diet selection (Appendix 3 .6). The results showed that 
preference did not change according to time of the day. The proportion of time spent 
grazing one of the two sward types in the morning of Day 2 was on average intermediate 
between the evening observations of Day 1 and Day 2 in each experiment. 
Based on preVIOUS studies (Cosgrove et al. ,  1996) ten minute intervals between 
observations were used to provide an estimate of grazing activity. Hull et al. (1960) did 
not find a significant difference between continuos recording and up to 30 minute 
intervals. Continuous recording is important mainly for minor behaviour patterns like 
drinking water, but not for a major behavioural activity such as grazing (Gary et al., 1970). 
Although continuous assessment is more precise, it is very difficult to carry out without 
automatic equipment CHodgson, 1982b). 
The rate of biting in BW and RC swards agrees with results summarised by Hodgson 
(1985): a range of 21 - 66 bites/minute for cattle. The 20 bites method described by Forbes 
and Hodgson (1985) usually over-estimates the long-term mean rate of biting because it 
Experiments 1, 2, 3 and 4 1 1 1  
does not consider the short-term interruptions to grazing activity. However the estimates 
of maximum rate of biting provided a standardised basis for comparison between swards. 
3.4.2. Definition: Selection and Preference 
The separate definition of diet selection and preference is important because it reflects the 
research context and clarifies the information obtained. This study follows the definition 
of Hodgson ( 1 979) where "preference is a general term that describes the discrimination 
exerted by animals between areas of sward or the components of a sward canopy" and 
selection is the "preference modified by opportunity for selection". Therefore preference is 
demonstrated by the animals when free choice and equal opportunity is offered. In the four 
experiments described in this section, conditions most closely approached "free choice" in 
Day 1 when the herbage mass and height was relatively high. However, in most situations 
in these studies the animals demonstrated selection. In contrast, in Experiments 5 and 6 
(see Chapter 4), where access to a range of spaced plants offered similar opportunity, 
preferential behaviour was mainly tested. 
3.4.3. Components of ingestive behaviour and selection 
Grazing time, rate of biting and intake per bite partly describe the ingestive behaviour and 
the multiplication of these three factors results in the estimation of herbage intake (Allden, 
1 962; Allden and Whittaker, 1 970). Each factor of this equation is a component of 
selective and intake behaviour, and all could be used as indices of preference. However 
there is a reciprocal relation between intake per bite and rate of biting (Hodgson et al., 
1 994). Animals usually have greater rate of biting for small bites than for large bites (Laca 
et al., 1 992). Thus, different rates of biting and intakes per bite can generate similar total 
herbage intake. Because of this compensation, it is difficult to determine selection using 
rate of biting without considering intake per bite. Therefore, the estimation of grazing time 
is one of the best parameters for determining animal selection when estimation of herbage 
intake is not possible 
Experiments 1, 2, 3 and 4 1 1 2 
In these studies because the estimation of herbage intake was only possible by the 
difference in herbage mass before and after grazing, not considering the variation between 
days, grazing time, rather than herbage intake, was used to estimate diet selection. The 
assessment of grazing time was a good reflection of the herbage intake estimated per 
sward in E l  and E3 (Tables 3.7, 3.28). Both intake and grazing time (for Days 2 and 3) in 
E l  were highly affected by the amount of area offered of each sward (Figures 3.7,3 .9). ill 
E3 the greater intake of tall rather than short swards reflected the fact that the animals 
spent more time grazing tall than short swards (Table 3.29). However, in E2, although the 
animals spent more time grazing RC, on average over the three days of grazing, the 
animals had similar intakes between the two species (Table 3. 1 6). ill this case the 
difference between grazing time and the estimation of intake was probably related to the 
fact that the estimation of intake was a mean value for three days and did not take into 
account the important variation between days. Therefore, it can be concluded that 
estimation of grazing time provided a good estimation of the animals diet selection during 
the three days of grazing. Several other studies (e.g. Penning et al. ,  1 99 1 ;  Parsons et al. ,  
1 994a; Cosgrove et al. ,  1 996) have also used grazing time as an important assessment of 
diet selection and animal preference. 
Rate of biting reflected morphological characteristics of the sward. Although in all three 
experiments rate of biting on BW was on average higher than on RC, the rate of biting 
varied according to the sward conditions and from Day 1 to Day 3. While in E l  the rate of 
biting increased from Day 1 to Day 3 (Table 3 . 1 0) ,  in E3 it decreased (Table 3.32). This 
difference was apparently related to the difference in sward height and physical 
characteristics. Because of the lower herbage mass in E3, the animals had much lower 
residual after two days of grazing in E3 than in E l .  In E l  the increase in rate of biting with 
progressive defoliation probably was one of the compensating responses of the animal to 
limitations in intake per bite (Hodgson, 1 98 1  b). The animals apparently were getting 
smaller bites. Small bites with a high proportion of leaves required less manipulative 
movements (Hodgson, 1 985), so the animals could swallow faster and graze 
uninterruptedly (Laca et al. ,  1 993). On the other hand, in E3 on Day 3, the animals had a 
slower rate of biting than on Day 1 ,  as a result of the fact that after two days of grazing, the 
Experiments 1, 2, 3 and 4 1 1 3 
sward had been trodden on the ground and the animals had greater difficulty in 
prehending. Allden and Whittaker ( 1970); Kenney and Black, ( 1 984a,b) and Colebrook et 
al., ( 1 987) explain that rate of biting is affected by ease of prehension. The animals had to 
spend time looking and trying to find where and how to graze. The animals probably also 
reduced their intake rate to avoid grazing close to their dung and urine (Gamer, 1 963; 
Arnold, 1 9 8 1 ) .  In E2 there was no apparent change in rate of biting across the three days. 
The increase in sward density, promoting greater bite size and slower rate of biting was 
balanced by the reduction in proportion of leaves, generating reduction of bite size and 
increasing the rate of biting (Penning 1 986; Laca et al., 1 993). 
The rate of biting apparently did not influence the diet selection between different swards, 
because, as shown in E2 and E3 (Tables 3.22, 3.33), the rate of biting on one sward type 
was not affected by the alternative sward condition. Rate of biting was more a 
consequence of sward physical characteristics than a cause of selection. The animals 
probably regulated the rate of biting according to the specific sward conditions in each bite 
(Penning, 1 986; Laca et al., 1 993), 
The greater rate of biting observed in BW in all three experiments, mainly in Day 1 (Table 
3 . 1 0, 3 .2 1 ,  3 .32), was probably the consequence of combined effects of a lower sward 
surface height and higher bulk density (Tables 3 .2, 3 . 1 1 , 3 .23). Several studies (Chacon 
and Stobbs, 1 976; Hodgson and Jamieson, 1 98 1 ;  Milne et. al., 1 982; Philips and Leaver, 
1 98 5 ;  Penning et al. 1 99 1 ;  Burlison et al. ,  1 99 1 ;  Laca et al., 1 992; Mitchell et al., 1 993) 
have observed a negative relationship between rate of biting and either sward height, 
herbage mass or bulk density. The animals apparently spent more time selecting leaf in RC 
swards than they did in BW swards. Although in E3 there were no important differences in 
height and bulk density between BW and RC within the same height treatment, the BW 
swards had a greater proportion of their mass closer to the ground than RC (as 
demonstrated by the point quadrat analysis), This greater density was related to the 
presence of white clover close to the ground in BW swards (Figures 3. 1 2, 3 . 1 3) .  
Experiments 1, 2, 3 and 4 1 14 
The discussion of animal selective response to variation in sward characteristics will be 
concentrated on the most important behaviour response observed. The discussion will 
focus first on an overview of the pattern of selection across experiments and then will 
provide discussions based on effects of different experiments, of sward changes from Day 
1 to Day 3 and of high herbage mass and height (Day 1) .  The discussion will then finish 
with more detailed comparisons within each sward. In this way the main effects of each 
swards characteristics on diet selection will be covered. 
3.4.4. Pattern of selection across experiments 
Indications of the seasonal variation in selective behaviour was obtained by averaging, 
within each experiment, the proportion of grazing time spent on BW and the physical 
swards characteristics during the three days of grazing observation (Table 3 . 36). The 
standard error of the mean (SEM) estimated for the proportion of grazing time in each 
experiment was calculated by averaging the variances obtained in the three days of 
observation within each experiment. This was a useful analysis for verifying patterns of 
selection across experiments, though interpretation was limited since there was no 
replication of experiments. 
Experiments 1, 2, 3 and 4 1 1 5 
Table 3 .36. Overall averages of the proportion of grazing time on BW swards (proPGT), 
physical sward characteristics, extractable condensed tannin concentration 
(EeT - leaves of birdsfoot trefoil) and formononetin concentration (leaves 
of red clover) in Experiments 1 ,  2 and 3 (E l ,  E2 and E3) carried out in 
November (Nov), Febru� (Feb) and AEril-Mai: (AEr-Mai:) 1 9951 1 99 6 .  
E l  E2 E3 
Nov Feb AEr-Ma� 
ProPGT on BW 0.48 0.39 0.47 
SEM I 0.02 1 0.027 0.030 
S ward height (cm) BW 1 3 . 0  1 3 .5 8 . 5  
RC 2 1 .0 1 8 .9 8 . 9  
Herbage mass BW 3 1 60 3 870 1 93 5  
(KgDM/ha) RC 3955 4030 2000 
Bulk density BW 2 . 6 1 3 .08 2 .45 
(mgDMJcm3) RC 2.09 2.9 1 2 .49 
Leafiness (%) BW 3 5 . 9  1 9. 8  3 3 .0 
RC 29.2 1 8 .8 3 1 .4 
Proportion of stems BW 2 1 .6 52.9 3 3 . 2  
(%) RC 4 1 .2 46.4 22.6 
ECT (%) BW 1 .69 0.58 0.54 
Formononetin (%) RC 0.6 1 0.50 0.70 
ISEM - Standard error of the means 
Comparing the three experiments (Table 3.36) ,  E2 showed, based on SEM, a significantly 
different pattern from the other two in terms of the proportion of grazing time spent on 
BW. This result shows that there was a pattern of seasonal variation in selective behaviour 
between BW and RC. Although E3 had the least sward surface height and herbage mass, 
there was no indication that the animals selective behaviour was, on average, significantly 
different from E l .  The difference in proportion of grazing time between E2 and the other 
two experiments was apparently related to differences in plant morphological 
Experiments 1, 2, 3 and 4 1 16 
characteristics. For both swards the highest proportion of stems, and the lowest proportion 
of leaves occurred in E2. The proportion of stem was particularly high in BW swards. This 
pattern of seasonal variation was apparently not related to the secondary compound 
concentrations. The greatest contrast found between Experiment 1 and the two later 
experiments in ECT concentration did not reflect the differences in proportion of grazing 
time, and the small contrast in formononetin concentration between experiments suggests 
limited opportunity for seasonal effect of formononetin on animal preference. There is, 
then, an indication that the animals grazed proportionally more RC to avoid the high 
concentration of stems in BW. Thus, it can be concluded that this pattern of grazing 
behaviour observed is mainly related to differences in plant morphological characteristics 
promoted by differences in plant maturity. Nevertheless, the possibility of this seasonal 
variation also being affected by the concentration of secondary compound needs to be 
examined (see Experiments 5 and 6 - Chapter 4). It can also be concluded that although 
the animals had different behaviour between days within each experiment, the overall 
selective behaviour was not affected by the difference in sward surface height and herbage 
mass. 
As shown in Table.3.36, the ECT concentration of birdsfoot trefoil leaves decreased from 
Experiment 1 (in spring) to Experiment 3 (in autumn) .  This seasonal pattern was also 
observed in other studies (Cope et al. ,  1 97 1 ;  Windham et al. ,  1988; Roberts et al., 1 993). 
However the decrease of condensed tannin concentration between late spring and late 
summer disagrees with reports for several plant species (Clark et al. ,  1939; Stitt and 
Clarke, 1 94 1 ;  Donnely, 1959; Cope et al., 1 97 1 ;  Windham et al. ,  1988 and Iason et al., 
1 995). These authors found that the concentration of condensed tannin increased during 
summer months. However the concentration of ECT can also vary for several other 
reasons such as site to site variation (Douglas et al. ,  1993), soil fertility (Barry, 1 989) and 
differences in plant morphology (Douglas et al. ,  1993; Iason et al., 1995). A better 
understanding of this variation can be achieved by comparing the EeT concentration of 
these three experiments with Experiments 5 and 6 (see Chapter 4). 
Experiments 1, 2, 3 and 4 1 17 
The variation in fOlIDononetin concentration in leaves of red clover (Table 3 .36) between 
experiments was relatively small compared with studies of Anwar ( 1 994), who found 
fOlIDononetin concentrations at various stages of vegetative leaf development for cultivar 
Pawera varied between 0.75 and 2. 1 6%. This small variation is probably related to the 
genotype. Anwar ( 1994) also found smaller variation of fOlIDononetin concentration in 
cultivar G-27 than in the cultivar Pawera. According to Nicollier and Thompson ( 1 982) a 
significant portion of the isoflavone variation in clover is related to the genetic variance. 
The use of two different genotypes in Experiments 5 and 6, helped to clarify the variation 
in fOlIDononetin in red clover between different periods of the year (see Chapter 4). 
3.4.5. Comparison of diet selection in Experiments 1, 2 and 3 
Comparing experiments, herbage mass had an important effect on selection. The animals 
showed selection for swards with greater herbage mass, except in Day 1 of El  and E2. In 
E l  Day 1 ,  when the herbage mass was on average greater than 3900 kglha, the animals 
spent proportionally more time grazing in the minority sward type than was justified by the 
proportion of area offered, independent of the herbage mass (Tables 3 .8,  3 .9; Figures 3 .7, 
3 .8) .  In a similar way, in E2 Day I the animals did not graze in proportion to herbage 
mass, but grazed for longer per kg of DM available in RC than in BW, except when BW 
was immature and RC was mature (Table 3 .20). This lower effect of herbage mass in Day 
1 in E l  and E2 contrasts with the greater effect of herbage mass in E3, where the animals 
grazed in proportion to herbage mass offered (section 3 .3 .3 . 1  - Grazing time per Kg of 
DM offered). This difference can be explained by the fact that swards in Experiment 3 had 
the lowest herbage mass and the lowest height. Although the animals '  intake should not 
have been constrained by the herbage mass available (Holmes, 1987; Hodgson, 1 990; 
Gibb et al. ,  1 997) selection behaviour did seem to be influenced by herbage mass levels 
found in E3. The increasing effect of declining levels of herbage mass on diet selection 
could also be observed over time in E l ,  when in Day 3 the animals grazed in proportion 
to herbage mass available (Table 3.9, Figure 3.8) .  However, as indicated above, in E2 the 
animals were also strongly affected by sward morphological composition. This 
Experiments 1, 2, 3 and 4 1 18 
observation suggests that above some limits (approaching 4000 kg DMlha in these 
studies), the influence of herbage mass on cattle diet selection is reduced. 
The greater grazing time on the sward with greater herbage mass (represented by sward 
height and bulk density) in Day 3 of El  (Tables 3.9, Figure 3.8), and Day 1 and 3 of E3 
(section 3.3 .3. 1 - Grazing time per Kg DM offered) also demonstrated animal preference 
for taller swards with greater bulk density. Gong ( 1993) also found greater effect of bulk 
density on ingestive behaviour when sheep grazed clover than when they grazed grass. 
According to Burlison et al. ( 1991 )  and Laca et al. ( 1 992) height and density effects on 
intake are independent and additive. 
As in other studies (Mitchell et al., 1 99 1 ,  lllius et al., 1 992; Laca et al., 1 992; Demment et 
al., 1 993; Gong et al., 1996; Gibb et al., 1997; Torres-Rodriguez, 1 997), sward surface 
height was one of the most important pasture characteristics that affected ingestive 
behaviour. Its importance was demonstrated mainly in Day 1 (when the contrast between 
the two swards was high) of E2 (except in the treatment: BW immature and RC mature) 
and E3. In these experiments the animals grazed proportionally more time in the tallest 
sward (Table 3. 1 1 , 3 . 1 9, 3.23, 3 .30). However in El  Day 1 the animals were strongly 
influenced by area, grazing proportionally more in the minor area (Table 3.8,  Figure 3.7). 
Two possible hypothesis can be suggested to explain the fact that the height effect was 
stronger in E2 and E3 than in El .  Firstly, the height effect in El  was reduced because it 
was above some limits of herbage mass (as discussed earlier), but sward maturity 
influenced the preference for red clover in E2. Secondly, selective behaviour that drives 
the animal to seek a mixed diet (see section 3 .4.7) was stronger than the preference for 
taller swards in El .  These two hypothesis were clarified with E5 and E6 where cattle 
preference was tested using spaced plants of birdsfoot trefoil and red clover of different 
heights but similar maturity. 
Experiments 1, 2, 3 and 4 1 19 
Effect of alternative sward 
The greater the physical constraints on a sward, the more the animals were affected by the 
alternative sward. In E2 and E3 the total grazing time on one sward was affected by the 
alternative sward only on Day 3 (Tables 3. 1 8, 3 .29). On this day the herbage mass and 
height was relatively low and prehension was more difficult. This result suggests that the 
effect of the alternative sward and, by inference, of alternative patches within a sward on 
diet selection is probably more important when there are physical limitations to ingestion 
such as occurs when height is low. This result also reinforces the conclusion drawn above 
that the animals were strongly influenced by the herbage mass available. However, it does 
not agree with Griffiths et al. ( 1 997), who did not find evidence that use of individual 
sward patches was influenced by conditions on adjacent patches. This fact may be 
explained by the fact that Griffith et al. ( 1 997) did not test the situation where all swards 
had low sward surface height and herbage mass. 
3.4.6. Diet selection changes over time (from Day 1 to Day 3) 
The animals were able to adjust their diet selection as the sward characteristics changed. In 
El ,  the animals progressively adjusted their selection so as to graze in proportion to area 
and herbage mass offered, having initially selected strongly for the minority sward (Tables 
3.8,  3.9, Figures 3.7, 3 .8). This result is confirmed by the results in E2 and E3. In this 
adjustment the selection for the sward with greater proportion of leaf was overridden by a 
stronger selection for greater sward height and bulk density. Although several studies have 
shown that animals prefer leaves and reject stems (O'Reagain and Mentis, 1 989; 
0' Reagain, 1 993), in Day 3 of E2 and E3 the animals selected taller swards with higher 
herbage mass and greater percentage of stems. As discussed earlier (section 3 .4.5), in Day 
3 of E2 and E3 the animals were affected by the alternative sward. In E2 Day 3 they 
grazed substantially less time in BW when they had mature RC as the alternative sward 
(Tables 3 . 1 8, 3. 19). The attraction for mature RC is probably related to the fact that this 
was one of the swards with greater herbage mass (more height and bulk density), and more 
stems left after two days of intense grazing (Tables 3 . 1 1 , 3 . 1 2). Similar behaviour 
Experiments 1, 2, 3 and 4 1 20 
happened in E3 Day 3, when the grazing time on BW was affected by both the sward 
height in BW and by the height of the alternative sward (Table 3 .29). Animals spent more 
time grazing tall than short BW, but this grazing time increased when the alternative sward 
was short and decreased when the alternative sward was tall. In these cases the animals 
selected swards with higher herbage mass (taller and higher bulk density) (Table 3 .23) and 
greater proportion of sterns (Table 3.24), once most of the leaves were lost from the 
canopy. The selection of swards with higher proportion of stems in Day 3 contrasts with 
the behaviour observed when comparing the pattern of selection across experiments (see 
section 3 .4.4) and in Day 1 of the three experiments (see section 3 .4.7) where the animals 
showed selection for swards with higher leaf/stem ratio. 
3.4.7. Diet selection in Day 1 (high herbage mass and height) 
Results from Day 1 in each of the three experiments, when the herbage mass and height 
was high, suggests that selection was determined more by physical sward characteristics 
than by the preference of a specific species. In E l  the animals allocated grazing activity 
preferentially to the minor sward component on each treatment (Tables 3 . 8 ,  3 .9, Figure 
3 .7). In E2 the animals allocated a greater proportion of time grazing RC probably because 
of its greater height and proportion of leaves (Tables 3. 1 1 , 3 . 1 2, 3 . 1 6).  In E3 the animals 
grazed according to the sward height and in proportion of herbage mass available (Tables 
3 . 23, 3 .28, section 3 . 3 .3 . 1 - Grazing time per Kg of DM offered). These results confirmed 
the findings of Torres-Rodriguez ( 1 997) where heifers did not show preference between 
monocultures of red clover and birdsfoot trefoil, but their preference was affected by 
sward physical characteristic (in this case, sward surface height). 
This behaviour in Day 1 also shows that the animals consistently grazed both swards on 
offer, even though they could have met intake requirements by grazing only one sward 
type. Parsons et al. ( l994a) and Cos grove et al. ( 1 996) described this behaviour where 
animals do not have a unique preference for a species or sward as partial preference. 
Newman et al., ( 1 992) suggested some explanation for this behaviour: (a) animals try to 
obtain a balanced diet; (b) animals try to maximise intake rate; (c) physiological responses 
Experiments 1, 2, 3 and 4 1 2 1  
to a novel diet. In E l  Day 1 the fact that the animals spent proportionally more time 
grazing on the minority sward than the proportion of area offered to them, demonstrated 
that the animals either needed to sample other species (mius and Gordon, 1 990) or 
preferred a mixed diet (Parson et al. ,  1 994a Cos grove et al., 1 996; Torres-Rodriguez, 
1 997). mius and Gordon ( 1 990) and mius et al., ( 1987) explain that animals need to 
sample to learn about alternative food on offer. Although Edwards ( 1994), working in 
controlled conditions demonstrated that sheep quickly learn about the location of food 
"patches" and do not need to continue the sampling strategy, he also pointed out that 
changeable sward conditions could make the memory of limited use. According to 
Freeland and Janzen ( 1974) the preference for mixed diet might indicate an evolutionary 
adaptation to reduce the possibility of toxicity. In this context, the effect of secondary 
compounds concentration might be important and a discussion in the context of E5 and E6 
is important to clarify tins behaviour (see General Discussion - Chapter 5). 
Dwell time was calculated from the grazing time observations (each ten minutes) of Day 1 
to clarify the behaviour observed where the animals had access to alternative swards 
(Table 3 .37). The dwell time corresponds to the amount of time each animal spent grazing 
one sward before moving to the alternative one. It was calculated based on the group mean 
value. Although tills analysis provided a good relative impression of the animal movement 
between swards witilln each plot, it is limited by the fact that the grazing observation was 
done only each 10 minutes and no other comparable evidence was found in the literature. 
The results show that in all three experiments the animals changed swards relatively 
frequently, and the dwell time of both swards were similar (each about 20 minutes). Tills 
result indicates that the animals were constantly moving from one sward to another and 
acilleving a mixed diet. However, the proportion of grazing time changed as the sward 
conditions changed. In Day 1 of E l  and E3, for example, the proportion of grazing time 
spent on BW was significantly affected by the treatment (Tables 3 .8  and 3.30). On the 
other hand, in Day 1 of E2 the proportion of grazing time was not affected by treatment, 
the animals grazed about 40-50% of the total grazing time on BW in all treatments. 
Experiments 1, 2, 3 and 4 1 22 
Table 3.37. Amount of time (dwell time - minutes) the animal spent grazing in swards 
of BW or RC before moving to the alternative sward in Experiments 1 ,  2 
and 3 (El ,  E2 and E3, respectively), Day 1 .  
E2 
SEM 
E3 
SEM 
BW RC 
23.2 19.6 
2.2 1 1 .94 
20.3 25.5 
2.52 2.60 
1 9. 1 22.8 
1 .93 2.2 1 
ISEM - Standard error of the mean 
calculated from the variation among 
the five group of animals in four 
replications 
As discussed earlier in relation to the pattern of selection across experiments (section 
3 .4.4), sward morphological characteristics also had important effects on allocation of 
grazing activity between swards in Day 1 of E2. The small difference in morphological 
characteristics between swards (Table 3.3, 3 .24) in E l  and E3, reflected the small 
influence of this component on selection. However in E2, when the sward surface height 
and herbage mass was high, the animals spent more time grazing in RC because of it was 
taller and had greater proportion of leaves. In Day 1 ,  the animals avoided grazing BW 
swards probably because of the lower percentage of leaves and higher percentage of stems 
(Table 3 . 1 2). In fact, the treatment that had the highest proportion of grazing time on BW 
was the treatment in which BW had the highest proportion of leaves and the lowest 
proportion of stems (Treatment B :  BW immaturelRC mature) (Table 3.20). This result 
agrees with other studies where animals showed preference for swards with a greater 
percentage of leaves (Theron and Booysen, 1966; O'Reagain and Mentis, 1989; 
O'Reagain, 1993). 
The animal preference observed in each experiment was unlikely to be influenced by the 
ECT concentration. Although the highest ECT concentration was observed in E l  (Table 
3.4), in that experiment the animals seemed to be more influenced by the area ratio than by 
the species composition (Tables 3.7, 3.8, 3 .9, Figure 3.7). The greatest contrast in ECT 
Experiments 1, 2, 3 and 4 123 
concentration between treatments was found in E2. In this case the small effect of plant 
maturity on diet selection in Days 1 and 2 (Table 3.18), and the small variation between 
days in the proportion of grazing time devoted to graze BW (Table 3.19) demonstrated 
that ECT concentration probably had small effect on diet selection. The variation in ECT 
concentration observed between mature and immature birdsfoot trefoil leaves (ECT 
concentration of 0.37 and 0.79 %, respectively - Table 3.13) was probably not enough to 
affect selection. In addition, the difference in amount of leaf removed in E2 between 
mature and immature plants (Table 3.12) was not so clear as the difference in ECT 
concentration. The higher concentration of condensed tannin in immature plants than in 
mature is also found in other plant species. Coley (1983) and Furstenburg and van Hoven 
(1994), for example, found with tropical trees and shrubs, a higher percentage of tannin in 
young leaves than in old. However, Donnelly (1959) showed that tannin concentration 
increased with maturity in Serica lespedeza. 
As expected, there was a significantly higher concentration of formononetin in red 
clover than in birds foot trefoil or white clover (Tables 3.5, 3.14, 3.26). There was also a 
consistently higher concentration of formononetin in leaves of red clover than in stems 
or petioles across the three experiments. This result agrees with the observations of 
McMurray et al. (1986) and Anwar (1994). However, because there were relatively 
small differences for the percentage of leaves and stems between RC swards within each 
experiment (Appendix 3.3, Tables 3.12, 3.24), it is very unlikely that the formononetin 
concentration had a significant effect on the variation of grazing preference among RC 
swards. On the other hand, the fact that the concentration of formononetin was not 
affected by the variation in red clover maturity in E2 disagrees with the observation of 
Keogh (1995) and Anwar (1994). These authors reported that in red clover there is 
higher concentration of formononetin in young leaves, declining progressively as the 
leaves age. However, Anwar (1994) also observed that the difference in concentration 
between old and young leaves changes according to genotype. The small effect of 
maturity in the currently studies probably reflected the characteristic of the cultivar 
Colenso. No previous study was found in the literature comparing the effect of maturity 
on formononetin concentration of this cultivar. 
Experiments 1, 2, 3 and 4 1 24 
White clover, birdsfoot trefoil and red clover species showed a high nutritive value 
compared to standards described by NRC ( 1989, 1996), particularly in relation to leaves 
(Tables 3.6, 3. 1S, 3.27). Leaves had higher quality based on a lower percentage of fibre 
and a higher percentage of protein and digestibility, than the petioles and stems. However, 
grazing preference was unlikely to have been affected by the sward nutritive value, 
because there were only small differences in quality between species .  The effect of 
nutritive value on diet selection was also examined in ES and E6 (see Chapter 4). 
3.4.8. Diet selection within each sward 
Defoliation caused a substantial decrease in the proportion of leaves and an increase in the 
proportion of stem and dead matter of both swards in all treatments. The increased 
proportion of dead material and stems found in both manual morphological separation 
(Tables 3.3, 3. 1 2, 3.24) and point quadrat analysis (Figures 3.S, 3 .6, 3 . 10, 3. 1 1 , 3 . 12, 3 . 13)  
was explained probably by the selection of green leaf material by animals. Animals usually 
select leaf from stem (Arnold, 1960; Juko and Bredon, 1 96 1 ;  Arnold, 1 964; van Dyne and 
Heady, 1965; Guy et al., 1 98 1 ;  Arnold, 1 98 1 ;  L'Huiller et al. 1 984; L'Hiuller, 1986; 
Edwards, 1 994), young components in relation to old (Arnold, 1960), and green in 
preference to dead material (Arnold, 1960; Juko and Bredon, 1961 ). However as discussed 
in section 3.4.6, in conditions of low herbage mass and sward surface height in Day 3, the 
animals demonstrated selection for swards with high proportion of stems. 
In most of the BW swards in the three experiments, birdsfoot trefoil / white clover ratio 
(B/W ratio) was lower after defoliation than before (Tables 3.3,  3. 12 ,  3 .24), indicating a 
greater intake of birdsfoot trefoil than of white clover. This result together with the point 
quadrat data, showing larger contribution of birdsfoot trefoil in the upper strata (in relation 
to white clover - Figures 3.5, 3 . 10, 3. 12), suggested that the animals grazed more 
birds foot trefoil because it was in the upper strata. This agrees with the results of Gammon 
and Roberts ( 1978), Hodgson ( 1 98 1b), Milne et al. ( 1982), Bircham and Hodgson ( 1 983) 
Briske ( 1986) and mius et al ( 1992), all of which suggest that the variation in the 
Experiments 1, 2, 3 and 4 1 25 
frequency or severity of defoliation of individual sward component is likely to be directly 
related to the size of individual plants and their proximity to the surface of the vegetation 
canopy. 
3.5. CONCLUSIONS 
Sward physical characteristics described by contrasts (in E l ,  E2 and E3) of sward area, 
maturity and structure (represented mainly by height and herbage mass) were important 
determinants of selective behaviour. However their effects were also altered by the 
reduction in herbage mass and height from Day 1 to Day 3 .  Conclusions from these three 
experiments are summarised below: 
1 )  The use of strips of specific swards provided a flexible condition to test animal diet 
selection. The small difference between groups of animals, the absence of any 
previous treatment effect and the agreement between intake and grazing time in E l  
and E 3  showed a good reliability of the results. 
2) Rate of biting apparently did not influence diet selection between swards. Rate of 
biting was more a consequence of sward physical characteristics than a cause of 
selection. 
3) The results suggest that there was a seasonal variation in diet selection between BW 
and RC and that this variation was influenced mainly by the effect of sward 
matumy. This seasonal variation and maturity effect could be mainly explained by 
the preference for leaves and rejection of stems of each species .  
4) The effect of the alternative sward on cattle diet selection was more important as  the 
sward physical limitations increased. This result may be extrapolated to contrasts 
between patches in a heterogeneous sward. 
Experiments 1, 2, 3 and 4 1 26 
5) The animals had the capacity to adjust their diet selection as sward characteristics 
changed. 
6) Results from Day 1 of E l ,  E2 and E3 suggest that selection between BW and RC 
swards was determined more by physical sward characteristics than by the 
preference for a specific species. 
7) Even though the animals could meet intake requirements by grazing only one sward 
type, they grazed both swards, and frequently changed between swards, to achieve a 
mixed diet. 
8) Sward area was an important determinant of selective behaviour between two 
alternative swards, but its importance varied according to the herbage mass and 
height. 
9) Herbage mass and sward surface height had a strong effect on animal diet selection, 
however these experiments suggest that when both swards had high herbage mass 
and height their influence on diet selection was reduced. 
10) Although there was variation between experiments, and between mature and 
immature swards in ECT and formononetin concentration, preference did not seem 
to be affected by the levels of ECT and formononetin concentration found in this 
experiment. 
1 1 ) The animals showed greater intake of birdsfoot trefoil than white clover in BW 
swards. This behaviour can be explained by the fact that birdsfoot trefoil was 
distributed higher in the sward canopy. 
CHAPTER 4 
EXPERIMENTS 5 AND 6 
4.1. INTRODUCTION 
Experiments 5 and 6 were carried out in order to provide explanation for the selective 
behaviour pattern observed in Experiments 1 ,  2 and 3 in relation to how physical and 
biochemical characteristics of plants influence animal preference, and to what extent 
this influence affected seasonal variations in preference. The effects of plant 
morphological characteristics were related to the effects of concentrations of extractable 
condensed tannin in birdsfoot trefoil and to formononetin in red clover. Although 
condensed tannin (e.g. Barry and Blaney, 1 987; Barry, 1 989; Waghorn et al. , 1 990) and 
formononetin (e.g. Marshall, 1 973;  Keogh, 1 995) are recognised to be important 
secondary compounds in pastures, very little is known of their influence on animal 
appraisal and selective behaviour. 
In these experiments dairy cows grazed spaced plants of two genotypes of birds foot 
trefoil with high and low concentrations of extractable condensed tannins, and two 
genotypes of red clover with high and low formononetin concentrations.  For 
convenience, the different cultivars (birdsfoot trefoil: Goldie; red clover: Pawera and G-
27) and accession (birdsfoot trefoil: PI273938) used in these experiments will be 
described as different "genotypes".  Experiment 6, using a rumen content modification 
approach, also had the objective of determining the effect of post-ingestive feedback 
from secondary compounds on cattle diet selection. These experiments were based on 
recent reports and studies that show that the selective response to secondary compounds 
is related to post-ingestive feedback (e.g. Provenza, 1 995) and that sheep can make 
rapid changes in their diet selection as a result of manipulation of rumen environment 
(Carter and Grovum, 1 990; Hou, 199 1  in Cooper et al., 1 995). 
Experiments 5 and 6 1 28 
4.2. MATERIAL AND METHODS 
Two experiments (Experiments 5 and 6) were carried out to investigate the influence of 
condensed tannin in birdsfoot trefoil and formononetin in red clover on diet selection by 
dairy cows. In both experiments sequences of spaced plants of two genotypes of birdsfoot 
trefoil (Lotus corniculatus L.), providing low (cultivar Ooldie) and high (accession PI 
273938) concentration of condensed tannin (CT) and two of red clover (Trifolium 
pratense L.) providing low (cultivar 0-27) and high (cultivar Pawera) concentration of 
formononetin, were offered to dairy cows. The same field design was used in each 
experiment to determine dietary preference, grazing behaviour and trade-off decisions of 
cows. In Experiment 6, rumen conditions in fistulated cows were modified by the 
addition of plant material to investigate the effect of rumen feed back on selective 
behaviour. In Experiment 5 treatments were run three times while in Experiment 6 they 
were run twice. The methodology used in each experiment is described below. 
4.2.1. Experimental site 
Both experiments were located in Palmerston North (400 23 ' S ,  1 750 3 7'E), Manawatu, 
with an average annual rainfall of approximately 1 000 mm . The weather conditions 
monitored at an adjacent site at AgResearch showed that the average monthly soil 
temperature in the year of the experiments (from July 1 996 to June 1 997) ranged from 
8 .0°C (July 1 997) to 1 7.7°C (February 1 997). The monthly rainfall and mean soil 
temperature (la cm depth) from July 1 996 to June 1 997 compared with 60-year average 
values at the site are presented in Appendix 4. 1 .  
4.2.1.1. Experiment 5 
Experiment 5 was carried out at the Dairy Cattle Research Unit (Dairy 3), Massey 
University. The research area was a flat paddock with a soil classified as Tokomaru silt 
loam, a yellow-grey earth which has poor internal drainage (Cowie, 1 972). A sward of 
Experiments 5 and 6 129 
perennial ryegrass with white clover formed this paddock before the experiment, and 
was rotationally grazed by dairy cows. The experiment was repeated in the same area 
three times (three periods): 1 7  January, 1 9  March and 24 April. 
4.2.1.2. Experiment 6 
Experiment 6 was carried out at the AgResearch Grasslands Research Centre. The soil is 
classified as a recent Manawatu silt loam (Cowie, 1972). Before this trial, the area was 
sown to a perennial ryegrass - white clover sward. This sward had been grazed by sheep, 
and occasionally by cattle. The trial was carried out in two periods on 22, 23, 28, 29 
January and on 3, 4, 8, 9 April 1 997. 
4.2.2. Glasshouse sowing and management 
A random sample of seeds of each genotype was sown in a glasshouse on 23 August 1 996. 
Five hundred and sixty plastic pots ( 14  cm diameter and 15 cm height) were sown ( 140 
pots per genotype). Four seeds per pot were sown in a medium composed of 60% 
peatl40% pumice, and a mixture of WOg agricultural lime+300g dolomite+60g of 
Micromax1+300g pa2 mix per 100 1 of medium. Rhizobium trifolii were applied to the 
pots with red clover and Rhizobium loti to the pots with birds foot trefoil. They were 
applied in slurry with water. Glasshouse temperature was regulated between 25°C and 
15°C and the pots were watered daily. 
Thinning started on 1 3  September with the appearance of the first complete leaf (leaf with 
3 leaflets) to eliminate the weaker plants. When the plants had about 4-5 leaves (23 
September 1 996) they were thinned finally to one plant per pot. The plants stayed in the 
glasshouse until transplanted to the field. 
1 The Micromax constituents were 1 2 %  Fe, 2.5% Mn, 1 % Zn, 0.5% Cu, 0. 1 % B ,  0.05% Mo and 1 5 %  of 
combined sulphur. 
2 The PG mix constituents were: 14%N, 1 6%P20s, 1 8%K20, O.03%B, 0. 1 2 %  Cu, 0.02% Mo, 0. 1 6% Mn, 
0.04% Zn, O.09%Fe .  
Experiments 5 and 6 1 30 
4.2.3. Site preparation and management 
Before planting the experiments, the existing perennial ryegrass/white clover sward was 
killed with a broad spectrum herbicide (Round-up+ Granstar). The areas were planted 
on 25 and 2 6  October (Experiment 6) and on 29 and 30 October 1 99 6  (Experiment 5). 
Before the first grazing on both sites, the weeds were hand sprayed with Buster (broad 
spectrum herbicide) and 9 g (90 kglha) of a 1 5 - 1 0- 1 0  (N-P-K) fertiliser were applied per 
plant. After each post-grazing assessment, all the plants were cut at 7 cm and the same 
amount of fertiliser was then applied per plant. Before the third grazing in Experiment 5 
and the second grazing in Experiment 6, a combination of a broad spectrum herbicide 
(Round-up), a broad leaf weed control (Granstar), and a residual herbicide (Simazine) 
was used to control the weeds present between the plants. The information about 
Experiments 5 and 6 is summarised in Table 4. 1 .  
Experiments 5 and 6 1 3 1  
Table 4. 1 .  Management schedule, agrochemical and fertiliser applied in Experiment 5 
and 6. 
Management Schedule 
Agrochemical and Fertiliser 
Area 
preparation 
Planting 
Fertiliser 
Weed 
control 
Experiment 
runs 
Experiment 5 
1 6110/96 
29-30/10196 
23/12/96 and 
after grazing 
23/1 2/96 
02/04/97 
17/01197 
19/03/97 
24/04196 
I 360 gIl Glyphosate 
2 750 glkg Tribenuron 
3 200 g/l Glufosinateammonium 
4 500 gIl Simazine 
4.2.4. Design 
Experiment 6 
1 5/10/96 
25-26/10/96 
applied 
3 1Jha Round-upl + 40 glha 
Gransta? 
20/12/96 and 9 g of 1 5- 10- 10  (N-P-K)/plant 
after grazing 
20/12/96 
0 1/04/97 
22-23/01/97 
28-29101/96 
03-04/04/96 
08-09/04/96 
Buster3 30 mlIl 
0. 1 mlIl Round-up + 33 mgll 
Granstar + 1 ml/l Simazine4 
The trials were formed by isolated plants spaced at one meter intervals in sequences of 
26 plants, each providing blocks (replicates) of sequences of eight treatments in a 
balanced design. The experimental design was balanced with enough replication in 
order to allow plant morphological modification. Differences in plant morphology were 
anticipated between genotypes within species, and this effect was balanced by a 
trimming treatment. In each block one plant of each genotype was trimmed to make the 
two plants of the same species similar in height, area and leafiness. This resulted in 8 
treatments (2 species x 2 genotypes x (± trimming)) allocated to individual position in 
balanced sequences (Table 4.2.). 
Experiments 5 and 6 1 32 
Table 4. 2. Descriptions of the 8 treatments. 
Treatm. Species Genotype Concentration of Plant 
Number secondary compounds morphology 
modification 
1 Birdsfoot trefoil Goldie Low - Condensed Not trimmed 
Tannin 
2 Birdsfoot trefoil Goldie Low - Condensed Trimmed 
Tannin 
3 Birdsfoot trefoil PI273938 High -Condensed Not trimmed 
Tannin 
4 Birdsfoot trefoil PI273938 High -Condensed Trimmed 
Tannin 
5 Red clover G-27 Low - Formononetin Not trimmed 
6 Red clover G-27 Low - Formononetin Trimmed 
7 Red clover Pawera High -Formononetin Not trimmed 
8 Red clover Pawera Hi�h -Formononetin Trimmed 
The eight treatments were arranged at random in linear series with three replications (3 
blocks), forming a sequence of 24 plants. The treatment order was balanced across 
sequences so that each treatment was followed by each other treatment at least once. All 
of the treatments were present with equal frequency down the sequence (Figure 4. 1 .) .  
The first and last plants in each sequence were included to minimise end effects and 
were not included in the analysis. Before pre-grazing assessment any flower present was 
hand removed from all plants. 
Experiments 5 and 6 1 33 
Figure 4. 1 .  Distribution of the set of four sequences with the eight treatments 
(treatments: 1 to 8 - see the description of the 8 treatments in Table 4.2) 
arranged in 3 blocks. 
SEQUENCE J SEQUENCE 2 SEQUENCE 3 SEQUENCE 4 
7 1 3 5 
4 2 4 6 
1 8 3 7 
7 5 1 3 
5 1 6 5 BLOCK 1 
6 7 8 4 
2 4 7 2 
3 6 2 1 
8 3 5 8 
1 3 5 8 
4 6 2 1 
8 4 7 2 
3 7 8 4 BLOCK 2 
2 1 6 5 
6 5 1 3 
5 8 3 7 
7 2 4 6 
7 4 3 5 
5 5 6 4 
6 2 4 2 
2 7 7 1 BLOCK 3 
3 8 1 8 
8 6 5 6 
4 1 8 7 
1 3 2 3 
1 7 5 3 
Experiments 5 and 6 1 34 
4.2.5. Animals and sequence allocations 
4.2.5.1. Experiment 5 
Animals 
Four mature lactating Friesian cows were used. The animals were trained for four weeks 
before starting the trial. Initially the cows were trained to graze down a narrow raceway 
and to be accustomed to fences and people. In order to get used to plant species, one 
week before starting the trial, the cows were trained in an adjacent area with similar 
plant sequences to the experimental area. 
Because the cows in the first run (January/96) did not graze birdsfoot trefoil, two weeks 
before the cows entered the adjacent areas in the second run, they were offered every 
day pre-c1ipped birdsfoot trefoil to get used to this species. No extra pre-clipped red 
clover was offered to the cows because the animals had already experience of grazing 
swards containing red clover and they did not show any rejection of this species in the 
first run. 
Sequences allocation 
One set of four sequences (see Figure 4. 1 .) was used. Each sequence of 26 plants was 
randomly allocated and grazed by one cow. Four spare sequences were planted to be 
used for training responses. Two more sequences were used for probe and visual 
assessment calibration (see Section 4.2.8). 
Experiments 5 and 6 1 35 
4.2.5.2. Experiment 6 
Animals 
Two rumen-fistulated dry Friesian cows were used. The cows were trained for 6 weeks 
before starting the trial. The training procedure was done as in Experiment 5: one week 
before starting the trial the cows were trained in adjacent plant sequences with the same 
plant species and genotypes. 
The cows had previous experience in grazing birdsfoot trefoil and red clover and they 
did not show any rejection in grazing these species during the training period. 
RUMEN MODIFICATION 
The rumen contents of the cows were modified to provide contrasts between high and low 
concentrations of tannin, and high and low concentrations of formononetin. Plants of 
birdsfoot trefoil cv. Goldie (Lotus comiculatus L. - low condensed tannin content), lotus 
maku (Lotus pedunculatus L.- high tannin content), red clover cv. Pawera (Trifolium 
pratense - high formononetin content) and red clover cv. Astred (low formononetin 
content) were applied directly into the cow's rumen through a fistula of approximately 100 
mm diameter. 
