Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An evaluation of initial rates of digestion in a strain of Alfalfa selected to prevent bloat in grazing… Berg, Bjorn Peter 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-486052.pdf [ 9.36MB ]
Metadata
JSON: 831-1.0089771.json
JSON-LD: 831-1.0089771-ld.json
RDF/XML (Pretty): 831-1.0089771-rdf.xml
RDF/JSON: 831-1.0089771-rdf.json
Turtle: 831-1.0089771-turtle.txt
N-Triples: 831-1.0089771-rdf-ntriples.txt
Original Record: 831-1.0089771-source.json
Full Text
831-1.0089771-fulltext.txt
Citation
831-1.0089771.ris

Full Text

A N E V A L U A T I O N O F I N I T I A L R A T E S O F D I G E S T I O N I N A S T R A I N O F A L F A L F A S E L E C T E D T O P R E V E N T B L O A T I N G R A Z I N G R U M I N A N T S by B J O R N P E T E R B E R G B.Sc, The University of Alberta, 1976 M.Sc, The University of Alberta, 1983 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Animal Science) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 2000 © Bjorn Peter Berg, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of /Jn fyn & / ^/ The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract The practice of grazing alfalfa (Medicago sativa L.) is limited by frothy bloat. Experiments were conducted to verify that AC Grazeland B r (ACG), an alfalfa selected for a low initial rate of digestion (LIRD), has an effect on digestion and bloat that does not differ from a parental standard, Beaver (BVR). The rate of maturity of ACG, measured morphologically (mean stage by stem count, MSC), was greater (P< .05) than BVR in the field and in a controlled environment. ACG stems matured at a rate of 0.09 MSC d"1 compared to 0.06 MSC d"1 for BVR. Kinetic models of in vivo dry matter disappearance (DMD) on fresh stem tips in the rumens of steers showed that DMD rates declined with advancing maturity. The average hourly rate of DMD did not differ (P>.05) between cultivars. The postulated requirement to prevent bloat was a 25% reduction in DMD from that of a conventional standard alfalfa. Immature ACG had a DMD 4.9% lower than immature BVR. Mature ACG had a DMD 22.2% less than the same standard. Gas production (GP) from in vitro incubations of ACG leaves in rumen fluid, decreased with advancing maturity (P<.05). BVR leaf GP remained the same (P>.05) with advancing maturity. Maximum rates of GP from fresh leaves did not differ between cultivars. Late bud ACG stems produced less gas (P^.05) than any other stem class including those of BVR. A predecessor of ACG, LIRD-3, was fed to wethers. Its neutral detergent fibre (NDF) was less than BVR (?<, .05, 39.4% and 43.6%, respectively) when LIRD-3 was less mature, and similar to BVR (P>.05, 42.8 and 44.1%, respectively) when the maturities were equivalent. Digestibility of the LIRD-3 cell wall was similar (P>.05) to BVR. When immature ACG and BVR were fed, bloat frequencies were similar (P>.05). Bloat incidence in grazing steers was reduced (P<.05) in only one autumn trial of six trials conducted over two years. Bloat distensions in that trial occurred 29.6% less often (P<.05) and incidents of severe froth and pressure were 55.2% lower (P<.05) in cattle grazing ACG. Selection for LIRD produced ACG, an early maturing cultivar, with a digestion profile otherwise similar to a standard, BVR. Maturity is more crucial than the cultivar for preventing bloat. Table of Contents Abstract ii List of Tables vii List of Figures ix List of Abbreviations xi Acknowledgements xii Preface xiii Chapter 1 A Review of Alfalfa Grazing and Bloat Theory 1 1.1 INTRODUCTION 1 1.2 GRAZING ALFALFA: AN OPPORTUNITY 2 Current Status 2 Potential for Improvement 3 1.3 BLOAT MANAGEMENT 5 Diagnosis 5 C A U S E S OF B L O A T 5 T Y P E S OF B L O A T 6 Sub-acute bloat 6 Acute bloat 7 Frothy and free-gas bloat 8 Theory and Practice of Bloat Prevention 9 T H E E X C E S S I V E C O N S U M P T I O N HYPOTHESIS 9 T H E O R Y OF A N I M A L SUSCEPTIBILITY 10 T H E F O A M HYPOTHESIS 12 T H E C E L L R U P T U R E HYPOTHESIS 15 1.4 ALFALFA FOR GRAZING 17 Precursors of a Grazing Alfalfa 18 Evaluating Low Initial Rates of Digestion in Alfalfa 19 JUSTIFICATION F O R R E S E A R C H 19 H Y P O T H E S I S A N D R E S E A R C H O U T L I N E 2 0 Chapter 2 Rates of Dry Matter Disappearance 2 2 2.1 INTRODUCTION 2 2 2.2 MATERIALS AND METHODS 2 5 Morphological Stage of Development 2 5 M A T U R I T Y IN T H E FIELD 2 6 M A T U R I T Y IN A C O N T R O L L E D E N V I R O N M E N T 2 7 iii Dry Matter Disappearance 2 7 A N I M A L M A N A G E M E N T 2 7 N Y L O N B A G P R O C E D U R E S 2 8 F R E S H S T E M TIP CUTTINGS ( E X P E R I M E N T 1) 2 9 F R E S H M A S T I C A T E F R O M G R A Z I N G C A T T L E ( E X P E R I M E N T 2) 31 D R I E D , G R O U N D S T E M TIPS OF LIRD A L F A L F A PROGENITORS ( E X P E R I M E N T 3). 