Areas with birdsfoot trefoil, lotus maku and red clover were cut in the evening before 
the day of the experiment. After cutting, the material was kept in a chiller until it was 
used. Next morning the material was minced to provide a better distribution of 
secondary compounds. In order to get a good minced material, alfalfa pellets had to be 
mixed to increase the dry matter content. The botanical composition and chemical 
analyses of the minced materials are presented in Appendix 4. 10. 
After mincing, the material was then put into the cow's rumen through the fistula, about 
one hour before running the experiment. The cows were fasted overnight and rumen 
content was removed before the minced material was inserted. The objective was to 
Experiments 5 and 6 1 36 
remove 20-25 kg fresh rumen contents and replace with 1 5-20 kg minced material, 
though the amount of rumen content removed and the material inserted varied according 
to the cow's size, amount of minced material available and previous amount of rumen 
content (see Appendix 4 . 10). It was important not to create an impression of satiety in 
the cows by adding the minced forage to the filled rumen, and it was also important to 
achieve some mixing of mince material with the digesta in order to facilitate feedback 
response. Previous studies with these cows suggested that the minced material 
accounted for about 25% of rumen digesta wet weight (G. C. Waghom, pers. comm.) 
(see Appendix 4. 10) .  
Sequence allocation 
Two sets of the four sequences (see Figure 4. 1 .) were used. Each sequence of 26 plants 
was randomly allocated and grazed in each period by one cow. Two spare sequences 
were planted to be used for training responses. 
The rumen contents of the two cows were modified in balanced sequence according to 
Tables 4.3 and 4.4 in the first and second period, respectively. Each set of four sequences 
in each period corresponded to a run. The cows grazed the first set (in the first two days -
first run) when they had rumen contents modified for forrnononetin concentration. The 
second set of sequences was grazed when the cows had their rumen content modified for 
tannin concentration. In the second period, the same approach was used, however in this 
case the first four sequences were grazed by cows that had their rumen content modified 
for tannin concentration, and the second four sequences for forrnononetin concentration. 
High and low rumen modifications for each secondary compound were balanced between 
cows, runs and periods as shown in Table 4.3. To avoid residual effects, there was at least 
5 days interval between when the rumen was modified with forrnononetin and when it was 
modified with tannin. 
Experiments 5 and 6 
Table 4. 3. Schedule of rumen modification in the first period. 
Run 
1 
2 
Date 
22/0 1197 
23/01197 
28/0 1197 
29101197 
Period 1 
Cow 95 
Low Formononetin 
High Formononetin 
High Tannin 
Low Tannin 
Cow 223 
High Formononetin 
Low Formononetin 
Low Tannin 
High Tannin 
Table 4. 4. Schedule of rumen modification in the second period. 
Run 
3 
4 
Date 
03/04/97 
04104197 
08/04/97 
09/04/97 
Period 2 
Cow 95 
High Tannin 
Low Tannin 
High Formononetin 
Low Formononetin 
Cow 223 
Low Tannin 
High Tannin 
Low Formononetin 
High Formononetin 
4.2.6. Grazing behaviQur assessment 
1 37 
In both experiments (Experiment 5 and 6), the cows during the trial were allowed to 
walk undisturbed down the sequences. The animals were trained for not reversing 
direction or going backwards. They were allocated to a holding area as soon as the 
sequence was completed. The time required to complete the sequence was on average 5 
minutes (ranging from 1 to 9 minutes). During grazing the cow's  behaviour was 
assessed in terms of the time (seconds) spent grazing and the number of bites taken from 
each grazed plant. The assessment was registered manually and by video camera. 
General views of the plant sequences and animal grazing observation are shown in Plate 
4. 1 and 4.2, respectively. 
Experiments 5 and 6 1 38 
Plate 4. 1 .  General view of the sequence of plants, before grazing 
Plate 4. 2. Animal grazing observation. 
Experiments 5 and 6 1 39 
4.2.7. Plant assessment 
A series of descriptive measurements were made on individual plants before and after 
each grazing: 
Plant height 
Plant height was measured with a sward stick (Bircham, 1 98 1 ;  Barthram, 1 986). The 
highest point of each plant was measured in half centimetre increments. 
Plant diameter 
Two diameters (perpendicular and parallel to the sequence) of each plant were measured 
with a ruler in centimetre units. 
Plant density 
Visual assessments of the number of sterns were made for each plant before grazing, and 
given a score from 1 to 10 in 0.5 units, 10 being the densest plant available at the time of 
the leafiness and herbage mass calibration (Real-Ferreiro, 1 997). 
Leafiness 
An eye estimation was made of the percentage of leaf per plant, in intervals of 5%, with 
respect to the total plant material (Real-Ferreiro, 1997) .  The eye estimation was firstly 
calibrated using plants on spare sequences in Experiment 5 ,  offering contrasting ranges 
in size and leafiness. These plants had their leaves (leaf lamina + petioles) weighed 
separated from the other plant parts. 
Experiments 5 and 6 140 
The plants were visually compared with angles drawn on paper (Real-Ferreiro, 1 997): 
Score 1 - angle between the main stems and the horizontal (parallel to the soil) in the 
interval of 0° - 1 8° interval - plant completely prostrate. 
Score 2 - angle between the main stem and the horizontal in the interval of 1 8°_36°. 
Score 3 - angle between the main stem and the horizontal in the interval of 36°_54°. 
Score 4 - angle between the main stem and the horizontal in the interval of 54°-720 
Score 5 - angle between the main stem and the horizontal in the interval of 72°-90° 
Herbage mass 
Herbage mass was assessed with a pasture probe GrassMaster (manufactured by Tru­
Test®) pre and post-grazing. The Pasture probe was calibrated with spare plants in 
Experiment 5 to provide contrasts in morphology (size, leafiness, density and habit) for 
the calibration. Separate equations were used to estimate herbage mass per plant for 
each species. The two equations are presented in Appendix 4.2. The average of three 
random readings per plant was used for estimation of herbage mass. The herbage 
removed from each plant was assessed through the difference between pre and post 
grazing assessments. 
4.2.8. Morphological and chemical analysis 
Four intact stems were harvested from each plant (not considering the first and the last 
plant in each sequence) before grazing. The cuttings were done in a way to avoid 
modification of plant structure. Two stems were randomly chosen for morphological 
determination and two for chemical analysis. The samples were then bulked across blocks 
within each sequence according to genotype and trimming condition. In this way, there 
were 4 replicates per treatment (8 treatments) in each run. 
Experiments 5 and 6 141 
4.2.8.1. Morphological analysis 
The samples for morphological analysis were taken to the laboratory and separated into 
leaf, petiole (only in clovers), stem, flower and dead material. All the samples were dried 
to constant weight in a forced-draught oven, at a temperature of 70-80°C. The samples 
then were weighed individually. 
4.2.8.2. Chemical analysis 
The samples for laboratory analysis were stored at -20°C. The samples were then freeze 
dried and ground to pass through a 1 mm diameter screen. The red clover samples were 
analysed for formononetin content using the methodology described on Section 3 .2.6. 1 
(Chapter 3) and the birdsfoot trefoil samples were analysed for extractable condensed 
tannin content using the methodology described on Section 3 .2.6. 1 (Chapter 3) and 
Appendix 3 . 1 . 
After taking sub-samples for extractable condensed tannin and formononetin analysis, 
similar amounts of the residual samples were bulked across sequences to obtain 8 samples 
(4 different cultivars x trimmed and not trimmed plants) for each run in Experiment 5 ,  and 
8 samples for each different rumen modification run in Experiment 6. These samples were 
analysed, using Near Infrared Reflectance Spectroscopy (NIRS) (Shenk and Westerhaus, 
1994) for estimation of the content of crude protein, neutral detergent fibre (NDF), acid 
detergent fibre (ADF), carbohydrates (soluble sugars plus starch), ash and lipid content. In 
vitro dry matter digestibility (IVDMD) was calculated from ADF, according to equations 
presented in Section 3 .2.6.2 (Chapter 3). 
4.2.9. Statistical analysis 
The plant and animal data were analysed using the statistical package SAS (SAS Institute 
Inc. ,  1985 and 1990). Analyses of variance were carried out to obtain information on 
treatment contrasts in the number of bites taken and grazing time per plant, balanced for 
Experiments 5 and 6 142 
previous treatment. Analyses of variance were also performed for comparison of the sward 
characteristics. Least square means and standard errors of differences were used to 
quantify the contrasts in grazing behaviour and sward characteristics. Analyses of variance 
were performed for Experiments 5 and 6 to verify specific effects in each experiment: 
4.2.9.1. Experiment 5 
A combined analysis between the second and third periods was first performed to clarify 
the effects of the interactions between period and treatment (species, secondary 
compound concentration and trimming). Analyses of variance were also carried out 
within periods to check animal preference between different legumes. Because the 
variation within each species in relation to number of bites was different, a separate 
analysis of each species was performed to clarify the trimming and secondary compound 
concentration effect on the number of bites taken from each plant. 
4.2.9.2. Experiment 6 
As for Experiment 5 ,  combined analyses were used first to check the interaction 
between period and treatment (rumen chemical, species, plant secondary compound 
concentration and trimming) effects and the interaction between rumen chemical and 
species effects. Separate analyses, combining specific rumen chemical with specific 
species over periods were performed to obtain the rumen concentration and plant 
genotype effect on number of bites. Finally, particular analyses were carried out in each 
period, combining specific rumen chemical with specific legume species, to determine 
the trimming and genotype effects on number of bites taken in each plant. 
4.2.9.3. Correlation Analyses 
Correlation analyses were used to clarify the relationships among plant biochemistry, 
plant morphology and number of bites per plant. Correlation analysis was firstly carried 
out to investigate the importance of general plant chemical characteristics on number of 
Experiments 5 and 6 143 
bites. Correlation analyses were also performed to clarify the relationship of ECT and 
formononetin concentration with number of bites and morphological characteristics of 
birds foot trefoil and red clover plants, and the relationship between plant morphology 
and number of bites. In this analysis, logarithmic transformation of the data was used. 
The correlation analysis, followed by covariance analysis, was used to separate the 
effects of plant morphology and ECT concentration on number of bites. 
4.2.9.4. Co variance analyses 
Analysis of co variance was used to separate the plant morphological effect from the 
secondary compound effect on the number of bites taken per plant. Because the effect of 
secondary compounds on number of bites was significant only in plants of birdsfoot 
trefoil, the analysis of covariance was concentrated on number of bites taken in 
birds foot trefoil plants. These analyses were performed to partition differences between 
the two genotypes (cultivar Goldie and accession PI273938) in terms of the effects of 
volume (height and area), leafiness, or secondary compound concentration. As in the 
correlation analyses, logarithmic transformation of the data was used. The regression 
coefficient (R -square) of each analysis was then compared to quantify the morphological 
and biochemical effects. 
Three different covariance analyses were carried out in each experiment. Covariance 
analysis was firstly carried out using individual data of all experimental plants. In this 
case plant morphological characteristics (leafiness and volume) were used as covariates 
while genotype effect was used as a class variable (described as high and low 
concentration of ECT). Genotype effect (class variable) and one covariate, either 
Ieafiness or volume, were added to a basic model to determine their effect on R-square 
changes. A second analysis was then performed considering only the untrirnrned plants 
to test the natural plant morphological variation. A third analysis was carried out 
including both plant morphology and ECT concentration as covariates .  In this analysis 
values of plant volume and leafiness were averaged according to sets of plants bulked 
for ECT chemical determination. 
Experiments 5 and 6 144 
The basic model changed according to  the experiment. In Experiment 5 the basic model 
used contained the effects of plant sequences (= effect of cow), blocks (within 
sequences) and trimming. In Experiment 6 the basic model contained the effects of day, 
cow, rumen chemical added to the rumen, rumen chemical concentration, plant 
sequence, block (within sequence) and trimming. However in the second covariance 
analysis of both experiments the effect of trimming was not included in the basic model, 
and in the third analysis, because values of plant volume and leafiness were averaged 
according to sets of plants bulked for ECT chemical determination, the basic model did 
not include the effect of blocks within sequence. 
4.3. RESULTS 
4.3.1. Experiment 5: Effect of condensed tannin in birdsfoot trefoil (Lotus 
corniculatus L) andformononetin in red clover (Trifolium pratense 
L.) on preference and grazing behaviour of dairy cows. 
4.3.1.1. Plant characteristics 
The results are firstly presented for comparison between birds foot trefoil (BT) and red 
clover (RC) (Table 4.5). Because birdsfoot trefoil was not grazed in Period 1 ,  only 
results of Periods 2 and 3 will be presented here. Attention will then be focused on the 
main trimming and genotype (secondary compound concentration) effects for 
comparisons within each species in Tables 4.6 and 4.7. 
Comparisons between species 
The plant characteristics of the two genotypes of birdsfoot trefoil (BT) and the two 
genotypes of red clover (RC), with and without trimming in Period 2 and 3 are 
presented in Table 4.5.  There were significant interactions involving plant species, 
Experiments 5 and 6 145 
genotypes and trimming effects. Plants were generally bigger in Period 2 than in Period 
3. BT plants were taller and with larger area than RC, but RC had a higher proportion of 
leaves and were more erect and denser than BT. On average, the untrimmed plants had 
more herbage mass than the trimmed ones. Comparing the eight treatments, untrimmed 
plants of cultivar Goldie had the greatest height, mass and area. 
The estimation of the amount of area, height and leafiness removed by grazing 
demonstrated that RC had significantly greater reduction of these characteristics than 
BT. The estimation of herbage mass removed showed that in Period 2, accession 
PI273938 and cultivar Goldie had significantly the lowest and highest, respectively, 
amount of mass removed by the animals, and in Period 3 trimmed plants had 
significantly more mass removed than untrimmed. 
More detailed analyses considering trimming and genotype differences within each 
species are given below. 
Comparisons between cultivars of the same species 
Attention in this section is focus sed on effects of trimming and secondary compound 
concentration (differences between genotypes) within each species and their repeated 
measurements in different periods (Table 4.6 and 4.7). 
BIRDSFOOT TREFOIL 
Since birdsfoot trefoil was not grazed in Period 1 ,  only Periods 2 and 3 are presented 
here (Table 4.6). In Period 2, the genotype with low extractable condensed tannin (EeT) 
concentration was significantly taller and had greater herbage mass, percentage of leaves 
and density, but did not differ in plant area and habit with the high ECT concentration 
genotype. The un trimmed plants were significantly taller with greater mass, area, 
leafiness and density, but did not differ in habit. 
Experiments 5 and 6 146 
Table 4. 5 .Characteristics of untrimmed (Ntrim) and trimmed (Trim) plants of birdsfoot 
trefoil cultivar Goldie (Low Tannin) and accession PI273938 (High Tannin), 
and red clover cultivars G27 [Low Form (formononetin)] and Pawera [High 
Form (formononetin)] in Periods 2 and 3 of EXEeriment 5 .  
Birdsfoot trefoil(BT) Red clover(RC) 
Low Tannin High Tannin Low Form High Form SED] 
Ntrim Trim Ntrim Trim Ntrim Trim Ntrim Trim 
Height (cm) 
pre-grazing 1 8.9 1 5 . 1 1 3 .7 1 3.7 1 5 .6 1 2. 9  1 5 .0 1 3.5 1 .27 
removed2 2 . 1  0.6 -0.2 -0.2 4.4 1 .3 5 .0 2.9 0.97 
Plant mass (gDMlml) 
pre-grazing 494 4 1 7  389 330 393 369 35 1 369 20.4 
removed 1 30 8 8  6 1  1 6  8 6  69 77 79 23.7 
M Plant area (cm2) � 
0 pre-grazing 5540 3 2 1 0  45 1 0  3 1 40 1 790 1 2 1 0  2260 1 240 536.6 -
= removed 3 80 60 1 20 60 570 1 80 530 570 1 63 . 2  � Leafiness (% ) =--
pre-grazing 50 42 40 37 75 7 2  7 5  74 2.4 
removed 8 5 2 1 2 1  1 2  29 1 2  3 . 8  
Habit 
pre-grazing 2.6 2.5 2.2 2.3 2.7 3.0 2.3 3. 1 0.267 
Density 
Ere-gEazing 7.3 6.4 6.2 5 . 5  7 . 9  7 . 8  7 . 7  7.6 0.297 
Height (cm) 
pre-grazing 14. 1 1 1 .0 1 5.5 1 1 .0 1 2.7 1 0. 2  1 1 .2 1 0.2 1 .077 
removed lA  004 1 .0 0.3 3 . 1  1 .2 2. 1 1 .5 0.707 
Plant mass (gDMlm2) 
pre-grazing 430 394 400 361 363 340 356 325 24. 8  
removed 92 99 87 55 93 77 1 03 55 22. 1 
('t') Plant area (cm2) � 
0 pre-grazing 3 1 60 1 770 3440 1 740 1440 970 1 930 970 370.6 - removed 1 80 40 60 40 1 30 40 1 50 60 1 1 8 .0 = 
� Leafiness (%  ) =--
pre-grazing 47 43 42 40 75 69 73 69 2 . 3  
removed 5 2 4 2 1 8  1 5  24 1 7  3 . 3  
Habit 
pre-grazing 204 2.2 2.2 204 3.0 2 . 8  2 . 2  3 . 0  0.206 
Density 
Ere-gEazjn� 6.9 6.5 6.0 
I SED - Standard error for differences of means. 
6.5 7.8 7.7 7.4 7 . 2  0.353 
2 Removed = pre-grazing minus post-grazing assessment 
Experiments 5 and 6 147 
Table 4.6. Characteristics of birdsfoot trefoil plants before grazing, and difference 
between before and after grazing (removed from height, mass, area and 
leafiness) according to trimming (Ntrim = untrimmed; Trim = trimmed) and 
genotype [secondary compound concentration (sec. comp. conc.) :  high 
(accession PI273938), low (cultivar Goldie)] effects in Period 2 and 3 of 
EXEeriment 5 .  
Birdsfoot trefoil 
Trimming Sec. comE. cone SED1 
Ntrim Trim P-yalue� High Low p-yalue� 
Height (cm) 
pre-grazing 1 6.3 14.4 0.0437 1 3.7 1 7.0 0.0006 0.87 
removed3 0.9 0.2 0. 1444 -0.2 1 .4 0.0034 0.50 
Plant mass (gDMfm2) 
pre-grazing 441 373 0.0005 359 455 0.0001 1 7.8  
removed 95 52 0.0386 38 109 0.0015  20.3 
N Plant area (cm2) Q 
0 pre-grazing 5020 3 1 80 0.0001 3820 4380 0. 1 764 402 -
� removed 250 60 0. 1073 90 220 0.2736 1 15 � Leafiness (%)  � 
pre-grazing 45 40 0.0002 39 46 0.0001 1 .3 
removed 5 3 0. 1 6 19  1 .5 6.5 0.0016  1 .5 
Habit 
pre-grazing 2.4 2.4 0.7998 2.2 2.5 0.0827 0. 1 6  
Density 
Ere-grazing 6.8 6 .0 0.0001 5.9 6.9 0.0001 0. 1 7  
Height (cm) 
pre-grazing 14.8 1 1 .0 0.0001 1 3 .3 12.6 0.3388 0.73 
removed 1 . 2  0.4 0.01 23 0.7 0.9 0.4547 0.33 
Plant mass (gDMfm2) 
pre-grazing 4 15  377 0.09 13  380 4 12  0. 1533 2 1 .7 
removed 89 77 0.4761 7 1  95 0. 1 685 1 7.4  
l"') Plant area (cm2) Q 
0 pre-grazmg 3300 1 760 0.0001 2590 2470 0.6575 286 -
� removed 120 40 0.3055 50 1 10 0.4373 78  � Leafiness (%) � 
pre-grazing 45 42 0.0032 4 1  45 0.0003 1 .0 
removed 5 2 0 .0053 2.7 3 .7 0.2589 0.9 
Habit 
pre-grazing 2.3 2.3 0.7042 2.3 2.3 0.7042 0. 1 1  
Density 
Ere-�azing 6.5 6.2 0 . 1 19 1  6 .0 6.7 0.0001 0 . 1 6  
I SED - Standard error for differences of means when comparing either trimmed with untrimmed 
plants or high and low sec. comp. conc. genotypes. 
2 P-value of the treatment main effect. 
3 Removed ::: pre-grazing minus post-grazing assessment 
Experiments 5 and 6 148 
In Period 3 the genotype with low ECT concentration had a significantly (P<O.05) 
greater percentage of leaves and density than the high ECT concentration genotype. 
Similarly to Period 2, the untrimmed plants were significantly taller and had larger area 
and percentage of leaves, but they did not differ in herbage mass, habit and density from 
the trimmed plants. 
The amount removed with grazing was also affected by trimming and genotype effects. 
In Period 2, the animals removed significantly (P<O.05) more herbage mass from 
untrimmed plants, however there was no significant trimming effect in the removal of 
height, area or leafiness. In the same period, the genotype with low ECT concentration 
had greater reduction in height, mass and leafiness than the genotype with high ECT 
concentration, but they did not differ significantly in area reduction. In Period 3, 
untrimmed plants had significantly greater reduction in height and leafiness, but there 
was no significant trimming effect on mass and area. In this period there were no 
significant differences between genotypes in relation to area, mass and percentage of 
leaves removed by grazing. 
RED CLOVER 
The trimming and genotype (secondary compound concentration) effects on RC plants 
are presented in Table 4.7. In Period 1 ,  untrimmed plants differed from trimmed mainly 
because of the larger area. In Period 2, untrimmed plants were taller, had larger area and 
were more prostrate (had proportionally more stems close to the ground) . In Period 3 ,  
untrimmed plants were significantly taller than trimmed plants, with more herbage 
mass, area and leafiness, but did not differ in habit and density. There were no 
significant differences between cultivars with high and low formononetin concentration 
in each of the three periods (P>O.05). 
149 
Table 4. 7 .  Characteristics of red clover plants before grazing and difference between before 
and after grazing (removed from height, mass ,  area and leafiness) according to 
trimming (Ntrim == untrimmed; Trim == trimmed) and genotype [secondary 
compound concentration (Sec. Comp. Conc.): High (accession PI273938), Low 
(cultivar Goldie)] effects in Period 1 , 2 and 3 of Experiment 5.  
Height (cm) 
pre-grazing 
removed3 
Plant mass (gDMlm2) 
pre-grazing 
removed � Plant area (cm2) 
o pre-grazing 
; removed 
g: Leafiness (%)  
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
pre-grazing 
Height (cm) 
pre-grazing 
removed 
Plant mass (gDMlm2) 
pre-grazing 
removed � Plant area (cm2) 
sa pre-grazing 
� removed 
g: Leafiness (%  ) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
pre-grazing 
Height (cm) 
pre-grazing 
removed 
Plant mass (gDMlm2) 
pre-grazing 
removed � Plant area (cm2) 
o pre-grazing 
-
� removed 
g: Leafiness (%  ) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
Red clover (RC) 
Trimming 
Ntrim Trim P-value2 
23.6 
4.9 
343 
28 
3720 
2 1 0  
76 
3 1  
3.2 
7.7 
15.2 
4.7 
372 
82 
2030 
550 
75 
25 
2.5 
7.8 
12.0 
2.6 
359 
98 
1 680 
1 40 
74 
2 1  
2.6 
22.4 
4.5 
3 1 6  
7 
2680 
1 20 
74 
3 1  
3 . 0  
7.4 
1 3.2 
2 . 1 
369 
74 
1 220 
1 70 
73 
1 6  
3.0 
7.7 
1 0.2 
1 .4 
333 
66 
970 
5 0  
69 
1 6  
2.9 
0.47 8 8  
0.7695 
0.3038 
0.3743 
0.0062 
0.583 1 
0.2 1 1 9 
0.9406 
0.3445 
0.4690 
0.0096 
0.00 1 5  
0.80 1 1 
0.4948 
0.0004 
0.00 1 1 
0.3656 
0.0088 
0.02 1 2  
0.7 108 
0.0045 
0.0432 
0.0 1 87 
O.oI 1 6  
0.0024 
0.3 1 46 
0.0299 
0.0808 
0.0990 
Sec comp cone. 
High Low P-value2 
23.0 
4.4 
3 10 
30 
3 1 20 
1 20 
74 
29 
3 . 1  
7.9 
1 4.2 
4.0 
360 
78 
1 750 
350 
74 
24 
2.7 
7.6 
1 0.7 
1 . 8  
340 
79 
1 450 
1 1 0 
7 1  
20 
2 . 6  
23 . 1  
5 . 1  
350 
320 
3280 
2 1 0  
76 
33 
3.2 
7.2 
1 4. 2  
2.8 
3 8 1  
7 8  
1 500 
370 
74 
1 7  
2.9 
7.9 
1 1 . 5  
2. 1 
350 
85 
1 200 
80 
72 
1 7  
2.9 
0.9328 
0.6330 
0. 1 28 8  
0.2389 
0.6599 
0.5850 
0.2 1 1 9 
0. 1 245 
0.849 1 
0.2308 
0.9780 
0. 1409 
0.0694 
0.9730 
0.223 1 
0.8069 
0.6967 
0.0228 
0.4258 
0.2700 
0.2040 
0.5964 
0.2962 
0.6365 
0.2684 
0.7997 
0.5744 
0.2384 
0.0990 
SED 
1 .7 2  
1 .3 3  
25.5 
23.9 
357 
1 67 
1 .5 
2 . 8  
0.2 1  
0.5 1 
0.75 
0.75 
1 1 .2 
I Ll  
204 
1 05 
1 . 6  
3. 1 
0 . 2 1  
0.22 
0.59 
0.58 
1 0.7 
1 1 .9 
2 1 7  
9 1  
2 . 2  
2 . 8  
0. 1 7  
pre-grazing 7.6 7 .5 0.60 1 6  7.3 7.7 0. 1 966 0.32 
lSED - Standard error for differences of means when comparing either trimmed with untrimmed plants or high and low sec. 
comp. cone. genotypes. 
2 P-value of the treatment main effect. 
3 Removed :;;; pre-grazing minus post-grazing assessment 
Experiments 5 and 6 1 50 
The amount removed by grazing from cultivars with high or low formononetin 
concentration did not differ significantly in the three periods of measurements, except in 
Period 2 where the cultivar with high formononetion concentration had greater 
reduction in percentage of leaves. In Period 1 ,  there were no significant differences 
between trimmed and untrimmed plants in reduction of height, mass, area and leafiness. 
In Period 2, untrimmed plants had significantly greater height, area and leafiness 
removed by grazing. In Period 3 there was a significantly larger decrease only in height 
and mass of un trimmed plants, compared with trimmed plants. 
4.3.1.2. Sward chemical composition 
Extractable condensed tannin concentration 
There was a significant interaction (in absolute terms) in concentration of extractable 
condensed tannin (ECT) between genotype and period in plants of birdsfoot trefoil 
(Table 4.8.) .  There was a larger difference between genotypes in Period 2 than in Period 
3 .  In proportional terms, accession PI273938 had on average 4.2 times more tannin than 
cultivar Goldie. Comparing within each genotype, the concentration of ECT was greater 
in Period 2 than in Period 3 .  There was no significant interaction between trimming and 
genotype effect (Table 4.8) and no difference between trimmed and untrimmed 
(trimming main effect) birds foot trefoil plants in ECT content (P=0.3984). 
Table 4. 8 .  Extractable condensed tannin (ECT) concentration (%DM) of birds foot 
trefoil genotypes considering the interactions with period and trimming 
effects (untrimmed: Ntrim; trimmed: Trim) of Experiment 5.  
Goldie Pl273938 SED1 P-value2 
Period 
Trimming 
2 
3 
Ntrim 
Trim 
0.72 3.28 
0.47 1 . 8 1  
0.48 
0.72 
2.47 
2.62 
I SED - Standard error for differences of means 
0.214  0.0007 
0.2 14 0.8063 
2 P-value of the interactions : period *genotype or trimming*genotype. 
Number of observation contributing for each mean (n::::8) 
Experiments 5 and 6 1 5 1  
Formononetin concentration 
There was a significant interaction between trimming and genotype effect in 
fonnononetin content in red clover (Table 4.9). Pawera had higher concentration of 
fonnononetin when trimmed, though G-27 did not show difference between trimmed 
and untrimmed. There were no significant interactions with period (Table 4.9). In 
proportional tenns, overall Pawera had 2.3 times more fonnononetin than G-27. 
Table 4. 9. Formononetin concentration (%DM) of red clover genotypes considering the 
interactions with period and trimming (untrimmed: Ntrim; Trimmed: Trim) 
effects of Experiment 5 .  
1 
Period 2 
3 
Ntrim 
Trimming 
Trim 
G-27 Pawera SED] P-valui 
0.28 0 . 67 
0.30 0.72 
0.30 0.65 
0.29 0.63 
0.26 0.73 
0.03 1 0. 1 5 5 1 
0.025 0.0086 
ISED - Standard error for differences of means 
2 P-value of the interactions : period *genotype or trimming*genotype. 
Number of observation contributing for each mean of the interaction 
genotype*period (n=8) and genotype*trimming (n::::: 12) 
General chemical composition 
The general chemical composition of each genotype is presented in Table 4 . 1 0. There 
were significant differences between genotypes in relation to the percentage of protein, 
lipids, ADF and in vitro DM digestibility (IVDMD). The accession PI273938 had 
significantly the lowest proportion of protein compared to the other genotypes, and the 
cultivar Goldie had significantly the highest percentage of lipid and IVDMD. In this 
case, the two cultivars of red clover had the lowest percentage of lipid and IVDMD. 
Experiments 5 and 6 1 52 
Red clover had significantly higher NDF (27.0 vs 24. 1 ,  SED 0.922, P=0.007 1 )  and ash 
( 1 0.5 vs 8 .7,  SED 0. 1 1 3,  P=O.OOOl )  than birdsfoot trefoil, independent of genotype. 
Table 4. 1 0. Percentage of crude protein (CP), lipid, acid and neutral detergent fibre 
(ADF, NDF), carbohydrates (soluble sugars plus starch)(CHO), ash and in 
vitro dry matter digestibility (IVDMD) determined by Near Infrared 
Reflectance Spectroscopy (NIRS) of plants of birdsfoot trefoil and red 
clover of Experiment 5 (percentage of DM basis). 
Birdsfoot trefoil Red clover p-SED] SED2 SEd value 4 
Goldie PI273938 G-27 Pawera 
Protein 2 1 .7 17.5 23 .6 24.0 0.456 0.345 0.404 0.0001 
Lipid 3.7 3.4 2.8 2.9 0.066 0.050 0.058 0.0010  
ADF 20.8 23.9 25.2 25.5 0.729 0.55 1 0.646 0.0033 
NDF 25.5 22.6 27.0 26.9 1 .297 0.98 1 1 . 1 50 0. 1 002 
CHO 16.0 1 6.6 14.2 14.4 0.9 1 2  0.689 0 .808 0.7899 
Ash 8.7 8.8 10.4 1 0.5 0. 173 0. 130 0. 1 53 0.5696 
IVDMD 77.2 74.4 73.3 73.0 0.658 0.497 0.583 0.0033 
ISED - standard error for differences of means when comparing means between birdsfoot trefoil genotypes. 
2SED 
- standard error for differences of means when comparing means between red clover genotypes. 
3SED - standard error for differences of means when comparing means of species. 
4P-value of genotypes within each species. 
Number of observation contributing for each mean (n=6). 
4.3.1.3. Number of bites per plant 
There were no significant interactions between period and treatment (species, trimming 
and secondary compound effects) (P>O.05) in relation to number of bites per plant, so 
attention is focused within each period. 
In Period 1 the animals grazed only red clover, therefore comparisons including the four 
genotypes are made only for Period 2 and 3 .  There were no significant interactions with 
trimming effect. However, there was a significant main effect of trimming in both 
periods (Period 2 and Period 3). Plants that had been trimmed had less bites (Period 2 -
7.6 vs 3 . 1 ,  SED 1 .00, P=O.OOOl ;  Period 3 - 8.4 vs 3.6, SED 1 . 2 1 ,  P=0.0002). The 
number of bites per plant was marginally greater for RC than BT in Period 2 (6.3 vs 4.4, 
Experiments 5 and 6 1 53 
SED 1 .00, P==0.0526), but there was no significant difference in Period 3 (P==O.6 198). 
However there was a significant interaction between plant secondary compound 
concentration and species effects on the number of bites (Table 4. 1 1 ) .  In Period 2, the 
BT genotype with high concentration of ECT had significantly the least number of bites 
(P==0.0046). In Period 3, there was similar selective behaviour, but the difference was 
only marginal (P==0.06 16) .  
Table 4.  1 1 . Average of number of bites per plant in birdsfoot trefoil (BT) and red 
clover (RC), in relation to plant secondary compound concentration 
(Sec.Comp.Conc.)  and trimming (Trim == trimmed; Ntrim == untrimmed) 
characteristics in Period 1 and Period 2, Experiment 5 .  
Sec. Camp. Cone. Trimming SED] 
Species Low High P-value2 Trim. Ntrim. P-value2 
Period 
1 RC 10.3 9.4 0.5 1 3 8  8 .0 1 1 .7 0.0 1 26 1 .39 
Period BT 6.6 2. 1 0.0068 1 .9 6 .8 0.0035 1 .57 
2 RC 5.5 7.2 0. 1 305 4.3 8 .3  0.0007 1 .07 
Period BT 7.5 3 .9 0.0693 3.0 8 .4 0.0086 1 .9 1  
3 RC 5.3 7.3 0. 1 9 1 2  4.2 8 .4 0.0096 1 .53 
SED3 1 .422 P-value3 0.0046 
SED4 1 .7 15  P-vaIue4 0.06 16  
I SED - Standard error for differences of means when comparing means with the same level of 
treatment (sec. comp. conc. or trimming) 
2 P-value of the sec. comp. conc. or trimming main effect. 
3 SED - Standard error for differences of means and P-vaIue when comparing means of the four 
genotypes (interaction sec. comp. conc. * species) in Period 2. 
4 SED - Standard error for differences of means and P-vaIue when comparing means of the four 
genotypes (interaction sec. comp. cone. *species) in Period 3 .  
Individual analyses for comparisons of genotypes within each species were carried out 
to clarify genotype and trimming effects. The analyses are presented in Table 4. 1 1 . The 
interaction between secondary compound concentration and trimming effects was not 
significant in all three periods, so results are presented as main effects only. 
The animals took significantly more bites from plants with low ECT concentration than 
with high ECT concentration,  and from untrimmed than trimmed plants of birdsfoot 
Experiments 5 and 6 1 54 
trefoil. However the difference between high and low ECT concentration in Period 3 
was only marginal. 
In red clover, trimming had a major influence on bite number. There was no significant 
effect of formononetin concentration. There were significantly more bites in untrimmed 
plants than trimmed in all three periods of measurement. 
Correlation and covariance analysis using number of bites per plant were carried out for 
better understanding of the animal preferential grazing. The analyses are presented in 
section 4.3. 1 .5 and 4.3 . 1 .6. 
4.3.1.4. Rate of Biting 
The values of number of bites per minute in birdsfoot trefoil and red clover plants 
according to the secondary compound concentration and trimming effect are presented 
in Tables 4. 1 2  and 4. 1 3. Comparing only the grazed plants, there was consistently no 
significant difference between birdsfoot trefoil (BT) and red clover (RC) in Periods 2 
(P2) and 3 (P3) (42.5 vs 36.9 bites/min, SED 7.47, P=0.5542 in P2; 38.2 vs 42.4 
bites/min, SED 2.76, P=O.0856 in P3; for BT and RC, respectively). There was also no 
significant effect of secondary compound concentration and trimming (Tables 4. 1 2, 
4. 1 3.), and no significant interactions of these effects with period. 
Experiments 5 and 6 1 55 
Table 4. 1 2. Average of number of bites per minute (biting rate) in birdsfoot trefoil in 
relation to secondary compound concentration (Sec.Comp.Conc.) and 
trimming (Ntrim = untrimmed; Trim = trimmed) characteristics in Period 2 
and Period 3 .  
Sec. comp. conc. 
Low 
High 
SED 
P-value 
Trimming 
Ntrim 
Trim 
SED! 
P-value2 
Birdsfoot trefoil 
Period 2 Period 3 
43. 1 
49.2 
4.050 
0. 1246 
46.9 
45.7 
4. 1 17 
0.5527 
39.8 
4 1 .3 
3 .872 
0.6022 
38.4 
42.7 
4.001 
0. 1970 
ISED - Standard error for differences of means when comparing means with the same 
level of treatment (sec. comp. conc. or trimming) 
2 P-value of the sec. comp. conc. or trimming main effect. 
Table 4. 1 3 .  Average of  number of bites per minute in  red clover, i n  relation to 
secondary compound concentration (Sec.Comp.Conc.) and trimming (Trim 
::: trimmed; Ntrim ::: untrimmed) characteristics in Periods 1 ,  2 and 3,  
EXEeriment 5 .  
Red clover 
Period 1 Period 2 Period 3 
Sec. comp. conc. 
Low 3 1 . 1  36. 1 42.8 
High 33.7 39.5 42.3 
SEDl 2.260 2.700 3 .892 
P-value2 0.2032 0.7448 
Trimming 
Ntrim 3 1 .0 38.3 42.3 
Trim 33.8 37.3 42.8  
SEDl 2 .260 2.700 3 .849 
P-value2 0. 1992 0.65 1 1 0.85 10 
lSED - Standard error for differences of  means when comparing means with the same 
level of treatment (sec. comp. conc. or trimming) 
2 P-value of the sec. comp. conc. or trimming main effect. 
4.3.1.5. Correlation Analyses 
Correlation analysis was firstly carried out to investigate the importance of general plant 
chemical characteristics on number of bites per plant. Correlation analyses were also 
performed to clarify the relationship of ECT and formononetin concentration with 
number of bites and morphological characteristics of birdsfoot trefoil and red clover 
Experiments 5 and 6 1 56 
plants, and the relationship between plant morphology and number of bites. The 
correlation analysis, followed by covariance analysis (section 4.3 . 1 .6), was used to 
separate the effects of plant morphology and ECT concentration on number of bites. The 
full correlation coefficient matrices of each analysis is presented in Appendices 4.3, 4.4, 
4.5, 4.6. 
Number ot bites vs General plant chemical characteristic 
The correlation coefficients and the probability of significance between general plant 
chemical characteristic and number of bites are given in Table 4. 14. The number of 
observations for this analysis was limited to the number of samples used in the chemical 
analysis: one sample of each treatment in each period. Because in Period 1 only red 
clover plants were chemically analysed, correlation analysis was based on 1 2  
observations of red clover and 8 o f  birdsfoot trefoil. The number of bites was averaged 
according to the sets of samples bulked for chemical analysis. Because of the limited 
number of observations the differences between periods were not considered. 
Lipid in birdsfoot trefoil was the only compound to have a significant (P<0.05) 
correlation with number of bites. Carbohydrates (soluble sugars plus starch) and ash in 
red clover showed marginal significance. The other plant chemical characteristics did 
not have significant (P>0.05) correlation with number of bites. 
Table 4. 14. Pears on Correlation coefficients (r) from correlation analysis between 
number of bites per plant and percentage of protein, lipid, acid and neutral 
detergent fibre (ADF, NDF), carbohydrates (soluble sugars plus 
starch)(CHO), ash and in vitro dry matter digestibility (IVDMD) of plants of 
birdsfoot trefoil and red clover of Experiment 5 (percentage of DM basis). 
Number of Bites Protein Lipid ADF NDF CHO Ash IVDMD 
Birdsfoot trefoil 
r 0.6287 1 0.7905 -0.3072 0.5888 -0.5275 
P-value 0.0950 0.0195 0.4592 0. 1246 0. 1791 
Red clover 
r -0.46502 -0. 1 737 -0. 1932 0.3327 0.5523 
P-value 0.1277 0.5892 0.5474 0.2906 0.0626 
lNumber of observation contributing for each correlation of birdsfoot trefoil (n:::8)  
2Number of observation contributing for each correlation of red clover (n=1 2) 
0.3839 0.3 1 15 
0.3477 0.4526 
-0.5405 0. 1955 
0.0696 0.5426 
Experiments 5 and 6 1 57 
ECT concentration vs Plant morphology and Number of bites in birdsfoot trefoil plants 
The correlation analyses between ECT concentration and either birdsfoot trefoil 
morphological characteristics or number of bites taken from birds foot trefoil plants in 
Periods 2 and 3 are presented in Table 4. 1 5 .  In these analyses values of plant 
morphology and number of bites were averaged according to sets of plants bulked for 
ECT chemical determination. There were significant negative correlations (P<0.05) 
between ECT concentrations and plant height, leafiness and number of bites in Period 2, 
but in Period 3 (Table 4. 1 5) there was only a significant negative correlation between 
ECT concentration and leafiness. 
Table 4. 1 5 .Pearson Correlation coefficients (r) from correlation analysis between 
extractable condensed tannin concentration and plant area, height, volume, 
leafiness and number of bites per plant (N. Bites) of birdsfoot trefoil in 
Periods 2 and 3 of Experiment 5 .  
Period 
2 
r 
P-value 
Area Height 
-0.0973 -0.6054 
0. 7199 0.0130 
Volume 
-0.2683 
0.3151 
Period r 0.0399 0. 1448 0.0926 
3 P-value 0.8878 0.6065 0. 7428 
Number of observation contributing for each correlation (n=1 6) 
Leafiness 
-0.6728 
0.0043 
-0.4860 
0.0662 
N.Bites 
-0.503 1 
0.0470 
-0.0 1 38 
0.9609 
Formononetin concentration vs Pant morphology and Number of bites in red clover 
plants 
The correlation analyses between formononetin concentration and red clover 
morphological characteristics, and between formononetin concentration and number of 
bites taken from red clover plants in Periods 1 ,  2 and 3 are presented in Table 4. 1 6. In 
this analysis values of plant morphological characteristics and number of bites were 
averaged according to sets of plants bulked for formononetin chemical determination. 
There was no significant correlation (P>0.05) between formononetin concentration and 
plant height, area, volume and leafiness, or number of bites, in the three periods. 
Experiments 5 and 6 1 58 
Table 4. 16.Pearson Correlation coefficients (r) from correlation analysis between 
formononetin concentration and plant area, height, volume, leafiness and 
number of bites per plant (N. Bites) of red clover in Periods 1 and 2 of 
Experiment 5 .  
Area Height Volume Leafiness N.Bites 
Period r -0. 1 596 -0. 1063 -0. 1 7 1 8  -0. 1 854 -0. 1024 
1 P-value 0.5548 0.6951 0.5245 0.4919 0.7058 
Period r 0. 141 1 -0.0492 0.0962 0.0926 0. 1 725 
2 P-value 0.6021 0.8565 0. 7230 0. 7330 0.5229 
Period r 0.01 19  -0. 1795 -0.0522 -0. 1 048 0.2558 
3 P-value 0.9650 0.5060 0.8477 0.6993 0.3389 
Number of observation contributing for each correlation (n=1 6) 
Number o[bites vs Plant more.hoiogy 
BIRDS FOOT TREFOIL 
The correlation analysis between number of bites per plant and morphological 
characteristics of birdsfoot trefoil plants in Periods 2 and 3 is shown in Table 4. 1 7. In 
this analysis individual data of all experimental plants was used. In both periods there 
were significant and positive correlations between number of bites and plant area, 
height, volume and leafiness. 
Table 4. 17 .Pearson Correlation coefficients (r) from correlation analysis between 
number of bites per plant and area, height, volume and leafiness of 
birds foot trefoil plants in Periods 2 and 3 of Experiment 5 .  
Area Height Volume Leafiness 
Period r 0.2847 0.4398 0.3945 0.5749 
,., P-value 0.0499 0.0018 0.0055 0.0001 ... 
Period r 0.438 1  0.5089 0.5049 0.434 1 
3 P-value 0.0018 0.0002 0.0003 0.0020 
Number of observation contributing for each correlation (n:::48) 
Experiments 5 and 6 1 59 
RED CLOVER 
The correlation analysis between number of bites per plant and morphological 
characteristics of red clover plants in Periods 1 ,  2 and 3 is shown in Table 4. 1 8 .  As in 
the previous analysis ,  this analysis was carried out using individual data of all 
experimental plants. In the three periods there were significant and positive correlations 
between number of bites and plant area, height and volume. The correlation between 
number of bites and leafiness was only significant in Period 3. 
Table 4. 1 8 .Pearson Correlation coefficients (r) from correlation analysis between 
number of bites per plant and area, height, volume and leafiness of red 
cloveq;!lants in Periods 1 , 2 and 3 of EXEeriment 5 .  
Area Heis.ht Volume Leafiness 
Period r 0.5595 0.4473 0.6525 0. 1 1 26 
1 P-value 0.0001 0.0014 0.0001 0.4462 
Period r 0.7837 0.4789 0.7842 0.0485 
2 P-value 0.0001 0.0006 0.0001 0.743 1 
Period r 0.5676 0.573 1 0.6379 0.3497 
3 P-value 0.0001 0.0001 0.0001 0.0148 
Number of observation contributing for each correlation (n:::48) 
The correlation matrices showed that the coefficients derived from the smaller number 
of observations used in the analyses involving ECT and Forrnononetin concentration 
were not substantially different from the analyses carried out with individual plant data 
(more number of observations) . There was only a notable greater correlation coefficient 
between number of bites and height of red clover in Period 2 (0. 14  vs 0.48) when more 
observations were used. However the significance between number of bites and either 
area (P= 0.0499 vs 0.365 1 )  or volume ep= 0.0055 vs 0. 1654) of BT in Period 2 was also 
greater when more observations were used. 
Experiments 5 and 6 1 60 
4.3.1.6. Use of Covariates 
Co variance analyses were used to separate the plant morphological effect from the 
secondary compound concentration effect on bite number. Because the effect of 
secondary compounds on number of bites was significant only in plants of birdsfoot 
trefoil, the analysis of covariance is concentrated on number of bites taken from 
birds foot trefoil plants. 
Covariance analysis was firstly carried out using individual data of all experimental 
plants (number of observations (n) = 48). Plant morphological characteristics (leafiness 
and volume) were used as covariates while genotype effect was used as a class variable 
(described as high and low concentration of ECT). Genotype effect (class variable) and 
one covariate, either leafiness or volume, were added to a basic model to determine their 
effect on R-square changes. The basic model of the analysis of variance considered as 
causes of variation the effects of sequences, blocks (within sequences) and trimming. A 
second analysis was then performed considering only the untrimmed plants (n=24) to 
test the natural plant morphological variation. In this case, the effect of trimming was 
not included in the basic model. A third analysis was carried out including both plant 
morphology and ECT concentration as covariates (n= 16) .  In this analysis values of plant 
volume and leafiness were averaged according to sets of plants bulked for ECT 
chemical determination, and the basic model did not include the effect of blocks within 
each sequence. These three analyses are presented below. 
Genotype effect (class variable) vs Plant morphological characteristics (covariate) 
TRIMMED AND UNTRIMMED PLANTS 
The R-square changes when adding genotype and either the percentage of leaves or 
plant volume to a basic model in the Periods 2 and 3 are given in Figure 4.2 and 4.3 . 
The basic R-square (without the addition of genotype effect or any covariate) in Period 2 
was lower than in Period 3 and the variation in R-square was greater in Period 2 than in 
Period 3 .  The basic model explained about 40% in Period 2, and 50% in Period 3 of the 
Experiments 5 and 6 1 6 1  
total variation. The improvement i n  R-square after fitting covariates and genotype effect 
varied from 12  to 27% (adding both leafiness and genotype effects in Periods 3 and 2, 
respectively). Leafiness explained more variation than volume in Period 2 (27 vs 14 %), 
but they had similar effect on R -square variation in Period 3 ( 1 5  vs 1 7  %). The genotype 
effect improved the magnitude of R-square more in Period 2 than in Period 3 ( 14 vs 3 
%). 
In both periods there was a significant effect of genotype (marginal in Period 3), 
leafiness and volume when added to the basic model . In Period 2, leafiness explained 
more variation in number of bites than genotype. There was only a small residual effect 
of genotype after leafiness had been added to the model. In this period plant volume 
explained a similar amount of variation as genotype. In Period 3, leafiness was much 
more important than the genotype effect to explain variation in number of bites. After 
adding leafiness to the model, the genotype effect was not significant (P=O.9525). 
Experiments 5 and 6 
Genotype effect 
before Leafiness 
Leafiness effect after 
Genotype 
Gentype effect before 
Volume 
Volume effect after 
Genotype 
162 
PERIOD 2 
0.4270 Leafiness effect before 
Genotype 
.. 13887 � .. 2654(0.0000/ 
0.5657 0.6924 
•• 1283 � / .... 17 (0 •. ,.::,67:....;4..:;;.,1 ):...-_-::-::----::---, 
Genotype effect after 
•. 13887 
0.5658 
•. M96� 
0.694 1 Leafiness �------------� 
0.4270 Volume effect before 
Genotype 
� •• 1450(0.0060) 
0.5720 
/ •• 0433 �.1J665 
Genotype effect after 
0.6154 Volume 
Figure 4.2. Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
or genotype effect had been added. Considering all experimental plants of 
birdsfoot trefoil in Period 2 of Experiment 5.  
l P-value for the differences o f  R-squares. 
Experiments 5 and 6 
Genotype effect 
before Leafiness 
Leafiness effect after 
Genotype 
Genotype effect before 
Volume 
Volume effect after 
Genotype 
1 63 
PERIOD 3 
0.5453 Leafiness effect before 
Genotype 
0.03517 � 0.1243(0.0013/ 
0.5804 0.6696 
0.0892 � / 0.OOOO(0'r95:::.::2:.::..5,--_�_.,.---, Genotype effect after 
0.035117 
0.5804 
0.1946� 
0.6696 Leafiness �------------� 
0.5453 Volume effect before 
Genotype 
� 0.1680(0.0001) 
0.7133 
/0.0617 (i.OO57 
Genotype effect after 
0. 7750 Volume 
Figure 4.3 .  Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
or genotype effect had been added. Considering all experimental plants of 
birdsfoot trefoil in Period 3 of Experiment 5 .  
1 P-value for the differences of  R-squares. 
Experiments 5 and 6 164 
UNTRIMMED PLANTS 
The R-square changes when adding genotype effect and either the plant leafiness or 
volume in Period 2 and 3 ,  considering only untrimmed plants of birdsfoot trefoil are 
presented in Figure 4.4. and 4.5. In both periods the basic model, without the covariate 
and genotype effects, explained between 45 and 47 % of the total variation. Leafiness 
and genotype effects together explained around 35 % of the total number of bites 
variation in both periods. Plant volume together with genotype effect explained much 
more variation in number of bites in Period 3 than Period 2. The improvement in R­
square in Period 2 was about 25%, but in Period 3 was about 43%. In Period 3, the basic 
model together with genotype and volume effects explained most of the variation (R­
square == 0.9055) in number of bites. 
In Period 2, leafiness had an important effect on the variation in number of bites, but it 
did not have a significant effect after genotype had been added to the model. In this 
period leafiness was more important than volume to explain the changes of R -square. 
Volume had a marginal effect (P=O.0969) on changes of R -square, but also did not have 
significant effect after genotype had been added to the modeL In both cases, after 
adding leafiness or volume to the model, genotype effect was not significant. The effect 
of genotype was lower than leafiness and greater than volume, though the differences 
between genotype effect and the effect of the covariates were small. 
In Period 3, leafiness and volume had similar effects on the variation of number of bites. 
The variation in number of bites explained by differences in genotypes seems to be 
independent of plant volume, but related to plant leafiness. The genotype effect after 
volume had been added to the model, and the volume effect after genotype had been 
added to the model were highly significant. On the other hand, the genotype effect was 
not significant after leafiness had been added to the model. 
Experiments 5 and 6 
Genotype effect 
before Leafiness 
Leafiness effect after 
Genotype 
Genotype effect before 
Volume 
Volume effect after 
Genotype 
1 65 
PERIOD 2 
0.4506 Leafiness effect before 
Genotype 
0.2348 7 � 0.2885(0.0071/ 
0.6854 0. 7391 
0.0614 � / 0.0077(0 .;:.,59::..:;5:.;:.9'--_...,..,-_..,.......----, Genotype effect after 
0'��7 
0.6854 
0.0124� 
0. 7468 Leafiness �------------� 
0.4506 Volume effect before 
Genotype 
� 0.2055(0.0969) 
0. 6561 
/ 0.0417 0.2925 
Genotype effect after 
0.6978 Volume 
Figure 4.4. Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
(leafiness or volume) or genotype effect had been added. Considering only 
untrimmed birdsfoot trefoil plants in Period 2 of Experiment 5. 
l P-value for the differences of R-squares. 
Experiments 5 and 6 
Genotype effect 
before Leafiness 
Leafiness effect after 
Genotype 
Genotype effect before 
Volume 
Volume effect after 
Genotype 
166 
PERIOD 3 
0.5410  
0.4706 Leafiness effect before 
Genotype 
0. 7160 
0."'75 � / 0.0325(0,r::;28:=2::...5'-:-_-;;---:--;:----, Genotype effect after 
0'��7 
0.54 10 
03645� 
0. 7485 Leafiness L..-______ -' 
0.4706 Volume effect before 
Genotype 
� 0.2397(0,(1034) 
0. 7103 
/0.1952(0,0020 
Genotype effect after 
0.9055 Volume 
Figure 4.5 . Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
(Ieafiness or volume) or genotype effect had been added. Considering only 
untrimmed birdsfoot trefoil plants in Period 3 of Experiment 5. 
l P-value for the differences of R-squares. 
Experiments 5 and 6 1 67 
ECT concentration (covariate) vs Plant morphological characteristics (covariate) 
The R-squares and the differences of R-squares when plant leafiness or volume and 
BCT concentration were added to the model in the analysis of variance, considering 
either Period 2 or Period 3, are given in Figures 4.6 and 4.7. The basic model in Period 
2 explained less variation (59%) in number of bites than in Period 3 (80%). Because a 
large amount of the variation was explained by the basic model, the improvement in R­
square after fitting the covariates was much smaller in Period 3 (from 2 to 1 2%) than in 
Period 2 (about 20%). 
In Period 2, there were significant effects of both ECT concentration and leafiness on R­
square, but there was no significant effect (P>0.05) of volume. However BCT 
concentration and leafiness were not independent. After adding one covariate, the effect 
of the other became non significant. In Period 3, ECT concentration and plant 
morphological characteristics were not important in explaining the number of bites . In 
Period 3 there were no significant effects of ECT concentration and plant morphology 
(leafiness and volume) on R-square changes. 
The effect of the covariate ECT concentration increased relatively to the differences 
between genotypes. Comparing the three different covariate analyses, the ECT 
concentration (covariate) and genotypes (class variable) had stronger effects in Period 2 
than Period 3 .  In all analyses leafiness had the strongest effect on changes of R-square. 
The effects of genotype (class) or ECT concentration (covariate) were never significant 
after leafiness had been added to the model. 
Experiments 5 and 6 168 
PERIOD 2 
ECT conc. effect before 0.5872 Leafiness effect before 
Leafiness ECT conc. 
'---------' 
•• 1522 7' � 0.2163 (D OIl 6) , 
0. 7395 0.8036 
Leafiness effect after 
ECT conc 
ECT cone. effect before 
Volume 
~ •• 064' /'0.0008 (0.8;;::5..:.;17...:.,) __ .....",_-..,...---, ECT conc. effect after 0.8044 Leafiness '-- ----_ ..... 
0.5872 Volume effect before 
ECT conc. 
•• 1522 7 � 0.1390 (0. 1545) 
0. 7395 0. 7262 
Volume effect after ECT 
conc. 
0 •• 274� 
0. 7669 
Figure 4.6. Effect of extractable condensed tannin concentration (covariate) (BCT cone . )  and 
plant morphology (covariate) (Leafiness or Volume) o n  R-squares (from the 
analysis of variance of number of bites), when added to the model before 
(independent) or after plant morphology or ECT cone. had been added. 
Considering plants of birdsfo o t  trefoil in Period 2 of Experiment 5 .  
l P-value for the differences o f  R-squares. 
Experiments 5 and 6 1 69 
PERIOD 3 
ECT conc. effect before 
Leafiness 
0.8137 
0.8041 Leafiness effect before 
ECT conc. � 0.0917 (0 8205) ' 
0.8958 
leafiness effect after 
ECT conc 
0.1079 � /0.0257 (0.4;.;3":,:35;;,},--_-..,,,_-..,...--, ECT conc. effect after 
ECT conc. effect before 
Volume 
0.8137 
Volume effect after ECT 
conc. 
0.013� 
0.9216 Leafiness '-- ----_ ..... 
0.804 1 
0.8267 
Volume effect before 
ECT conc. 
� 0.0185 (0.7035) 
0.8226 
Figure 4.7. Effect of extractable condensed tannin concentration (covariate) (BCT cone.) and 
plant morphology (covariate) (Leafiness or Volume) on R-squares (from the 
analysis of variance of number of bites), when added to the model before 
(independent) or after plant morphology or ECT cone. had been added. 
Considering plants of birdsfoot trefoil in Period 3 of Experiment 5 .  
lP-value for the differences ofR-squares. 
Experiments 5 and 6 170 
4.3.2. Experiment 6: Effect of condensed tannin in birdsfoot trefoil (Lotus 
corniculatus L.) and formononetin in red clover (Trifolium pratense 
L) on diet selection and grazing behaviour of dairy cows: rumen 
content modification approach 
4.3.2.1. Plant characteristics 
Comparisons between species 
Plant characteristics of the different genotypes of birdsfoot trefoil (BT) and red clover 
(RC), trimmed and not trimmed are presented in Table 4. 19  and Table 4.20. Table 4. 1 9  
corresponds to the first period of assessment, and Table 4.20 to the second period. Each 
table is subdivided into runs according to sets of four plant sequences (each sequence 
was formed by 26 plants). In each run, one rumen chemical modification was tested: 
Run 1 in Period 1 and Run 2 in Period 2 - formononetin was tested; Run 2 in Period 1 
and Run 1 in Period 2 - tannin was tested (see rumen modification procedure section 
4.2.6.2 and Appendix 4. 1 0). 
Species, genotype and trimming effects influenced the difference in plant morphological 
characteristics. There were significant interactions involving plant species, genotypes 
and trimming effects. On average RC plants had greater proportion of leaves and 
density than BT, and untrimmed had greater size than trimmed plants in most plant 
characteristics. Comparing all experimental plants, BT had the greatest variation 
between genotypes and between trimmed and un trimmed plants. Untrimmed plants of 
cultivar Goldie generally had the greatest height, area and herbage mass and were the 
most erect. On the other hand, in most runs accession PI273938 had the smallest height, 
herbage mass, proportion of leaves, and density. 
The amount of height, herbage mass, area and leafiness removed also varied according 
to species, genotype and trimming effects. Red clover had, in most runs ,  greater 
Experiments 5 and 6 1 7 1  
reduction in leafiness than BT. Untrimmed cultivar Goldie had, in most runs, the 
largest reduction of height, herbage mass and area, and accession PI273938 had one of 
the lowest reductions in height. More detailed analyses comparing genotypes within 
each species are presented in the next section. 
Table 4. 1 9. Characteristics of untrimmed (Ntrim) and trimmed (Trim) plants of 
birdsfoot trefoil Goldie (Low Tannin) and accession PI273938 (High 
Tannin), and red clover cultivars G27 (Low Form) and Pawera (High Form) 
in Period 1 ,  Runs 1 and 2 (rumen content modified with formononetin and 
tannin, respectively) . 
Birdsjoot trejoil(BT) Red clover(RC) SEvI 
Low Tannin High Tannin Low Form High Form 
Ntrim Trim Ntrim Trim Ntrim Trim Ntrim Trim 
Height (cm) 
pre-grazing 26.2 16.5 
removed2 5.2 1 .9 
Plant mass (gDMlm2) 
pre-grazing 4 1 9  297 
removed 77 1 1  
14.9 1 4.4 2 1 .4 17.7 23.3 2 1 .0 1 .8 85 
1 . 2  0.8 2.6 2 . 8  4.0 4.2 1 .6 1 8  
248 274 324 295 3 3 3  3 1 1  2 1 .2 
-2 28 35 1 6  1 4  29 29.0 
.... Plant area (cm2) 
Z pre-grazing 
� removed 
Leafiness (%)  
pre-grazing 
removed 
Habit 
3 1 1 0 1 570 2430 1 6 1 0  3640 3030 3780 2850 499. 6  
7 1 0  1 90 4 1 0  1 40 8 3 0  3 10 990 8 3 0  33 1 .5 
40 36 32 3 2  74 74 74 72 1 .83 
7 2 2 1 25 22 2 3  26 2.53 
pre-grazing 
Density 
pre-grazing 
Height (cm) 
3 .5 2 . 8  2.3 2.6 2.8 2.6 2.7 2.8 0.242 
5.9 5 . 1  4.4 4.5 7.3 7.2 7.3 7 . 1 0.334 
pre-grazing 30.4 1 7 .9 1 5.7 1 5 . 2  25.0 2 1 .9 25. 9  2 1 .0 2.558 
removed 7.5 3.0 1 . 8  2.2 3.2 2.2 1 .7 1 . 3  1 .479 
Plant mass (gDMlm2) 
pre-grazing 400 284 232 230 383 352 376 343 25.2 
removed 1 06 34 23 32 54 27 29 2 1  1 9 .9 
N Plant area (cm
2) 
Z pre-grazing 
;;J removed 
50 1 0  2750 3340 2620 5500 4 8 1 0  6880 4640 68 1 .2 
� 940 270 520 2 1 0  -30 1 80 - 1 60 -4 1 0  3 84 . 1  Leafiness (%  ) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
38 
3 
3.4 
34 
o 
2.5 
32 
3 
2.2 
3 1  
2 
2.2 
74 
7 
2.5 
pre-grazing 6.2 5 . 1  4.7 4.4 7.9 
i SED - Standard error for differences of means 
2 Removed ::: pre-grazing minus post-grazing assessment 
7 1  
6 
2.4 
7.6 
73 
5 
2.7 
7.9 
7 2  1 .94 
5 2.30 
2.2 0.367 
7.7 0.352 
Experiments 5 and 6 172 
Table 4.20. Characteristics of un trimmed (Ntrim) and trimmed (Trim) plants of 
birdsfoot trefoil Goldie (Low Tannin) and accession PI273938 (High 
Tannin), and red clover cultivars G27 (Low Form) and Pawera (High Form) 
in Experiment 6 Period 2, Runs 1 and 2 (rumen content modified with 
tannin and formononetin, res2ectivel�). 
Birds/oot tre/oil(BT) 
Low Tannin High Tannin 
Ntrim Trim Ntrim Trim 
Height (cm) 
pre-grazing 1 5 .5 9.7 1 2  9.7 
removed2 4.0 1 .0 0.3 0.2 
Plant mass (gDMlm2) 
pre-grazing 392 284 228 232 
removed 1 1 3 1 7  1 3 1  
..... Plant area (cm2) 
Z pre-grazing 2060 1 070 27 1 0  1 070 
;;;;J removed 560 200 320 1 00 � 
Leafiness (% ) 
pre-grazing 42 39 32 3 1  
removed 8 5 2 2 
Habit 
pre-grazing 3 . 1  2 .7 2. 1 2.6 
Density 
£re-�azing 5 . 8  5.6 4.5 3.9 
Height (cm) 
pre-grazing 1 6.4 1 1 .4 1 1 .5 1 1 .4 
removed 4.5 1 .3 0.9 0.4 
Plant mass (gDMlm2) 
pre-grazing 386 328 255 242 
removed l l O 57 29 24 
N Plant area (cm
2) 
Z pre-grazing 4930 1 390 2050 1 380 
;;;;J removed 1 770 250 280 1 60 � Leafiness (%  ) 
pre-grazing 45 3 8  3 1  3 1  
removed 1 3  5 2 2 
Habit 
pre-grazing 2.6 2.8 2.4 2.7 
Density 
3 . 8  £re-t[azin� 6.4 5.6 4.0 
t SED - Standard error for differences of means 
2 Removed :::::: pre-grazing minus post-grazing assessment 
Comparisons between genotypes within each species 
BIRDSFOOT TREFOIL 
Red clover(RC) 
Low Form High Form SED} 
Ntrim Trim Ntrim Trim 
1 1 . 3  1 0.5 1 4 . 1 2  1 0.5 1 .466 
0.9 1 .3 1 .6 0.4 0.76 1 
3 2 1  338 322 306 30.4 
30 4 1  40 25 25.9 
1 040 980 1 630 980 359.0 
10 30 90 0 1 7 7 .4 
74 69 67 67 2.92 
1 0  7 7 5 3 .06 
2.7 2.7 2.8 2.7 0.259 
7.2 7.0 6.6 6.7 0.566 
1 7.0 1 5 .0 1 8.4 1 5. 0  1 .377 
3 .5 1 .8 2.7 2.3 0.624 
4 1 7  379 387 367 24. 1 
I l 5 49 6 2  6 7  2 2 . 2  
2930 1 920 3250 1 920 569.0 
230 1 20 370 1 00  262.3 
75 70 67 67 2 . 69 
22 1 3  1 3  1 3  3.4 1 
2.7 2.9 2.7 2.8 0.278 
7.7 7.3 7.0 6.9 0.3 1 1  
The characteristics of birdsfoot trefoil plants in the different periods and runs, according 
to trimming and genotype (secondary compound concentration) effects are presented in 
Experiments 5 and 6 1 73 
Table 4.2 1 and 4.22. Comparing the genotypes of birds foot trefoil (BT) in Runs 1 and 2 
of Period 1 ,  there was a significant interaction between trimming and genotype effects 
for height, mass and habit. Untrimmed Goldie was significantly taller with more mass 
and a more erect habit than the other plants in both runs. In Run 1 untrimmed plants had 
a significantly greater area and accession PI273938 a significantly lower percentage of 
leaf and density. In Run 2, untrimmed plants had more area, leafiness and were denser, 
and cultivar Goldie had significantly more area, leafiness and density than accession 
PI273938 .  
In Run 1 of Period 1 there was a significant interaction between trimming and genotype 
effects in relation to reduction of area and mass, and a significant trimming effect for 
reduction of leafiness. Untrimmed cultivar Goldie had significantly the greatest area and 
mass removed, and trimmed plants the least leafiness reduction. In Run 2 there were 
significant interactions between trimming and genotype effects in relation to reduction 
of height and mass, and a significant trimming effect for area removed. Un trimmed 
cultivar Goldie had significantly the greatest height and mass removed, and untrimmed 
plants had greater reduction of area than trimmed plants. 
In Period 2 there was, in Run 1 ,  a significant trimming effect for height and area; a 
significant genotype effect for percentage of leaves, habit and density; and a significant 
interaction between trimming and genotype for mass. In this case untrimmed plants 
were taller and had more area. Cultivar Goldie had greater percentage of leaves, density 
and were more erect than accession PI273938 ,  and untrimmed cultivar Goldie had 
significantly more mass. In the last run there were significant interactions between 
trimming and genotype effects in relation to height, area and leafiness. There was also a 
significant main effect of trimming for mass, and a main effect of genotype for mass 
and density. Untrimmed cultivar Goldie had (P<O.05) greater height, area and 
percentage of leaves. Untrimmed plants had significantly (P<O.05) more mass, and 
cultivar Goldie had more mass and were denser than accession PI273938. 
Experiments 5 and 6 174 
Table 4.2 1 .  Characteristics of birdsfoot trefoil plants before grazing and difference 
between before and after grazing (removed from height, mass, area and 
leafiness) according to trimming (Ntrim == untrirnrned; Trim = trimmed) and 
genotype [secondary compound concentration (Sec. Comp. Conc.) :  high 
(accession PI273938), low (cultivar Goldie)] effects in Period 1 ,  Runs 1 and 
2 of Experiment 6. 
Birdsfoot trefoil (BT) 
Trimming Sec comp conc. 
Ntrim Trim P-value2 High Low P-value2 SEDl 
Height (cm) 
pre-grazing 20.6 
removed3 3 . 2  
Plant mass (gDMlm2) 
pre-grazing 333 
removed 3 8  
... Plant area (cm2) 
� pre-grazing � removed 
Leafiness (% ) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
pre-grazing 
Height (cm) 
2770 
560 
36 
5 
2.9 
5 . 1  
pre-grazing 23.0 
N 
removed 4.6 
Plant mass (gDMlm2) 
pre-grazing 3 1 6  
removed 65 
Plant area (cm2) 
� pre-grazing � removed 
4 1 80 
730 
Leafiness (% ) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
35 
3 
2.8 
1 5 .4 0.0001 14.6 
1 .3 0. 1 942 1 .0 
286 0.0205 2 6 1  
1 9  0.3690 1 3  
1 590 0.0001 2020 
1 70 0. 1433 280 
34 0. 1 1 93 32 
1 0.0533 2 
2.7 0.2756 2.5 
4.8 0. 1 390 4.5 
1 6.6 0.00 1 2  1 5 .4 
2.6 0.0582 2.0 
257 0.0029 23 1 
33 0.0503 28 
2690 0.004 2980 
240 0.2 1 9  370 
32 
1 
2.4 
0.0337 
0.0932 
0.0807 
3 1  
2 
2.2 
2 1 .3 0.000 1 1 .056 
3.6 0.0769 1 .4 1 5  
358 0.0001 1 9.6 
44 0. 1 325 20. 1 
2340 0 . 1 800 2 3 3 .4 
450 0.5037 2 6 1 .7 
3 8  0.0002 1 .43 
4 0.09 1 8  1 .5 6  
3 .2 0.00 1 0  0 . 1 5 1  
5.5 0.0001 0.282 
24. 1 0.0001 1 .828 
5.3 0.0032 1 . 1 05 
342 0.0001 1 8. 3  
70 0.0 1 03 1 5. 5  
3880 0.024 1 3 80.9 
6 1 0  0.25 3 1 205. 9  
36 
2 
3.0 
0.0001 1 . 1 3  
0.73 1 7  1 .20 
0.0027 0.23 1 
pre-grazing 5.4 4.8 0.0140 4.5 5.7 0.0001 0.257 
I SED _ Standard error for differences of means when comparing means with the same level of 
treatment (trimming or sec. comp. conc.) 
2 P-value of the treatment (trimming or sec.comp. conc.) main effect. 
3 Removed :::: pre-grazing minus post-grazing assessment 
Experiments 5 and 6 1 75 
Table 4.22. Characteristics of birds foot trefoil plants before grazing and difference 
between before and after grazing estimates (removed) of height, mass, area 
and leafiness according to trimming (Ntrim :: untrimmed; Trim :: trimmed) 
and genotype [secondary compound concentration (Sec. Comp. Conc.): 
High (accession PI273938), Low (cultivar Goldie)] effects in Period 2 ,  Runs 
1 and 2 of Experiment 6. 
Birdsfoot trefoil (BT) 
Trimming Sec comp conc. 
Ntrim Trim P-value2 High Low P-value2 SEDl 
Height (cm) 
pre-grazing 1 3 .8 
removed3 2. 1 
Plant mass(gDMlm2) 
pre-grazing 3 1 0  
removed 57 
.... Plant area (cm2) 
Z pre-grazing 2380 � removed 440 
Leaflness (%  ) 
pre-grazing 37 
removed 5 
Habit 
pre-grazing 2.7 
Density 
pre-grazing 5.2 
Height (cm) 
pre-grazing 14.0 
removed 2.7 
Plant mass (gDMlm2) 
pre-grazing 3 2 1  
removed 70 
N Plant area (cm
2) 
Z pre-grazing 3490 � removed 1 020 
Leaflness (%)  
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
3 8  
7 
2.5 
9.7 0.00 1 1 1 0.9 
0.6 0.0224 0.2 
258 0.0399 230 
24 0. 1 1 83 1 6  
1 070 0.0001 1 890 
ISO 0. 1 085 2 1 0  
35 0.2268 32 
4 0.3390 2 
2.6 0.8208 2.3 
4.7 0.2984 4.2 
1 1 .4 0.003 1 1 1 .5 
0.9 0.0030 0.7 
285 0.0327 249 
40 0.0549 27 
1 380 0.0001 1 7 1 0  
200 0.0020 220 
35 
4 
2.7 
0.0482 3 1  
0.0062 2 
0. 1 504 2.54 
1 2.6 0. 1 24 1  1 . 1 22 
2.5 0.00 1 8  0.66 1 
338 0.000 1 24.2 
65 0.0232 20.7 
1 560 0.2845 300.8 
380 0. 3454 1 77.2 
4 1  0.0001 1 .69 
6 0.0209 1 .72 
2.9 0.0055 0. 1 82 
5.7 0.0006 0.394 
1 3 . 9  0.0053 0.809 
2 .9  0.0004 0.5 7 1  
357 0.0001 1 6.0 
83 0.0006 1 4 . 8  
3 1 60 0.0054 484.6 
1 0 1 0  0.0027 243 . 1 
4 1  0.000 1 1 .73 
9 0.000 1 1 .35 
2.70 0.3334 0. 1 7  
pre-grazing 5.2 4.7 0. 1 649 3.92 6.00 0.0001 0.35 
i SED - Standard error for differences of means when comparing means with the same level of 
treatment (trimming or sec. comp. conc.). 
2P-value of the treatment (trimming or sec. comp. cone.) main effect. 
3 Removed == pre-grazing minus post-grazing assessment 
In Run 1 of Period 2, there were significant interactions between trimming and 
genotype for the amount of height and mass removed by grazing. Cultivar Goldie 
untrimmed had the greatest reduction in height and mass. There was also a significantly 
larger reduction in leafiness, independent of trimming effect, of cultivar Goldie. In Run 
Experiments 5 and 6 176 
2 of Period 2 there were significant interactions between trimming and genotype effect 
in relation to reduction in height, area and leafiness. There was also a significant main 
effect of genotype for mass. Untrimmed cultivar Goldie had significantly (P<O.05) 
greater reduction of height, area and percentage of leaves. Cultivar Goldie had, 
independent of trimming effect, significantly (P<O.05) larger reduction of mass than 
accession PI273938. 
RED CLOVER 
The red clover plant characteristics in different periods and runs, according to trimming 
and genotype (secondary compound concentration) effects, are presented in Table 4.23 
and Table 4.24. In Run 1 of Period 1 ,  there were no significant differences in plant 
characteristics and in the amount removed of each plant characteristic by grazing. In 
Run 2 there was a significant trimming effect only on height and area. Untrimmed 
plants had larger areas and were taller than trimmed plants. In this run there were no 
significant effects of either trimming or genotype for the amount removed of any plant 
characteristic. 
In the Period 2 there were more significant differences between trimmed and untrimmed 
plants and between cultivars G-27 and Pawera. In Run 1 there was a significant 
interaction between trimming and genotype effects for area. Un trimmed plants of 
cultivar Pawera had the largest area. Untrimmed plants also were significantly taller 
than the trimmed ones. In the last run ,  untrimmed plants were significantly taller and 
with more area, and cultivar G-27 had a higher percentage of leaves than cultivar 
Pawera, independent of the trimming effect. 
In Run 1 of Period 2, there were significant interactions between trimming and 
genotype for the amount removed by grazing of height and area. Trimmed plants of 
Pawera had the lowest reduction of area and untrimmed plants of Pawera had the least 
reduction of height. The amount of mass removed by grazing in Run 2 was also affected 
by an interaction between trimming and genotype effects. Untrimmed G-27 had the 
greatest reduction in mass and G-27 trimmed had the lowest reduction.  
Experiments 5 and 6 1 77 
Table 4. 23. Characteristics of red clover plants before grazing and difference between 
before and after grazing estimates (removed from height, mass, area and 
leafiness) according to trimming (Ntrim = untrimmed; Trim = trimmed) and 
genotype [secondary compound concentration (Sec. Comp. Cone.) :  High 
(cultivar Pawera) and Low forrnononetin (cultivar G-27)] effects in Period 
1 ,  Runs 1 and 2 of Experiment 6.  
Red clover (RC) 
Trimming Sec comp conc. 
Ntrim Trim P.value2 High Low P.value2 SEDl 
Height (cm) 
pre-grazing 22.3 
removed3 3 . 3  
Plant mass (gDMlm2) 
pre-grazing 352 
removed 24 
..... Plant area (cm2) 
Z pre-grazing ; removed 
Leafiness (%  ) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
pre-grazing 
Height (cm) 
37 1 0  
9 1 0  
74 
24 
7 .3 
2 . 8  
pre-grazing 25 .5 
removed 2.5 
Plant mass (gDMlm2) 
pre-grazing 379 
removed 42 
('<I Plant area (cm2) 
Z pre-grazing 
;:! removed � Leafiness (%) 
pre-grazing 
removed 
Habit 
pre-grazing 
Density 
6 1 90 
- 1 00 
74 
6 
2 . 6  
1 9.3 0.0579 22. 1 
3 . 5  0.8 1 2 1  4 . 1 
326 0. 1 075 343 
22 0 . 8999 2 1  
2940 0.074 1 3320 
570 0. 1 1 43 9 1 0  
73 0.3 1 48 73 
24 1 .0000 25 
7 .2 0.5800 2 . 8  
2.7 0.5333 7.2 
2 1 .5 0.0340 23.4 
1 . 8  0.5 1 28 1 .5 
348 0.0679 360 
24 0.6440 250 
4720 0.0 1 1 6  5760 
- 1 1 0  0.9534 -290 
72 0.255 1 73 
5 0.6200 5 
2.3 0.2594 2.5 
1 9.5 0. 1 00 1  1 .527 
2.7 0. 1 539 0.956 
335 0.6090 1 6 . 1 
25 0.7846 1 4. 0  
3330 0.9697 4 1 7.0 
570 0. 1 1 44 2 1 0.4 
74 0.3 1 48 1 .224 
24 0.4895 1 .7 8 8  
2.7 0.7395 0. 1 86 
7.3 0.7080 0. 1 65 
23.5 0.98 1 8  1 .809 
2.7 0.2453 1 .039 
367 0. 1 426 1 6. 8  
400 0.2024 1 1 . 9  
5 1 60 0.2799 550. 1 
80 0.2 1 55 284.3 
72 0.8984 1 .6 1  
6 0. 1 426 1 .25 
2.5 1 .0000 0.29 
pre-grazing 7.9 7.7 0.3394 7.8 7.8 0.9303 0.24 (SED - Standard error for differences of means when comparing means with the same level of 
treatment (trimming or sec. comp. conc. ). 
2 P-value of the treatment (trimming or sec. comp. conc.) main effect. 
3 Removed = pre-grazing minus post-grazing assessment 
Experiments 5 and 6 178 
Table 4.24. Characteristics of red clover plants before grazing and difference between 
before and after grazing (removed from height, mass, area and leafiness) 
according to trimming (Ntrim = untrimmed; Trim = trimmed) and genotype 
[secondary compound concentration (Sec. Comp. Cone.) :  High (cultivar 
Pawera), Low formononetin (cultivar G-27)] effects in Period 2, Runs 1 and 
2 of EXEeriment 6. 
Red clover (RC) 
Trimming Sec comp conc. 
Ntrim Trim P-value2 High Low P-value2 SED
! 
Height (cm) 
pre-grazing 12 .7 10.5 0.01 6  1 2.3 1 0.9 0. 1 302 0.886 
removed3 1 .3 0.9 0.3 1 76 1 .0 1 . 1  0.79 1 1 0.390 
Plant mass (gDMlm2) 
pre-grazing 32 1 322 0.9544 3 1 4  329 0.3733 1 6.9 
removed 35 33 0.8809 33 36 0.8584 1 5.3 
..... Plant area (cm
2) 
Z pre-grazing 1 340 970 0.01 22 1 3 1 0  10 10  0.0336 1 33.4 
;;;J removed 50 1 5  0. 1 284 50 20 0.2325 22.3 1 � 
Leafiness (%  ) 
pre-grazing 7 1  68 0.3021 67 7 1  0.0793 2. 1 9  
removed 9 6 0.2737 6 8 0.3604 2.25 
Habit 
pre-grazing 2 .7 2.7 1 .0000 2.8 2.7 0.6465 0. 1 80 
Density 
Ere-�azing 6.9 6.8 0.8 153 6.6 7 . 1  0. 1 670 0.354 
Height (cm) 
pre-grazing 1 7.7 15 .0 0.0044 1 6.7 1 6.0 0.4360 0.898 
removed 3. 1 2.0 0. 1 140 2.5 2.6 0.7967 0.642 
Plant mass (gDMlm2) 
pre-grazing 402 373 0. 1 179 377 397 0.2592 1 79.9 
removed 89 58 0.0655 65 82 0.2790 16.2 
N Plant area (cm
z) 
Z pre-grazing 3090 1 920 0.0001 2590 2420 0.554 1 27 1 .3 
;;;J removed 300 1 10 0.001 8  230 170 0.2887 55.7 � 
Leafiness (%  ) 
pre-grazing 7 1  68 0. 1 223 67 72 0.0083 1 .7 1  
removed 1 7  1 3  0. 1 375 1 3  1 7  0. 1 375 2.87 
Habit 
pre-grazing 2.7 2.9 0.4069 2.8 2.8 1 .0000 0. 198  
Density 
Ere-�azins 7.4 7 . 1  0.2472 7.0 7 .5 0.0096 0.2 1 2  
j SED - Standard error for differences of means when comparing means with the same level of 
treatment (trimming or sec. comp. conc.). 
2 P-value of the treatment (trimming or sec. comp. conc.) main effect. 
3 Removed = pre-grazing minus post-grazing assessment 
Experiments 5 and 6 179 
4.3.2.2. Sward chemical composition 
Extractable condensed tannin (ECT) 
The averages of ECT concentration according to genotype, period and trimming effects 
are given in Table 4.2S. There was no significant (P>O.OS) interaction between genotype 
and period, and there was no significant difference in ECT concentration between runs. 
Cultivar Goldie had significantly lower concentration of ECT than accession PI273938 
(Table 4.2S). In proportional terms, accession PI273938 had on average 3.5 times more 
tannin than cultivar Goldie. Trimming did not affect significantly (P>O.OS) the 
percentage of ECT (Table 4.25). 
Table 4.25.Extractable condensed tannin (ECT) concentration (%DM) of birdsfoot 
trefoil genotype main effect and interactions with period and trimming 
effects (untrimmed: Ntrim; Trimmed: Trim) of Experiment 6.  
Genotype 
2 
Period 
3 
Ntrim 
Trimming 
Trim 
Goldie P1273938 SED] P-valui 
0.43 1 .52 0 .071 0 .0001 
0.32 
0.53 
1 .42 
1 .62 
1 .49 
1 .55 
0.35 
O.S I 
0. 1 39 0.4677 
0. 1 39 0.0734 
1 SED - Standard error for differences of means 
2 P-value of genotype main effect and interactions : period *genotype or 
trimming*genotype. 
Number of observation contributing for each mean of the genotype main 
effect (n=32) and interactions (n=1 6) 
F ormononetin concentration 
There was also a strong effect of genotype in relation to formononetin content. Cultivar 
Pawera had significantly higher concentration of formononetin than G-27. In 
proportional terms, overall Pawera had 2.3 times more formononetin than G-27. There 
was no effect of trimming and no interactions with different periods. The averages of 
formononetin content according to genotype, period and trimming effects are given in 
Table 4.26. 
Experiments 5 and 6 1 80 
Table 4. 26. Formononetin concentration (%DM) of red clover genotype main effect 
and interactions with period and trimming effects (untrimmed: Ntrim; 
Trimmed: Trim) of Experiment 6. 
G-27 Pawera P-valui 
Genotype 0.26 0.61 0.017  0.0001 
2 0.23 0.60 
Period 0 .024 0. 1434 
3 0.29 0.62 
Ntrim 0.26 0.59 
Trimming 0.024 0.3355 
Trim 0.26 0.63 
ISBD - Standard error for differences of means 
2 P-value of genotype main effect and interactions : period *genotype or 
trimming*genotype. 
Number of observation contributing for each mean of the genotype main 
effect (n=32) and interactions (n=16) 
General chemical composition 
The general chemical composition of each genotype is presented in Table 4.27. There 
were significant differences between genotypes in relation to the percentage of protein 
and neutral detergent fibre (NDF). The accession PI273938 had the lowest (P<0.05) 
proportion of protein and NDF compared to the other genotypes. However most of the 
component concentration differences were explained by differences in species main 
effect. Red clover had significantly (P<0.05) higher percentage of acid detergent fibre 
(ADF) and ash. Birdsfoot trefoil had significantly (P<0.05) higher concentration of lipid 
and in vitro dry matter digestibility (IVDMD). There was a significant effect of 
trimming for CHO (soluble sugars plus starch) content. Untrirnrned plants had a 
significantly (P<O.05) higher concentration of CHO. 
Experiments 5 and 6 1 8 1  
Table 4.27. Percentage of dry matter (DM), crude protein (CP), lipid, acid and neutral 
detergent fibre (ADF, NDF), carbohydrates (soluble sugars plus 
starch)(CHO), ash and in vitro dry matter digestibility (IVDMD) determined 
by Near Infrared Reflectance Spectroscopy (NIRS) of plants of birdsfoot 
trefoil and red clover of Experiment 6 (percentage of DM basis). 
Protein 
Lipid 
ADF 
NDF 
eRa 
Ash 
IVDMD 
Birdsjoot Red clover p-trefoil SED1 SED2 value3 
Goldie PI273938 G-27 Pawera 
22.0 
3 .3  
2 1 .5 
28.2 
14.2 
8 .4 
77.7 
17.5 
3.2 
23. 1 
2 1 . 8  
16.4 
8 .5  
76.0 
23 . 1  
2.7 
25.3 
27.4 
14.6 
10.2 
74.6 
23.0 
2.6 
25.7 
29.2 
14.9 
10 . 1 
74.3 
0.736 
0. 1 04 
1 .221 
2.095 
1 .049 
0.359 
1 . 1 02 
0.521 0.0001 
0.073 0.2675 
0.863 0.4092 
1 .482 0.0169 
0.742 0. 1 303 
0.254 0.9527 
0.779 0.3265 
ISED 
- standard error for differences of means when comparing means between genotypes. 
2SED - standard error for differences of means when comparing means of species. 
3P-value of genotypes within each species. 
Number of observation contributing for each mean (n=6). 
4.3.2.3. Number of bites per plant 
The results of the number of bites taken from each plant are shown in Tables 4.28, 4 .29, 
4.30, and 4.3 1 .  Comparing all the treatments in the two different periods, there were no 
significant interactions between period and treatment [type of chemical added into the 
rumen, rumen chemical concentration (within each type of chemical) ,  plant species, 
trimming and plant secondary compound concentration effects] (P>0.05) in relation to 
number of bites. However there was a significant interaction between trimming and 
secondary compound effects (Table 4.28) and between type of chemical inserted into 
the rumen and plant species (Table 4.29) . 
The interaction between trimming and plant secondary compound effects show that the 
highest number of bites was taken from un trimmed plants of cultivar Goldie (low ECT 
concentration) and untrimmed plants of both red clover cultivars (Table 4.28). 
Experiments 5 and 6 1 82 
Table 4.28. Average of number of bites in birdsfoot trefoil (BT) and red clover (RC), in 
relation to secondary compound concentration (High and Low) and 
trimming (Ntrim = untrimmed plants Trim = trimmed plants) effect in 
Ex.e,:riment 6. 
Birdsfoot trefoil Red clover 
High Low High Low 
Ntrim 2.5 10 .4 8 .2 8 .0 
Trim 2. 1 3 . 1  4.9 4.7 
SED! 1 . 1 12 P-value2 0.0001 
I SED - Standard error for differences of means 
2P-value of the interaction: secondary compound concentration*trimming effect. 
The interaction between type of material added into the rumen and plant species shows 
a greater number of bites in red clover when formononetin (red clover material) was 
added to the rumen, independent of the rumen chemical concentration (within each type 
of chemical) and plant secondary compound concentration (Table 4 .29). There was no 
significant effect of rumen chemical concentration (within each type of chemical) on 
number of bites. 
Table 4.29. Average of number of bites in birdsfoot trefoil (BT) and red clover (RC) 
plants in relation to rumen chemical modification (Tannin and 
Formononetin) effect in Experiment 6 .  
Type of rumen chemical 
modification 
Tannin 
Formononetin 
SED] 0.786 
BT RC 
4.4 
4.7 
4. 1 
8 .8  
P-value2 0 .0001 
I SED - Standard error for differences of means 
2P-value of the interaction specie*rumem chemical modification effect. 
In order to clarify the effect of rumen concentrations (within each type of chemical 
modification: either tannin or formononetin) on number of bites, separate statistical 
analyses were performed for the two types of rumen chemical modifications (either 
tannin or formononetin) over both periods . There was no significant effect of rumen 
chemical concentration (within each type of rumen chemical) and no significant 
interaction with treatments (species, trimming and plant secondary compound effects). 