33 Statistics 3 4 M O R P H O L O G I C A L S T A G E OF D E V E L O P M E N T 3 4 F R E S H CUTTINGS ( E X P E R I M E N T i) 3 4 F R E S H M A S T I C A T E ( E X P E R I M E N T 2) 3 6 L I R D A L F A L F A PROGENITORS ( E X P E R I M E N T 3) 3 6 K I N E T I C A N A L Y S I S 3 6 2.3 RESULTS AND DISCUSSION 3 7 Morphological Stage of Development 3 7 I N T H E FIELD 3 8 I N A C O N T R O L L E D E N V I R O N M E N T 3 9 Dry Matter Disappearance 41 F R E S H S T E M TIP C U T T I N G S ( E X P E R I M E N T 1) 4 2 F R E S H M A S T I C A T E F R O M G R A Z I N G C A T T L E ( E X P E R I M E N T 2) 4 7 D R I E D , G R O U N D S T E M TIPS OF LIRD A L F A L F A PROGENITORS ( E X P E R I M E N T 3) . 4 7 Kinetic Models 4 8 2.4 CONCLUSIONS 5 5 Chapter 3 Gas Production Kinetics 5 8 3.1 INTRODUCTION 5 8 3.2 MATERIALS AND METHODS 6 0 Gas Production Experiments (Alfalfa Comparisons) 6 0 E F F E C T OF C U L T I V A R ( E X P E R I M E N T 1) 6 0 E F F E C T OF L E A F T R E A T M E N T ( E X P E R I M E N T 2) 61 E F F E C T O F M A T U R I T Y ( E X P E R I M E N T 3) 6 3 Gas Production Methods 63 R U M E N F L U I D C O L L E C T I O N 6 4 M A N O M E T R I C M E T H O D 65 H O E N H E I M M E T H O D ( M E N K E ' S A P P A R A T U S ) 6 6 Statistics 6 7 D E S C R I P T I V E STATISTICS 6 7 KINETIC ANALYSIS 6 7 3.3 RESULTS AND DISCUSSION 6 9 Effect of Cultivar on Gas Production (Experiment 1) 6 9 P A R E N T A L C U L T I V A R S (TRIAL 1) 6 9 L I R D S Y N T H E T I C C U L T I V A R S (TRIAL 2) 7 5 Effect of Leaf Treatment on Gas Production (Experiment 2 ) 7 9 G R O U N D O R W H O L E L E A V E S 7 9 C R U S H E D , P E R F O R A T E D OR C U T L E A V E S : 81 iv Effect of Maturity on Gas Production (Experiment 3) 85 L E A F L E T S 85 S T E M S A N D PETIOLES : 88 Kinetic Models 91 3.4 CONCLUSIONS 94 Chapter 4 Feed Composition, Digestibility and Bloat 96 4.1 INTRODUCTION 96 4.2 MATERIALS AND METHODS 99 Digestibility Experiments 100 F I E L D A N D SITE DESCRIPTION 100 DIGESTIBILITY OF LIRD PROGENITORS ( E X P E R I M E N T 1) 100 Trial feeding periods 100 Animal husbandry and feed allowance 100 Sample collection and processing 101 DIGESTIBILITY OF A C G R A Z E L A N D ( E X P E R I M E N T 2) 102 Trial feeding periods and maturity 102 Animal husbandry and feed allowance 103 Bloat . . 104 Sample collection and processing 105 Laboratory Methods 105 E X P E R I M E N T 1 105 E X P E R I M E N T 2 106 Statistics 106 4.3 RESULTS AND DISCUSSION 108 Digestibility of LIRD Progenitors (Experiment 1) 108 F E E D COMPOSITION 108 F E E D DIGESTIBILITY 112 Digestibility of AC Grazeland (Experiment 2) 114 A L F A L F A M A T U R I T Y 115 F E E D COMPOSITION 116 F E E D DIGESTIBILITY 121 B L O A T INCIDENCE 124 4.4 CONCLUSIONS 126 Chapter 5 Bloat Incidence in Grazing Steers 128 5.1 INTRODUCTION 128 Bloat in Feeding Trials 128 Bloat in Grazing Trials of LIRD Alfalfa 129 5.2 MATERIALS AND METHODS 131 Field Conditions and Management 131 SITE DESCRIPTIONS 131 v F I E L D M A N A G E M E N T 131 A L F A L F A M A T U R I T Y 132 Animal Management 132 G R A Z I N G PERIODS 133 B L O A T M E A S U R E M E N T 133 Statistical Analysis 135 5.3 RESULTS AND DISCUSSION 136 Bloat Frequency Distribution 136 Effect of Cultivar and Plant Maturity on Bloat 139 5.4 CONCLUSIONS 141 Chapter 6 Analyses and Implications 143 6.1 SYNTHESIS 143 In Summary 143 On a LIRD Alfalfa 145 On Bloat 147 On a Unified Theory 148 C E L L R U P T U R E HYPOTHESIS. 148 P A S S A G E R A T E T H E O R Y 149 6.2 FUTURE DIRECTIONS 151 References 153 Appendix 168 I EXPONENTIAL MODELS FOR KINETIC ANALYSIS 168 Introduction 168 Analysis and Interpretation 168 Generic Equations 169 The Gompertz Growth Equation 170 References 172 II DMD AND GP TABLES 175 List of Tables 175 III METEOROLOGICAL DATA 191 vi List of Tables 1.1 The potential economic benefit of grazing alfalfa to calf production in Alberta . . 