Experiments 5 and 6 1 83 
Because there was a significant interaction between type of rumen chemical and plant 
species (Table 4.29), individual analyses considering only one type of rumen chemical 
with one specific species were carried out. The results of these analyses are presented in 
Table 4.30 and Table 4.3 1 .  All analyses show that there was no significant effect of 
rumen concentration or interaction with period effect on number of bites (P>0.05). 
The number of bites taken from birdsfoot trefoil plants when the cows had either tannin 
or formononetin in the rumen were affected by a significant interaction between 
trimming and plant secondary compound concentration (Table 4.30.). Plants with low 
concentration of ECT and untrimmed had significantly higher number of bites, 
independent of the rumen chemical modification. The accession PI273938 (high 
concentration of ECT) also had the lowest number of bites in both rumen modification 
cases. 
Table 4.30.A verage of number of bites in birdsfoot trefoil (BT) in relation to plant 
secondary compound concentration (Plant Sec. Comp. Conc.: High and Low 
tannin concentration), plant trimming characteristic (Ntrim = untrirnrned; 
Trim = trimmed plants), type of rumen chemical modification [Tannin and 
Formononetin (Form.)] and rumen concentration effect (within each type of 
rumen chemical modification) in Experiment 6. 
Plant Sec. Camp. Canc. 
Low High SED] P -value2 
NTrim Trim NTrim Trim 
Type of rumen chemical 
Tannin 9.4 3 . 1  2.9 2.3 
Form. 1 1 .5 3.2 2.2 1 .9 
Rumen concentration effect 
Tannin Low 8.3 3.0 3.2 2.6 High 10.4 3 .2 2.6 2 . 1 
Form. Low 12.6 2. 1 0.4 1 . 8  
1 .24 1 0.00 1 8  
1 .629 0.00 1 0  
1 .755 0.5869 
2.303 0.0942 High 10.3 4.2 4.0 2.0 
ISED - Standard error for differences of means when comparing means with the same level of type o f  
rumen chemical modification. 
2P-value of the interaction plant sec. comp. conc. *trimming, or plant sec. comp. 
conc . *trimming*rumen chemical concentration. 
In relation to number of bites in red clover when modifying the rumen content either 
with tannin or formononetin, there was a significant trimming main effect in both 
Experiments 5 and 6 1 84 
situations (Table 4.3 1 ). In this case untrimmed plants had a significantly higher number 
of bites than trimmed ones. There was also a marginally significant interaction when 
modifying the rumen with tannin between period and trimming effects. There was a 
larger number of bites in the first (P 1 )  than in the second (P2) period with larger number 
of bites for un trimmed plants (PI :  7.9 vs 3.7 bites/plant; P2: 3.0 vs 1 . 8  bites/plant, SED 
1 .076, P==0.0502). 
Table 4.3 1 .  Average of number of bites in red clover in relation to type of rumen 
chemical modification [Tannin and Formononetin (Form.)] and plant 
trimming characteristic (Ntrim == untrimmed plants; Trim = trimmed plants) 
and, in relation to rumen concentration effect (within each type of rumen 
chemical modification), plant secondary compound concentration (Plant 
Sec. Comp. Conc. )  and plant trimming characteristic, in Experiment 6. 
Plant Characteristics SED' P-value 
Type of rumen chemical 
Tannin 
Form. 
Rumen concentration effect 
NTrim 
5 .4 
1 0.7  
Trim 
2 .7 
6.9 
Plant Sec. Comp. Conc. 
Low High 
NTrim Trim Ntrim 
Tannin Low 3.8 2.5 4.7 High 7. 1 3.9 6. 1 
Trim 
1 .7 
2.8 
Form. Low 1 0.9  5.0 1 0. 1  6.4 
0.761 
1 .304 
1 .5219  
2 .6078 
0.0008 
0.0047 
0.6048 
0.6220 High 10.0 7.3 1 1 .7  8.7 
ISED - Standard error for differences of means when comparing means with the same level of type of 
rumen chemical modification. 
2P-value of the trimming main effect or interaction plant sec. comp. conc. *trimming*rumen chemical 
concentration (within each type of rumen chemical modification). 
For better understanding of the animals preferential grazing, correlation and covariance 
analyses were performed using number of bites per plant. These analyses are presented 
in section 4.3.2.5 and 4.3 .2.6, respectively. 
Experiments 5 and 6 1 85 
4.3.2.4. Rate of biting 
An analysis was performed to verify the significant effects of interactions of period and 
rumen chemical modification, and to clarify differences between treatments (species, 
trimming and plant secondary compound effects) in rate of biting. There were no 
significant interactions between period and treatment (species ,  trimming and plant 
secondary compound effects). However there was a significant interaction between 
trimming and species effects, and a marginally significant difference between the four 
different genotypes (independent of trimming effect) (Table 4.32). Untrimmed plants of 
birdsfoot trefoil had significantly the lowest rate of biting, independent of ECT 
concentration effect. On the other hand, the low ECT concentration genotype (cultivar 
Goldie) had the lowest rate of biting, independent of trimming effect. There was no 
significant difference between RC plants in rate of biting. 
Table 4.32. Average of number of bites per minute (rate of biting) in birdsfoot trefoil 
and red clover in relation to plant trimming characteristic (Ntrim = 
untrimmed plants; Trim = trimmed plants) and plant secondary compound 
concentration (Sec. Corn . Conc . :  Hi h a.T1d Low) in Ex eriment 6. 
Trimming 
Ntrim 
Trim 
SED! 
P-value3 
Sec. Comp. Cone. 
High 
Low 
SEDl 
P-value3 
Birdsfoot trefoil 
28. 1 
34.2 
2.5 1 
0.0285 
34.0 
28.3 
2.50 
0.058 1 
Red clover 
36.0 
35.4 
2 . 14 
35 .5 
35.8 
2. 1 1  
2.2 1 
2 .46 
2.48 
2. 1 3  
ISED - Standard error for differences of means when comparing means with the same level of 
plant specie 
2SED - Standard error for differences of means when comparing means with the same level of 
treatment (trimming or secondary compound concentration) 
3P-value of the interaction treatment (trimming or secondary compound concentration)*specie 
effect. 
Experiments 5 and 6 1 86 
4.3.2.5. Correlation Analyses 
Correlation analyses were perfonned to investigate the relationships between general 
plant chemical characteristic, secondary compound concentration, number of bites and 
plant morphology. As in Experiment 5 (see section 4.3. 1 .5 .) ,  four correlation analyses 
were carried out: the correlations between number of bites and general plant chemical 
characteristics ;  secondary compounds (ECT and fonnononetin concentration) and either 
plant morphology or number of bites; and number of bites and plant morphology. The 
number of observations in each correlation was related to the number of samples bulked 
for the chemical analysis. In this case the average of the number of bites was used. The 
correlation between number of bites and plant morphology used all experimental plants. 
As in Experiment 5 ,  correlation analyses together with covariance analyses were used to 
separate the effects of plant morphology and ECT concentration on number of bites. 
The full correlation coefficient matrices of each analysis is presented in Appendices 4.7, 
4.8 and 4.9. 
Because no significant interaction between rumen chemical with plant chemical 
characteristics was found for number of bites, the correlation analyses were perfonned 
within each period. In this way the effect of rumen manipulation was balanced within 
period and comparisons with Experiment 5 were possible. 
Number of bites vs General plant chemical characteristics 
The correlation coefficients and the probabilities of significance between general plant 
chemical characteristic and number of bites taken from birdsfoot trefoil and red clover 
plants are shown in Table 4.33. The number of bites in birdsfoot trefoil did not show 
significant correlation with any general plant chemical characteristic (P>0. 1 ) .  However 
the number of bites in red clover plants had significant (P<0.05) negative correlation 
with acid detergent fibre and ash, and positive correlation with carbohydrates (soluble 
sugars plus starch) and in vitro dry matter digestibility. 
Experiments 5 and 6 1 87 
Table 4.33.  Pearson Correlation coefficients (r) from correlation analysis between 
number of bites per plant and percentage of protein, lipid, acid and neutral 
detergent fibre (ADF, NDF), carbohydrates (soluble sugars plus 
starch)(CHO), ash and in vitro dry matter digestibility (IVDMD) of plants 
of birdsfoot trefoil and red clover of Experiment 6 (percentage of DM 
basis). 
Protein LieJd ADF NDF CHO Ash lVDMD 
Birdsfoot trefoil 
r 0.3970 0.2923 -0.3385 0. 1 25 1  0 . 1025 0.0252 0.3962 
P-value 0. 1589 0.2719 0. 1996 0.6444 0. 7056 0. 9262 0.1287 
Red clover 
r -0.4169 -0.423 1 -0.6093 -0.3043 0.5849 -0.5567 0.5982 
P-value 0. 1082 0.1025 0.0122 0.2518 0.0173 0.0251 0.0144 
Number of observation contributing for each correlation (n:::: 1 6) 
EeT concentration vs Plant morphology and Number of bites in birdsfoot trefoil plants 
The correlation analyses between ECT concentration and plant morphological 
characteristics and between ECT concentration and number of bites in Periods 1 and 2 
are shown in Table 4.34. As in Experiment 5, there were significant negative 
correlations (P<0.05) between ECT concentrations and plant height, leafiness and 
number of bites in Period 1 .  However in Period 2 there was only a significant negative 
correlation between ECT concentration and leafiness. There was no significant (P>O.05) 
correlation between ECT concentration and number of bites in Period 2. 
Table 4.34. Pearson Correlation coefficients (r) from correlation analysis between 
extractable condensed tannin concentration and plant area, height, volume, 
leafiness and number of bites per plant (N. Bites) of birds foot trefoil in 
Periods 1 and 2 of Experiment 6 .  
Period 
1 
r 
P-value 
Period r 
2 P-value 
Area Height 
-0.0925 -0.401 1 
0.6 145 
0.0705 
0.7012  
0.0229 
-0. 14 19  
0.4384 
Number of observation contributing for each correlation (n=32) 
Volume 
-0.2540 
0. 1607 
-0.003 1 
0.9865 
Leafiness 
-0.63 1 2  
0.0001 
-0.6014  
0.0003 
N.Bites 
-0.3628 
0.0413  
-0.2694 
0. 1 359 
Experiments 5 and 6 188  
Formononetin concentration vs Plant morphology and Number of bites in red clover 
plants 
The correlation coefficient between forrnononetin concentration and either plant 
morphological characteristics or number of bites in Periods 1 and 2 are presented in 
Table 4 .35 . There was only a significant correlation between forrnononetin 
concentration and percentage of leaves in Period 2. 
Table 4.35.Pearson Correlation coefficients (r) from correlation analysis between 
forrnononetin concentration and plant area, height, volume, leafiness and 
number of bites per plant (N. Bites) of red clover in Periods 1 and 2 of 
Experiment 6 .  
Area Height Volume Leafiness N.Bites 
Period r -0.0036 0.0596 0.0134 -0.0557 -0.0323 
1 P-value 0.9846 0.7458 0.9421 0. 7621 0.8608 
Period r 0.0148 -0.00 10  0.0 190 -0.4746 -0.0733 
2 P-value 0.9360 0.9955 0.9179 0.0061 0.6902 
Number of observation contributing for each correlation (n:::32) 
Number of Bites vs Plant Morphology 
BIRDSFOOT TREFOIL 
The correlation analysis between number of bites and morphological characteristics of 
birdsfoot trefoil plants in Periods 1 and 2 are given in Table 4.36. As in Experiment 5, 
there were significant (P<0.05) and positive correlations between number of bites and 
plant area, height, volume and leafiness in both periods. 
Table 4.36.Pearson Correlation coefficients (r) from correlation analysis between 
number of bites per plant and area, height, volume and leafiness of 
birdsfoot trefoil plants in Periods 1 and 2 of Experiment 6. 
Area Height Volume Leafiness 
Period r 0.6060 0.43 10  0.6374 0.6222 
1 P-value 0.0001 0.0001 0.0001 0.0001 
Period r 
2 P-value 
0.6674 
0.0001 
0.5 1 10 
0.0001 
Number of observation contributing for each correlation (n=%) 
0.6968 
0.0001 
0.7397 
0.0001 
Experiments 5 and 6 1 89 
RED CLOVER 
The correlation analysis between number of bites per plant and morphological 
characteristics of red clover plants in Periods 1 and 2 is shown in Table 4.37. In Period 
1 the number of bites was significantly correlated to plant leafiness but not to plant area, 
height and volume. In Period 2 there were significant correlations between number of 
bites and all four plant morphological characteristics. 
Table 4.37.Pearson Correlation coefficients (r) from correlation analysis between 
number of bites per plant and area, height, volume and leafiness of red 
clover plants in Periods 1 and 2 of Experiment 6.  
Area Height Volume Leafiness 
Period r 0. 1260 0.0608 0. 1 1 93 0.2476 
1 P-value 0.2214 0.5563 0.2469 0.0150 
Period r 
2 P-value 
0.7779 
0.0001 
0.6989 
0.0001 
Number of observation contributing for each correlation (n:::96) 
0.7889 
0.0001 
0.4090 
0.0001 
The correlation matrices showed that correlation coefficient results from the smaller 
number of observations used in the analyses involving ECT and Formononetin 
concentration was not substantially different from the analyses carried out with 
individual plant data (more number of observations). Small differences observed in 
correlation coefficient did not alter the significance of the relationship. 
4.3.2.6. Use of covariates 
The analyses of covariance were performed to distinguish the plant morphological 
characteristic effect from the extractable condensed tannin (ECT) concentration effect 
on number of bites in birdsfoot trefoil. As in Experiment 5, in the first analysis, 
individual morphological characteristics (leafiness or volume) data of all experimental 
plants (number of observations (n) :::: 96) were used as covariate and genotype effect 
was a class variable (described as high and low concentration of ECT). The effect of 
covariates or class variable added to the analysis of variance was assessed in terms of 
the R -square changes. The covariates were added to a basic model formed by the effect 
of day variation, individual cow variation, rumen chemical added to the rumen, rumen 
Experiments 5 and 6 190 
chemical concentration, sequence of plants, block of plants within each sequence and 
trimming. Leafiness and plant volume were used as covariates. In the second analysis, 
the number of samples was restricted to the untrimmed plants (n=48) to verify the 
effects of the natural variation. In this case the trimming effect was not included in the 
basic model. In the third, ECT concentration, and plant leafiness and volume were all 
used as covariates. The degrees of freedom of this analysis were limited by the number 
of samples used for the chemical determination of ECT concentration (n=32). 
Genotype effect (class variable) vs Plant morphological characteristics (covariate l 
TRIM:MED AND UNTRIMMED PLANTS 
The R-square changes when adding genotype effect (class variable) and either the 
percentage of leaves or plant volume in the first and second periods are given in Figure 
4 .8  and 4.9. The basic model explained about 43% in Period 1 ,  and 30% in Period 2 of 
the total variation. The improvement in R-square after fitting covariates and genotype 
effect varied from 27 to 42% (adding both plant volume and genotype effects) in 
Periods 1 and 2, respectively. Leafiness explained more variation than volume in both 
periods (Period 1 :  33 vs 26 % and Period 2 :  39 vs 33 %). However both leafiness and 
volume effects explained important amounts of the variation in number of bites, more 
than the genotype effect, in both periods. 
Experiments 5 and 6 
Genotype effect 
before Leafiness 
Leafiness effect after 
Genotype 
Genotype effect before 
Volume 
Volume effect after 
Genotype 
1 9 1  
PERIOD 1 
0.4348 Leafiness effect before 
Genotype 
0.1292
7 
� 0.3310(0.0000/ 
0.5641 0. 7658 
0.2025� / 0.0007(0.r=::64;.;;.O;;.J.5 __ 7::'""'""........,,_-, Genotype effect after 
0.1292
7 
0.5641 
0.1402� 
0. 7666 Leafiness �------------� 
0.4348 Volume effect before 
Genotype 
� 0.Ul3(O.0000) 
0.6961 
/0.0082 V.l760 
Genotype effect after 
0. 7043 Volume 
Figure 4.8. Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
or genotype effect had been added. Considering all experimental plants of 
birdsfoot trefoil in Period 1 of Experiment 6. 
I P-value for the differences of R-squares. 
Experiments 5 and 6 192 
PERIOD 2 
Genotype effect 
before leafiness 
0.2964 leafiness effect before 
Genotype 
0.1618 7 � 0.3909(0.0000/ 
0.4581 0.6873 
leafiness effect after 
Genotype 
0.2295 � / 0.0003 (0.;.:.,,79:..=2:.;.,1:....-_�----:_--, Genotype effect after 
Genotype effect before 
Volume 
0.6876 Leafiness L-____________ � 
0.2964 Volume effect before 
Genotype 
0.1618 7 � 0.3272(0.0000) 
0.4581  
r.:-:-- -:::--�::-- ---, 
0.2537 � 
Volume effect after 
Genotype 
0.6235 
0. 71 18 
Figure 4.9. Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
or genotype effect had been added. Considering all experimental plants of 
birdsfoot trefoil in Period 2 of Experiment 6. 
lP-value for the differences of R-squares .  
Experiments 5 and 6 193 
In both periods there was a highly significant effect of genotype, plant leafiness and 
volume when initially added to the model. However in Period 1 leaflness and volume 
were important for explaining the variation between genotypes in relation to number of 
bites. After adding the covariates Ieafiness or volume, the genotype effect became non 
significant (P>O.05). In Period 2 , although volume had a substantial effect on changes 
of R-square, it did not seem to be important for explaining the variation between 
genotypes. After adding volume, the genotype effect was still highly significant. In this 
period leafiness was more effective than volume for explaining the differences between 
genotypes in relation to number of bites. After adding the covariate leafiness, the 
genotype effect became non significant (P>O.05). Plant morphological characteristics 
(leafiness and volume) had greater effect on R-square changes than genotype effect in 
both periods.  
UNTRIMMED PLANTS 
The R-square changes when adding genotype effect and either the percentage of leaves 
or volume in Periods 1 and 2, considering only untrimmed plants, are presented in 
Figure 4. 1 0  and 4. 1 1 . The basic R-square (without the addition of genotype effect or 
any covariate) in Period 1 was greater than in Period 2. The R-squares of the basic 
models in both periods were smaller than in the first analyses . The improvement in R­
square after fitting covariates and genotype effect varied from 40 to 65% (adding both 
plant volume and genotype effects in Periods 1 and 2, respectively). As in the previous 
analysis, the variation in R-square was also greater in Period 2 than in Period 1 .  In both 
periods, the changes in R-square promoted by the covariates and genotype effect were 
greater than the basic R-square. However morphological characteristics had greater 
effect on changes in R-square than the genotype effect in both periods .  
Experiments 5 and 6 
Genotype effect before 
Leafiness 
Leafiness effect after 
Genotype 
Genotype effect before 
Volume 
194 
PERIOD 1 
0.3072 Leafiness effect before 
Genotype 
0.3179  � 0.4805 (0.0000/ 
0.6251 0. 7877 
0.1667 � / 0.0041 (0r.:.5�18:.:::.3,"-:) _---::::--:--::-_-, 
Genotype effect after 
0. 7918 Leafiness �------------� 
0.3072 Volume effect before 
Genotype 
0.3179  � 0.3722 0·0002) 
0.6251 0.6794 
Volume effect after 
Genotype 0. 7050 
Figure 4. 1 0. Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R-squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
(leafiness or volume) or genotype effect had been added. Considering only 
untrimmed birdsfoot trefoil plants in Period 1 of Experiment 6. 
l P-value for the differences of R-squares .  
Experiments 5 and 6 
Genotype effect before 
Leafiness 
Leafiness effect after 
Genotype 
Genotype effect before 
Volume 
Volume effect after 
Genotype 
1 95 
PERIOD 2 
0.2128 Leafiness effect before 
Genotype 
0.3570  � 0.5341 (0.0000/ 
0.5699 0. 7469 
0.1845 � / 0.OO74(O.4r:2:.:.5:::..<3�_-:-_-:-----.. 
Genotype effect after 
0. 7543 Leafiness �------------� 
0.2128 Volume effect before 
Genotype 
0.3570  � 0.4564(0.0000) 
0.5699 0.6692 
0.2�  
0.8623 
Figure 4. 1 1  Effect of genotype (class variable) and plant morphology (covariate) (Leafiness 
or Volume) on R -squares (from the analysis of variance of number of bites), 
when added to the model before (independent) or after plant morphology 
(leafiness or volume) or genotype effect had been added. Considering only 
un trimmed birdsfoot trefoil plants in Period 2 of Experiment 6. 
l P-value for the differences of R-squares .  
Experiments 5 and 6 196 
There were significant effects of genotype, leafiness and volume in both periods. 
However genotype effect was not independent of leafiness in either period, and of 
volume in Period 1 .  In this case, after adding leafiness or volume, the genotype effect 
became non significant. The genotype effect was independent of volume in Period 2. In 
both periods, plant leafiness and volume had greater effects than genotype in R-square 
change, but leafiness was more important than volume to explain the R -square changes. 
EeT concentration (covariate) vs Plant morphological characteristics (covariate) 
The R-squares and the differences of R-squares when plant leafiness or volume and 
EeT concentration were added to the model as covariate in the analysis of variance, 
considering either Period 1 or Period 2, are given in Figures 4. 1 2  and 4. 13 .  The basic 
model (without the covariate effects) explained a higher percentage of the variation in 
Period 1 (56%) than in Period 2 (27%). The improvement in R-square after fitting 
covariates was about 28 and 43% in Periods 1 and 2, respectively. Leafiness, volume 
and EeT concentration explained important variation in number of bites. However 
leafiness and volume had greater effect on R-square change than EeT concentration. 
Although EeT concentration showed similar effect in both periods, leafiness and 
volume effects explained higher variation of the R -square in Period 2 than in Period 1 .  
In both periods, addition of EeT concentration, leafiness or plant volume to a basic 
model significantly increased R-square. However the EeT concentration was not 
independent of the plant leafiness because after adding leafiness to the model the effect 
of EeT concentration became non significant (P>O.05) .  However after adding volume, 
the EeT concentration effect was still marginally significant. 
Experiments 5 and 6 197 
PERIOD 1 
ECT conc. effect before 
Leafiness 
0.5638 Leafiness effect before 
ECT conc. 
� 0.2"'. (0.00001 ' 
0. 7296 0.8547 
0.1253 � /0.0002 (0.8641) 
Leafiness effect after r::E:-:::C=T:-c
-
o
-
nc
-
.
-
e7."ffe
-
c�t
-
aft':':"e
-
r'" 
ECT conc 0.8549 Leafiness L..-______ ...J 
ECT cone. effect before 0.5638 Volume effect before 
Volume ECT conc. L..-______ ....J 
0.1658 7' � 0.24% (0.0000) 
0. 7296 
Volume effect after ECT 
cone. 
0.'151� 
0.8134 
0.8446 
Figure 4. 1 2.Effect of extractable condensed tannin concentration (covariate) (ECT conc) 
and plant morphology (covariate) (Leafiness or Volume) on R-squares 
(from the analysis of variance of number of bites), when added to the 
model before (independent) or after plant morphology or ECT conc. 
had been added. Considering plants of birdsfoot trefoil in Period 1 of 
Experiment 6 .  
lP-value for the differences ofR-squares. 
Experiments 5 and 6 
ECT cone. effect before 
Leafiness 
PERIOD 2 
0.2773 
198 
Leafiness effect before 
ECT conc. 
� 0.4273 (0.0000) ' 
0.4274 0. 7045 
Leafiness effect after 
ECT conc 
ECT cone. effect before 
Volume 
02838 � /0.0067 (0.4 r.,9::=:16::::J:--_-:::--:---:-:--., 
ECT cone. effect after 
0. 71 12 Leafiness 1...-_------' 
0.2773 Volume effect before 
ECT conc. 
0.1502 7' � 0.3694 (0.0003) 
0.4274 0. 6466 
0.2803 � /0.0611 (0.Or;;5;...:.;43::..:..... _____ ....., 
Volume effect after ECT ECT cone. effect after 
conc. 0. 7077 Volume L..-______ .....l 
Figure 4. 1 3 .Effect of extractable condensed tannin concentration (covariate) (ECT cone.) and 
plant morphology (covariate) (Leafiness or Volume) on R-squares (from the 
analysis of variance of number of bites), when added to the model before 
(independent) or after plant morphology or ECT cone. had been added. 
Considering plants of birdsfoot trefoil in Period 2 of Experiment 6. 
l P-value for the differences of R-squares .  
Experiments 5 and 6 1 99 
This analysis was similar to the analysis where individual plant data and genotype 
effects (rather than ECT concentration) were used. In both periods, the effect of 
leafiness was greater than the effect of volume. In both analyses plant morphology 
(leafiness and volume) had greater effects on the R-square changes than ECT 
concentration (or genotype effect). The difference between the effects of plant 
morphology and ECT concentration (or genotype effect) was greater in Period 2 than in 
Period 1 .  
4.4. DISCUSSION 
4.4.1. Evaluation of experimental procedures 
The grazing preference of cows was assessed using a procedure based on techniques 
developed by Laca et al. ( 1 993) and Griffiths et al. ( 1996) to determine grazing 
behaviour on patches of ryegrass pasture, and on techniques developed by Real-Ferreiro 
( 1 997) to assess pre and post grazing characteristics of spaced plants of red clover. 
These experiments were designed to determine animal preference, rather than selection, 
by maximising opportunity and offering free choice (see Chapter 3, section 3.4.2 for 
definition of preference and selection) .  The technique of assessing behaviour as an 
animal grazes along a sequence of spaced plants differs from the conventional technique 
where groups of animals demonstrate their preference by selecting from several plants 
or swards in a field (Simon, 1 974; Marten and Jordan, 1 974; Hedges et al. 1 978; 
MacGraw et al. ,  1 989; van Santen, 1 992; Shewmaker et al. ,  1 997) . With this technique 
a close assessment of animal preference was possible. It allowed the precise 
measurement of the number of bites, grazing time and rate of biting for each plant. In 
addition it provided information about behaviour at a specific locus in relation to 
conditions at the preceding and the succeeding loci in a sequence, and the animals were 
not influenced by previous grazing of each plant, deposition of feces and treading. The 
other advantage of this technique is that the assessment period is short. However it 
requires the participation of at least two people for monitoring behaviour and it is very 
Experiments 5 and 6 200 
time consuming in relation to animal training and pre and post grazing assessment The 
other constraint is that the animals need to be accustomed to handling and close grazing 
control. 
Using this technique it was possible to explain a substantial proportion of the variation 
in the number of bites per plant. From 62 to 92 % of the total variability in number of 
bites was explained when analysing only one plant species and including covariates in 
the model (see covariance analyses, section 4.4.4). The combined analysis of number of 
bites per plant including all periods within each experiment explained approximately 
40% of the total variability. Both Experiments 5 (E5) and 6 (E6) also explained similar 
variability (similar R-square). The accuracy reached with this technique, explaining up 
to 92% of the variation, is better than that achieved by Shewmaker et al. ( 1 997) using 
visual preference score (44% of variation) when assessing cattle preference for eight tall 
fescue cultivars planted in swards. However, comparison of this technique with other 
techniques is difficult because of other sources of variation among the different 
experiments (such as differences in animal species; plant species, maturity, morphology, 
biochemistry) and because of the scarcity of information published. The small 
variability between E5 and E6 is in agreement with what was found by van Santen 
( 1992), assessing cattle preference of twenty-five tall fescue cultivars and populations, 
where preference rating agreed very closely within and between years. 
Preferential behaviour was assessed mainly by the comparison of number of bites the 
animals took per plant. Although the variability of animal response was explained 
equally well by either number of bites or grazing time, number of bites was more 
precisely measured than grazing time. Grazing time per plant was used in the analyses 
of rate of biting. 
Plant morphology had an important effect on preference. The measurement of plant 
morphology was a useful measurement for better understanding of choice. The 
assessment of plant characteristics, other than herbage mass, before and after grazing 
followed a modified technique described by Real-Ferreiro ( 1997). This technique 
provided individual assessment of the plants in terms of what was offered to the animal, 
Experiments 5 and 6 201 
and the reduction of each plant variable by grazing. The use of visual assessment of 
plant leafiness,  habit and density provided non-destructive and useful information with 
minimum use of time and resource. Real-Ferreiro ( 1997) when assessing the amount 
removed from several cultivars of red clover by sheep concluded that visual assessment 
of post-grazing leafiness was the best estimation of forage removal. 
The assessment of herbage mass was one of the main difficulties in this research. It was 
only possible to have an indirect assessment for each plant. Prediction equations 
developed for the pasture probe from plants in the spare sequences showed R-squares of 
0.37 for birds foot trefoil (BT) and 0.20 for red clover (RC) (see Appendix 4.2). 
Although the R-squares were relatively low for both species, the herbage mass 
measurements agreed relatively weII with visual comparisons between plants. Big and 
dense plants usually had greater herbage mass than small and sparse plants. This 
observation was confirmed by correlation analyses that showed that in most periods of 
both experiments, herbage mass had a significant correlation (R-square generally better 
than 0.5) with plant area, height, volume and leafiness (Appendices 4.4, 4.5, 4.6, 4.8, 
4.9). The only exceptions happened with RC plants in Periods 1 and 2 of E5. However, 
there were significant correlations in Period 1 between herbage mass and leafiness and 
in Period 2 between herbage mass and height. In these cases the lowest correlation 
found in both periods was between herbage mass and area. This low correlation is 
probably related to the fact that the probe device used is mainly sensitive to pasture 
height and density. 
Plant trimming was used as a way to assess, and separate into morphological and 
chemical components, the effects of plant genotype on preferential behaviour. In both 
experiments, trimming made the alternative genotypes within each species 
morphologically uniform. However trimming usually reduced the plant size and 
leafiness, which in part also altered the animal preference and needed to be considered 
in the analyses. Trimming did not affect the ECT concentration of BT plants in either 
experiment (Tables 4.8, 4.25) or the formononetin concentration in E6 (Table 4.26). 
However, in E5 trimmed plants of cultivar Pawera had higher concentration of 
formononetin than untrimmed plants (Tables 4.9), though this difference did not affect 
Experiments 5 and 6 202 
the substantial contrast between cultivars G-27 and Pawera. The higher formononetin 
concentration of trimmed plants of Pawera may be explained by the fact that the old 
leaves were trimmed and there was a proportionate increase of young leaves. Rossiter 
and Beck ( 1 967) and Keogh ( 1995) demonstrated that concentration of formononetin in 
individual leaves decreased from emergence to senescence. In addition Anwar ( 1994) 
observed that Pawera have greater variation in formononetin concentration between 
young and old leaves than did G-27. 
Previous experience was necessary for the animals to get used to the plant species and 
sequences. Although Lascano et al. ( 1 988) affirm that in cafeteria trials where the 
objective is to rank forage species in terms of palatability it is not necessary to subject 
animals to short-term previous experience on the individual species under evaluation, it 
was important in E5 that the cows had previous experience of grazing BT. During the 
training period, one week before running the experiment, the animals were not willing 
to graze BT. This result contrasts with E6, using different cows, where the animals had 
only the normal training with the same plant genotypes one week before starting the 
trial, but they did not reject BT. The probable reason is that the cows used in E6 were 
used to a greater variety of foods and had experienced BT before. 
The use of sequences of spaced plants with one animal per sequence did not allow 
assessment of between-animal variation in preference. Because different animals grazed 
different sequences, the animal effect was confounded with sequence variation. 
However, the variation of plant sequence and cow were considered in the statistical 
model before treatment effects. In this way variation of cows or plant sequence were not 
included in the treatment effect. 
The analysis of extractable condensed tannin (ECT) was done mainly on samples of BT 
plants, and formononetin analysis on samples of RC plants. This was because 
concentrations of ECT in samples of RC, and of formononetin in samples of BT were 
negligible. Other studies have also shown that the amount of ECT in RC is very small 
(Jackson et al . ,  1996). The analysis of ECT and formononetin concentration followed 
the same procedures described in Chapter 3 (see section 3 .4. 1 ). 
Experiments 5 and 6 203 
The addition of two different materials (Lotus species and red clover cultivars) with 
different ECT and formononetin concentrations into the rumen of cows in E6 
demonstrated the effects of manipulation of rumen content on cattle diet selection. 
Although other studies (e.g. Cooper et al. ,  1995 ; Carter and Grovum, 1 990) using 
penned animals have been carried out to determine the effect of rumen manipulation on 
diet selection, this study is believed to be one of the first attempts to use rumen 
manipulation in grazing trials. The cows had a significantly higher number of bites in 
RC when RC (independent of the cultivar) was inserted into the rumen (see discussion 
section 4 .4 .4) . However, comparing the concentration effect within each material 
inserted in the rumen (Lotus species or red clover cultivars), there was no significant 
effect of the different concentrations on cattle preference (Tables 4.30, 4.3 1 ). This 
result was apparently a reflection of the low concentrations and small contrast between 
high and low concentrations of either tannin or formononetin added to the rumen 
(Appendix 4. 1 0) .  The low concentrations and small contrasts were related in part to the 
addition of pellets of alfalfa to facilitate the mincing process (see section. 4.2.6.2) and in 
part to the amount of weeds present in some of the materials used. 
The contrasts between different plant species and genotypes provided the opportunity 
for investigation of the effect of plant morphological and biochemical characteristics on 
cattle preference. The range of variation found in each experiment is summarised in 
Table 4.38.  Although ECT concentration and plant area had the greatest range of 
variation, the correlation analysis showed that the sward morphological characteristics 
were significantly correlated among themselves (Appendices 4.4, 4 .5 ,  4 .6,  4 .8 ,  4.9). The 
correlations of plant morphological and biochemical characteristics and their 
importance on cattle preference are discussed in section 4.4.4. 
Experiments 5 and 6 204 
Table 4.38.  Range of values (minimum and maximum) of individual plant 
morphological and biochemical characteristics observed in Experiments 5 
and 6. These values were extracted from averages presented in Tables 4.5, 
4.6, 4.7, 4. 19, 4.20, 4.2 1 , 4.22, 4.23 and 4.24. 
EXl!.eriment 5 EXl!.eriment 6 
Plant Characteristics Minimum Maximum Minimum Maximum 
Height (cm) 1 0.2  1 8.9 9.7 
Area (cm2) 970 5540 980 6880 
Leafiness (%) 37 75 3 1  75 
Herbage mass (gDMlm2) 325 494 228 4 1 9  
ECT (%) 0.47 3.28 0.35 1 .62 
Formononetin (%) 0.29 0.73 0.26 0.63 
Although the cows and experimental site differed between E5 and E6, the behavioural 
responses in these two experiments were similar. The similar approach for each 
experiment, the limited effect of rumen manipulation, and the similar animal response 
justify a combined discussion for the two experiments. Differences between 
experiments will be explained where appropriate. 
4.4.2. Plant morphological characteristics 
There were important variations in plant morphology in both experiments (Tables 4.5, 
4.6, 4.7, 4. 1 9  - 4.24). Red clover (RC) had greater density and percentage of leaves than 
birdsfoot trefoil (BT) in all runs and periods. Within species, BT genotypes showed 
greater morphological contrast than RC genotypes. Untrimmed plants of Goldie were on 
average the largest plants and accession PI273938 were the smallest. Trimmed plants 
usually were smaller and had less leaves and density than untrimmed. In fact, the 
animals were faced with a complex decision involving variations not only in plant 
biochemistry but also in plant morphology. 
The analysis of the reduction by grazing in magnitude of each plant variable showed 
that in part the animals responded to the plant morphology (Tables 4.5, 4. 1 9, 4.20). The 
Experiments 5 and 6 205 
animals tended to remove more from the larger and leafier plants, showing a preference 
for plants with a greater proportion of leaves (as in Theron and Booysen, 1 966; 
O'Reagain and Mentis ,  1989, 0' Reagain, 1993), and greater height and density (as in 
Illius et al. ,  1 992; Demment et al. ,  1 993). This result confirms what was found by Clark 
(reported in Illius and Gordon, 1990) where cattle were very sensitive to sward height 
and they could discriminate well between alternative swards and select across a broad 
range of height contrasts the taller one. It also indicates, as demonstrated by Clark (in 
Illius and Gordon, 1 990) and Illius et al. ( 1992), that the preference was affected by 
higher levels of plant height and herbage mass than values reported in other studies as 
limiting intake (Holmes, 1987; Hodgson, 1990; Gibb et al. ,  1 997). More detailed 
analyses including analysis of variance of number of bites, correlation and covariance 
analyses were carried out for a better understanding of the effect of plant morphology 
on animal preference (see section 4.4.4) . 
4.4.3. Plant chemical composition 
There were significant contrasts in ECT concentration between genotypes of BT in both 
experiments (Tables 4.8, 4.25, summarised in Table 4.39). As expected, the ECT 
concentration was significantly higher in accession PI273938 than in the cultivar 
Goldie. The accession PI273938 had, on average, 4.2 times (in E5) and 3 .5  times (in 
E6) higher ECT concentration than the cultivar Goldie (Table 4.39). The concentration 
of ECT in E6 either in January or April was similar to the concentration found in April 
in E5. However the concentration found in March (E5) was notably higher. The increase 
in concentration of tannin during summer months and a decline from summer to autumn 
was also reported in several studies for different species (Clark et al. , 1 939; Stit and 
Clark 1 94 1 ;  Donnelly, 1959; Cope et al. 197 1 ;  Windham et al. 1988;  Iason et al. ,  1 995). 
The smaller difference in ECT concentration found in E6 between January and April 
can therefore be explained by the fact that the comparison was done in the beginning of 
summer and in autumn. The higher percentage of ECT in March (E5) showed that 
probably the ECT concentration increased from January to March and decreased from 
March to April. The lowest concentration in autumn was also found in E l ,  E2 and E3 
(Tables 3 .4, 3. 1 3, 3 .25, summarised in Table 4.39). However, in these three previous 
Experiments 5 and 6 206 
experiments the cultivar Goldie had higher concentration in November (spring) than in 
February and April-May. 
Table 4.39. Extractable condensed tannin (ECT) concentration (%) in birdsfoot trefoil 
(cultivar Goldie and accession PI273938) between November and April­
May of Experiments 1 ,  2 and 3 (only leaves) and Experiments 5 and 6 
(intact stems: leaf and stems).  
GenoW!.e EXl!.erimentt November January February March Ap'ril-Ma� 
1 , 2 and 3 1 .69 0.58 0.54 
Goldie 5 0.72 0.47 
6 0.32 0.54 
PI273938 5 3 .28 1 . 8 1  
6 1 .49 1 .55 
f The Experiments 1 ,2 and 3 were carried out in 1 99511 996, and the Experiments 5 and 6 in 1 997. 
Comparing the ECT concentration of cultivar Goldie in the five experiments it can be 
observed in Table 4.39 that there was a decrease in concentration from spring to 
summer, an increase during the summer months and a decline from summer to autumn. 
However the variation observed between experiments and within each experiment can 
also be affected by several factors discussed below. 
Previous studies showed variation of ECT concentration for BT between 0.25 to 3 .58 % 
of DM (Douglas et al . ,  1993, 1 995; Jackson et al. ,  1996; John and Lancashine, 1 98 1 ;  Li 
et al. ,  1996; Lowther et al. ,  1 987; Terrill et al . ,  1 992; Wang et al . ,  1995 ; Waghom et al. 
1987ab). However analyses of cultivar Goldie reported by Douglas et al. ( 1993, 1995) 
showed greater concentrations of ECT ( 1 . 1 8  to 2.52 % DM) than those found in E5 and 
E6 (0.32 to 0.72 % DM). This difference is probably explained by seasonal variation 
(Roberts et al. ,  1 993), site variation (Douglas et al. ,  1993), soil fertility (Barry, 1989), 
tannin analysis procedures (Furstenburg and van Hoven, 1994; McNabb, pers. comm.) 
or differences in plant morphological characteristics (lason et al . ,  1 995; Douglas et al. ,  
1 993). The fact that the plants in E5 and E6 were sown in a glasshouse and were planted 
out in the field probably influenced the percentage of ECT. Studies with Lotus 
pedunculatus and Lotus corniculatus (John and Lancashire, 198 1 ;  Barry and Forss, 
Experiments 5 and 6 207 
1 983;  Barry, 1985;  Lowther et al, 1 987) showed that growing conditions have an 
important effect on condensed tannin concentration. In addition, the DMACA-HCL 
procedure used in these studies is more specific for condensed tannin than butanol-HCL 
(McNabb, pers. comm.)  used by Douglas et al. ( 1 993, 1995), so values of ECT 
measured using butanol-HCL protocol tend to be higher than with DMACA-HCL. 
Douglas et al. ( 1 993, 1995) also used condensed tannin from Lotus pedunculatus as a 
standard curve, while in the analyses for these studies condensed tannin extracted from 
Lotus corniculatus was used (Appendix 3 . 1 ). The difference in absorbance curves 
between differing species was reported by Furstenburg and van Hoven ( 1 994), who 
explain that condensed tannin composition is diverse and species specific. Oiven the 
differences in the standard curves, there is always a tendency to predict a higher 
concentration when using Lotus pedunculatus as the standard compared to Lotus 
comiculatus (W.e. McNabb, pers. comm.) .  
The concentration of  formononetin was very similar in  both experiments (Table 4 .9  and 
4.26, summarised in Table 4.40). As expected, there was a substantially higher 
formononetin concentration in cultivar Pawera than in cultivar 0-27. The concentrations 
found in both cultivars were similar to levels reported in other studies (Kelly et al. ,  
1 979; Anwar, 1994). In both experiments Pawera had on average 2.3 times more 
formononetin than 0-27, but the concentration of formononetin in cultivar Pawera 
seemed more variable than the concentration in 0-27. Anwar ( 1 994) explained that the 
decline in formononetin concentration of leaves with age was greater in Pawera than in 
0-27. However, there was no significant variation between different periods. This result 
agrees with the results found in E 1 ,  E2 and E3 with cultivar Colenso (Tables 3.5,  3 . 14, 
3 .26, summarised in Table 4.40). The variation of formononetin concentration in 
cultivars Pawera and 0-27 seemed more related to the age of each plant part (Rossiter 
and Beck, 1967; McMurray et al. ,  1 986; Keogh, 1995) than to seasonal variation. The 
plant parts probably had a similar age in different periods and experiments. 
Experiments 5 and 6 208 
Table 4.40. Formononetin concentration (%) in leaves of red clover cultivar Colenso 
and intact stems (leaf, petiole and stems) of cultivar G-27 and Pawera 
between November and April-May of Experiments 1 ,  2 and 3 and 
Experiments 5 and 6. 
Cultivars EXl!.erimentt November January February March Ae.ril-May' 
Colenso 1 , 2 and 3 0.61 0.50 0.70 
G-27 5 0.28 0.30 0.30 
6 0.29 0.23 
Pawera 5 0.67 0.72 0.64 
6 0.60 0.62 
f The Experiments 1 , 2  and 3 were carried out in 1 99511996, and the Experiments 5 and 6 in 1 997. 
The four genotypes showed a relatively high nutritional value in terms of general 
chemical composition in both experiments (Tables 4. 10, 4.27). All four genotypes had 
digestibility above 74 %. The data showed lower ADF and NDF than standard values 
published by NRC ( 1 989, 1996), but similar concentrations of the other components. 
The high nutritive values for BT and RC were also found in E l ,  E2 and E3 (Tables 3 .6, 
3 . 1 5, 3 .27). Correlation analyses were carried out to investigate the influence of the 
plant nutritive value on animal preference (Tables 4. 14, 4.33). Although there was a 
significant correlation between the number of bites and lipid concentration in BT (E5, 
Table 4. 14), and between the number of bites and ADF, carbohydrates and ash in RC 
(E6, Table 4.33), the four genotypes in both experiments did not differ significantly in 
concentrations of lipid, ADF, carbohydrate and ash. It can be concluded, therefore, that 
nutritive value did not have an important effect on preference in relation to the four 
different cultivars. 
Experiments 5 and 6 209 
4.4.4. Grazing behaviour 
The rate of biting showed little variation between different plant genotypes in E5, but 
was affected by plant morphology in E6. In both experiments there were also no 
significant interactions with the Period effect. In E5 the difference in plant 
characteristics, mainly density, was not large enough to affect rate of biting (Tables 
4. 1 2, 4. 1 3) .  This result suggests that the time spent grazing each plant was proportional 
to the number of bites taken (Griffiths et al., 1 996). However in E6 the rate of biting 
demonstrated that the animals spent more time to obtain the same number of bites from 
untrimmed plants of BT, and from BT cultivar Goldie, than from the other plants (Table 
4.32). This result apparently reflected the plant morphology. Untrimmed plants of 
cultivar Goldie were large, but with lower density than RC plants. As several studies 
(Chacon and Stobbs, 1976; Hodgson and Jamieson, 1 98 1 ;  Milne et. Al., 1982; Philips 
and Leaver, 1985;  Penning et al. 1 99 1 ;  Mitchell et al., 1 993) have also shown, there was 
a negative relationship between rate of biting and sward height, or rate of biting and 
herbage mass. In this case the animals had lower rate of biting in tall sparse swards than 
on short dense ones of equal mass. 
Comparing the number of bites taken per plant in BT and in RC within E5, there was a 
significant preference for RC plants in Period 2, but not in Period 3 (Table 4. 1 1 ) .  The 
preference for RC in the second period might be influenced by the higher concentration 
of ECT in BT in this period. In E6 the preferential behaviour between BT and RC was 
modified by an interaction between trimming and secondary compound concentration 
effects (Table 4.28) and by the rumen content manipulation (Table 4 .29). The fact that 
the animals took more bites from untrimmed plants of RC and untrimmed plants of 
cultivar Goldie reflected the preference the animals had for bigger, denser plants and 
with greater proportion of leaves (O'Reagain and Mentis, 1989; lllius et al. 1992; 
O'Reagain, 1993; Demment et al . ,  1993) . The fact that in both experiments the animals 
always left a great portion of herbage mass behind and never completely grazed a plant 
(Tables 4.5, 4.6, 4.7, 4. 1 9  - 4.24) probably demonstrated that the amount of herbage 
mass did not limit the number of bites. On the other hand, the fact that the animals took 
significantly fewer bites from accession PI273938, independent of the trimming effect, 
Experiments 5 and 6 210  
could also indicate that the animals were affected by  the concentration of  ECT. 
Individual analyses of variance including only one species at a time, followed by 
correlation and covariance analyses were carried out to clarify these results. In E6 the 
animals were also influenced by the type of material inserted into the rumen. The 
animals took more bites from RC plants than BT when RC material was inserted in the 
rumen. The effects of rumen manipulation on animals diet selection will be discussed 
later. 
The variation In number of bites between RC genotypes in both experiments 
demonstrated that morphological effects were more important than formononetin 
concentration in determining preferential behaviour (Tables 4. 1 1 , 4.3 1 ). In all analyses 
of variance, trimming had a substantially more important effect than the difference 
between genotypes. Cows were apparently not directly influenced by the formononetin 
concentration in RC. This result disagrees with results reported by Frands ( 1973).  This 
author, working with variants of subterranean clover cultivar Geraldton with different 
isoflavone composition (formononetin concentration varied from 0 . 1 5  to 1 .05 % DM),  
postulated that unhydrolyzed glycosides in subterranean clover could contribute to 
unpalatability. However Harbome ( 1993) argued that isoflavones are not sufficiently 
repellent in taste to deter feeding. 
In both experiments there was substantially more contrast between genotypes in number 
of bites per plant in BT than in RC. This seemed to be related either to the contrasts in 
morphological characteristics of the two genotypes of BT or to the contrasts in ECT 
concentrations. In E5 and E6 the number of bites taken from BT was affected by both 
trimming and genotype effects (Tables 4. 1 1 , 4.30). The fact that trimmed plants had 
lower number of bites is probably associated with the smaller size and lower percentage 
of leaf found in trimmed plants than in un trimmed plants. As in other studies, the 
animals seemed to show an immediate reaction against small and stemmy plants or 
patches in a sward (Arnold, 1960, Arnold, 198 1 ;  L'Huiller et al. ,  1 984; L'Hiuller, 1 986; 
Q'Reagain and Mentis, 1989; lllius et al. 1 992; Q'Reagain, 1993; Edwards, 1 994). 
However, in E5 Period 2 ECT concentration (represented by differences in genotypes), 
independent of trimming effect, had an immediate effect on selection, and had a 
Experiments 5 and 6 2 1 1  
substantially greater effect on diet selection than did formononetin. This result probably 
reflects the higher concentration of ECT in this period than in Period 3 and in both 
periods of E6. However, further analyses are important to clarify the effects of plant 
morphology and plant biochemistry and the effect of their relationship with preference. 
Correlation analysis between number of bites and either ECT or formononetin 
concentration was one of the first steps to clarify the effect of plant morphology and 
biochemistry on grazing preference. In both experiments there was low correlation 
between ECT concentration and number of bites in April, but there were significant 
correlations in March (E5) and January (E6) (Tables 4. 15 ,  4.34). However, number of 
bites were also highly correlated with plant morphological characteristics (area, height, 
volume (area x height) and leafiness of BT plants) (Tables 4. 1 7, 4.36). This result 
reflects the difference in plant morphology between BT genotypes. The significant 
morphological contrasts between genotypes with high and low concentration of ECT 
can also be detected by the significant correlation between ECT concentration and plant 
morphology. Covariance analysis then was necessary to determine the relative 
magnitude of the ECT concentration and plant morphology effects on selective 
behaviour and how they were interrelated. 
In contrast, formononetin did not have an important effect on preferential behaviour. 
There was no significant correlation between number of bites and formononetin 
concentration in RC plants in either experiment (Tables 4 . 1 6  and 4.35). Although there 
were no significant correlations between formononetin concentration and RC 
morphological characteristics (except in Period 2 of E6), there were, in both 
experiments, significant correlations between number of bites and plant morphology 
(Tables 4. 1 8  and 4.37). The low correlation between formononetin concentration and 
plant morphology reflects the fact that the RC genotypes (with high and low 
concentration of formononetin) were morphologically similar. However, most of the 
variation in number of bites probably could be explained by the variation in RC 
morphology. Thus, it is concluded that morphological characteristics were much more 
important than formononetin content in influencing preference between cultivars of RC. 
Experiments 5 and 6 2 12  
Covariance analyses were carried out to clarify the effects o f  ECT concentration and 
plant morphology on preference in BT which were observed in the analyses of variance 
and correlation analyses. The effect of BT genotype and ECT concentration on animal 
preference was tested in the covariance analysis against two of the most important plant 
morphological characteristics (leafiness and volume) that affected preference in BT. 
Similar analyses were not necessary for RC since formononetin clearly had a very 
minor influence on preference.  Although the covariance analyses did not separate 
completely the effect of plant morphological characteristics from the effects of ECT 
concentrations in BT, the use of three different analyses of covariance helped in 
understanding of this relationship. The addition of plant morphology and genotype 
effect to a basic model had greater effect (greater R-square improvement) in the analysis 
where only untrimmed plants were considered than in the analysis where all 
experimental plants were used (Figures 4.2 - 4.5, 4.8 - 4 . 1 1 ). This result reflects the 
greater morphology differences between un trimmed than trimmed plants. Therefore, 
plant morphology had an important effect on selective behaviour in both experiments 
and all periods. The use of means of the subset of plants, according to the sets of plants 
bulked for ECT chemical determination, reduced the R-square improvement with the 
addition of the covariates to the basic model (Figures 4.6, 4.7, 4. 1 2, 4. 1 3) .  This 
reduction can be explained by the greater variability in number of bites explained by the 
basic model. Although the average number of bites within each sequence reduced the 
variation within and between sequences of plants, ECT concentration as a covariate had 
a similar effect to that of the genotype effect in the analysis where all experimental 
plants were used. The effect of the covariate ECT concentration increased as the 
contrast between genotypes increased. 
Although the two experiments had similar design, they need to be discussed separately 
because of the difference in causes of variation in each experiment. While the basic 
model of E5 was formed by the variation of sequences of plants, blocks (within 
sequences) and trimming, in E6 it was formed by the variation of day, cow, chemical 
inserted in the rumen, rumen chemical concentration, sequence of plants, blocks (within 
sequences) and trimming. The significance of the effect of the covariates on R-square 
changes in E5 and E6 are summarised in Tables 4.4 1 ,  4.42, respectively. 
Experiments 5 and 6 2 1 3  
Table 4.4 1 .  Summary of the R-square improvement when the covariates plant leafiness, 
volume or ECT concentration were added before and after fitting the 
alternative covariate to the model in analyses 1 ,  2 and 3 for Periods 2 and 3 
of EXEeriment 5 .  
AnaL-l.sest Period Leafiness Volume ECytt 
Before After B efore After Before After After 
ECT ECT ECT ECT leaf. leaf. vol. 
1 2 +++' ++ ++ + ++ + + 
3 ++ + ++ ++ + + + 
2 2 +++ + +++ + +++ + + 
3 + + +++ ++++ + + ++ 
3 2 ++ + ++ + ++ + + 3 + ++ + + + + + 
tAnalyses 1 - full number of observation were used (n::::48);  Analyses 2 - only untrimmed plants were 
used (n=24) ;  Analyses 3 - averages according to the bulking for EeT concentration 
analyses were used (n=16) .  
tt  EeT concentration in analyses 1 and 2 corresponded to the genotype class variable 
IR-square changes: + - from 0 to 10%; ++ - from 1 1  to 20%; +++ - from 2 1  to 30%; ++++ - > 30% 
Table 4.42. Summary of the R-square improvement when the covariates plant leafiness, 
volume or ECT concentration were added before and after fitting the 
alternative covariate to the model in analyses 1 ,  2 and 3 for Periods 1 and 2 
of EXEeriment 6. 
Anal"l.sest Period Leafiness Volume ECytt 
Before After Before After Before After After 
ECT ECT leaf. leaf. vol. 
1 1 ++++1 ++ +++ ++ ++ + + 
2 ++++ +++ ++++ +++ ++ + + 
2 1 ++++ ++ ++++ + ++++ + + 
2 ++++ ++ ++++ +++ ++++ + ++ 
3 1 +++ ++ +++ ++ ++ + + 
2 ++++ +++ ++++ +++ ++ + + 
tAnalyses 1 - full number of observation were used (n=96); Analyses 2 - only untrimmed plants were 
used (n=48) ;  Analyses 3 - averages according to the bulking for EeT concentration 
analyses were used (n=32). 
tt EeT concentration in analyses 1 and 2 corresponded to the genotype class variable 
IR-square improvement: + - from 0 to 10%;  ++ - from 1 1  to 20%; +++ - from 2 1  to 30%; ++++ - > 30% 
In E5, the first covariance analysis, where the full number of observations were used 
and the genotype effect was tested, leafiness was shown to be more effective (greater 
improvement in R-square) in explaining variation in number of bites between genotypes 
than volume (Table 4.41 ,  Figures 4.2, 4.3). In the second analysis (only untrimmed 
Experiments 5 and 6 214  
plants - plant natural variation) Period 2 ,  both leafiness and volume had important and 
similar effects on the variation in number of bites, but they did not have a significant 
effect after genotype had been added to the model (Figures 4 .4,  4.5). This result 
demonstrates that the variations between genotypes were unlikely to be completely 
explained by either leafiness or volume. In the third analysis (ECT as covariate) Period 
2, ECT concentration and leafiness both had important effects on the improvement of 
R-square (Table 4.4 1 ). However, both variables reduced the effect of the other, so that 
after adding one covariate to the model the other covariate was not significant (Figure 
4.6). In this analysis, Period 3 ,  variation between R-squares was small and ECT 
concentration and morphological characteristics were not important in explaining the 
number of bites (Figure 4.7). Most of the variation was probably explained by the 
variation between sequences of plants, which had the greatest sums of square in the 
analysis of variance in this period. 
In E6, the first and second analyses showed that in the first period either leafiness or 
volume were important to explain the variation between genotypes in number of bites 
(Table 4.42, Figures 4.8,  4.9, 4. 10,  4. 1 1) .  However in the second period leafiness was 
more effective than volume in explaining the difference between genotypes. Volume in 
this case was important to explain the variation in number of bites but was not related to 
the genotype effect. The third covariance analysis showed that most of the variation in 
number of bites between BT genotypes apparently was related more to plant 
morphological characteristics than to ECT concentration (Table 4.42, Figures 4. 12,  
4. 1 3) .  ECT concentration did not have a significant effect when added after plant 
morphological characteristics. However, when adding either leafiness or volume after 
ECT concentration, they still showed a significant effect on R-square improvement. 
Although volume also had an important effect on variation in number of bites, it did not 
seem to be related to ECT concentration. 
Conclusions from the covariance analysis can be summarised as follows. In E5 Period 
2, plant morphology and ECT concentration both had important effects on number of 
bites. The modification of one characteristic might change the effect of the other in 
cattle diet selection. In E5 Period 3 and E6, plant morphology had the major effect on 
Experiments 5 and 6 2 1 5  
variation in number of bites. In this case, most of the apparent effect o f  ECT 
concentration on number of bites could be accounted for by correlated differences in 
plant morphology. 
The correlation and covariance analyses showed similar results . Both analyses 
demonstrated that the greater part of the variation in number of bites between genotypes 
was explained by leafiness. They also agreed in showing the greatest importance of the 
ECT concentration effect on number of bites in Period 2 of E5, which was associated 
with the highest ECT concentration found over the two experiments. Condensed tannins 
have been thought to decrease forage intake by inhibition of digestion (Fenny, 1 969; 
Roades and Cates, 1 976; Swain, 1979; Barry, 1989). However, this result indicates that 
they also have an immediate effect on preference. Provenza and Malechek ( 1984) also 
observed rejection by goats of shrubs containing high concentrations of tannin and high 
levels of energy. They concluded that goat nutrition was affected more by the adverse 
effects that tannins had on palatability than by the negative effects they had on 
digestibility. 
A concentration of 3 .2% of ECT was apparently high enough to reduce preference. 
Studies, involving different animal and plant species and different management 
conditions have shown different responses to tannin concentration. Donelly and 
Anthony ( 1 969), for example, reported that the condensed tannin level required for 
rejection by grazing animals was as low as 0.2% of DM. Barry ( 1989), summarising 
several studies reported that concentrations of condensed tannin at or above 0.63 % 
depress food intake by herbivores. More recently, Waghorn et al. ( 1 990) suggested 
values exceeding 5 .5% depress voluntary intake. However, there is no evidence in the 
literature suggesting levels of condensed tannin that promote an immediate effect on 
cattle preference. 
According to previous studies the immediate effect of condensed tannin observed in 
these experiments could be explained either by the taste of tannin (Arnold et al. 1 980; 
Van Soest, 1994) or by an association between the taste and aversive post-ingestive 
consequences (Provenza and Malechek, 1984; Provenza et al. ,  1990; Provenza, 1 995). 
Experiments 5 and 6 2 1 6  
There is no evidence in the literature that animals can smell tannin. Provenza et al. 
( 1 990) explain that animals learn to avoid plants with high concentration of condensed 
tannin (CT) because of the internal malaise promoted by CT and not because of its 
flavour. According to Provenza ( 1995) this post-ingestive feedback is a rapid process, 
being less than one hour in goats. However Van Soest ( 1994) argued that tannin has 
astringent properties caused by the precipitation of salivary mucoprotein, which has an 
important effect on taste. According to what was reported by Provenza ( 1 995) the 
difference of one hour between rumen manipulation and the preferential behaviour 
observation in this study was probably enough for the animal to show aversive 
consequences of tannin. However the low concentration of ECT and small contrast 
between the materials added to the rumen in E6 (see section 4.4. 1 and Appendix 4. 10), 
precluded a conclusive explanation of the influence of rumen manipulation on 
preferential behaviour. In addition, the animal training experience with similar plant 
sequences ,  one week before running each trial probably did not provide each cow with 
enough herbage to cause post-ingestive feedback. Puther research is still needed to 
explain how the immediate effects of condensed tannin content influence animal 
preferential behaviour. 
In the context of rumen manipulation, as discussed in section 4.4. 1 ,  the low response to 
different concentrations of each material (either Lotus species or red clover cultivars) 
inserted into the rumen is probably the reflection of the limited secondary compound 
concentrations and contrasts found in each material (see Appendix 4. 10). However, 
higher preference for RC when red clover was added to the rumen (Table 4.29) might 
show that cattle diet selection could be influenced positively by the rumen content 
composition. The positive post-ingestive feedback in cattle diet selection observed in 
this research was also reported by Pfister et al. ( 1997), where highly nutritious plants 
caused positive animal response toward a greater intake of the same plant. The influence 
of rumen content on animal diet selection has been demonstrated by recent studies with 
sheep. Carter and Grovum ( 1990b) and Hou ( 199 1 )  in Cooper et al. ( 1 995) have shown 
that sheep can make rapid changes in their diet selection as a result of manipulation of 
rumen environment. Cropper ( 1987) in Cooper et al. ( 1 995) and Parsons et al. ( 1994a) 
suggested that ruminants appear to select, from two feeds that differ in nutrient density 
Experiments 5 and 6 2 17  
or digestibility, a diet that enables their rumens to remain in a fi t  and adaptive state. 
However, future studies need to be carried out to confirm and give a better 
understanding of these effects. 
4.5. CONCLUSIONS 
1 )  The technique using plant sequences with substantial genotype contrasts in 
morphology and secondary compound concentration was effective in determining 
the preference of cattle. 
2) Plant morphology played a very important role in diet selection. The animals 
seemed to prefer large, dense and leafy plants to small, sparse and stemmy plants, 
and showed an immediate discrimination between plants on the basis of 
morphology. 
3) Differences in formononetin concentration between red clover plants were 
influenced more by plant morphological characteristic than by seasonal effects. 
Plant morphology was substantially more important than formononetin 
concentration in affecting diet selection by cattle. 
4) The concentration of 3 .2% of ECT in birdsfoot trefoil was apparently high enough 
to cause an immediate negative effect on preference. Concentrations below 1 .8 % 
apparently did not have an immediate influence on cattle preference. However this 
effect was also confounded with plant morphology characteristics (mainly leafiness) 
between genotypes. On the other hand, the contrast of animal preference between 
periods was associated with variation in EeT concentration. This suggests a basis 
for seasonal variations in preference for birdsfoot trefoil. 
5) The composition of rumen content affected diet selection by cattle. The preference 
for red clover when red clover was inserted in the rumen suggests that there is a 
Experiments 5 and 6 2 1 8  
positive post-ingestive feedback that affects cattle preference. Although no effects 
of tannin and fonnononetin concentration were found when inserted in the rumen, 
this conclusion needs to be treated with caution since only low concentrations and 
contrasts of these secondary compounds were used in these studies. Further research 
is required to clarify this result. 
CHAPTER S 
GENERAL DISCUSSION 
The main objective of Experiments 5 and 6 (E5 and E6) was to provide an explanation 
for the selective behaviour observed in Experiments 1 ,  2 and 3 (El,  E2 and E3) in 
relation to how physical and biochemical plant characteristics influenced animal 
preference and to what extent this influence explained seasonal variations in preference. 
This general discussion will therefore focus on these issues. However, it is important to 
distinguish animal response between experiments. In E l ,  E2 and E3 the animals 
expressed their selection ("preference modified by opportunity", Hodgson, 1 979). In E5 
and E6, although preferential behaviour was also modified by variations in plant 
morphology or structure, similar opportunity for access to a range of spaced plants was 
provided and the animals could express their preference with little constraint. This 
discussion will deal first with the general pattern of selective behaviour observed in E l ,  
E2 and E3, and then on the deviation from the general behaviour in response to specific 
sward characteristics and possible explanations for this behaviour. 
The overall partial preference demonstrated between the birds foot trefoil/white clover 
(BW) and red clover (RC) swards was close to 50:50 (based on grazing time) in E l  and 
E3 and 40:60 in E2 (section 3.4.4). There was thus relative stability of partial preference 
across experiments, matching well with the non-significant difference in preference 
between birdsfoot trefoil (BT) and RC demonstrated in an overall analysis of E5 (5.9 vs 
5.4 ± 0.637 bites/plant of BT and RC respectively, P = 0. 1066) and to the value close to 
50:50 reported by Torres-Rodiguez ( 1 997). However, the partial preference was shown 
to be sensitive to manipulations of sward height, herbage mass and plant morphology. 
This pattern of general stability but responsiveness to specific manipulation is also 
apparent in the series of studies on combinations of perennial ryegrass and white clover 
swards by Newman et al. ( 1 992); Parsons et al. ( 1994a) and Cos grove et al., ( 1 996). 
General Discussion 220 
Observations showed that animals did not graze randomly but modified behaviour to 
meet their preference for a mixed diet. This was demonstrated on Day 1 of E l ,  when 
animals allocated a disproportionately large time to graze in the minor swards (Table 
3.8,  Figure 3 .3), even though they could have satisfied appetite by grazing only in one 
sward. This result is in agreement with Parsons et al. ( 1 994a) who found that sheep did 
not graze at random when different proportions in ground area of white clover in 
relation to perennial ryegrass were offered. In addition, although E5 and E6 showed that 
the preference between BT and RC was largely influenced by plant height and herbage 
mass, in E l  Day 1 ,  when herbage mass and height of both swards were relatively high 
and both sward were in similar maturity stage, the animals allocated proportionally 
more time (in relation to herbage mass offered) grazing on the sward with lower mass 
(Table 3.9, Figure 3 .8). This behaviour also demonstrated that, with high levels of 
herbage mass and height, animals preference for a mixed diet was stronger than the 
desire to graze swards which were taller or with greater herbage mass. 
The objective of obtaining a mixed diet was also modified by sward physical 
characteristics. E5 and E6 clarified these effects. In E l ,  E2 and E3, the animals 
demonstrated selection for swards with greater herbage mass and height (except in E l  
Day 1 when the swards were in similar maturity stage, and sward height and herbage 
mass was high - see above) and proportion of leaves. This selective behaviour was 
illustrated by the selection of swards with greater herbage mass in Day 3 of E l  (Table 
3.9, Figure 3.8), and Day 1 and 3 of E3 (see section 3.3.3 . 1  - Grazing time per kg of DM 
offered), by the selection of swards with greater height in Day 1 of E2 (Tables 3. 1 1 , 
3 . 1 9) and E3 (Tables 3.23, 3.30) and by the selection of swards with greater leaf/stem 
ratio in E2, mainly in Day 1 (Tables 3. 1 2, 3 . 1 9). In E5 and E6, when given a greater 
opportunity to demonstrate their preference, the animals also showed preference for 
plants with greater height, volume, density and higher proportion of leaves (Tables 4.5, 
4. 19, 4.20). These results therefore confirmed that the objective of obtaining a mixed 
diet in E l ,  E2 and E3 was strongly influenced by the sward physical characteristics. 
The preference observed in E5 and E6 for taller plants, with larger area and a greater 
proportion of leaves can explain the sward maturity effect on selection in E2. The 
General Discussion 22 1 
selection for RC rather than BW sward in E2 was probably related to the greater 
herbage mass, height and higher proportion of leaves of RC than of BT. The high 
correlation in E5 and E6 between leafmess and number of bites, mainly for BT plants 
(Tables 4. 17, 4.36), the substantial effect of leafiness when included as a covariate in 
the analysis of variance of number of bites taken from BT plants (covariance analyses 
Figures, 4.2 - 4. 13) and the pattern of behaviour across experiments in the discussion of 
Chapter 3 (Table 3.36) showed that plant leafiness (or leaf/stem ratio) had a strong 
effect on selection. Therefore, considering the fact that variation in selective behaviour 
between seasons was closely related to plant maturity (see discussion in Chapter 3), it 
can be concluded that in swards formed mainly by BT or RC the variation in selection 
across seasons is mainly related to variation of leaf/stem ratio contrasts between swards, 
though this effect may be modified by sward structure (mainly height and herbage 
mass). This complex of physical effects on diet selection agrees with the conclusion of 
Real-Ferreiro ( 1997), where preference of RC cultivars could not be determined by 
simple morphological characters alone. 
In addition to sward physical characteristics, biochemical characteristics could also have 
affected selection (Hodgson et al., 1994; Provenza, 1995; Launchbaugh, 1996). In El,  
E2 and E3 it was not clear how secondary compounds affected selective behaviour, how 
they were related to the sward physical characteristics or whether they made any 
contribution to the variation in selective behaviour observed in different periods of the 
year. E5 and E6 clarified the effect of the formononetin concentration in RC and ECT 
concentration in BT on animal preference. 
In E5 and E6 the animals were substantially more directly influenced by the 
morphological characteristics of RC than by the concentration of formononetin, 
agreeing with the conclusions of Harborne ( 1993). In both sets of experiments there was 
a relatively small variation, within each RC genotype among different experiments and 
periods, in formononetin concentration, smaller than that observed by Anwar ( 1994), 
suggesting that formononetin concentration would not have influenced the seasonal 
variation in selection observed in El ,  E2 and E3. In addition to the small direct effect of 
formononetin concentration observed in E5 and E6, it also did not have an important 
General Discussion 222 
post-ingestive feedback effect in E6: considering only the case where RC was inserted 
into the rumen, there was no significant effect of fonnononetin concentration on 
preference for RC cultivars. Therefore, it can be concluded that the concentration of 
fonnononetin apparently did not contribute to the variation in selective behaviour 
observed in E 1 ,  E2 and E3. 
E5 and E6 also provided an understanding of the effect of ECT concentration on 
preference and its relation with plant morphological characteristics. The correlation and 
covariance analyses carried out in E5 and E6 showed that animals responded 
immediately by selecting against the BT genotype containing high ECT concentration. 
However, the negative effect of ECT concentration on preference was confounded with 
variation in plant morphology. The main confounding effect observed was between 
plant leafiness and ECT concentration: plant leafmess had a positive effect on 
preference, ECT concentration had a negative effect. These experiments also showed 
that the effect of ECT concentration on animal preference was observed only in 
concentrations of approximately 3 .2% (Period 2 of E5). Concentrations below 
approximately 1 .8% (Period 3 of E5) did not have an immediate effect on preference. In 
addition to these results, the modification of rumen content in E6 also demonstrated that 
low levels of ECT concentration in the rumen did not have an important effect on 
preference for BT genotypes (Table 4.30). Although several studies and reports 
(Provenza and Malechek, 1984; Provenza et al., 1990; Provenza et al., 1994; Provenza, 
1995) argue that condensed tannin causes a negative post-ingestive feedback on 
preference, in E6 preference was not affected by the low rumen loading with ECT. It 
can be concluded that the selective behaviour in E l ,  E2 and E3, where there was a 
relatively low range of variation of ECT in BT (0.54 to 1 .69 %) (Table 3 .36), was 
unlikely to have been affected by ECT concentration. However, further studies on the 
effect of higher ECT concentrations in the rumen on the post-ingestive feedback of 
condensed tannin need to be carried out for a better understanding of the cattle selective 
response to condensed tannin. 
Whittaker and Feeny ( 1971 ), Roades and Cates ( 1976) and Lindroth ( 1 989) suggested 
that condensed tannin acts as a protection against herbivory. The higher concentration 
General Discussion 223 
of ECT in leaves than in stems (Table 4.8, 4.25), demonstrated also by the significant 
correlations between ECT concentration and leafiness (Tables 4. 15, 4.34), and the fact 
that relatively high concentrations of ECT reduce preference for BT, suggest a 
hypothesis that the plant produces ECT concentration in the most preferred plant part to 
inhibit grazing. Thus, it can be inferred that relatively high ECT concentration in leaves 
of BT may increase the persistence of birdsfoot trefoil in a mixed sward by reducing the 
preference for this species, if persistence of BT is adversely affected by selective 
grazing. However, the fact that the animals were attracted by leaves of BT and rejected 
stems in E l ,  E2 and E3 indicated that the low ECT concentration was not effective in 
protecting the plants against grazing. 
The importance of leaf/stem ratio and secondary compound concentration in affecting 
the selective and preferential behaviour in these studies shows scope for manipulation 
of preferential behaviour. In erect legumes like BT and RC, stems make an important 
contribution to the total herbage mass, but also affect negatively preferential behaviour. 
Therefore, it can be inferred that sward managements or plant breeding programmes 
that increase the leaf/stem ratio of either BT or RC will result in an improvement in 
animal preference. However the increase in proportion of leaves in BT needs to be 
associated with low ECT concentration. Leaves of BT have much higher concentration 
of ECT than stems (Tables 4.8, 4.25), indicating a potentially negative effect on 
preference. 
The selective behaviour observed in El ,  E2 and E3, where the animals grazed to obtain 
a mixed diet and were also influenced by sward physical characteristics, may be 
explained by several hypotheses proposed in the literature. The hypothesis that selection 
is related to the secondary compound concentration present in the sward and that the 
animals prefer a mixed diet to minimise the risk of toxicity (Freeland and Janzen, 1974; 
Laycock et al. ,  1988) was not relevant in this experiment because the two most 
important secondary compounds in the swards did not have any apparent effect on diet 
selection. However it is important to recognise that there is a potential effect on 
preference of ECT when in high concentrations in BT and that the current studies were 
General Discussion 224 
limited only to the effects of the major secondary compounds (BeT and formononetin) 
of BT and RC. 
The other possible hypothesis is that animals graze to obtain a balanced diet (Newman 
et al. ,  1992). The overall stability of partial preference across E l ,  E2 and E3, discussed 
in the beginning of this chapter, associated with the fact that the animals grazed the 
minor sward component when both swards had relatively high mass and height and 
were in a similar stage of maturity, suggests that the animals may have targeted a 
balanced diet reflecting partial preference for BW and RC close to 50:50. However, at 
relatively low herbage mass and sward surface height, the animals modified their 
behaviour and selected the taller and greater herbage mass sward, grazing in proportion 
to the herbage mass offered (El -Table 3.9, Figure 3.7; E3 - section 3.3.3 . 1 - Grazing 
time per kg of DM offered). This behaviour may indicate that the animals adjusted their 
selection towards swards that provided potentially greater intake rate. This result also 
suggests that there is a trade off between preference and intake. Parsons et al. ( 1994b) 
and Newman et al. ( 1995) models explain that changes in diet selection behaviour, to 
include more of the less preferred species in the diet, may allow animals to maintain 
intake as the total herbage mass available declines. This interpretation also agrees with 
studies showing that a choice between alternative forages or patches strongly favours 
those with greater potential intake rate (Laca et al. ,  1993; Kenney and Black, 1984; and 
Black and Kenney, 1984). 
The hypothesis that the animal sampled in both swards to constantly reinforce 
awareness of the sward conditions (mius and Gordon, 1990; filius et al., 1987) could 
also help to explain selective behaviour in these studies. Edwards ( 1994), working in 
carefully controlled conditions demonstrated that when food "patches" (bowls 
containing varying types of pellets) remained in the same location, sheep learned the 
location quickly and did not need to continue a sampling strategy. However, he also 
pointed out that changeable sward conditions could make reliance on memory alone of 
limited effectiveness. In fact, the current studies suggest that sampling was a consistent 
feature that helped animals to adjust their selection according to changing sward 
conditions. 
GENERAL CONCLUSIONS 
The overall conclusions in relation to the most significant findings from these studies on 
the effects of morphological and biochemical characteristics of birds foot trefoil and red 
clover on cattle preference are summarised below. 
1 )  The methodologies used in the two sets of experiments were effective in 
determining and explaining grazing behaviour in studies involving animal response 
to physical and biochemical sward characteristics. 
2) There was overall a relative stability in preference between birdsfoot trefoil and red 
clover across experiments with an average partial preference of approximately 
50:50. However, this preference was shown to be flexible once it was modified by 
sward height, herbage mass, plant morphology and EeT concentration. 
3) Birdsfoot trefoil/white clover and red clover sward structures had important effects 
on selective behaviour. The animals were attracted by the tallest swards with the 
greatest herbage mass. However, on tall and high mass swards (in these studies, 
approaching 4000 kglha) this selection was modified by the preference for a mixed 
diet. 
4) Plant maturity had an important effect on preferential grazing behaviour, mediated 
through effects on sward structure and leaf/stem ratio. 
5) Formononetin concentration between 0.26 and 0.73 % on a DM basis did not affect 
preference for red clover. Differences in physical characteristics of the two 
genotypes of red clover had greater effects on preference than differences in 
formononetin concentration. 
General Conclusions 226 
6) Preference of cattle was affected by ECT concentration in birdsfoot trefoil. However 
this effect was confounded with associated contrasts in plant morphology, mainly 
leafiness, and was only apparent in relatively high concentrations of EeT (in these 
studies 3.2% on a DM basis, but not below 1 .8% on a DM basis). 
7) The absence of discrimination between red clover and a birdsfoot trefoil genotype 
with low ECT concentration contrasts with the lower preference for a birdsfoot 
trefoil genotype with high EeT concentration. It may be inferred that animals would 
not show preference between red clover and birdsfoot trefoil cultivars with low EeT 
concentration if the physical characteristics were similar. Therefore, it can be 
concluded that birdsfoot trefoil and red clover were suitable species for determining 
the effects of sward physical characteristics on selective behaviour. 
8) Leaf/stem ratio was one of the most important plant morphological characteristics 
that affected selection between swards based on birdsfoot trefoil and red clover 
species. The fact that animals usually preferred leaves and rejected stems caused 
variations in selection between swards differing in maturity. The preference for leaf 
also influenced the effect of EeT concentration on preference. 
9) Seasonal variability in selection between swards formed by strips of birdsfoot trefoil 
and red clover species could be explained by the effect of plant maturity where the 
animals demonstrated preference mainly for swards with high leaf/stem ratio, 
provided the EeT concentration in BT was not high enough to adversely affect this 
preference. 
10) In addition to the fact that the animals were attracted by swards with greater 
leaf/stem ratio, three hypotheses provide possible explanations for the selective 
behaviour observed in El ,  E2 and E3: (i) animals tried to obtain a balanced diet; (ii) 
animals selected swards that provided the potentially higher rate of intake; (iii) 
animals sampled to constantly reinforce awareness of the sward conditions. 
General Conclusions 227 
1 1 ) Sward managements or plant breeding programmes that increase the leaf/stem ratio 
of either birdsfoot trefoil or red clover will result in an improvement in animal 
preference. However this improvement in proportion of leaves in birdsfoot trefoil 
needs to be associated with low EeT concentration. 
REFERENCES 
AlIden, W.R. 1962. Rate of herbage intake and grazing time in relation to herbage 
availability. Proceeding of Australian Society of Animal Production 4: 163- 166. 
ADden, W.G., and Whittaker, I.A. 1970. The detennination of herbage intake by 
grazing sheep: the interrelationship of factors influencing herbage intake and 
availability. Australian Journal of the Agricultural Research 2 1 :  755-766. 
Allison, Mol. 1978. The role of ruminal microbes in the metabolism of toxic 
constituents of plants. In: Effects of Poisonous Plants on Livestock. Keeler, R.F., 
VanKampen, K.R., and James, L.F. (eds.). Academic Press, New York, pp. 101-
120. 
Anwar, M. 1994. Formononetin content in selected red clover strains and its effects on 
reproduction in ewes. Ph.D. Thesis, Massey University. 
Arave, C.W., PureeD, D., and Engstrom, M. 1989. Effect of feed flavours on 
improving choice for a ten per cent meat and bone meal dairy concentrate. 
Journal of Dairy Science 72 (Suppl 1 ), 563. 
Arditi, R., and Daeorogna, B. 1988. Optimal foraging on arbitrary food distributions 
and the definition of habitat patches. The American Naturalist 1 3 1 :  837-846. 
Armstrong, R.H., Robertson, E., and Hunter, E.A. 1995. The effect of sward height 
and its direction of change on the herbage intake, diet selection and performance 
of weaned lambs grazing ryegrass swards. Grass and Forage Science 50: 389-
398. 
References 229 
Armstrong, R.H., Robertson, E., Lamb, C.S •• , Gordon, I.J., and Elston, D.A. 1993. 
Diet selection by lambs in ryegrass-white clover swards differing in the 
horizontal distribution of clover. Proceedings of the XVII International 
Grassland Congress, New Zealand I: 7 15-7 16. 
Arnold, G.W. 1960. Selective grazing by sheep of two forage species at different stages 
of growth. Australian Journal of Agricultural Research 1 1 : 1026-1033. 
Arnold, G.W. 1964. Some principles in the investigation of selective grazing. 
Proceedings of the Australian Society of Animal Production 5 :  258-27 1 .  
Arnold G.W. 1966a.The special senses in grazing animals I. Sight and dietary habits in 
sheep. Australian Journal of Agricultural Research 17 :  521 -529. 
Arnold G.W. 1966b.The special senses in grazing animals IT. Smell, taste, and touch 
and dietary habits in sheep. Australian Journal of Agricultural Research 17 :  
53 1 -542. 
Arnold. G.W. 1970. Regulation of food intake in grazing ruminants. In: Physiology of 
Digestion and Metabolism in the Ruminant. Phillipson, A.T.(ed.). Oriel Press, 
Newcastle, pp. 264-276. 
Arnold G.W. 1981. Grazing behaviour. In: Grazing Animals. Morley F.H.W. (ed.). 
Elsevier, Netherlands, pp. 79-104. 
Arnold G.W. 1987. Influence of the biomass, botanical composition and sward height 
of annual pastures on foraging behaviour by sheep. Journal of Applied Ecology 
24: 759-772. 
References 230 
Arnold, G.W., and Hill, J.L. 1972. Chemical factors affecting selection of food plants 
by ruminants. In: Photochemical Ecology. Annual Proceedings of the 
Photochemical Society, N° 8. Harbone, J.L. (ed.), pp.7 1 - 1 0 1 .  
Arnold, G.W., and Maller, R.A. 1977. Effects of nutritional experience in early life 
and adult life on the performance and dietary habits of sheep. Applied Animal 
Ecology 3:  5-26. 
Austin, Pol., Suchar, L.A., Robbins, C.T., and Hagerman. 1989. Tannin-binding 
proteins in saliva of deer and their absence in saliva of sheep and cattle. Journal 
of Chemical Ecology 15(4) : 1335-1347. 
Bailey, D.W., Rittenhouse, R.H., Hart, R.H., and Richards, R.W. 1989. 
Characteristics of spatial memory in cattle. Applied Animal Behaviour Science 
23: 33 1-340. 
Barnard, C..J. 1980. Flock feeding and time budgets in the house sparrow (Passer 
domesticus L. ). Animal Behaviour 28: 295-309. 
Barry, T.N. 1985. The role of condensed tannins in the nutritional value of Lotus 
pedunculatus for sheep. 3. Rates of body and wool growth. British Journal of 
Nutrition 54: 2 1 1 -217. 
Barry, T.N. 1989. Condensed tannins: their role in ruminant protein and carbohydrate 
digestion and possible effects upon the rumen ecosystem. In: The Roles of 
Protozoa and Fungi in Ruminant Digestion. Nolan, J.V.; Leng, R.A. ; Demeyer, 
DJ. (eds.). Penambul Books, Armidale, pp. 153- 169. 
Barry, T.N., and Blaney, T.R. 1987. Secondary compounds of forages. In: The 
Nutrition of Herbivores. Hacker, J.B., and Temouth, J.H.(eds.). Academic Press, 
Sydney, N.S.W, pp. 92- 1 19. 
References 23 1 
Barry, T.N., and Forss, D.A. 1983. The condensed tannins in the nutritional value of 
Lotus pedunculatus, its regulation by fertiliser application, and effects upon 
protein solubility. Journal of the Science of Food and Agriculture 34: 1047-
1 056. 
Barry, T.N., and Manley, T.R. 1984. The role of condensed tannins in the nutritional 
value of Lotus pedunculatus for sheep. 2. Quantitative digestion of 
carbohydrates and proteins. British Journal of Nutrition 5 1 :  493-504. 
Barry, T.N., and Manley, T.R. 1986. Interrelationships between the concentrations of 
total condensed tannins, free condensed tannins and lignin in Lotus spp. and 
their possible consequences in ruminant nutrition. Journal of the Science of 
Food and Agriculture 37: 248-254. 
Barthram, G.T. 1986. Experimental techniques: The HFRO sward stick. Biennal Report 
1984-1985. Hill Farming Research Organisation, Penicuik, pp. 29-30. 
Barthram, G.T., and Grant, S.A. 1984. Defoliation of ryegrass-dominated swards by 
sheep. Grass and Forage Science 39: 2 1 1 -2 19. 
Bate-Smith, E.C. 1972. Attractants and repellents in higher animals. In: Phytochemical 
Ecology. Harbome, J.B. (ed.).  Academic Press, New York, pp.45-56. 
Bate-Smith, E.C. 1973. Haemanalysis of tannins : the concept of relative astringency. 
Phytochemistry 12:  907-9 12. 
Bazely, D.R. 1990. Rules and cues used by sheep foraging in monocultures. In: 
Behavioural mechanisms offood selection. Hughes R.N. (ed.). NATO ASI series 
Vol. G20, Springer-Verlag, Berlin, pp. 343-367. 
References 232 
Bazeley, D.R., and Ensor, C.V. 1989. Discrimination learning in sheep with cues 
varying in brightness and hue. Applied Animal Behaviour Science 23 : 293-299. 
Bedell, T .E. 1973. Botanical composition of sub clover pastures as affected by single 
and dual grazing by cattle and sheep. Agronomy Journal 65: 502-504. 
Bell, W J. 1991. Searching Behaviour. The behavioural ecology of finding resources. 
University Press, Cambridge. 
Bell, F.R., and Sly, J. 1976. The assessment of sodium appetite in calves using operant 
conditioning procedures. Journal of Physiology 263 : 1 78- 1 79. 
Bell, F.R., and Sly, J. 1977. The specificity of sodium appetite in calves. Journal of 
Physiology 272: 60-6 1 .  
Bell, F.R., and Sly, J. 1983. The olfactory detection of sodium and lithium salts by 
sodium deficient cattle. Physiology and Behavior 3 1 :  307-3 1 3. 
Bircham, J.M. 1981. The effects of change in herbage mass on herbage growth, 
senescence and net production rates in a continuously stocked mixed species 
swards. In: Plant Physiology and Herbage Production. Wright, C.E. (ed). 
Occasional Symposium N° 1 3 ,  British Grassland Society, pp. 85-87. 
Bircham, J.M., and Hodgson, J. 1983. The influence of sward conditions on rates of 
herbage growth and senescence in mixed swards under continuous stocking 
management. Grass and Forage Science 38:  323-33 1 .  
/ 
Black, J.L., and Kenney, P.A. 19 • Factors affecting diet selection by sheep. IT. 
Height and density of pasture. Australian Journal of Agricultural Research 35: 
565-578. 
References 233 
Brelin, B. 1979. Mixed grazing with sheep and cattle compared with single grazing. 
Swedish Journal of Agricultural Research 9: 1 1 3- 1 20. 
Briske, D.D. 1986. Plant responses to defoliation: Morphological considerations and 
allocation priorities. In: Proceedings Second International Rangelands 
Congress. PJ.Joss, P.W. Lynch, and O.B. Williams (ed.). Australian Academy 
of Sciences, Canberra, pp.425-427. 
Briske, D.D. 1996. Strategies of plant survival in grazed systems: A functional 
interpretation. In: The Ecology and Management of Grazing Systems. Hodgson, 
J., and mius, A.W. (ed.). CAB Intematonal, Wallingford, pp. 37-67. 
Brown, G.S. and Gass, C.L. 1993. Spatial association learning by hummingbirds. 
Animal Behaviour 46: 487-497. 
Bryant, J.P., Reichardt, P.B., and Clausen, T.P. 1992. Chemically mediated 
interactions between woody plants and browsing mammals. Journal of Range 
Management 45: 1 8-24. 
Burritt, E.A., and Provenza, F.D. 1989a. Food aversion learning: conditioning lambs 
to avoid a palatable shrub (Cerocarpus montanus). Journal of Animal Science 
67: 650-653. 
Burritt, E.A., and Provenza, F.D. 1989b. Food aversion learning: ability of lambs to 
distinguish safe from harmful foods. Journal of Animal Science 67: 1732- 1 739. 
Burlison A.J., Hodgson J., and Illius A.W. 1991. Sward canopy structure and the bite 
dimensions and the bite weight of grazing sheep. Grass and Forage Science 46: 
29-38. 
References 234 
Burrit, E.A., and Provenza, F.D. 1991. Ability of lambs to learn with a delay between 
food ingestion and consequences given meals containing novel and familiar 
foods. Applied Animal Behaviour Science 32: 1 79- 1 89. 
Bush L., and Burton H. 1994. Intrinsic chemical factors in forage quality. In: Forage 
quality, evaluation, and utilization. Fahey G.c. Jr. (ed.). American Society of 
Agronomy, Inc., Crop Science Society of America, Inc. ,  and Soil Science 
Society of America, Inc. ;  Madison, Wisconsin, pp. 367-405. 
Butler, B.M. 1991. The packages of tissue turnover analysis and point quadrat 
analysis. Massey University, New Zealand. 
Cahn, M. G., and Harper, J. L. 1976. The biology of the leaf mark polymorphism in 
Trifolium repens L. 2. Evidence for the selection of leaf marks by rumen 
fistulated sheep. Heredity 37(3): 327-333. 
Carter, R.R., and Grovum, W.L. 1990. A review of the physiological significance of 
hypertonic body fluids on feed intake and rumen function: salivation, motility 
and microbes. Journal of Animal Science 68: 28 1 1 -2832. 
Chacon E., and Stobbs T.H. 1976. Influence of progressive defoliation of a grass 
sward on the eating behaviour of cattle. Australian Journal of Agricultural 
Research 27: 709-727. 
Chapman, R.F., and Blaney, W.M. 1979. How animals perceive secondary 
compounds. In: Herbivores: their interactions with secondary plant metabolites. 
Rosenthal, G.A., and Janzen, D.H. (eds.). Academic Press, New York, pp. 1 6 1 -
1 98.  
References 235 
Church, D.C. 1979. Taste, appetite and regulation of energy balance and control of 
food intake. In: Digestive physiology and nutrition of ruminants. Church, D.C. 
(ed.). Nutrition, Vol. 2, 0& B Books Inc, Corvallis, pp. 28 1 .  
Clark D.A., and Harris P.S. 1985. Composition of the diet of sheep grazing swards of 
differing white clover content and spatial distribution. New Zealand Journal of 
Agricultural Research 28: 233-240. 
Clark, D.A., Lambert, M.G., Rolston, M.P., and Dymock, N. 1982. Diet selection by 
goats and sheep on hill country. Proceedings of the New Zealand Society of 
Animal Product 42: 155- 1 57. 
Clarke, I.D., Frey, R.W, and Hyland, H.L. 1939. Seasonal variation in tannin content 
of lespedeza sericea. Journal of Agricultural Research 58: 1 3 1 - 139. 
Coley, P.D. 1983. Herbivory and defensive characteristics of tree species in a lowland 
tropical forest. Ecological Monographs 53: 209-233.  
Collins, M. 1989. Single and mixed grazing o f  cattle, sheep and goats. Ph.D. Thesis, 
Lincoln College. 
Collins, W.B., Urness, P.J.m and Austin, D.D. 1978. Elk diets and activities on 
different lodgepole pine habitat segments. Journal of Wildlife Management 42: 
799-8 10. 
Coon, E.E. 1979. Cyanids and cyanogenic glycosids. In: Herbivores: Their Interaction 
with Secondary Plant Metabolites. Rosenthal, G.A., and Janzen, D.H. (eds). 
Academic Press, New Yourk, pp. 387-412.  
References 236 
Cooper, S.D.B., Kyriazakis, I., and Nolan, J.V. 1995. Diet selection in sheep: the role 
of the rumen environment in the selection of a diet from two feeds that differ in 
their energy density. British Journal of Nutrition 74: 39-54. 
Cope, W.A., Bell, T.A., and Smart, Jr., W.W.G. 1971. Seasonal changes in an 
enzyme inhibitor and tannin content in Serica Lespedeza. Crop Science 1 1 : 893-
895. 
Cope, W.A., and Burns, J.C. 1971. Relationship between tannin levels and nutritive 
value of sericea lespedeza. Crop Science 1 1 : 23 1-233. 
Cosgrove, G.P., Anderson, C.B., and F1etcher, R.H. 1996. Do cattle exhibit a 
preference for white clover? In: White Clover: New Zealand's competitive edge. 
Woodfield (ed.). Proceedings of a joint symposium between Agronomy S ociety 
of New Zealand and New Zealand Grassland Association held at Lincoln 
University, New Zealand 2 1 -22 November, 1 995, pp. 83-86. 
Cosgrove, G.P., Anderson C.B., and F1etcher R.H. 1997. Species preference 
influences on cattle grazing behaviour. XVIII International Grassland Congress 
I: section 5-7. 
Cosgrove, G.P., and Mitchell, R.J. 1995. Effect of sward type on intake rate 
parameters during progressive defoliation by lambs. Annales de Zootechnie 44 
(Suppl): 249. 
Coughenour, M.B. 1991. Spatial components of plant-herbivore interactions in pastoral 
ranching and native ungulate ecosystems. Journal of Range Management 44: 
530-542. 
Cowie, J.D. 1972. Soil map and extended legend of Kairanga county. New Zealand Soil 
Bureau Publication. 538 p. 
References 237 
Crawley, MJ. 1983. Herbivory: the dynamics of animal-plant interactions, Blackwell 
Scientific Publications, London. 437 p. 
Crawley, W., and Pacala, S. 1991. Herbivores, plant parasites and plant diversity. 
In: Parasitism: conflict of coexistence. Toft, C. (ed.). Oxford University Press, 
pp. 157- 174. 
Cumming, D.H.M. 1982. The influence of large herbivores on savanna structure in 
Africa. In: The ecology of tropical savannas. Huntley, B.J., Walker, B .H. (ed.). 
Ecological Studies Vol. 42, Berlin, Springer-Verlag, pp. 2 17-245. 
CurD, M.L., and Gleeson, A.C. 1987. The introduction of red or white clover into a 
perennial grass sward. Grass and Forage Science 42: 397-403. 
Curll, M.L., and Wilkins, R.J. 1980. The relationship between selective grazing by 
sheep and the botanical composition of a grass/clover sward. European 
Grassland Federation General Meeting Proceedings, Forage Production Under 
Marginal Conditions, pp. 7. 17-7.23. 
Davies, H. L.; Rossiter, R.e., and MaUer, R. A. 1970. The effects of different 
cultivars of subterranean clover (T. subterraneum L.) on sheep reproduction in 
south-west of Western Australia. Australia Journal of agricultural resesarch 
2 1 :  359-369. 
Demment, M.W., Distel., R.A., Griggs, T.C., Laca, E.A., and Deo, G.P. 1993. 
Selective behaviour of cattle grazing ryegrass swards with horizontal 
heterogeneity in patch height and bulk density. Proceedings of the XVII 
International Grassland Congress I: 7 12-7 14. 
References 238 
Demment, M.W., Peyraud, J.-L., and Laca, E.A. 1995. Herbage intake at grazing: a 
modelling approach. In: Recent Developments in the Nutrition of Herbivores. 
Journnet, M., Grenet, E., Farce, M-H., Theriez, Demarquilly, C.(eds). INRA, 
Paris, pp. 137- 14 1 .  
Donnelly, E.D. 1959. The effect of season, plant maturity, and light on the tannin 
content of sericea lespedeza, L. cuneata. Agronomy Journal 5 1 :  7 1 -73. 
Donelly, E.D., and Anthony, W.B. 1969. Relationship of tannin, dry matter 
digestibility and crude protein in Serica lespedeza. Crop Science 9: 36 1-362. 
Dougherty C.T., Collins M., Bradley N.W., Lauriault M.,and Cornelius P.L. 1992. 
The proloxalene on ingestion by cattle grazing lucerne. Grass and Forage 
Science 47: 1 80- 199. 
Douglas, G.B., Donkers, P., Foote, A.G., and Barry, T. 1993. Determination of 
extractable and bound condensed tannins in forages species. Proceedings of the 
VXII International Grassland Congress I: 204-206. 
Douglas, G.B., Wang, Y., Waghorn, G.C., Barry, T.N., Purchas, R.W., Foote, 
A.G.,and WilsoD, G.F. 1995. Liveweight gain and wool productionn of sheep 
grazing Lotus corniculatus and lucerne (Medicago sativa). New Zealand Journal 
of Agricultural Research 38: 95- 104. 
Dudzinski, M. L., and Arnold, G. W. 1973. Comparisons of diets of sheep and cattle 
grazing together on sown pastures on the southern tablelands of New South 
Wales by principal components analysis. Australian Journal of Agricultural 
Research 24: 899- 9 1 2. 
References 239 
Edwards G.R. 1994. The creation and maintenance of spatial heterogeneity in plant 
communities: the role of plant-animal interactions. Ph.D. Thesis, University of 
Oxford. 1 80 p. 
Edwards, G.R., Newman, J.A., Parsons, A.J., and Krebs, J.R. 1996a. Effects of the 
total, vertical and horizontal availability of the food resource on diet selection 
and intake sheep. Journal of Agricultural Science Cambridge 1 27: 555-562. 
Edwards, G.R., Parsons, A.J., Newman, J.A., and Wright, I.A. 1996b. The spatial 
pattern of vegetation in cut and grazed grass/ white clover pastures. Grass and 
Forage Science 5 1 :  2 1 9-23 1 .  
Fenny, P. 1969. Inhibitory effect of oak leaf tannins on the hydrolysis of proteins by 
trypsin. Phytochemistry 8 :  2 1 1 9-2 1 26. 
Fisher, D.S., Burns, J.C., and Mayland, H.F. 1997. Variation in preference for 
morning of afternoon harvested hay in sheep, goats, and cattle. Journal of 
Animal Science 75 (Supplement). 
Forbes, J.M. 1996. Voluntary food intake and diet selection in farm animals. CAB 
International, Wallingford, 532p. 
Forbes J.M., and Kyriazakis I. 1995. Food preferences in farm animals: why don't  
they always choose wisely? Proceedings of the Nutrition Society 54: 429-440. 
Forbes, T.D.A., and Hodgson, J. 1985. Comparative studies of the influence of sward 
conditions on the ingestive behaviour of cows and sheep. Grass and Forage 
Science 40: 69-77. 
Frame, J., and Newbould, P. 1986. Agronomy of white clover. Advances in Agronomy 
40: 1 - 1 88. 
References 240 
Francis, C.M. 1973. The influence of isoflavone glycosides on the taste of subterranean 
clover leaves. Journal of the Science of Food and Agriculture 24: 1 235- 1 240 
Francis, C.M., and Millington, A.l. 1965. Varietal variation in the isoflavone content 
of subterranean clover: its estimation by a microtechnique. Australian Journal of 
Agricultural Research 16 :  557-564. 
Francis, C.M., Millington, A.l., and Bailey, E.T. 1967. The distribution of 
oestrogenic isoflavones in the genus Trifolium. Australian Journal of 
Agricultural Research 18 :  47-54. 
Freeland, W.l., Calcott, P.H., and Anderson, L.R. 1985. Tannins and saponins: 
interaction in herbivore diets. Biochemical Systematic Ecology 1 3 :  1 89- 1 93.  
Freeland, W.l., and janzen, D. 1974. Strategies in herbivory by mammals: The role of 
plant secondary compounds. The American Naturalist 1 08(96 1 ): 269-289. 
Furstenburg, D., and van Hoven, W. 1994. Condensed tannin as anti-defoliate agent 
against browsing by giraffe (Giraffa camelopardalis) in the Kruger National 
Park. Comparative Biochemistry and Physiology 107A: 425-43 1 .  
Gammon, D.M., and Roberts,B.R. 1978. Patterns of defoliation during continuous 
and rotational grazing of the Matopos Sandveld of Rhodesia. 1 .  Selectivity of 
grazing. Rhodesia Journal Agricultural Research 16:  1 17- 1 3 1 .  
Ganskopp, D.A., Angell, R., and Rose, J. 1993. Effect of low densities of senescent 
stems in crested wheat-grass on plant selection and utilisation by beef cattle. 
Applied Animal Behaviour Science 38: 227-233. 
References 241 
Garcia, J. 1989. Food for Tolman: cognition and cathexis in concert. In: Aversion 
Avoidance and Anxiety. Archer, T., and Nilsson, L. (eds.). Lawrence Erlbaum 
and Associates, HiUsdade, New Jersey, pp. 45-85. 
Garcia, J., and Hankins, W.G. 1975. The evolution of bitter and the acquisition of 
toxiphobia. In: Olfaction and Taste. Denton, D., and Coghlan, J. (eds). 
Academic Press, Vol. 5, New York, pp. 39-4 1 .  
Garcia, J., and Hankins, W.G. 1977. On the origin of food aversion paradigms. In: 
Learning mechanisms in food selection. Barker, L., Best M., Domjan, M. (eds.). 
Baylor University Press, Waco. 
Gardener, C.J. 1980. Diet selection and live weight performance of steers on 
Stylosanthes hamata-native grass pastures. Australian Journal of Agricultural 
Research 3 1 :  379-392. 
Garner, F.H. 1963. The palatability of herbage plants. Journal of the British Grassland 
Society 1 8 :  79-89. 
Gary, L.A., Sherrit, G.W. and Hale, E.B. 1970. Behaviour of charolais cattle on 
pasture. Journal of Animal Science. 30: 203-206. 
Georgiadis, N.J., and McNaughton, S.J. 1988. Interactions between grazers and a 
cyanogenic grass, Cynodon plectostachyus. Oikos 5 1 :  343-350. 
Gibb, M.J., Huckle, C.A, Nuthall, R., and Rook, A.J. 1997. Effect of sward surface 
height on intake and grazing behaviour by lactating Holstein Friesian cows. 
Grass and Forage Science 52: 309-3 2 1 .  
References 242 
Gibb, MJ., and Ridout, M.S. 1988. Application of double nonnal frequency 
distributions fitted to measurements of sward height. Grass and Forage Science 
43: 1 3 1 - 1 36. 
Goatcher, W.D., and Church, D.e. 1970a. Taste responses in ruminants. 1. Reactions 
of sheep to sugars, saccharin, ethanol and salts. Journal of Animal Science 30: 
777-783. 
Goatcher, W.D., and Church, D.C. 1970b. Taste responses in ruminants. ID. 
Reactions of pygmy goats, nonnal goats, sheep and cattle to sucrose and sodium 
chloride. Journal of Animal Science 3 1 :  364-372. 
Gong, Y. 1994. Comparative studies of effects of sward structure on ingestive 
behaviour of sheep and goats grazing grasses and legumes. Ph.D. Thesis, Massey 
University. 
Gong, Y., Hodgson, J., Lambert, M.G., Chu, A.C.P., and Gordon, I.L. 1993. 
Comparisons of response patterns of bite weight and bite dimensions between 
sheep and goats grazing a range of grasses and clovers. Proceedings of the XVII 
International Grassland Congress, New Zealand I: 726-727. 
Gong, Y., Lambert, M.G., and Hodgson, J. 1996. Effects of contrasting sward heights 
within forage species on short-tenn ingestive behaviour of sheep and goats 
grazing grasses and legumes. New Zealand Journal of Agricultural Research 39: 
83-93. 
Gordon, I.J. 1988. Facilitation of red deer grazing by cattle and its impact on red deer 
perfonnance. Journal of Applied Ecology 25: 1-9. 
Gordon, IJ. 1989a. Vegetation community selection by ungulates on the Isle of Rhum. 
I. Food supply. Journal of Applied Ecology 26: 35-52. 
------------
- -
References 243 
Gordon, I.j. 1989b. Vegatation community selection by ungulates on the Isle of 
Rhum. IT. Vegetation community selection. Journal of Applied Ecology 26: 53-
64. 
Gordon, I.j. 1989c. Vegatation community selection by ungulates on the Isle of Rhum. 
rn. Detenninants of vegetation community selection. Journal of Applied Ecology 
26: 65-79. 
Gordon I.j., and Lascano C. 1993. Foraging strategies of ruminant livestock on 
intensively managed grasslands: potential and constraints. Proceedings of the 
XVII International Grassland Congress, New Zealand I: 68 1 -690. 
Gosden, A.F., and Jones, R. 1978. A routine method for predicting the formononetin 
content of red clover. Journal of the Science of Food and Agriculture 29: 925-
929. 
Grant, S.A., Suckling, D.E., Smith, H.K., TorveD, L., Forbes, T.D.A., and Hodgson, 
J. 1985. Comparative studies of diet selection by sheep and cattle: The hill 
grasslands. Journal of Ecology 73: 987- 1004. 
Griffiths, W.M., Hodgson, J., Amold, G.C., Hoimes, C.W. 1997. Influence of 
vegetation patch characteristics on discriminatory grazing. Proceeding of the 
XVIII International Grassland Congress, Canada I: section 5- 1 .  
Griffiths, W.M., Hodgson, J., Holmes, C.W., Amold, G.C. 1996. The use of a novel 
approach to detennine the influence of sward characteristics on the 
discriminatory grazing behaviour of dairy cows. Proceedings of the New 
Zealand Society of Animal Production 56: 1 22- 1 29. 
References 244 
Grill, HJ., Berridge, K.C., and Ganster, DJ. 1984. Oral glucose is the prime elicitor 
of preabsorptive insulin secretion. American Journal of Physiology 246: R88-
R95. 
Guy, M.C., Watkin, B.R. , and Clark, D.A. 1981. Effects of season , stocking rate and 
grazing duration on the diet selected by hoggets grazing mixed grass-clover 
pastures. New Zealand Journal of Experimental Agriculture 9: 141- 146. 
Harborne, J.B. 1993. Introduction in Ecological Biochemistry, 3rd edn. Academic 
Press, San Diego California. 3 1 8p. 
Hart, B.L. 1985. The behavior of domestic animals. W.H. Freeman, New York. 390 p. 
Hedges, D.A., Wheeler, J.L., Mulcahy, C., and Vincent, M.S. 1978. Composition and 
acceptability to sheep of twelve summer forage crops. Australian Journal of 
Experimental Agriculture and Animal Husbandry 18 :  520-526. 
Hendricksen, R.W., and Minson, DJ. 1981. The feed intake and grazing behaviour of 
cattle grazing a crop of Lablab purpureus cv. Rongi. Journal of Agricultural 
Science, Cambridge 95: 547-554. 
Herms, D.A., and Mattson, W.J. 1992. The dilemma of plants: to grows or defend. 
Quaterly Review of Biology 67: 283-3 13 .  
Hodge, R.W., and Doyle, JJ. 1967. Diet selected by lambs and yearling sheep grazing 
on anual and perennial pastures in southern Victoria. Australian Journal of 
Experimental Agriculture and Animal Husbandry 7: 141- 143. 
Hodgson, J. 1979. Nomenclature and definitions in grazing studies. Grass and Forage 
Science 34: 1 1- 18 .  
---------- -
References 245 
Hodgson, J. 1981a. Testing and hnprovement of Pasture Species. In: Grazing animals. 
Morley, F.H.W. (ed.). Elsevier Scientific Publishing Company, B.,  pp. 309-3 17.  
Hodgson, J. 1981b. Variations in the surface characteristics of the sward and the short­
term rate of herbage intake by calves and lambs. Grass and Forage Science 36: 
49-57. 
Hodgson, J. 1982a. Influence of sward characteristics on diet selection and herbage 
intake by the grazing animal. In: Nutritional limits to animal production from 
pastures. Hacker, J.B. (ed.). Commonwealth Agricultural Bureaux, London, pp. 
1 53- 1 66. 
Hodgson, J. 1982b. Ingestive behaviour. In: Herbage intake handbook. Leaver J.D. 
(ed.). The British Grassland Society, Berkshire, pp. 1 1 3- 1 38. 
Hodgson J. 1985. The control of herbage intake in the grazing ruminant. Proceedings 
of the Nutrition Society 44: 339-346. 
Hodgson, J. 1990. Grazing management. Science into practice. United States: Longman 
Scientific & Technical. 224 p. 
Hodgson, J., Clark, D. A., and Mitchell, R. J. 1994. Foraging behaviour in grazing 
animals and its impact on plant communities. In: Forage Quality, Evaluation, 
and Utilization. Fahey G.c. et al. (eds.). American Society of Agronomy, Inc. 
Crop Science Society of America, Inc. Soil Science Society of America, Inc., pp. 
796-828. 
Hodgson, J., Cosgrove, G.P., Woodward, S.J.R. 1997. Research on foraging 
behaviour: progress and priorities. Proceedings of the XVIII International 
Grassland Congress, Canada (in press). 
References 246 
Hodgson, J., and Jamieson, W.S. 1981. Variations in herbage mass and digestibility, 
and the grazing behaviour and herbage intake of adult cattle and weaned calves. 
Grass and Forage Science 36: 39-48. 
Hodgson, J., and Ollerenshaw, J.H. 1969. The frequency and severity of defoliation 
of individual tillers in set stocked swards. Journal of the British Grassland 
Society 24: 226-234. 
Hoffman, R.R. 1989. Evolutionary steps of ecophysiological adaptation and 
diversification of ruminants: a comparative view of their digestive system. 
Oecologia 78: 443-457. 
Holland, O. 1994. Seasonal dynamics of digestion in relation to diet quality and intake 
in European roe deer. (Capreolus). Oecologia 99: 274-279. 
Holmes, C.W. 1987. Pastures for dairy cattle. In: Livestock Feeding on Pasture. New 
Zealand Society of Animal Production. Occasional Publication N° 10, pp. 1 33-
142. 
Hughes, T.P., Sykes, A.R., and Poppi, D.P. 1984. Diet selection of young ruminants 
in late spring. Proceedings of the New Zealand Society of Animal Production 44: 
109- 1 12. 
Hughes, T.P., Sykes, A.R., Poppi, D.P., and Hodgson, J. 1991. The influence of 
sward structure on peak bite force and bite weight in sheep. Proceeding of the 
New Zealand Society of Animal Production 5 1 :  153- 158. 
Hull, J.L., Lofgreen, G.P., and Meyer, J.H. 1960. Continuous versus intermittent 
observations in behaviour studies with grazing cattle. Journal of Animal Science 
19: 1204-1207. 
References 247 
Iason, G.R., Hodgson, J., and Barry, T. 1995. Variation in condensed tannis 
concentration of a temperate grass (Holcus lanatus) in relation to season and 
reproductive development. Journal of Chemical Ecology 2 1 :  1 103- 1 1 12. 
Illius, A.W., Clark, D.A., and Hodgson, J. 1992. Discrimination and patch choice by 
sheep grazing grass-clover swards. Journal of Animal Ecology 6 1 :  1 83- 194. 
Illius, A. W., and Fitzgibbon, C. 1994. Costs of vigilance in foraging ungulates. Animal 
Behaviour 47: 48 1 -484. 
Illius A.W., and Gordon IJ. 1987.The allometry of food intake in grazing ruminants. 
In: Herbivore nutrition research. Second international symposium on the 
nutrition of herbivores. Australian Society of Animal Production. Brisbane, pp 
103-104. 
Illius, A.W., and Gordon, IJ. 1990. Constraints on diet selection and foraging 
behaviour in mammalian herbivores. In: Behavioural Mechanisms of Food 
Selection. Hughes, R.N.(ed.). NATO ASI Series, Vol. G 20, pp. 369-393. 
Illius, A.W., and Hodgson, J. 1996. Progress in understanding the ecology and 
management of grazing systems. In: The Ecology and Management of Grazing 
Systems. Hodgson, J., and lllius, A.W. (eds.). CAB Intematonal, Wallingford, 
pp. 429-457. 
Illius, A.W., Wood-Gush, D.G.M., Eddison, J.C. 1987. A study of the foraging 
behaviour of cattle grazing patchy swards. Biology of Behaviour 12: 33-44. 
Jackson, P., Hodgson, J., and Rook, J.A.f. 1968. The voluntary intake of acetate by 
dairy cows given ammonium salts of short chain fatty acids their drinking water. 
Animal Production 10: 473-48 1 .  
References 248 
Jackson F.S., McNabb, W.C., Barry, T.N., Foo, Y.L. and Peters J.S. 1996. The 
condensed tannin content of an range of subtropical and temperate forages and 
the reactivity of condensed tannin with ribulose l ,5-his-phosphate caboxylase 
(Rubisco) Protein. Journal of Science of Food and Agriculture 72: 483-492. 
Jaramillo, V.J., and Detling, J.K. 1992. Small-scale patch heterogeneity in a semi­
arid North American grassland. IT. Cattle grazing of simulated urine patches. 
Journal of Applied Ecology 29: 9- 1 3. 
Joblin, A.D.H. 1960. The influence of night grazing on the growth rates of Zebu cattle 
in East Africa. Journal of the British of Grassland Society 15 :  2 12-2 15 .  
John, A., and Lancashire, J.A. 1981. Aspects of feeding value of Lotus spp. 
Proceddings of the New Zealand Grassland Association 42: 152- 159. 
Jones, W.T., Anderson, L.B., and Ross, M.D. 1973. Bloat in cattle. New Zealand 
Journal of Agricultural Research 16: 441-446. 
Jones, W.T., Broadhurst, R.B., and Lyttleton, J.W. 1976. The condensed tannins of 
pasture legume species. Phytochemistry 15 :  1407- 1409. 
Jones, W.T., and Lyttleton, J.W. 1971. A survey of legume forages that do and do not 
produce bloat. New Zealand Journal of Agricultural Research 14: 101- 107 .  
Juko, C.D., and Bredon, R.M. 1961. The chemical composition of leaves and whole 
plant as an indicator of the range of available nutrients for selective grazing by 
cattle. Tropical Agriculture 38: 179- 187. 
Kacelnik, A., and Bernstein, C. 1988. Optimal foraging and arbitrary food 
distributions:  patch models gain a lease of life. Trends in Ecology and Evolution 
3(10):  25 1-253. 
References 249 
Kalat, J.W. 1974. Taste salience depends on novelty, not concentration in taste­
aversion learning in the rat. Journal of Comparative and Physiological 
Psychology 86(1 ) : 47-50. 
Kelly, R.W., Hay, R.J.M., and Scbackell, G.H. 1979. Formononetin content of 
'Grasslands Pawera' red clover and its oestrogenic activity to sheep. New 
Zealand Journal of Experimental Agriculture 7: 1 3 1 - 134. 
Kendrick, K.M. 1992. Cognition . In: Farm Animals and the Environment. Phillips, 
C.J.c., and Piggins, D. (eds). CAB International, Wallingford, pp.209-23 1 .  
Kenney P.A., and Black J.L. 1984a. Factors affecting diet selection by sheep. I. 
Potential intake rate and acceptability of feed. Australian Journal of Agricultural 
Research 35: 55 1-563. 
Kenney P.A., and Black J.L. 1984b. Factors affecting diet selection by sheep. IT. 
Height and density of pasture. Australian Journal of Agricultural Research 35: 
565-578. 
Kenney, P.A., and Black, J.L. 1986. Effect of simulated sward structure on the rate of 
intake of subterranean clover by sheep. Proceeding of Australian Society of 
Animal Production 35: 55 1-563. 
Keogb, R.G. 1995. Oestrogens in Pastures. Proceedings of the Nutrition Society of New 
Zealand 20: 52-61 .  
Keogb, R.G., Kramer, R., McDonald, Blewman, A., and Crabb, J. 1996. The use of 
blood equol values in devising grazing management strategies for red clover­
based pastures. Proceedings of the New Zealand Grassland Association 58: 
265-269. 
References 250 
Kyriazakis, I., Anderson, D.H., and Duncan, AJ. 1998. Conditioned flavour 
aversions in sheep: the relationship between the dose rate of a secondary plant 
compound and the acquisition and persistence of aversions. British Journal of 
Nutrition 79: 55-62. 
L'Huillier, P. J., Poppi, D. P., and Fraser, T. J. 1984. Influence of green leaf 
distribution on diet selection by sheep and the implications for animal 
performance. Proceedings of the New Zealand Society of Animal Production 44: 
105-1 07. 
L'Hulller, PJ., Poppi, D.P., and Fraser, TJ. 1986. Influence of structure and 
composition of ryegrass and prairie grass-white clover swards on the grazed 
horizon and diet harvested by sheep. Grass and Forage Science 4 1 :  259-267. 
Laca E.A. and Demment M. W. 1996. Foraging strategy of grazing animals. In: The 
Ecology and Management of Grazing Systems. Hodgson, J., and lllius, A.W. 
(eds.). CAB Intematonal, Wallingford, pp. 1 37-158. 
Laca, E.A., Demment M.W., Distel R.A., and Griggs T.e. 1993. A conceptual model 
to explain variation in ingestive behaviour within a feeding patch. Proceedings 
of the XVII International Grassland Congress, New Zealand I: 7 10-7 12. 
Laca, E.A., Distel, R.A., Griggs, T.e., and Demment, M.W. 1994. Effects of canopy 
structure on patch depression by grazers. Ecology 75: 706-7 16. 
Laca, E.A., Distel, R.A., Griggs, T.e., Deo, G., and Demment, M.W. 1993. Field 
test of optimal foraging with cattle: the marginal value theorem successfully 
predicts patch selection and utilisation. Proceedings of the XVII International 
Grassland Congress, New Zealand I: 709-7 10. 
References 25 1 
Laca E.A., Ungar E.D., Seligman N., and Demment M.W. 1992. Effects of sward 
height and bulk density on bite dimensions of cattle grazing homogeneous 
swards. Grass and Forage Science 47: 91- 102. 
Lascano, C.E. and Thomas, D. 1988. Forage quality and animal selection of Arachis 
pintoi in association with tropical grasses in the castern plains of Colombia. 
Grass and Forage Science 43: 433-439. 
Launchbaugh, K.L. 1996. Biochemical aspects of grazing behaviour. In: The Ecology 
and Management of Grazing Systems. Hodgson, J., and mius, A.W. (eds.). CAB 
Intematonal, Wallingford, pp. 159- 184. 
Launchbaugh and Provenza, 1993. Can plants practice mimicry to avoid grazing by 
mammalian herbivores? Oikos 66: 501-506. 
Launcbbaugh, K.L., Provenza, F.D., and Burrit, E.A. 1993. How herbivores track 
variable environments: responses to variability of toxins. Journal of Chemical 
Ecology 19:  1047- 1056. 
Laycock, W.A., Young, J.A., and Ueckert, D.N. 1988. Ecological status of poisonous 
plants on rangelands. In: The Ecology and Economic Impact of Poisonous Plants 
on Livestock Production. James, L.F., Ralphs, M.H., and Nielson, D.B. (eds). 
Westview Press, Boulder, Colorado, pp. 27-42. 
Lechner-Doll, M., Rutagwenda, T., Schwartz, HJ., Schultka, W., and Engelbardt, 
W. v. 1990. Seasonal changes of ingesta mean retention time and forestomach 
volume in indigenous grazing camels, cattle, sheep and goats on a thornbush 
savanna pasture. Journal of Agricultural Science, Cambridge 1 15 :  409-420. 
References 252 
Ledgard, S.F., Steele, K.W., and Saunders, W.H.M. 1982. Effects of cow urine and 
its major constituents on pasture properties. New Zealand Journal of 
Agricultural Research 25: 61-68. 
Leigh, J.H., and Holgate, M.D. 1978. Effects of pasture availability on the 
composition and quanlity of the diet selected by sheep grazing nativem 
degenerate and improved pastures in the Upper Shoalhaven Valley, New South 
Wales. Australian Journal of Experimental Agriculture and Animal Husbandry 
18 :  38 1 -390. 
Leigh, J. H., and Mulham, W. E. 1966a. Selection of diet by Sheep grazing semi-arid 
pastures on the Riverine Plain 1 .  A bladder Saltbush (Atriplex vesicaria) -
Cotton Bush (Kochia aphylla) community. Australian Journal of Experimental 
Agriculture and Animal Husbandry 6: 460-467. 
Leigh, J. H., and Mulham, W. E. 1966b. Selection of diet by Sheep grazing semi-arid 
pastures on the Riverine Plain 2. A Cotton Bush (Kochia aphylla) - grassland 
(Stipa variabilis - Danthonia caespitosa) community. Australian Journal of 
Experimental Agriculture and Animal Husbandry 6: 468-474. 
Lendrem, D.W. 1983. Predation risk and vigilance in the blue tit (Parus caeruleus). 
Behavioral Ecology and Sociobiology 14: 9- 13 .  
Li, Y -G., Tanner, G., and Larkin, P. 1996. The DMACA-HCL protocol and the 
threshold proanthocyanidin content for bloat saafety in forage legumes. Journal 
of the Science of Food and Agriculture 79: 89-101 .  
Lindroth, R.L. 1989. Mammalian herbivore-plant interactions. In :  Plant-Animal 
Interactions. Abrahamson, W.G. (ed.). McGraw-Hill Book Co., New York, pp. 
163-205. 
References 253 
Lowther, W.L., Manley, T.R., and Barry, T.N. 1987. Condensed tannin 
concentrations in Lotus corniculatus and L. pedunculatus cultivars grown under 
low soil fertility conditions. New Zealand Journal of Agricultural Research 30: 
23-25. 
Lubcheno, J. 1978. Plant species diversity in a marine intertidal community: 
importance of herbivore food preference and algal competitive abilities. The 
American Naturalist 1 12: 23-39. 
MacGraw, R.L., Beuselinck, P.R., and Marten, G.C. 1989. Agronomic and forage 
quality attributes of diverse. entries of birdsfoot trefoil. Crop Science 29: 1 160-
1 164. 
Malechek, J.C., and Balph, D.F. 1987. Diet selection by grazing and browsing 
livestock. In: The Nutrition of Herbivores: Second Interntional Symposium on 
I 
the Nutrition of Herbivores. Hacker, J.B., and Temouth, J.H.(eds.). Academic 
Press, Sydney, pp. 199-201 .  
Marshall, T. 1973. Clover disease - what we know and what we can do. Journal of 
Agriculture, Western Australia (Series 4) 14: 198-206. 
Marten, G. c., and Anderseo, R. N. 19'/5. Forage nutritive value and palatability of 
1 2  common annu� weeds. Crop Science 15 :  82 1-827. 
Marten, G.e. and Jordan, R.l\1. 1974. Significance of palatability differences among 
Phalaris arundinacea L., Bromus inermis Leyss. and Dactylis glomerata L. 
grazed by sheep. Proceedings of the XII International Grassland Congress, 
Moscow 3(1) :  305-3 12. 
Math Soft. 1996. S-plus 3.4 for Unix, Seattle. 
References 254 
McDonald, M.F. 1995. Effects of plant oestrogens in ruminants. Proceedings of the 
Nutrition Society of New Zealand, 20: 43-5 1 .  
McKey, D. 1979. The distribution of secondary compounds within plants. In: 
Herbivores, their interaction with secondary plant metabolites. Rosenthal, G.W. 
and Janzen, D.H. (eds.). Academic Press, New York, pp. 55- 1 33.  
McLaughlin, C.L., Baldwin, B.A., and Baile, C.A. 1974. Olfactory bulbectomy and 
feeding behavior. Journal of Animal Science 39: 1 36. 
McLeod, M.N. 1974. Plant tannins - their role in forage qUality. Nutrition Abstracts and 
Reviews 44: 803-8 15.  
McMurray, C.H., Laidlaw, A.S. and McElroy, M. 1986. The effect of plant 
development and environment on formononetin concentration in red clover 
(Trifolium pratense L.). Journal of the Science of Food and Agriculture 37: 333-
340. 
Mehansho, H., Butler, L.G., and Carlson, D.M. 1987. Dietary tannins and salivary 
proline-rich proteins: Interactions, induction, and defence mechanisms. Annual 
Review of Nutrition 7: 423-440. 
Metcalfe, N.B., and Fumess, R.W. 1984. Changing prioritities: the effect of pre­
migratory fattening on the trade-off between foraging and vigilance. Behavioral 
Ecology and Sociobiology 15 :  203-206. 
Millington, A..}., Francis, C.M., and McKeown, N.R. 1964. Wether bioassay of 
annual pasture legumes, n. The oestrogenic activity of nine strains of Trifolium 
subterraneun L. Australian Journal of Agricultural Research 15 :  527-536. 
References 255 
Milne J.A., Hodgson J., Thompson R., Souter W.G., and Barthram G.T. 1982. The 
diet ingested by sheep grazing swards differing in white clover and perennial 
ryegrass content. Grass and Forage Science 37: 209-2 18 .  
Mitchell R.J., Hodgson J., and Clark D.A. 1991. The effect of varying leafy sward 
height and bulk density on the ingestive behaviour of young deer and sheep. 
Proceedings of the New Zealand Society of Animal Production 5 1 :  1 59-1 65.  
Mitchell, RJ., Hodgson, J., Clark, D.A., and Anderson, C.B. 1993. The independent 
effects of sward height and bulk density on the bite parameters of Romney ewes 
and red deer hinds. Proceedings XVII International Grassland Congress, New 
Zealand I: 704-706. 
Montossi, F., Hu, Y.,Hodgson, J., and Morris, S.T. 1994. Herbage intake, ingestive 
behaviour and diet selection in sheep grazing Holcus lanatus and perennial 
ryegrass swards. Proceedings of the New Zealand Society of Animal Production 
54: 7 1-74. 
Motl, JJ. 1985. Mosaic grazing-animal selectivity in tropical savannas of northern 
Australia. Proceedings XV International Grassland Congress, Japan 1 129-1 130. 
Mursan, A., Hughes, T.P., Nicol, A.M., and Suguira, T. 1989. The influence of sward 
height on the mechanics of grazing in steers and bulls. Proceedings of the New 
Zealand Society of Animal Production 4 1 :  233-236. 
Neil, H.G., and Marshall, T. 1970. Superphosphate deficiency raises pasture 
oestrogens. Journal of Agriculture, Western Australia (Series 4) 1 1 (2) : 43-44. 
New Zealand, AgResearch Grassland Research Centre. 1995. Variety: 'Grassland 
G27'. Aplication N° 94/2 13 .  Plant Varities Journal 8: 1 , 29-30, 33.  
References 256 
Newman, J.A., and Parsons, A.J. 1993. A model of the interaction between grazing 
mammals and a two species sward. Journal of Agricultural Science, Cambridge 
1 2 1 :  284. 
Newman J.A., Parsons A.J., and Harvey A. 1992. Not all sheep prefer clover: diet 
selection revisited. Journal of Agricultural Science, Cambridge 1 19: 275-283. 
Newman J.A., Parsons A.J., Thornley J.H.M. and Penning P.D. 1995. Optimal diet 
selection by a generalist grazing herbivore. Functional Ecology 9: 255-268. 
Newman J.A., Penning P.D., Parsons A.J., Harvey A., and Orr R.J. 1994. Fasting 
affects intake behaviour and diet preference of grazing sheep. Animal Behaviour 
47: 1 85- 193. 
Nicol, A.M., and Nicoll, G.B. 1987. Pastures for beef cattle. In: Livestock feeding on 
pasture. New Zealand Society of Animal Production. Occasional Publication N° 
10, pp 1 19- 120. 
Nicol, A.M., Russel, A.J.F., and Wright, I.A. 1993. Integrating grazing of goats with 
sheep or cattle on continuously grazed pasture .Proceeding XVII International 
Grassland Congress, New Zealand I: 1320- 1322. 
Nicollier, G.F., and Thompson, A.C. 1982. Separation and quantitation of estrogenic 
isoflavones from clovers by high-performance liquid chromatography. Journal 
of Chromatography (note) 249: 399-402 
Nolan, T., and Connolly, J. 1977. Mixed stocking by sheep and steers - a review. 
Herbage abstracts 47: 367-374. 
References 257 
Norton, B.W., Kennedy, P.J., Hales, J.W. 1990. Grazing management studies with 
Australian cashmere goats. 3. Effect of season on the selection of diets by cattle, 
sheep and goats from two tropical grass-legume pastures. Australian Journal of 
Experimental Agriculture 30: 783-788. 
N.R.C. (National Research Council) 1989. Nutrient requirements of dairy cattle. 6th 
Revised edition. National Academy Press. Washington, D.C. 1 57p. 
N.R.C. (National Research Council) 1996. Nutrient requirements of beef cattle. 7th 
Revised edition. National Academy Press. Washington, D.C. 242p. 
O'Regain, P.J. 1993. Plant structure and the acceptability of different grasses to sheep. 
Journal of Range Management 46: 232-236. 
O'Reagain P.J., and Mentis M.T. 1989. The effect of plant structure on the 
acceptability of different grass species to cattle. Journal of the Grassland Society 
of South Africa 6: 163- 170. 
O'Reagain P.J., and Schwartz J. 1995. Dietary selection and foraging strategies of 
animals on rangeland. Coping with spatial and temporal variability. In: Recent 
developments in the nutrition of herbivores. M. Journet, E. Grenet, M.H. Farce, 
M.  T.heriez and C. Demarquilly (eds). INRA editions, Paris, pp 407-423. 
Owen-Smith, R.N. 1994. Foraging response of kudus to seasonal changes in food 
resources: elasticity in constraints. Ecology 75: 1050- 1062. 
Parsons A.J., Newman J.A., Penning P.D., Harvey A., and Orr R.J. 1994a. Diet 
preference of sheep: effects of recent diet, physiological state and species 
abundance. Journal of Animal Ecology 63: 465-478. 
References 258 
Parsons AJ., Thomley J.H.M., Newman J., and Penning P.D. 1994h. A mechanistic 
model of some physical detenninants of intake rate and diet selection in a two­
species temperate grassland sward. Functional Ecology 8: 1 87-204. 
Penning, P.D. 1986. Some effects of sward conditions on grazing behaviour and intake 
by sheep. In: Grazing Research as Northern Latitudes. Gudmundssin, O. (ed.). 
Proceedings of a NATO Advanced Workshop, Vol. 108. Hvanneyri, Iceland, pp. 
2 19-226. 
Penning, P.F., Parsons, AJ. , Newman, J.A., Orr, R., and Harvey, A. 1993. The 
effects of group size on time budgets in grazing sheep. Applied Animal Behavio'r 
Science 37: 101- 109. 
Penning P.D., Rook A.J., and Orr RJ. 1991. Patterns of ingestive behaviour of sheep 
continuously stocked on monocultures of ryegrass or white clover. Applied 
Animal Behaviour Science 3 1 :  227-250. 
Pfister, J.A., Provenza, F.D., Manners, G.D., Gardner, D.R., and Ralphs, M.H. 
1997. Tall larkspur ingestion: can cattle regulate intake below toxic levels? 
Journal of Chemical Ecology 23(3): 759-777. 
Philips, C.J.C., and Leaver, J.D. 1985. Seasonal and diurnal variation in the grazing 
behaviour of dairy cows. In: Grazing. Frame J. (ed.). British Grassland Society. 
Occasional symposium. N° 19, pp. 98-104. 
Poppi D.P., Hughes T.P., and L'Huillier PJ. 1987. Intake of pasture by grazing 
ruminants. In: Livestock feeding on pasture. New Zealand Society of animal 
Production. Occasional Publication No. 10. New Zealand, pp. 55-64. 
Provenza F.D. 1995. Postingestive feedback as an elementary detenninant of food 
preference and intake in ruminants. Journal of Range Management 48: 2- 17.  
References 259 
Provenza F.D. 1996a. Familiarity and novelty in animal diets : implications for 
management. Proceedings of the Australian Society of Animal Production 2 1 :  
12- 16. 
Provenza F .D. 1996b. Acquired aversions as the basis for varied diets of ruminants 
foraging on rangelands. Journal of Animal Science 74: 2010-2020. 
Provenza, F.D., and Balph, D.F. 1988. The development of dietary choice in livestock 
on ragelands and its implications for management. Journal of Animal Science 
66: 2356-2368. 
Provenza, F.D., and Balph, D.F. 1990. Applicability of five diet-selection models to 
various foraging challenges ruminants encounter. In: Behavioural mechanisms of 
food selection . .  Hughes R.N. (ed.). NATO ASI series Vol. G20. Springer-Verlag, 
Berlin, pp. 422-459. 
Provenza, F.D., Burrit, E.A., Clausen, T.P., Bryant, J.P., Reichardt, P.B., and 
Distel, R.A. 1990. Conditioned flavour aversion: A mechanism for goat to avoid 
condensed tannisin blackbrush. The American Naturalist 136(6): 8 10-828. 
Provenza, F.D., Lynch, J-J., and Nolan, J.V. 1993. The relative importance of mother 
and toxicosis in the selection of foods by lambs. Journal of Chemical Ecology 
19:  3 1 3-323. 
Provenza, F.D., Lynch, J.J., and Nolan, J.V. 1994. Food aversion conditioned in 
anaesthetized sheep. Physiology and Behavior 55: 429-432. 
Provenza, F.D., and Malechek, J.C. 1984. Diet selection by domestic goats in relation 
to blackbrush twig chemistry. Journal of Applied Ecology 2 1 :  83 1-841 .  
References 260 
Provenza, F.D., pnster, J.A., and Cheney, C.D. 1992. Mechanisms of learning in diet 
selection with reference to phytotoxicosis in herbivores. Journal of Range 
Management 45: 36-45. 
Pulliam, H.R., and Caraeo, T. 1984. Living in groups: is there an optimal group size? 
In: Behavioural Ecology. Krebs, 1.R., and Caraco, T. (eds.). Blackwell 
Scientific, Oxford, pp. 122-147. 
Rattray, P.V., Thompson, K.F., Hawker, H., and Summer, R.M.W. 1987. Pastures 
for sheep production. In: Livestock Feeding on Pasture. New Zealand Society of 
Animal Production. Occasional Publication N° 10. pp. 89-103.  
Real-Ferreiro, D. 1997. Quantitative genetics of sheep preference in red clover 
(Trifolium Pratense L.) under spaced plant and sward conditions. PhD Thesis, 
Massey University. 
Rhodes, I., and Collins, R. 1993. Canopy structure. In: Sward Measurement Handbook. 
2nd Edition. Davis, A, Baker, R.D., Grant, S.A, and Laidlaw, AS.(eds.). British 
Grassland Society, pp. 139. 3 19. 
Ridout, M.S., and Robson, M.J. 1991. Composition of the diet of sheep grazing 
swards of differing white clover content and spatial distribution: a re-evaluation. 
New Zealand Journal of Agricultural Research 34: 89-93. 
Roades, D.F., and Cates, R.G. 1976. Towards a general theory of plant antiherbivore 
chemistry. Recent Advanced Phytochemistry 10: 168-2 13.  
Roberts, C.A., Beuselinek, M.R., EUersieck, Davis, D.K., and MeGraw, R.L. 1993. 
Quantification of Tannis in Birdsfoot Trefoil Germplasm. Crop Science 33: 
675-679. 
References 261 
Roberts G.E., and Packman, R.G. 1983. Feed infonnation and animal production. 
Proceedings of the second symposium of the international network of feed 
information, 507pp. 
Robbins, C.T., Hanley, T.A., Hagerman, A.E., Hjeljord. 0., Baker, D.L., Schwart, 
C.C., and Mantz. 1987a. Role of tannis in defending plant against ruminants: 
Reduction in protein availability. Ecology 68(1) :  98- 107. 
Robbins, C.T., Mole, S., Hagerman, A.E., and Hanley, T.A. 1987b. Role of tannis in 
defending plant against ruminants: Reduction in dry matter digestion. Ecology 
68(6): 1607- 1615 .  
Rosenthal, G.A., and Bell, E.A. 1979. Naturally occurring, toxic nonprotein amino 
acids. In: Herbivores: Their Interactions with Secondary Plant Metabolites. 
Rosenthal, G.A., and Jazen, D.H. (eds). Academic Press, New York, pp. 353-
386. 
Rossiter, R.C. 1970. Factors affecting the oestrogen content of subterranean clover 
pastures. Australian Veterinary Journal 46: 141-144. 
Rossiter, R.C., and Becker, A.B. 1967. Physiological and ecological studies on the 
oestrogenic isoflavones in subterranean clover (T. subterraneum L. ) 5. 
Oestrogenic changes. ). Australian Journal of Agricultural Research 18: 561-
573. 
Rossiter, R.C., and Ozanne, P.G. 1970. In: Australian Grasslands. Moore, M.R. (ed.) .  
Australian National University Press, 218 p. 
SAS Institute Inc., 1985. SAS User's guide: Statistics .. Version 5. Sixth edition. SAS 
Inc, Cary, North Carolina, USA. 
References 262 
SAS Institute Inc. 1990. SAS User's guide. Version 6. Fourth edition . .  SAS Inc, Cary, 
North Carolina, USA. 
Senft, R.L., Coughenour, M.B., Bailey, D.W., Rittenhouse, L.R., Sala, O.E., and 
Swift, D.M. 1987. Large herbivore foraging and ecological hierarchies. 
BioScience 37: 789-799. 
Shenk, J.s., and Westerhaus, M.O. 1994. The application of near infrared reflectance 
spectroscopy (NIRS) to forage analysis. In: Forage quality, evaluation, and 
utilization. Fahey, G.C. Jr. (ed.). American Society of Agronomy, Inc. Madison, 
Wisconsin, pp. 406-449. 
Shewmaker, G.E., Mayland, H.F., and Hansen S.B. 1997. Cattle grazing preference 
among eight endophyte-free tall fescue cultivars. Agronomy Journal 89(4):695-
701 .  
Shutt, D.A., and Braden, A.W.H. 1968.The significance of equol in relation to the 
oestrogenic responses in sheep ingesting clover with a high formononetin 
content. Australian Journal of Agricultural Research 19:  545-553. 
Shutt, D.A., Weston, R.H., and Hogan, J.P. 1970. Quantative aspects of phyto­
oestrogen metabolism in sheep fed on subterranean clover (Trifolium 
subterranean cultivar Clare) or red clover (Trifolium pratense). Australian 
Journal of Agricultural Research 2 1 :  7 13-722. 
Simon, U. 1974. Palatability and voluntary intake of contrasting legume and grass 
varieties by grazing sheep. Proceedings of the XII International Grassland 
Congress, Moscow 3(1) :  487-494. 
Smith, G.S. 1992. Toxification and detoxification of plant compounds by ruminants: an 
overview. Journal of Range Management 45: 25-30. 
References 263 
Smith, G.R., Randel, R.D., and Bradshaw, C. 1986. Influence of harvest date, 
cultivar, and sample storage method on concentration of isoflavones in 
subterranean clover. Crop Science 26: 1013-1016. 
Spalinger, D.E., and Hobbs, N.T. 1992. Mechanisms of foraging in mammalian 
herbivores: new models of functional response. The American Naturalist 140: 
325-348. 
Stephens, D.W., and Krebs, J.K. 1986. Foraging Theory. Princeton University Press, 
Princetown, New Jersey, 247pp. 
Stitt, R.E., and Clarke, I.D. 1941. The relation of tannin content of sericea lespedeza 
to season. Journal of the American Society of Agronomy 33: 739-742. 
Stobbs, 1973a Stobbs T.R. 1973. The effect of plant structure on the intake of tropical 
pastures. 1 .  Variation in the bite size of grazing cattle. Australian Journal of 
Agricultural Research 24: 809-8 19. 
Stockdale, C.R., and DeIIow, D.W. 1995. The productivity of lactating dairy cows 
grazing White Clover and supplemented with maize silage. Australian Journal 
of Agricultural Research 46: 1205- 121 17.  
Stubbs, 0.1., and Kare, M.R. 1958. Taste preferences of cattle. Journal of Animal 
Science 17:  1 162. 
Stuth J.W. 1991. Foraging behaviour. In: Grazing management. An ecological 
perspective. Heitschmidt R.K. and Stuth J.W. (eds). Timber press, Oregon, pp. 
65-83. 
References 264 
Swain, T. 1979. Tannis and Lignins. In: Herbivores: Their Interaction with Secondary 
Plant Metabolites. Rosenthal, G.A., and Janzen, D.H.(eds). Academic Press, pp. 
657-682. 
Taylor, J.A. 1993. Foraging strategy. Proceedings of the XVII International Grassland 
Congress, New Zealand I: 739-740. 
Terrill, T.H., Rowan, A.M., Douglas, G.B., and Barry, T.N. 1992. Determination of 
extractable and bound eT concentrations in forage plants, protein concentrate 
meals and cereal grains. Journal of the Science Food and Agriculture 58: 321 -
329. 
Theron, E.P., and Booysen, P. de V. 1966. Palatability in grasses. Proceeding of the 
Grassland Society of South Africa. 1 :  1 1 1 - 120. 
Thorhallsdotir, T.E. 1990. The dynamics of a grassland community: a simultaneous 
investigation of spatial and temporal heterogeneity at various scales. Journal of 
Ecology 78: 884-908. 
Thornley, J.H.M., Parsons, AJ., Newman, J.A .. , and Penning, P.D. 1994. A cost­
benefit model of intake and selection in a two-species sward. Functional 
Ecology 8: 5- 16. 
Torres-Rodriguez, A. 1997. The Effect of Herbage Availability and Species Choice on 
Grazing Preference of Dairy Cattle. M.Sc.Thesis, Massey University. 
Tribe, D.E. 1949. The importance of the sense of smell to the grazing sheep. Journal of 
Agricultural Science 39: 309-3 12. 
Tribe, D.E., and Gordon, J.G. 1949. The importance of colour vision to the grazing 
sheep. Journal of Agricultural Science 39: 3 13-3 14. 
References 265 
Trudell, J., and White, R.G. 1981. The effect of forage structure and availability on 
food intake, biting rate, bite size and daily eating time of reindeer, Journal of 
Applied Ecology 1 8: 63-8 1 .  
Underwood, R. 1982. Vigilance behaviour in grazing African antelopes. Behaviour 79: 
79-107. 
Ungar, E.D., Genizi A., and Demment M.W. 1991. Bite dimensions and herbage 
intake by cattle grazing short hand-constructed swards. Agronomy Journal 83: 
973-978. 
Vallentine, J.F. 1990. Grazing management. Academic press. USA. 533 p. 
Van Dyne, G.M., and Heady H.F. 1965. Dietary chemical composition of cattle and 
sheep grazing in common on a dry annual range. Journal of Range Management 
1 8: 78-85. 
Van Niekerk, A.I., Greenhalgh, J.F.D., and Reid, G.W. 1973. Importance of 
palatability in determining the feed intake of sheep offered chopped and pelleted 
hay. British Journal of Nutrition 30: 95- 105. 
van Santen, E. 1992. Animal preference of Tall Fescue during reproductive growth in 
the spring. Agronomy Journal 84: 979-982. 
Van Soest, P.J. 1994. Nutritional Ecology of the Ruminant, 2nd edn. Comell University 
Press, Ithaca, New York, 476. 
Waghorn, G.C., John, A., Jones, W.T., and Shelton, I.D. 1987a. Nutritive value of 
Lotus corniculatus L. containing low and medium concentrations of condensed 
tannins for sheep. Proceedings of the New Zealand Society of Animal Production 
47: 25-30. 
References 266 
Waghom, G.C., Jones, W.T., Sheiton, I.D., and McNabb, W. 1990. Condensed 
tannins and the nutritive value of herbage. Proceedings of the New Zealand 
Grassland Association 5 1 :  1 7 1- 176. 
Waghom, G.C., Vlyatt, M.J., John, A., and Fisher, M.T. 1987b. The effect of CT on 
the sites of digestion of amino acids and other nutrients in sheep fed on Lotus 
corniculatus L. British Journal of Nutrition 57: 1 15- 126. 
Waldem, D.E., and Van Dyk, R.D. 1971. Effect of monosodium glutamate in starter 
rations on feed consumption of early weaned calves. Journal of Dairy Science 
54: 262-265. 
Wang, Y., Douglas, G.B., Waghorn, G.C., Barry, T.N., and Foote, A.G. 1995. The 
effect of condensed tannins upon the performance of lambs grazing Lotus 
corniculatus and lucerne (Medicago sativa). New Zealand Journal of 
Agricultural Research 38: 95- 104. 
Westoby, M. 1974. An analysis of diet selection by large generalist herbivores. The 
American Naturalist 108: 290-304. 
Westoby, M. 1978. What are the biological bases of varied diets? The American 
Naturalist 1 12: 627-63 1 .  
Whittaker, R.H., and Feeny, P.P. 1971. Allelochemicals: chemical interactions 
between species. Science 17 1 :  757-770. 
Wickstrom, M.L., Robbins, C.T., Hanley, T.A., Spalinger, D.E., and Parish, S.M. 
1984. Food intake and foraging energetics of elk and mule deer. Journal of 
Wildlife Management 48: 1285-1301 .  
References 267 
Willms W.D., Dormaar J.F., Scbaalje G.B. 1988. Stability of grazed patches on 
rough fescue grassland. Journal of Range Management 41 :  503-508. 
Windbam, W.R., Fales, S.L., and Hoveland, C.S. 1988. Crop utilization: analysis for 
condensed tannins in Serica Lespedeza by near infrared reflectance spectroscopy. 
Crop Science 28: 705-708. 
Wong, E. 1973. Plant phenolics. In: Chemistry and Biochemistry of Herbage. G.W. 
Buttler and RW. Bailey (eds.). Academic Press, Vol. 1, London, pp. 265-322. 
Zucker, W.V. 1983. Tannins: Does structure determine function? An ecological 
perspective. The American Naturalist 12 1 (3): 335-36. 
APPENDICES 
APPENDIX 3. 1. DMACA-HCL Protocol -using plate reader 
Li et al. ( 1996) developed a protocol using HCI-acidified 4-
dimethylaminnocinnamaldehyde (DMACA) for screening condensed tannins 
(proanthocyanidins) .  They developed a reliable and sensitive method to detect condensed 
tannin at concentrations lower than 0.4 mg g-I dry matter (DM). The DMCA-HCL 
protocol is recommended for the detection of condensed tannins in plants with low 
concentrations. 
Considering the methodology described by Li et al. ( 1996) and Terrill et al. ( 1992), a 
DMACA-HCL protocol was developed to utilize a plate reader with Softpro software. 
This protocol was used to determine the concentration of extractable condensed tannins 
in Lotus comiculatus L. 
FREE OR ACETONE EXTRACTABLE CONDENSED TANNIN 
Solutions: 
1 .  67 mM glycine HCI pH 3.0 
Dissolve 5.0 g of glycine in about 900 ml of milliQ H20. Adjust the pH to 3.0 with 
conc HCI and adjust the final volume to 1 litre. Filter through a 0.2 Jl.111 filter. 
2. Aqueous acetone (20mM glycine-HCI pH 3.0; acetone (30:70 v/v); ascorbic acid 
(lg r1» 
Per litre, combine 700 ml of acetone, 300 ml of 67 mM glycine HCI pH 3.0 and 1 g 
of ascorbic acid. Adjust pH to 3.0 with conc HCl. Store in a brown bottle away from 
direct sunlight. 
Method - Extraction of eT 
1 .  Weigh 500 mg of freeze dried and very finely ground sample (0.5 mm sieve) into a 
50 ml Oakridge centrifuge tube. Do a dry matter (DM) on each sample as well so that 
the CT concentration can be corrected for DM. 
2. Add 10  ml of aqueous acetone solution and homogenize on ice with the Utraturex for 
about 1 min. 
3.  Centrifuge at 15,000 rpm (about 22,OOOxg) for 10 min. 
4. Transfer the supematant to a 50 ml bluetop tube. Keep on ice. 
Appendices 269 
5.  Add 10 rnl of aqueous acetone solution to the residue remaining in the Oakridge tube, 
and re-homogenize. Centrifuge again and add the supematant to the bluetop tube 
from step (4). 
6. Add 20 rnl of methylene chloride. Vortex vigorously. 
7. Leave the tube standing until the phases are completely separated. 
8. Transfer the upper aqueous phase to a 50 rnl tube, leaving only the lower solvent 
phase as residual. 
9. Rewash the residue remaining in the tube with milliQ H20. Leave the phases to 
separate completely. The washing with milliQ H20 is important mainly in the fIrst 
wash to recover any condensed tannin trapped by lipid in the solvent phase. 
10. Rewash with the combined aqueous phases with methylene chloride. Continue 
washing the aqueous phases with methylene chloride until all the chlorophyll and 
lipids are removed. 
1 1 . Once the lower solvent phase is clear (transparent), transfer the aqueous phase to a 
250 rnl round-bottom flask and remove excess acetone and methylene chloride at 
40°C under reduced pressure. 
12.  Decant into a 50 ml transport storage tube and make up to 50 g with milliQ H20. The 
sample can be stored at -20°C at this stage. 
COLOURMETRIC DETECTION OF CONDENSED TANNINS - DMACA·HCL 
PROTOCOL 
The colourmetric detection of condensed tannins is performed using a standard curve 
with a range of condensed tannin concentrations. 
Solutions 
1 .  6M HCI 
Very slowly add 262 rnl of concentrated HCI to 200 rnl of milliQ H20. Make up to 500 
rnl with milliQ H20. 
2. Methanol:6M HCl (1: 1 v/v) 
Very slowly add 100 rnl of cold 6M HCI to 100 rnl of methanol. 
Appendices 270 
3 .  2% DMACA-HCl (w/v) 
Dissolve 0.2 g of 4-dimethylaminocinnamaldehyde (DMACA) in 10 g of methanol:6M 
HCI ( 1 : 1  v/v). Make fresh each time that the reagent is required and store in a dark bottle 
when in use. 
Method 
Thawing of samples 
It is possible to loose sample in the thawing process. Therefore it is essential that all 
tubes are well closed and no tubes have cracked during freezing. 
Standard Curve (applicable for Lotus comiculatus L.) 
1 .  Weigh 10 mg of Sephadex LH-20 extract (Lotus comiculatus condensed tannin 
extracted according to Jackson et al., 1996) into a 5 ml bluetop tube. Make up to 5 g 
with milliQ H20. This stock CT solution and is 2 mg ml-1 • 
2. Using the stock CT solution to make the following standards in 2 ml microcentrifuge 
tubes. 
3 .  0 Ilg ml-1 
12.5 Ilg ml-1 
25 1lg ml-1 
37.5 Jlg ml-1 
50 Ilg ml-1 
75 J.1g ml-1 
l00llg ml-1 
125 1lg ml-1 
1 50 Ilg ml-1 
175J.1g ml-1 
200 J.1g ml-1 
225 1lg ml-1 
250 J.1g ml-1 
275 1lg ml-1 
300llg ml-1 
Plate Reader 
o J.1l of the CT stock and 2000 J.1l of milliQ H20 
12.5 J.1l of the CT stock and 1987.5 J.1l ofmilliQ H20 
25 J.1l of the CT stock and 1975 J.1l of milliQ H20 
37.5 f.ll of the eT stock and 1 962.5 f.ll ofmilliQ H20 
50 J.1l of the CT stock and 1950 J.1l of milliQ H20 
75 J.1l of the CT stock and 1925 J.1l of milliQ H20 
100 J.1l of the CT stock and 1900 J.1l ofmilliQ H20 
125 J.1l of the CT stock and 1 875 J.1l of milliQ H20 
150 J.1l of the CT stock and 1850 J.1l of milliQ H20 
175 f.ll of the eT stock and 1 825 J.1l ofmilliQ H20 
200 f.ll of the CT stock and 1800 J.1l of milliQ H20 
225 J.1l of the eT stock and 1775 J.1l of milliQ H20 
250 f.ll of the CT stock and 1750 J.1l of milliQ H20 
275 J.1l of the CT stock and 1725 J.1l of milliQ H20 
300 J.1l of the eT stock and 1700 J.1l of milliQ H20 
1 .  Add 100 J.1l of each standard to a separate well in the culture or microtitre plate. Use 
the first 15  wells in the plate for the standard curve (AI-B3). 
2. Add 100 J.1l of each sample to a separate well in the culture plate. Analyze each 
sample in duplicate. 
Appendices 
3. Add 100 J.1l of milliQ H20 to each well containing standards and samples. 
4. Add 100 J.1l of methanol to each well containing standards and samples. 
27 1 
5. Add 50 J.1l of fresh DMACA-HCl reagent to each well containing standards and 
samples. 
6. With the addition of the DMACA-HCL reagent, ensure that the contents in each well 
are thoroughly mixed with a multi-channel pipette. 
7. Allow the assay to develop at room temperature for 20 minutes. 
8. Read the absobance at 643 nm. 
Computer and plate reader 
Turn on the computer and the plate reader at least 1 5  minutes before using. 
Adjust the wavelength to 643nm and enter the concentration of each standard and 
indicate the location of the standards and unknowns in the culture plate into the softpro 
software. 
A standard with a concentration of 25 Ilg ml-I will have 2.5 Ilg in the well. 
25 Ilg ml -I = 25 ng J.1l-1 
using 100 J.1l 
25ng J.1l- 1 x 100 J.1l = 2500 ng = 2.5 Ilg in the well 
Appendices 272 
APPENDIX 3. 2. Rainfall and soil temperature - Experiments 1, 2, 3 (Chapter 3) 
Table 3 . 1 .  Daily rainfall and average soil temperature ( 10  cm depth) during the 
experiment 1 at Flock House - from 30 October/95 to 27 November/95.  
Da te Average Rainfall  
soil temp . (mm) 
(oC) 
3 0 -0ct 1 1 . 2 0 . 0  
3 1 -0ct 1 1 . S 1 . S  
0 1 -Nov 1 1 . 7 2 B . 1  
02 -Nov 1 0 . B  1 6 . 9  
03 -Nov 1 0 . 4  0 . 0  
0 4 -Nov 1 1 . 1  4 . 6  
O S -Nov 1 1 . 0 1 . 8  
0 6 -Nov 1 0 . 9  1 . 8  
07 -Nov 1 0 . 6  0 . 0  
O B -Nov 1 1 . 3  0 . 0  
0 9 -Nov 1 1 . 9 6 . 7  
1 0 -Nov 12 . 6  l S . 4  
1 1 -Nov 12 . 3  0 . 3  
12 -Nov 1 1 . 7 0 . 8  
13 -Nov 1 1 . 7  0 . 0  
14 -Nov 12 . 0  0 . 0  
lS -Nov 1 1 . 6 0 . 0  
1 6 -Nov 11 . S 0 . 0  
17 -Nov 12 . 0  0 . 0  
1 B -Nov 11 . 9 0 . 0  
19 -Nov 11 . 7 0 . 0  
2 0 -Nov 1 1 . 3  B . 2 
2 1-Nov 1 0 . S  0 . 8  
22 -Nov 10 . 4  0 . 0  
2 3 -Nov 1 0 . S  0 . 0  
2 4 -Nov 1 1 . 1  3 . 6 
2 S -Nov 12 . 3  1 1 . B 
2 6 -Nov 1 1 . 9 S . 4  
2 7 -Nov 1 0 . 4  9 . 7  
Appendices 273 
APPENDIX 3.2. Rainfall and soil temperature - Experiments 1, 2, 3 (Chapter 3) 
Table 3.2 Daily rainfall and average soil temperature ( 10 cm depth) during the 
experiment 2 at Flock House - from 5 February/96 to 1 March/96. 
Average Rainfal l 
soil temp . (mm) 
(DC) 
S -Feb 14 . 10 0 . 6 0 
6 -Feb 14 . 10 0 . 0 0 
7 -Feb 1 4 . 4 0 1 6 . 10 
8 -Feb 1 4 . 5 0 0 . 0 0 
9 -Feb 14 . 0 0 0 . 0 0 
1 0 -Feb 1 4 . 00 0 . 0 0 
11 -Feb 13 . 50 0 . 0 0 
12 -Feb 1 4 . 80 0 . 0 0 
13 -Feb 15 . 3 0 1 .  0 0  
14 -Feb 15 . 10 0 . 0 0 
lS -Feb 1 4 . 60 0 . 0 0 
1 6 -Feb 13 . 9 0 0 . 0 0 
17 -Feb 13 . 8 0 0 . 0 0 
1 8 -Feb 14 . 3 0 4 . 9 0 
19 -Feb 15 . 3 0 4 . 40 
2 0 -Feb 1 4 . 5 0 1 8 . 9 0 
2 1 -Feb 13 . 2 0 1 .  5 0  
2 2 -Feb 11 . 8 0  1 8 . 7 0 
2 3 -Feb 11 . 9 0  3 . 60 
2 4 -Feb 1 0 . 8 0 0 . 0 0 
2 S -Feb 11 . 10 0 . 0 0 
2 6 -Feb 11 . 8 0 0 . 0 0 
2 7 -Feb 12 . 10 0 . 0 0 
2 8 -Feb 12 . 8 0 0 . 0 0 
2 9 -Feb 13 . 10 0 . 0 0 
1-Mar 13 . 10 0 . 0 0 
Appendices 274 
APPENDIX 3.2. Rainfall and soil temperature - Experiments 1, 2, 3 (Chapter 3) 
Table 3.3. Daily rainfall and average soil temperature ( 10 cm depth) during the 
experiment 3 at Flock House - from 15  AEril/96 to 10 Mal196. 
Average Rainfall  
soil temp . (mm) 
(DC) 
15 -Apr 1 1 . 8 0 3 . 8 0 
16 -Apr 11 . 7 0 2 . 60 
17 -Apr 1 1 . 0 0 0 . 0 0 
18 -Apr 9 . 9 0 0 . 0 0 
19 -Apr 1 0 . 2 0 8 . 7 0  
2 0 -Apr 1 1 . 0 0  9 . 5 0 
2 1 -Apr 11 . 4 0 1 9 . 4 0 
22 -Apr 1 0 . 1 0 0 . 0 0 
23 -Apr 9 . 3 0  0 . 0 0 
2 4 -Apr 9 . 2 0 0 . 0 0 
2 5 -Apr 8 . 3 0  0 . 0 0 
2 6 -Apr 7 . 7 0 0 . 0 0 
27 -Apr 8 . 60 0 . 0 0 
2 8 -Apr 9 . 4 0 0 . 8 0 
2 9 -Apr 9 . 3 0 0 . 3 0 
3 0 -Apr 9 . 6 0 6 . 4 0 
1 -May 8 . 9 0 3 . 60 
2 -May 6 . 9 0 0 . 0 0 
3 -May 6 . 9 0 0 . 0 0 
4 -May 6 . 4 0 0 . 0 0 
5 -May 6 . 7 0 0 . 0 0 
6 -May 6 . 8 0 0 . 0 0 
7 -May 7 . 6 0 0 . 0 0 
8 -May 8 . 10 2 . 2 0 
9 -May 7 . 8 0 2 . 1 0 
10 -May 6 . 1 0 0 . 0 0 
Appendices 275 
APPENDIX 3. 3. Experiment 1 (Chapter 3) 
Table 3. 4. Herbage mass (kg DMlha), sward height (cm) and bulk density (mg DMlcm3) 
before and after grazing, and estimation of the herbage mass removed (kg 
DMlha) of birdsfoot trefoil and white clover (BW) and red clover (RC) 
swards accordin� to treatment (area ratios) in EXEeriment 1 .  
Treatments 
A B C D SED1 p-
valui 
BW RC BW RC BW RC BW RC 
20% 80% 33% 67% 67% 33% 80% 20% 
Herbage mass 
(Kg DMlha) 
Pre-graz 3930 4500 4020 4580 3930 4670 3880 4530 328 0.9802 
Post graz 1930 3060 2450 3360 2560 3650 2580 3280 291 0.7261 
DM rem. 2000 1450 1560 1210 1370 1020 1 3 10 1250 412 0.8599 
Sward height (cm) 
Pre-graz. 19.8 26. 1 19.0 29. 1 19.3 29.2 1 8.4 26.3 2.68 0.7204 
after 1 day 1 1 .5 24.5 1 1 .5 23.3 12.6 2 1 .2 13.6 1 9.3 1 .7 1  0.0597 
Post graz. 6.8 1 3.8 6.8 13.4 8 .0 14.2 8.3 1 1 .7 0.70 0.0098 
Bulk density 
(mg DMlcm3) 
Pre-graz 2.00 1 .77 2. 12 1 .57 2.04 1 .60 2.10 1 .74 0.222 0.7732 
2.73 2. 16  3 .70 2.48 3 . 18  2.63 3 .08 2.78 0.433 0.5002 Post S!az 
i SED-Standard error for differences of means when comparing swards within each treatment. 
2 P-value of the interaction between treatment and sward type. 
Table 3.5. The effect of treatments (area ratios) on rate of biting (bites/minute) in the 
first, second and third days (Day 1 ,  2 and 3) (total 55 hours) of grazing 
assessment in EXEeriment 1 .  
Rate of biting 
(biteslmin) 
A 
BW RC BW 
Treatments 
B C 
RC BW RC 
D 
BW RC 
20% 80% 33% 67% 67% 33% 80% 20% 
SED1 p­
valui 
Day 1 49.5 45.4 52.7 45.0 48.2 44. 1 47.6 45.8 1 .45 0.0702 
Day 2 5 1 .2 43.8 53.8 46.9 53.5 45.9 52. 1 48.0 2.70 0.7468 
Day 3 55. 1 49.0 54.7 46.8 55.3 47.3 52.2 47. 1  5.60 0.9794 
i SED- Standard error for differences of means when comparing swards within each treatment. 
2 P-value of the interaction between treatment and sward type. 
APPENDIX 3. 3. Experiment 1 (Chapter 3) 
Table 3.6. Botanical characteristics of birdsfoot trefoil and white clover (BW), and red clover (RC) swards before and after grazing, 
according to the treatments (area ratios) (DM basis): (a) percentage of components in live fraction, (b) percentage of live matter 
in total DM of each sward and (c) ratio of the total live matter of birdsfoot trefoil and white clover (B:W) in the BW sward, 
EXEeriment 1 .  
Treatment A Treatment B Treatment C Treatment D 
BW RC BW RC BW RC BW RC SED1 P-value2 
20% 80% 33% 67% 67% 33% 80% 20% 
Pre-grazing 
(a) Leaf 4 1 .2 39.2 44.0 37.6 4 1 .6 40.4 49.2 47.2 5.78 0.9 135 
Petiole 13.6 17 .8 25.4 14.7 26.0 16.8 1 3 . 1  13 .8  5.7 1 0.3208 
Stem 15.9 34. 1 17.9 4 1 .9 12.9 33.9 1 5.4 27.2 5. 17 0.4 100 
Flower 0.7 0.7 0.8 1 .2 1 .3 0.8 0.4 1 . 1  0.61 0.6000 
Grass 10.6 1 . 1  2.7 0.0 8.9 1 .9 1 1 .2 4.4 5.73 0.8880 
Other species 16.7 5.7 7.8 2.9 1 1 .2 8.0 12.6 8.4 5.97 0.7900 
(b) Total live matter 97.3 9 1 .7 97.0 92.9 96. 1 9 1 .4 92.4 89.8  1 .85 0.7 185 
(c) B:W ratio 0.82 0.79 0.45 0.66 0.436 0.8220 
Post-grazing 
(a) Leaf 29.3  20.0 25.6 16.6 28. 1 17.7 28.2 14.8 5 . 12  0.9237 
Petiole 19.9 16.9 28.9 14.8 37.7 13 .8 30.0 19.4 3 .05 0.0 128 
Stem 23.2 50.4 42.3 5 1 .3 20.4 52.3 24.9 38.4 6.74 0.0707 
Flower 0.4 0.4 0.5 0.5 1 .4 2. 1 0.5 0.3 1 .06 0.9 101  
Grass 12.5 0.3 0.4 0. 1 5.3 5.3 2.9 16.4 7.63 0. 1522 
Other species 12.9 10.3 0.4 16.3 8.6 9.4 14.3 1 1 .6 5.8 1 0.0940 
(b) Total live matter 86.4 85.9 93.0 8 1 .3 94. 1 83.3 9 1 .7 86. 1 3.48 0. 1065 
(c) B:W ratio 0.47 0.47 0.25 0.41 0.221 
1 SED - Standard error for differences of means when comparing means with the same level of treatments. 
2 P-value of the interaction: treatment*sward type effect. 
N -...l 0\ 
Appendices 277 
APPENDIX 3.4. Experiment 2 (Chapter 3) 
Table 3.7. Formononetin concentration (%) of leaf, petiole, stem and flower of birdsfoot 
trefoil (BT) and white clover (WC) in birdsfoot trefoil and white clover 
sward (BW), and red clover (RC) in red clover sward (RC), according to the 
treatments (sward maturity: Imm = immature; Mat = mature) (DM basis) in 
Experiment 2. 
BWsward 
BT WC 
Imm Mat Imm Mat 
Leaf 0. 1 1  0. 13  0. 14 0.20 
Petiole 
Stem 
3 0. 18  0.24 
0. 1 1  0. 10 
RC sward 
RC 
Imm 
0.54 
0.50 
0.41 
Mat 
0.46 
.35 
0.25 
0.745 
0.064 
0.05 1 
0.021 (BT)3 
P-valuez 
0.9000 
0.7460 
0.0938 
Flower 0.07 0.06 0. 15 0. 1 1  0. 1 1  0. 10 0.018(WC)4 0.3840 
0.015(RC)5 
i SED - Standard error for differences of means when comparing within each species 
2 P-value of the interaction: species*maturity effect 
3 _ no sample 
Number of observations contributing to the mean of each specie (n=4), but n=3(immature) and 
n=2(mature) for birdsfoot trefoil, n=3 for white clover and n=4 for red clover in flower 
assessment. 
Table 3.8. The effect of treatments (maturity: Imm = immature, Mat = mature) on 
grazing time (minutes) in the fIrst, second and third days of grazing (Days 1 ,  
2 and 3), and average DM intake per animal per day (kg dmlhdlday) during 
55 hours of grazing in Experiment 2. 
A 
BW 
Imm 
Grazing time 
(min) 
Day 1 28.7 
Day 2 34.5 
Day 3 38.7 
Treatments 
B C D p­
valui 
RC BW RC BW RC BW RC 
Imm Imrn Mat Mat Imm Mat Mat 
42.9 39.5 41 .5 29.4 49.0 29.7 
49.0 3 1 .2 46.7 34.4 59.2 34.3 
53. 1 22.5 59 4 1 .2 5 1 .3 26.7 
46. 1 9.93 0.6236 
54.4 8.81 0.83 12 
58.9 10.23 0.2260 
Intake (Kg 4.29 2.33 5.20 5.69 5.28 3.30 4.66 4.47 1 .437 0.5293 
DM/hd/day) 
i SED- Standard error for differences of means when comparing swards within each treatment. 
2 P-value of the interaction between treatment and sward type. 
Appendices 278 
APPENDIX 3. 4. Experiment 4 (Chapter 3) 
Table 3. 9. Herbage mass (kg DMlha), sward height (cm) and bulk density (mg DMlcm3) 
pre-grazing (Pre-graz.), and sward height post grazing (post graz.) of 
birdsfoot trefoil and white clover (BW) and red clover (RC) according to 
treatment (sward position - side: close to the fence; central: in the middle of 
the plot; altern.= alternative sward) in Experiment 4. 
Treatments 
A B C D 
BW RC BW RC BW RC BW RC 
side altern. central altem. altem. side altem. central 
Herbage mass 
(KgDMlha) 
SED] p-
valui 
Pre-graz 2024 2280 2150 2270 21 10 2190 1910  2070 307 0.9829 
Sward height 
(cm) 
Pre-graz. 7.0 
Post graz 4.6 
Bulk density 
(mgDMlcm3) 
8.8 
5.0 
7 . 1  
4. 1 
7.0 
4.7 
7.0 
4. 1 
7.5 
4.0 
7.0 
4.4 
6.8 0.743 0.2655 
4. 1 0.5454 0.6565 
Pre-graz 2.82 2.64 3.05 3.36 3.09 2.96 2.76 2.97 0.4324 0.8 170 
i SED-Standard error for differences of means when comparing swards within each treatment. 
2 P-value of the interaction between treatment and sward type. 
Table 3 . 10. The effect of treatments (strip position - side: close to the fence; central: in 
the middle of the plot; altern.= alternative sward) on grazing time (minutes in 
three hours of observation) in Experiment 4. 
Grazing time 
(min) 
Treatments 
A B C D 
BW RC BW RC BW RC BW RC 
Imm Imm Imm Mat Mat Imm Mat Mat 
SED] p-
valui 
Day 1 46.3 95.0 44.5 103. 1 92.2 46.9 87.6 43.3 5.948 0.0000 
i SED- Standard error for differences of means when comparing swards within each treatment. 
2 P-value of the interaction between treatment and sward type. 
Appendices 279 
APPENDIX 3. 5. Grazing time - morning observation (Chapter 3) 
Table 3 . 1 1 .Treatment (20, 33, 67 and 80 % of the total area offered) effects on the 
proportion of grazing time (in relation to the total grazing time spent in plot) 
devoted to birdsfoot trefoil plus white clover swards (BW) in Experiment 1 .  
Day 1 - evening 
Day 2 -morning 
Day 2 - evening 
Day 3 - evening 
Means Day 1 and 
Day 2 evening 
20% 
0.35 
0.20 
0. 15  
0.09 
0.25 
Proportion of area of BW 
33% 67% 
0.46 0.60 
0.33 0.67 
0.26 0.70 
0.20 0.66 
0.36 0.65 
80% 
0.69 
0.80 
0.80 
0.79 
0.75 
Mean 
0.50 
0.50 
Table 3 .  12.Treatment effects on the proportion of grazing time (in relation to the total 
grazing time spent in the plot) devoted to birdsfoot trefoil and white clover 
swards (BW) in Experiment 2 Treatment A= BW and RC immature; 
Treatment B=BW immature and RC mature; Treatment C= BW mature and 
RC immature; Treatment D= BW and RC mature). 
Means Treatments A B C D 
Day l -evening 0.39 0.5 1 0.39 0.41 
Day 2 - morning 0.46 0.44 0.42 0.36 0.42 
Day 2 - evening 0.42 0.39 0.38 0.38 
Day 3 - evening 0.42 0.27 0.46 0.34 
Means Day 1 and 0.41 0.45 0.38 0.40 Day 2 evening 0.41 
Table 3 . 1 3 .Treatment effects on the proportion of grazing time devoted to birdsfoot 
trefoil and white clover swards (BW) in Experiment 3 Treatment A= BW and 
RC short; Treatment B=BW short and RC tall; Treatment C= BW tall and RC 
short; Treatment D= BW and RC tall) 
Treatments 
A B C D 
-------------------------------- Means 
Day 1 -evening 0.46 0.29 0.63 0.46 
Day 2 - morning 0.54 0.37 0.52 0.49 0.48 
Day 2 - evening 0.52 0.41 0.59 0.54 
Day 3 - evening 0.44 0.34 0.5 1 0.50 
Means Day 1 and 0.49 0.35 0.61 0.50 Day 2 evening 0.45 
Appendices 280 
APPENDIX 3. 6. Grazing time - morning observation (Chapter 3) 
ANALYSES OF V ARIANCES (ANOVA) 
ANOV A of total grazing time of morning observation (GTMORNl) of Day 2, 
Experiment 1 
DATA = GTMORN1 
Error: PERIOD:GROUP 
Df Sum of Sq 
TREATM 3 160.6480 
Mean Sq 
53.54934 
87.08849 
F Value 
0.6148842 
Pr(F) 
0.6224212 
Residuals 9 783.7964 
Error: Within 
Df 
SPECIE 1 
TREATM:SPECIE 3 
Residuals 16 
Sum of Sq 
444.443 
4202.219 
797.779 
Mean Sq 
444.443 
1400.740 
49.861 
F Value Pr(F) 
8.91 360 0.008738614  
28.09278 0.00000 1299 
ANOV A of total grazing time of morning observation (GTMORN2) of Day 2, 
Experiment 2 
DATA = GTMORN2 
Error: PERIOD:GROUP 
Df 
TREATM 3 
Residuals 9 
Error: Within 
Df 
SPECIE 1 
TREATM:SPECIE 3 
Residuals 16 
Sum of Sq 
270.78 1 
145 1 .649 
Sum of Sq 
2363.906 
480.747 
3799.444 
Mean Sq 
90.2604 
161 .2944 
Mean Sq 
F Value 
0.5596006 
Pr(F) 
0.6548896 
F Value Pr(F) 
2363.906 9.954745 0.0061291  
160.249 0.67483 1  0.579962 
237.465 
Appendices 28 1 
APPENDIX 3.6. Grazing time - morning observation (Chapter 3) 
ANOV A of total grazing time of morning observation (GTMORN3) of Day 2, 
Experiment 3 
DATA = GTMORN3 
Error: PERIOD:GROUP 
Df Sum of Sq 
TREATM 3 546.85 19 
Mean Sq 
182.2840 
104. 1 152 
F Value 
1 .75079 1 
Pr(F) 
0.226223 1 
Residuals 9 937.0370 
Error: Within 
Df Sum of Sq 
SPECIE 1 7 1 . 1 1 1  
TREATM:SPECIE 3 797.778 
Residuals 16 3453.333 
Mean Sq 
7 1 . 1 1 1 1  
265.9259 
215 .8333 
FValue 
0.329472 
1 .232089 
Pr(F) 
0.57395 1 3  
0.3306272 
Appendices 282 
APPENDIX 4. 1. Rainfall and temperature - Experiments 5 and 6 (Chapter 4) 
Table 4. 1. Monthly rainfall and average soil temperature ( 10  cm depth) from July/1996 
to June/1997 com ared with 60 ears avera e values at the site. 
Total monthly Total monthly Av. daily 10cm Av. daily 10cm 
rainfall rainfall soil temp. fC) soil temp. fC) 
av. 60- ears 
Jul-96 1 04.8 8.0 6.7 
Aug-96 82.3 8. 1 7.6 
Sep-96 1 02.8 75.0 1 1 .5 9.9 
Oct-96 95.6 88.0 1 3.7 1 2.5 
Nov-96 1 00.5 78.0 1 4.4 1 5. 1  
Oec-96 91 . 1  94.0 1 6.7 1 7.3 
Jan-97 68.0 79.0 1 7.6 1 8.5 
Feb-97 58.0 67.0 1 7.7 1 8. 1  
Mar-97 68. 1 69.0 1 5.8 1 6.3 
Apr-97 1 44.7 81 .0 1 2.5 13.2 
MaY-971 24.3 89.0 1 1 .6 1 0. 1  
JUn-971 60.4 97.0 8. 1 7.7 
Appendices 283 
APPENDIX 4. 2. Equations to estimate herbage mass per plants of birdsfoot trefoil 
and red clover, using probe GrassMaster reading (X) 
BIRDSFOOT TREFOIL 
Herbage mass = 1 .33 17 X - 9.6061 (R-square = 0.3653) 
RED CLOVER 
Herbage mass = 0.7902 X + 913.82 (R-square = 0. 1958) 
Appendices 284 
APPENDIX 4.3. Correlation matrices with Pearson Correlation Coefficients and p. 
values of Experiment 5 . plant nutritional characteristics 
BIRDSFOOT TREFOIL PLANTS 
Pearson Correlation Coefficients Prob > IR I under Ho : Rho=O I N = 8 
PROTEIN LIPID ADF NDF CHO ASH INVITRO BITES 
PROTEIN 1 . 00000 0 . 83385 -0 . 80969 0 . 70706 -0 . 46862 -0 . 31500 0 . 81069 0 . 62870 
0 . 0  0 . 0101 0 . 0149 0 . 0498 0 . 2415  0 . 4473  0 . 0146  0 . 0950 
LIPID 0 . 83385  1 . 00000 -0 . 77012 0 . 43637  -0 . 23585 0 . 08864  0 . 77082 0 . 79053 
0 . 0101 0 . 0  0 . 0254 0 . 2797  0 . 5739 0 . 8347 0 . 0252 0 . 0195 
ADF - 0 . 80969  - 0 . 77012 1 . 00000 -0 . 19217 -0 . 11968  0 . 33045  - 0 . 99992 - 0 . 30720 
0 . 0149  0 . 0254 0 . 0  0 . 6484 0 . 7777  0 . 4240 0 . 0001 0 . 4592 
NDF 0 . 70706 0 . 43637 -0 . 19217 1 .  00000 - 0 . 90728 -0 . 33505 0 . 19188 0 . 58885 
0 . 0498 0 . 2797  0 . 6484 0 . 0  0 . 0019 0 . 4172 0 . 6490  0 . 1246 
CHO - 0 . 46862  - 0 . 23585 -0 . 11968  -0 . 90728 1 . 00000 0 . 08362  0 . 12054 - 0 . 52746 
0 . 2415  0 . 5739  0 . 7777  0 . 0019 0 . 0  0 . 8439  0 . 7762 0 . 1791 
ASH -0 . 31500 0 . 08864 0 . 33045  -0 . 3 3505 0 . 08362 1 . 00000 - 0 . 32645  0 . 38395 
0 . 4473 0 . 8347 0 . 4240 0 . 4172  0 . 8439 0 . 0  0 . 4300 0 . 3477  
INVITRO 0 . 81069 0 . 77082 -0 . 99992 0 . 19188 0 . 12054 -0 . 32645 1 . 00000 0 . 31154  
0 . 0146  0 . 0252 0 . 0001 0 . 6490  0 . 7762 0 . 4300 0 . 0  0 . 4526  
BITES 0 . 62870 0 . 79053 -0 . 30720 0 . 58885 -0 . 52746 0 . 38395  0 . 31154  1 . 00000 
0 . 0950 0 . 0195 0 . 4592 0 . 1246  0 . 1791 0 . 3477  0 . 452 6 0 . 0  
RED CLOVER PLANTS 
Pearson Correlation Coefficients Prob > IR I under Ho : Rho=O / N = 12 
PROTEIN LIPID ADF NDF CHO ASH INVITRO BITES 
PROTEIN 1 . 00000 0 . 91763  -0 . 03276 0 . 11415  -0 . 77440 0 . 74333  0 . 01473 - 0 . 46503 
0 . 0  0 . 0001 0 . 9195 0 . 7239  0 . 0031 0 . 0056 0 . 9638  0 . 1277 
LIPID 0 . 91763  1 .  00000 - 0 . 03108 0 . 17897  - 0 . 66781 0 . 67739 0 . 01856 - 0 . 17374 
0 . 0001 0 . 0  0 . 9236  0 . 5779 0 . 0176 0 . 0155 0 . 9544 0 . 5892 
ADF -0 . 03276  -0 . 03108 1.  00000 0 . 53408 -0 . 51880 0 . 48006 -0 . 99925 - 0 . 19320 
0 . 9195  0 . 9236  0 . 0  0 . 0737 0 . 0839  0 . 1142 0 . 0001 0 . 5474 
NDF 0 . 11415  0 . 17897 0 . 53408 1 . 00000 -0 . 44129  0 . 20651  -0 . 55574 0 . 33273 
0 . 7239  0 . 5779 0 . 0737 0 . 0  0 . 1510 0 . 5196  0 . 0606  0 . 2906 
CHO -0 . 77440  - 0 . 66781 -0 . 51880 -0 . 44129  1 . 00000 - 0 . 90577 0 . 53698 0 . 55228 
0 . 0031  0 . 0176 0 . 0839  0 . 1510 0 . 0  0 . 0001 0 . 0718  0 . 0626 
ASH 0 . 74333  0 . 67739 0 . 48006 0 . 20651  -0 . 90577 1 . 00000 - 0 . 49066  - 0 . 54047 
0 . 0056  0 . 0155 0 . 1142 0 . 5196  0 . 0001 0 . 0  0 . 1053 0 . 0696 
INVITRO 0 . 01473 0 . 01856 -0 . 99925 -0 . 55574 0 . 53698 - 0 . 49066  1 . 00000 0 . 19550 
0 . 9638  0 . 9544 0 . 0001 0 . 0606 0 . 0718 0 . 1053 0 . 0  0 . 5426 
BITES - 0 . 46503 -0 . 17374 - 0 . 19320 0 . 33273 0 . 55228 -0 . 54047 0 . 19550 1 . 00000 
0 . 1277 0 . 5892 0 . 5474 0 . 2906 0 . 0626 0 . 0696 0 . 5426 0 . 0  
Appendices 285 
APPENDIX 4. 4. Correlation matrices with Pearson Correlation Coefficients and p. 
values of Experiment 5 Period 1 - characteristics of red clover plants. 
RED CLOVER MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Pr ob > J R J  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 21268  0 . 84572 - 0 . 32021 -0 . 01337 
0 . 0  0 . 1467 0 . 0001 0 . 0265  0 . 9281  
HEIGHT 0 . 21268  1 . 00000 0 . 70129 0 . 17703 0 . 22725 
0 . 1467 0 . 0  0 . 0001 0 . 2287 0 . 1203 
VOLUME 0 . 84572 0 . 70129 1 . 00000 -0 . 13693 0 . 11435  
0 . 0001 0 . 0001 0 . 0  0 . 3534  0 . 4390 
LEAF -0 . 32021 0 . 17703 - 0 . 13693 1 . 00000 0 . 45162 
0 . 0265 0 . 2287 0 . 3534 0 . 0  0 . 0013 
MASS -0 . 01337 0 . 22725 0 . 11435 0 . 45162 1 .  00000 
0 . 9281  0 . 1203 0 . 4390 0 . 0013 0 . 0  
RED CLOVER FORMONONETIN CONCENTRATION, PLANT MORPHOLOGY AND 
NUMBER OF BITES 
Pearson Correlation Coefficients Prob > J R J  under Ho : Rho=O / N = 1 6  
FORM AREA HEIGHT VOLUME LEAF BITES 
FORM 1 . 00000 -0 . 15965 - 0 . 10634 - 0 . 17185 -0 . 18536  - 0 . 10244 
0 . 0  0 . 5548 0 . 6951 0 . 5245  0 . 4919  0 . 7058 
AREA -0 . 15965  1 . 00000 0 . 2 6850 0 . 88169 -0 . 42843 0 . 70743 
0 . 5548  0 . 0  0 . 3147  0 . 0001 0 . 0978 0 . 0022 
HEIGHT - 0 . 10634 0 . 26850 1 . 00000 0 . 69124 0 . 25200 0 . 61915  
0 . 6951  0 . 3147 0 . 0  0 . 0030 0 . 3464  0 . 0105 
VOLUME -0 . 17185  0 . 88169 0 . 69124  1 . 00000 - 0 . 19796  0 . 83397  
0 . 5245 0 . 0001 0 . 0030 0 . 0  0 . 4624 0 . 0001 
LEAF -0 . 18536  - 0 . 42843 0 . 25200 - 0 . 19796 1 . 00000 0 . 04955  
0 . 4919  0 . 0978 0 . 3464  0 . 4624 0 . 0  0 . 8554 
BITES - 0 . 10244 0 . 70743 0 . 61915  0 . 83397 0 . 04955 1 . 00000 
0 . 7058 0 . 0022 0 . 0105 0 . 0001 0 . 8554 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS AND NUMBER OF 
BITES 
Pearson Correlation Coefficients / Prob > J R J  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 .  00000 0 . 21268 0 . 84572 - 0 . 32021 0 . 55954 
0 . 0  0 . 1467 0 . 0001 0 . 0265  0 . 0001 
HEIGHT 0 . 21268  1 . 00000 0 . 70129 0 . 17703 0 . 44735 
0 . 1467 0 . 0  0 . 0001 0 . 2287  0 . 0014 
VOLUME 0 . 84572 0 . 70129 1 . 00000 -0 . 13693 0 . 65253 
0 . 0001 0 . 0001 0 . 0  0 . 3534  0 . 0001 
LEAF - 0 . 32021 0 . 17703 -0 . 13693 1 . 00000 0 . 11256  
0 . 0265 0 . 2287 0 . 3534 0 . 0  0 . 4462 
BITES 0 . 55954 0 . 44735 0 . 65253 0 . 11256  1 . 00000 
0 . 0001 0 . 0014 0 . 0001 0 . 4462 0 . 0  
Appendices 286 
APPENDIX 4. 5. Correlation matrices with Pearson Correlation Coefficients and p. 
values of Experiment 5 Period 2 - characteristics of birdsfoot trefoil and 
red clover plants. 
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 48 
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 39755 0 . 93175 0 . 67672 0 . 41193  
0 . 0  0 . 0051 0 . 0001 0 . 0001 0 . 003 6 
HEIGHT 0 . 39755 1 . 00000 0 . 70358 0 . 50352 0 . 60727 
0 . 0051 0 . 0  0 . 0001 0 . 0003 0 . 0001 
VOLUME 0 . 93175 0 . 70358 1 .  00000 0 . 72332  0 . 55932 
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0001 
LEAF 0 . 67672 0 . 50352 0 . 72332 1 . 00000 0 . 71943  
0 . 0001 0 . 0003 0 . 0001 0 . 0  0 . 0001 
MASS 0 . 41193 0 . 60727 0 . 55932 0 . 71943 1 . 00000 
0 . 0036 0 . 0001 0 . 0001 0 . 0001 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 45853 0 . 94412 -0 . 13048 0 . 05364  
0 . 0  0 . 0010 0 . 0001 0 . 3767  0 . 7173  
HEIGHT 0 . 45853 1 . 00000 0 . 72581  0 . 00415  0 . 58554 
0 . 0010 0 . 0  0 . 0001 0 . 9776  0 . 0001 
VOLUME 0 . 94412 0 . 72581 1 . 00000 -0 . 09946  0 . 25869  
0 . 0001 0 . 0001 0 . 0  0 . 5012 0 . 0758 
LEAF - 0 . 13048 0 . 00415 -0 . 09946  1 . 00000 0 . 09735  
0 . 3767 0 . 9776  0 . 5012 0 . 0  0 . 5104  
MASS 0 . 05364 0 . 58554 0 . 25869  0 . 09735 1 . 00000 
0 . 7173 0 . 0001 0 . 0758  0 . 5104 0 . 0  
BIRDSFOOT TREFOIL ECT CONCENTRATION, MORPHOLOGICAL 
CHARACTERISTICS AND NUMBER OF BITES 
Pears on Correlation Coefficients / Pr ob > I R I  under Ho : Rho=O / N = 16  
TANNIN AREA HEIGHT VOLUME LEAF BITES 
TANNIN 1 . 00000 -0 . 09732 -0 . 60540 -0 . 26827 -0 . 67276  - 0 . 50307 
0 . 0  0 . 7199  0 . 0130  0 . 3151 0 . 0043 0 . 0470 
AREA -0 . 09732 1 . 00000 0 . 49337  0 . 95691 0 . 72687 0 . 24270 
0 . 7199  0 . 0  0 . 0521 0 . 0001 0 . 0014 0 . 3651  
HEIGHT -0 . 60540 0 . 49337 1 . 00000 0 . 72113 0 . 81835  0 . 53044 
0 . 0130 0 . 0521 0 . 0  0 . 0016 0 . 0001 0 . 0345 
VOLUME -0 . 26827 0 . 95691 0 . 72113 1 . 00000 0 . 83587 0 . 36427 
0 . 3 151  0 . 0001 0 . 0016 0 . 0  0 . 0001 0 . 1654 
LEAF -0 . 67276  0 . 72687 0 . 81835 0 . 83587 1 .  00000 0 . 59289 
0 . 0043 0 . 0014 0 . 0001 0 . 0001 0 . 0  0 . 0155 
BITES -0 . 50307 0 . 24270 0 . 53044 0 . 36427 0 . 59289 1 . 00000 
0 . 0470 0 . 3 651 0 . 0345 0 . 1654 0 . 0155 0 . 0  
Appendices 287 
RED CLOVER FORMONONETIN CONCENTRATION MORPHOLOGICAL 
CHARACTERISTICS AND NUMBER OF BITES 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 1 6  
FORM AREA HEIGHT VOLUME LEAF BITES 
FORM 1 . 00000 0 . 14113 -0 . 04916  0 . 09622 0 . 09261 0 . 17249  
0 . 0  0 . 6021 0 . 8565 0 . 7230  0 . 7330  0 . 5229 
AREA 0 . 14113  1 .  00000 0 . 51480 0 . 96082 0 . 17522 0 . 80100 
0 . 6021 0 . 0  0 . 0413 0 . 0001 0 . 5163 0 . 0002 
HEIGHT - 0 . 04916  0 . 51480 1 . 00000 0 . 73224 0 . 03475 0 . 14306  
0 . 8565  0 . 0413 0 . 0  0 . 0013 0 . 8983 0 . 5971  
VOLUME 0 . 09622 0 . 9 6082 0 . 73224 1 . 00000 0 . 15043 0 . 68256  
0 . 7230  0 . 0001 0 . 0013 0 . 0  0 . 5782 0 . 0036  
LEAF 0 . 09261  0 . 17522 0 . 03475 0 . 15043 1 . 00000 0 . 2 6445  
0 . 7330  0 . 5163 0 . 8983 0 . 5782 0 . 0  0 . 3223  
BITES 0 . 17249  0 . 80100 0 . 14306 0 . 68256 0 . 2 6445 1 . 00000 
0 . 5229  0 . 0002 0 . 5971  0 . 0036 0 . 3223 0 . 0  
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS AND NUMBER 
OF BITES 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 . 00000 0 . 39755 0 . 93175 0 . 67672 0 . 28467  
0 . 0  0 . 0051 0 . 0001 0 . 0001 0 . 0499 
HEIGHT 0 . 39755 1 . 00000 0 . 70358 0 . 50352 0 . 43985  
0 . 0051 0 . 0  0 . 0001 0 . 0003 0 . 0018 
VOLUME 0 . 93175 0 . 70358 1 .  00000 0 . 72332  0 . 39452 
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0055 
LEAF 0 . 67672 0 . 50352 0 . 72332  1 .  00000 0 . 57495 
0 . 0001 0 . 0003 0 . 0001 0 . 0  0 . 0001 
BITES 0 . 28467 0 . 43985 0 . 39452 0 . 57495  1 . 00000 
0 . 0499 0 . 0018 0 . 0055 0 . 0001 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS AND NUMBER OF 
BITES 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 48 
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 . 00000 0 . 45853 0 . 94412 - 0 . 13048 0 . 78368  
0 . 0  0 . 0010 0 . 0001 0 . 3767 0 . 0001 
HEIGHT 0 . 45853 1 . 00000 0 . 72581 0 . 00415 0 . 47889 
0 . 0010 0 . 0  0 . 0001 0 . 9776  0 . 0006 
VOLUME 0 . 94412 0 . 72581 1 . 00000 - 0 . 09946  0 . 78423 
0 . 0001 0 . 0001 0 . 0  0 . 5012 0 . 0001 
LEAF - 0 . 13048 0 . 00415 - 0 . 09946 1 . 00000 0 . 04855 
0 . 3767 0 . 9776 0 . 5012 0 . 0  0 . 7431  
BITES 0 . 78368 0 . 47889 0 . 78423 0 . 04855 1 . 00000 
0 . 0001 0 . 0006 0 . 0001 0 . 7431  0 . 0  
Appendices 288 
APPENDIX 4. 6. Correlation matrices with Pearson Correlation Coefficients and P­
values of Experiment 5 Period 3 - characteristics of birdsfoot trefoil and 
red clover plants. 
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS 
Pears on Correlation Coefficients / Prob > J R J  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 66337 0 . 95504 0 . 49994  0 . 43495  
0 . 0  0 . 0001 0 . 0001 0 . 0003 0 . 0020  
HEIGHT 0 . 66337 1 . 00000 0 . 85539 0 . 50360 0 . 62644  
0 . 0001 0 . 0  0 . 0001 0 . 0003 0 . 0001 
VOLUME 0 . 95504 0 . 85539 1 . 00000 0 . 54559 0 . 54928 
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0001  
LEAF 0 . 49994 0 . 50360 0 . 54559 1 . 00000 0 . 62960 
0 . 0003 0 . 0003 0 . 0001 0 . 0  0 . 0001 
MASS 0 . 43495 0 . 62644 0 . 54928 0 . 62960 1 .  00000 
0 . 0020 0 . 0001 0 . 0001 0 . 0001 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Prob > J R J  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 50765 0 . 95924 0 . 24888 0 . 56170 
0 . 0  0 . 0002 0 . 0001 0 . 0880 0 . 0001 
HEIGHT 0 . 50765 1 . 00000 0 . 73042 0 . 33649  0 . 71314 
0 . 0002 0 . 0  0 . 0001 0 . 0194 0 . 0001 
VOLUME 0 . 95924 0 . 73042 1 . 00000 0 . 30766  0 . 67918  
0 . 0001 0 . 0001 0 . 0  0 . 0334  0 . 0001 
LEAF 0 . 24888 0 . 33649  0 . 30766 1 . 00000 0 . 57524 
0 . 0880 0 . 0194 0 . 0334 0 . 0  0 . 0001 
MASS 0 . 56170 0 . 71314 0 . 67918  0 . 57524 1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0001 0 . 0  
BIRDSFOOT TREFOIL ECT CONCENTRATION, MORPHOLOGICAL 
CHARACTERISTICS AND NUMBER OF BITES 
Pearson Correlation Coefficients Prob > J R J  under Ho : Rho=O / N = 1 5  
TANNIN AREA HEIGHT VOLUME LEAF BITES 
TANNIN 1 . 00000 0 . 03988  0 . 14485 0 . 09258 - 0 . 48602 -0 . 01384 
0 . 0  0 . 8878 0 . 6065 0 . 7428 0 . 0662 0 . 9609 
AREA 0 . 03988 1 .  00000 0 . 86008 0 . 98304 0 . 59211 0 . 51227 
0 . 8878 0 . 0  0 . 0001 0 . 0001 0 . 0200 0 . 0509 
HEIGHT 0 . 14485 0 . 86008 1.  00000 0 . 93598 0 . 63244 0 . 60250 
0 . 6065 0 . 0001 0 . 0  0 . 0001 0 . 0114 0 . 0175 
VOLUME 0 . 09258 0 . 98304 0 . 93598  1 . 00000 0 . 61934 0 . 55326 
0 . 7428 0 . 0001 0 . 0001 0 . 0  0 . 0138 0 . 0324 
LEAF - 0 . 48602 0 . 59211 0 . 63244 0 . 61934  1 . 00000 0 . 55464 
0 . 0662 0 . 0200 0 . 0114 0 . 0138 0 . 0  0 . 0319  
BITES -0 . 01384 0 . 51227 0 . 60250 0 . 55326 0 . 55464 1 . 00000 
0 . 9609 0 . 0509 0 . 0175 0 . 0324 0 . 0319 0 . 0  
Appendices 289 
RED CLOVER FORMONONETIN CONCENTRATION, MORPHOLOGICAL 
CHARACTERISTICS AND NUMBER OF BITES 
Pearson Correlation Coefficients / Prob > I R I  under Ho: Rho=O / N = 16  
FORM AREA HEIGHT VOLUME LEAF BITES 
FORM 1 . 00000 0 . 01195  -0 . 17947 -0 . 05222 -0 . 10481 0 . 25581 
0 . 0  0 . 9650 0 . 5060 0 . 8477 0 . 6993 0 . 3389  
AREA 0 . 01195 1 . 00000 0 . 64001 0 . 96512 0 . 60020 0 . 70836 
0 . 9650 0 . 0  0 . 0076 0 . 0001 0 . 0140 0 . 0021 
HEICHT -0 . 17947 0 . 64001 1 .  00000 0 . 81886 0 . 52794 0 . 65241 
0 . 5060 0 . 007 6  0 . 0  0 . 0001 0 . 0356  0 . 0062 
VOLUME -0 . 05222 0 . 9 6512 0 . 81886  1 . 00000 0 . 62827  0 . 75148 
0 . 8477  0 . 0001 0 . 0001 0 . 0  0 . 0092 0 . 0008 
LEAF - 0 . 10481 0 . 60020 0 . 52794 0 . 62827 1 . 00000 0 . 5 0826 
0 . 6993  0 . 0140 0 . 0356  0 . 0092 0 . 0  0 . 0444 
BITES 0 . 25581  0 . 70836 0 . 65241 0 . 75148 0 . 50826 1 . 00000 
0 . 3389 0 . 0021 0 . 0062 0 . 0008 0 . 0444 0 . 0  
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS AND NUMBER 
OF BITES 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 . 00000 0 . 66337 0 . 95504 0 . 49994  0 . 43815  
0 . 0  0 . 0001 0 . 0001 0 . 0003 0 . 0018  
HEIGHT 0 . 66337 1 .  00000 0 . 85539 0 . 50360 0 . 50888 
0 . 0001 0 . 0  0 . 0001 0 . 0003 0 . 0002 
VOLUME 0 . 95504 0 . 85539 1 . 00000 0 . 54559 0 . 50491 
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0003 
LEAF 0 . 49994 0 . 50360 0 . 54559 1 . 00000 0 . 43414  
0 . 0003 0 . 0003 0 . 0001 0 . 0  0 . 0020  
BITES 0 . 43815 0 . 50888 0 . 50491 0 . 43414  1 . 00000 
0 . 0018 0 . 0002 0 . 0003 0 . 0020 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS AND NUMBER OF 
BITES 
Pears on Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 48  
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 . 00000 0 . 50765 0 . 95924 0 . 24888 0 . 56760  
0 . 0  0 . 0002 0 . 0001 0 . 0880 0 . 0001 
HEIGHT 0 . 50765 1 . 00000 0 . 73042 0 . 33649  0 . 57313  
0 . 0002 0 . 0  0 . 0001 0 . 0194 0 . 0001 
VOLUME 0 . 95924 0 . 73042 1 . 00000 0 . 30766 0 . 63793  
0 . 0001 0 . 0001 0 . 0  0 . 0334  0 . 0001 
LEAF 0 . 24888 0 . 33649  0 . 30766  1 . 00000 0 . 34966  
0 . 0880 0 . 0194 0 . 0334 0 . 0  0 . 0148  
BITES 0 . 56760 0 . 57313 0 . 63793 0 . 34966  1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0148 0 . 0  
Appendices 290 
APPENDIX 4. 7. Correlation matrices with Pearson Correlation Coefficients and P-
values of Experiment 6 - plant nutritional characteristics 
BIRDS FOOT TREFOIL PLANTS 
Pearson Correlation Coefficients Prob > I R I under Ho : Rho=O / N = 1 6  
PROTEIN LIPID ADF NDF CHO ASH INVITRO BITES 
PROTEIN 1 . 00000 0 . 82647 -0 . 75202 -0 . 18241  -0 . 04070 0 . 56706 0 . 74957 0 . 36957 
0 . 0  0 . 0001 0 . 0008 0 . 4989  0 . 8810 0 . 0220 0 . 0008 0 . 1589  
LIPID 0 . 82 647 1 .  00000 -0 . 85616  -0 . 62122 0 . 39394  0 . 71549  0 . 86038 0 . 2 9233  
0 . 0001 0 . 0  0 . 0001 0 . 0102 0 . 1311  0 . 0018 0 . 0001 0 . 2719  
ADF -0 . 75202 -0 . 85616 1 . 00000 0 . 57506 -0 . 51325 -0 . 49017 -0 . 97650 - 0 . 33854 
0 . 0008 0 . 0001 0 . 0  0 . 0198 0 . 0420 0 . 0539  0 . 0001 0 . 1996 
NDF - 0 . 18241 - 0 . 62122 0 . 57506 1 . 00000 -0 . 74256 - 0 . 66864 -0 . 58705 0 . 12507 
0 . 4989  0 . 0102 0 . 0198  0 . 0  0 . 0010 0 . 0046  0 . 0168 0 . 6444 
CHO - 0 . 04070 0 . 39394  - 0 . 51325 -0 . 74256  1 .  00000 0 . 26097 0 . 56136 0 . 10249 
0 . 8810  0 . 1311  0 . 0420 0 . 0010 0 . 0  0 . 3289  0 . 0237 0 . 7056 
ASH 0 . 56706  0 . 71549  - 0 . 49017 - 0 . 66864  0 . 2 6097 1 . 00000 0 . 58122 0 . 02521  
0 . 0220  0 . 0018 0 . 0539  0 . 0046 0 . 3289  0 . 0  0 . 0182 0 . 9262 
INVITRO 0 . 74957 0 . 86038 -0 . 97650 -0 . 58705 0 . 56136  0 . 58122 1 . 00000 0 . 39623  
0 . 0008 0 . 0001 0 . 0001 0 . 0168 0 . 0237 0 . 0182 0 . 0  0 . 1287 
BITES 0 . 36957 0 . 29233 - 0 . 3 3 854 0 . 12507 0 . 10249 0 . 02521 0 . 39623 1 . 00000 
0 . 1589  0 . 2719  0 . 1996  0 . 6444 0 . 7056 0 . 9262 0 . 1287 0 . 0  
RED CLOVER PLANTS 
Pearson Correlation Coefficients Prob > I R I under Ho : Rho=O / N = 1 6  
PROTEIN LIPID ADF NDF CHO ASH INVITRO BITES 
PROTEIN 1 . 00000 0 . 88656 0 . 58518 0 . 63669  -0 . 85542 0 . 86413 - 0 . 57851 - 0 . 41688  
0 . 0  0 . 0001 0 . 0173 0 . 0080 0 . 0001 0 . 0001 0 . 0189 0 . 1082 
LIPID 0 . 88656  1 . 00000 0 . 56582 0 . 39133  - 0 . 86009 0 . 90217 - 0 . 56364 - 0 . 42314 
0 . 0001 0 . 0  0 . 0223 0 . 1339  0 . 0001 0 . 0001 0 . 0230 0 . 1025 
ADF 0 . 58518 0 . 56582 1.  00000 0 . 68076 -0 . 85678 0 . 80302 - 0 . 99942 - 0 . 60927 
0 . 0173 0 . 0223 0 . 0  0 . 0037 0 . 0001 0 . 0002 0 . 0001 0 . 0122 
NDF 0 . 63669  0 . 39133 0 . 6807 6 1 .  00000 -0 . 62268 0 . 53448 - 0 . 67367 - 0 . 30432 
0 . 0080 0 . 1339  0 . 0037 0 . 0  0 . 0100 0 . 0329  0 . 0042 0 . 2518 
CHO -0 . 85542 - 0 . 86009 -0 . 85678 -0 . 62268 1 . 00000 -0 . 97046 0 . 85243 0 . 58488 
0 . 0001  0 . 0001 0 . 0001 0 . 0100 0 . 0  0 . 0001 0 . 0001 0 . 0173 
ASH 0 . 86413 0 . 90217 0 . 80302 0 . 53448 -0 . 97046 1.  00000 - 0 . 80038 - 0 . 55674 
0 . 0001 0 . 0001 0 . 0002 0 . 0329  0 . 0001 0 . 0  0 . 0002 0 . 0251  
INVITRO -0 . 57851 -0 . 56364 -0 . 99942 - 0 . 67367  0 . 85243 -0 . 80038 1 . 00000 0 . 59817 
0 . 0189 0 . 0230 0 . 0001 0 . 0042 0 . 0001 0 . 0002 0 . 0  0 . 0144 
BITES -0 . 4 1688 - 0 . 42314 -0 . 60927 -0 . 30432 0 . 58488 - 0 . 55674 0 . 59817 1 . 00000 
0 . 1082 0 . 1025 0 . 0122 0 . 2518 0 . 0173 0 . 0251  0 . 0144 0 . 0  
Appendices 29 1 
APPENDIX 4. 8. Correlation matrices with Pearson Correlation Coefficients and p. 
values of Experiment 6 Period 1 - characteristics of birdsfoot trefoil and 
red clover plants. 
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 9 6  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 .  00000 0 . 38059 0 . 88601  0 . 3 9506 0 . 2 9482 
0 . 0  0 . 0001 0 . 0001 0 . 0001 0 . 0035 
HEIGHT 0 . 38059 1 . 00000 0 . 76597 0 . 38425  0 . 68364 
0 . 0001 0 . 0  0 . 0001 0 . 0001 0 . 0001 
VOLUME 0 . 88601 0 . 76597 1 . 00000 0 . 46730 0 . 54773 
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0001 
LEAF 0 . 39506  0 . 3 8425 0 . 46730 1 . 00000 0 . 7 1891 
0 . 0001 0 . 0001 0 . 0001 0 . 0  0 . 0001 
MASS 0 . 29482 0 . 68364 0 . 54773 0 . 71891  1 . 00000 
0 . 0035 0 . 0001 0 . 0001 0 . 0001 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 9 6  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 48451 0 . 94801 0 . 04390 0 . 3 9031 
0 . 0  0 . 0001 0 . 0001 0 . 6710 0 . 0001 
HEIGHT 0 . 48451 1 . 00000 0 . 73771 -0 . 05826 0 . 56313 
0 . 0001 0 . 0  0 . 0001 0 . 5729  0 . 0001 
VOLUME 0 . 94801 0 . 73771 1 . 00000 0 . 01269 0 . 50608 
0 . 0001 0 . 0001 0 . 0  0 . 9024 0 . 0001 
LEAF 0 . 04390 -0 . 05826 0 . 01269 1 . 00000 0 . 3 6849 
0 . 6710 0 . 5729 0 . 9024 0 . 0  0 . 0002 
MASS 0 . 39031 0 . 56313 0 . 50608 0 . 36849  1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0002 0 . 0  
BIRDSFOOT TREFOIL ECT CONCENTRATION, MORPHOLOGICAL 
CHARACTERISTICS AND NUMBER OF BITES 
Pearson Correlation Coefficients / Pr ob > I R I  under Ho : Rho=O / N = 32 
TANNIN AREA HEIGHT VOLUME LEAF BITES 
TANNIN 1 . 00000 -0 . 09251 - 0 . 40110 - 0 . 25397 - 0 . 63116  - 0 . 36284 
0 . 0  0 . 6145 0 . 0229 0 . 1 607 0 . 0001 0 . 0413 
AREA - 0 . 09251 1 .  00000 0 . 63146  0 . 92830 0 . 34409 0 . 69709 
0 . 6145 0 . 0  0 . 0001 0 . 0001 0 . 0538  0 . 0001 
HEIGHT - 0 . 40110 0 . 63146  1 . 00000 0 . 86923 0 . 59623  0 . 60167 
0 . 0229 0 . 0001 0 . 0  0 . 0001 0 . 0003 0 . 0003 
VOLUME - 0 . 2 5397 0 . 92830 0 . 86923 1 . 00000 0 . 50190 0 . 73580 
0 . 1607 0 . 0001 0 . 0001 0 . 0  0 . 0034 0 . 0001 
LEAF -0 . 63116 0 . 34409 0 . 5 9623 0 . 50190 1 . 00000 0 . 68781 
0 . 0001 0 . 0538 0 . 0003 0 . 0034 0 . 0  0 . 0001 
BITES -0 . 3 6284 0 . 69709 0 . 60167 0 . 73580 0 . 68781 1 . 00000 
0 . 0413 0 . 0001 0 . 0003 0 . 0001 0 . 0001 0 . 0  
Appendices 292 
RED CLOVER FORMONONETIN CONCENTRATION, 
CHARACTERISTICS AND NUMBER OF BITES 
MORPHOLOGICAL 
Pears on Correlation Coefficients Prob > I R I  under Ho : Rho=O I N = 32  
FORM AREA HEIGHT VOLUME LEAF BITES 
FORM 1 . 00000 - 0 . 00356 0 . 05963 0 . 01337 -0 . 05570 0 . 03227 
0 . 0  0 . 9846  0 . 7458 0 . 9421 0 . 7 62 1  0 . 8608 
AREA - 0 . 00356  1 . 00000 0 . 60621 0 . 95593 -0 . 08904 - 0 . 24196  
0 . 9846  0 . 0  0 . 0002 0 . 0001 0 . 6279  0 . 1821 
HEIGHT 0 . 05963 0 . 60621 1 . 00000 0 . 80188 -0 . 12831 -0 . 24736  
0 . 7458  0 . 0002 0 . 0  0 . 0001 0 . 4840 0 . 1723 
VOLUME 0 . 01337  0 . 95593 0 . 80188 1 . 00000 - 0 . 13525 - 0 . 28522 
0 . 9421  0 . 0001 0 . 0001 0 . 0  0 . 4 605 0 . 1136  
LEAF -0 . 05570 -0 . 08904 - 0 . 12831 -0 . 13525 1 . 00000 0 . 33105 
0 . 7621  0 . 6279  0 . 4840 0 . 4605 0 . 0  0 . 0642 
BITES 0 . 03227 - 0 . 24196  -0 . 24736  - 0 . 28522 0 . 33105 1 . 00000 
0 . 8 608 0 . 1821  0 . 1723 0 . 1136  0 . 0642 0 . 0  
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS AND NUMBER 
OF BITES 
Pearson Correlation Coefficients I Prob > I R I under Ho : Rho=O I N = 9 6  
AREA HEIGHT 
AREA 1 . 00000 0 . 3 8059 
0 . 0  0 . 0001 
HEIGHT 0 . 38059 1 . 00000 
0 . 0001 0 . 0  
VOLUME 0 . 88601 0 . 76597  
0 . 0001 0 . 0001 
LEAF 0 . 39506  0 . 38425 
0 . 0001 0 . 0001 
BITES 0 . 60601 0 . 43104 
0 . 0001 0 . 0001 
RED CLOVER MORPHOLOGICAL 
BITES 
VOLUME LEAF BITES 
0 . 88601 0 . 39506  0 . 60601 
0 . 0001 0 . 0001 0 . 0001 
0 . 7 6597 0 . 38425 0 . 43104 
0 . 0001 0 . 0001 0 . 0001 
1 . 00000 0 . 46730  0 . 63741 
0 . 0  0 . 0001 0 . 0001 
0 . 46730 1 . 00000 0 . 62218 
0 . 0001 0 . 0  0 . 0001 
0 . 63741 0 . 62218  1 . 00000 
0 . 0001 0 . 0001 0 . 0  
CHARACTERISTICS AND NUMBER OF 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 9 6  
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 .  00000 0 . 48451 0 . 94801 0 . 04390 0 . 12597 
0 . 0  0 . 0001 0 . 0001 0 . 6710 0 . 2214  
HEIGHT 0 . 48451  1 . 00000 0 . 73771 - 0 . 05826 0 . 06079 
0 . 0001 0 . 0  0 . 0001 0 . 5729  0 . 5563  
VOLUME 0 . 94801 0 . 73771 1 .  00000 0 . 01269  0 . 11933  
0 . 0001 0 . 0001 0 . 0  0 . 9024 0 . 2469  
LEAF 0 . 04390 - 0 . 05826 0 . 01269 1 . 00000 0 . 24763 
0 . 6710  0 . 5729 0 . 9024 0 . 0  0 . 0150 
BITES 0 . 12597 0 . 06079 0 . 11933 0 . 24763  1 .  00000 
0 . 2214  0 . 5563 0 . 2469  0 . 0150 0 . 0  
Appendices 293 
APPENDIX 4.9. Correlation matrices with Pearson Correlation Coefficients and P­
values of Experiment 6 Period 2 - characteristics of birdsfoot trefoil and 
red clover plants. 
BIRDSFOOT TREFOIL MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients / Prob > I R I  under Ho : Rho=O / N = 96  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 . 00000 0 . 50705 0 . 9 6024 0 . 53047 0 . 40525 
0 . 0  0 . 0001 0 . 0001 0 . 0001 0 . 0001 
HEIGHT 0 . 50705 1 . 00000 0 . 72752 0 . 3 6859  0 . 58281 
0 . 0001 0 . 0  0 . 0001 0 . 0002 0 . 0001 
VOLUME 0 . 96024 0 . 72752 1 .  00000 0 . 54164  0 . 51136  
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0001 
LEAF 0 . 53047 0 . 36859 0 . 54164  1 .  00000 0 . 75514 
0 . 0001 0 . 0002 0 . 0001 0 . 0  0 . 0001 
MASS 0 . 40525 0 . 58281 0 . 51136  0 . 75514  1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0001 0 . 0  
RED CLOVER MORPHOLOGICAL CHARACTERISTICS 
Pearson Correlation Coefficients Prob > I R I  under Ho : Rho=O / N 96  
AREA HEIGHT VOLUME LEAF MASS 
AREA 1 .  00000 0 . 79143 0 . 97611  0 . 26509  0 . 56918 
0 . 0  0 . 0001 0 . 0001 0 . 0090  0 . 0001 
HEIGHT 0 . 79143  1 . 00000 0 . 90534 0 . 30731  0 . 67775 
0 . 0001 0 . 0  0 . 0001 0 . 0023  0 . 0001 
VOLUME 0 . 97611  0 . 90534 1 .  00000 0 . 29342  0 . 63638 
0 . 0001 0 . 0001 0 . 0  0 . 0037 0 . 0001 
LEAF 0 . 26509  0 . 30731 0 . 29342 1.  00000 0 . 4 9951 
0 . 0090 0 . 0023 0 . 0037 0 . 0  0 . 0001 
MASS 0 . 56918  0 . 67775 0 . 63638  0 . 49951  1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0001  0 . 0  
BIRDSFOOT TREFOIL ECT CONCENTRATION, MORPHOLOGICAL 
CHARACTERISTICS AND NUMBER OF BITES 
Pearson Correlation Coefficients Prob > I R I  under Ho : Rho=O / N = 32 
LNTANNIN LNAREA LNHEIGHT LNVOLUME LNLEAF LNBITES 
LNTANNIN 1 . 00000 0 . 07054 -0 . 14193 -0 . 00312 -0 . 60142 - 0 . 2 6943 
0 . 0  0 . 7012  0 . 4384 0 . 9865  0 . 0003 0 . 1359 
LNAREA 0 . 07054 1 . 00000 0 . 58124 0 . 97079 0 . 32934  0 . 63113 
0 . 7012 0 . 0  0 . 0005 0 . 0001 0 . 0657 0 . 0001 
LNHEIGHT - 0 . 14193 0 . 58124 1 . 00000 0 . 74500 0 . 47503 0 . 59118 
0 . 4384 0 . 0005 0 . 0  0 . 0001 0 . 0060  0 . 0004 
LNVOLUME -0 . 00312 0 . 97079 0 . 74500 1 . 00000 0 . 39643  0 . 68226 
0 . 9865 0 . 0001 0 . 0001 0 . 0  0 . 0247  0 . 0001 
LNLEAF -0 . 60142 0 . 32934  0 . 47503 0 . 39643 1 . 00000 0 . 69065 
0 . 0003 0 . 0657 0 . 0060 0 . 0247 0 . 0  0 . 0001 
LNBITES -0 . 26943 0 . 63113 0 . 59118 0 . 68226 0 . 69065 1 . 00000 
0 . 1359  0 . 0001 0 . 0004 0 . 0001 0 . 0001 0 . 0  
Appendices 294 
RED CLOVER FORMONONET:IN CONCENTRAT:ION MORPHOLOG:ICAL 
CHARACTER:IST:ICS AND NUMBER OF B:ITES 
Pearson Correlation Coefficients Prob > I R I  under Ho : Rho=O I N = 32 
FORM AREA HEIGHT VOLUME LEAF BITES 
FORM 1 . 00000 0 . 01478 -0 . 00105 0 . 01898 - 0 . 47464 - 0 . 07327 
0 . 0  0 . 9360  0 . 9955  0 . 9179  0 . 0061 0 . 6902 
AREA 0 . 01478 1 . 00000 0 . 86200 0 . 97773 0 . 15808 0 . 84889 
0 . 9360  0 . 0  0 . 0001 0 . 0001 0 . 3875 0 . 0001 
HEIGHT -0 . 00105 0 . 86200 1.  00000 0 . 94392 0 . 09680 0 . 76122 
0 . 9955  0 . 0001 0 . 0  0 . 0001 0 . 5982 0 . 0001 
VOLUME 0 . 01898  0 . 97773 0 . 94392 1 . 00000 0 . 12955 0 . 85153 
0 . 9179  0 . 0001 0 . 0001 0 . 0  0 . 4798 0 . 0001 
LEAF - 0 . 47464 0 . 15808 0 . 09680 0 . 12955 1 . 00000 0 . 27593 
0 . 0061 0 . 3875 0 . 5982 0 . 4798  0 . 0  0 . 1263  
BITES -0 . 07327 0 . 84889 0 . 7 6122 0 . 85153 0 . 27593 1 . 00000 
0 . 6902 0 . 0001 0 . 0001 0 . 0001 0 . 1263 0 . 0  
B:IRDSFOOT TREFO:IL MORPHOLOG:ICAL CHARACTER:IST:ICS AND NUMBER 
OF B:ITES 
Pearson Correlation Coefficients I Prob > I R I  under Ho : Rho=O I N = 96  
LNAREA LNHEIGHT LNVOLUME LNLEAF LNBITES 
LNAREA 1 . 00000 0 . 50705 0 . 96024 0 . 53047 0 . 66744 
0 . 0  0 . 0001 0 . 0001 0 . 0001 0 . 0001 
LNHEIGHT 0 . 50705 1.  00000 0 . 72752 0 . 3 6859 0 . 51102 
0 . 0001 0 . 0  0 . 0001 0 . 0002 0 . 0001 
LNVOLUME 0 . 96024 0 . 72752 1 . 00000 0 . 54164 0 . 69680 
0 . 0001 0 . 0001 0 . 0  0 . 0001 0 . 0001 
LNLEAF 0 . 53047 0 . 3 6859 0 . 54164 1 . 00000 0 . 73967 
0 . 0001 0 . 0002 0 . 0001 0 . 0  0 . 0001 
LNBITES 0 . 66744 0 . 51102 0 . 69680 0 . 73967  1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0001 0 . 0  
RED CLOVER MORPHOLOG:ICAL CHARACTER:IST:ICS AND NUMBER OF 
B:ITES 
Pearson Correlation Coefficients I Prob > I R I  under Ho : Rho=O I N = 9 6  
AREA HEIGHT VOLUME LEAF BITES 
AREA 1 . 00000 0 . 79143 0 . 97611  0 . 2 6509 0 . 77790 
0 . 0  0 . 0001 0 . 0001 0 . 0090  0 . 0001 
HEIGHT 0 . 79143 1 . 00000 0 . 90534 0 . 30731  0 . 69892 
0 . 0001 0 . 0  0 . 0001 0 . 0023 0 . 0001 
VOLUME 0 . 97611  0 . 90534 1.  00000 0 . 2 9342 0 . 78892 
0 . 0001 0 . 0001 0 . 0  0 . 0037 0 . 0001 
LEAF 0 . 26509 0 . 30731 0 . 29342 1 . 00000 0 . 40899 
0 . 0090 0 . 0023 0 . 0037 0 . 0  0 . 0001 
BITES 0 . 77790 0 . 69892 0 . 78892 0 . 40899 1 . 00000 
0 . 0001 0 . 0001 0 . 0001 0 . 0001 0 . 0  
Appendices 295 
APPENDIX 4. 10. Rumen manipulation - Experiment 6 
Table 4. 2. Botanical composition (% in DM of total material) and tannin and 
formononetin concentrations (%) of the minced material added to cow's  
rumen in the Period 1 .  
Period 1 Leaf stem Flower Dead matter Weeds DM Tannin Formononetin 
Day ] 
Red clover 
Astred 61  39 0 0 0 28 0.04 0.39 
Pawera 53 46 0 0 1 59 0.04 0.59 
Lotus spp. 
Goldie 19 1 8  4 8 5 1  32 0.07 0. 1 5  
Maku 29 59 9 2 1 29 0.59 0.2 
Day 2 
Red clover 
Astred 6 1  39 0 0 0 28 0.05 0.39 
Pawera 53 46 0 0 1 59 0.05 0.40 
Lotus sp. 
Goldie 42 46 5 3 4 30 0.23 0.2 
Maku 3 1  47 12 1 9 29 0.35 0.2 
Table 4.3. Botanical composition (% in DM of total material) and tannin and 
formononetin concentrations (%) of the minced material added to cow's 
rumen in Period 2. 
Period 2 Leaf Stem Flower Dead matter Weeds DM Tannin Formononetin 
Day ] 
Red clover 
Astred 7 1  6 0 16 7 29 0.04 0.45 
Pawera 79 1 0 10 10 32 0.04 0.50 
Lotus sp. 
Goldie 56 28 0 2 14 28 0. 19 0. 1 6  
Maku 47 2 1  0 3 29 28 0. 16  0.2 1 
Day 2 
Red clover 
Astred 77 8 2 1 1  2 19  0.04 0.43 
Pawera 86 3 0 3 8 1 6  0.06 0.50 
Lotus sp. 
Goldie 60 33 0 1 6 29 0.22 0. 1 8  
Maku 52 2 1  0 2 25 26 0. 15 0.23 
Appendices 296 
APPENDIX 4.10. Rumen manipulation - Experiment 6 
Table 4. 4. Kilograms (fresh weight) of rumen content taken out and of minced material 
put into cow's rumen with correspondent amount of lucerne pellets (kg) and 
secondary compound [extractable condensed tannin (ECT) or formononetin] 
(�) added to the minced material in Period 1 .  
Rumen content modification (kg) 
Period 1 Take Put Lucerne Secondary 
out into p'ellets Comp'ound 
Dayl (Kg) (Kg) (Kg) (g) 
Lotus sp. 
ECT 
Goldie 2 1 .2 14.5 2.5 3 
Maku 27.3 15.2 3.0 26 
Day2 
Lotus sp. 
Goldie 24.8 17.0 2.8 12 
Maku 20.0 15.0 2.8 15  
Dayl 
Red clover Fonnononetin 
Astred 28.0 15 . 1  3.0 16  
Pawera 28.0 16.0 3.5 56 
Day 2 
Red clover 
Astred 15.5 14.4 3 .0 16 
Pawera 17.7 15.5 3.0 62 
Appendices 297 
APPENDIX 4.10. Rumen manipulation - Experiment 6 
Table 4. 5. Kilograms (fresh weight) of rumen content taken out and of minced material 
put into cow's rumen with correspondent amount (kg) of lucerne pellets and 
secondary compounds [extractable condensed tannin (BeT) or formononetin] 
(g) added to the minced material in Period 2. 
Period 2 
Dayl 
Lotus sp. 
Goldie 
Maku 
Day2 
Lotus sp. 
Goldie 
Maku 
Dayl 
Red clover 
Astred 
Pawera 
Day 2 
Red clover 
Astred 
Pawera 
Take 
out 
(Kg) 
1 1 .7 
23.3 
10.4 
6.5 
14.9 
15 .8 
24.3 
9.9 
Rumen content modification (kg) 
Put Lucerne Secondary 
into pellets Compound 
(Kg) (Kg) (g) 
15.4 
16.0 
17.2 
14.0 
6.7 
1 1 .6 
6.7 
13 .9 
2.5 
3.0 
2.7 
2.5 
1 .2 
2.2 
1 . 1  
2.2 
ECT 
7 
9 
9 
8 
Formononetin 
9 
19 
5 
1 1