3 2.1 Standing crop dry matter and morphological stage of development of alfalfa (cv. AC Grazeland and Beaver) in the spring and after harvest 38 2.2 Rates of dry matter disappearance after a 4 hour digestion in vivo of fresh cuttings, masticate and genetic precursors of the cultivar AC Grazeland, and maturity relative to the standard alfalfa, Beaver 43 2.3 Parameter estimates of in vivo dry matter disappearance from fresh cuttings and masticate of the LIRD cultivar AC Grazeland and Beaver alfalfa 49 2.4 Parameter estimates of in vivo dry matter disappearance of AC Grazeland, its precursors (LIRD-1, LIRD-2, LIRD-3), and the standard alfalfa, Beaver 50 3.1 Rates of gas production during in vitro digestion of the LIRD cultivar, AC Grazeland and its parental lines, cv. Anchor, Beaver, Kane and Vernal 70 3.2 Parameter estimates of gas production from 48 hour in vitro incubations of the LIRD cultivar AC Grazeland and its parental lines, cv. Anchor, Beaver, Kane and Vernal 71 3.3 Rates of gas production during in vitro digestion of the alfalfa cultivars, LIRD-1 and AC Grazeland (LIRD-4) compared to the standard alfalfa, Beaver 75 3.4 Parameter estimates of gas production from 24 hour in vitro incubations of the LIRD cultivars AC Grazeland, LIRD-1, and the standard alfalfa, Beaver 77 3.5 Parameter estimates of in vitro gas production from fresh leaves and stems of the LIRD cultivar AC Grazeland and Beaver alfalfa, and the effect of maturity and treatment 78 3.6 Cumulative gas production from leaflets in 3 stages of development during in vitro digestion of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 87 3.7 Cumulative gas production from stems and petioles in 3 stages of development during in vitro digestion of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 90 4.1 Feed fractions of LIRD-3 and Beaver alfalfa, during two feeding periods coinciding with the seasonal maturity during a first (spring) and second (regrowth) hay cut 109 4.2 Calculated feed fractions of LIRD-3 and Beaver alfalfa, during two feeding periods coinciding with the seasonal maturity during the first and second hay cuts 109 4.3 Apparent digestibility of LIRD-3 and Beaver alfalfa feed fractions in wethers, during two feeding periods coinciding with the seasonal maturity of the first and second hay cuts I l l 4.4 Apparent digestibility of calculated LIRD-3 and Beaver alfalfa feed fractions in wethers, during two feeding periods coinciding with the seasonal maturity of the first and second hay cuts . 112 4.5 Feed fractions of AC Grazeland and Beaver alfalfa during four consecutive feeding periods beginning 26 days after a hay cut 118 4.6 Calculated feed fractions of AC Grazeland and Beaver alfalfa during four consecutive feeding periods beginning 26 days after a hay cut 120 4.7 Apparent digestibility of fresh feed fractions of AC Grazeland and Beaver alfalfa when fed to wethers during four consecutive feeding periods 122 4.8 Apparent digestibility of calculated feed fractions of AC Grazeland and Beaver alfalfa when fed to wethers during four consecutive feeding periods 123 4.9 Observed and expected (Poisson) frequency distributions of bloat in sheep fed two alfalfa cultivars (cv. AC Grazeland and Beaver) 125 5.1 Distension and rumen condition indexes used to estimate bloat incidence and severity in studies of alfalfa grazing 134 5.2 Bloat incidents in cattle grazing two alfalfa cultivars (cv. AC Grazeland and Beaver) 136 5.3 Observed and expected (Poisson) frequency distributions of bloat in cattle grazing two alfalfa cultivars (cv. AC Grazeland and Beaver) at Lethbridge, AB 137 5.4 Effect of alfalfa maturity on bloat (total incidents / days of grazing) in cattle grazing two alfalfa cultivars (cv. AC Grazeland and Beaver) at Lethbridge, AB 140 viii List of Figures 2.1 Change in maturity (Mean Stage number by stem Count, MSC) of two alfalfa cultivars (cv. AC Grazeland, Beaver) grown in a controlled environment over a 35 day period 39 2.2 Relationship between MSC (Mean Stage number by stem Count) and MSW (Mean Stage number by stem Weight) of two alfalfa cultivars (cv. AC Grazeland, Beaver) grown under irrigation . 40 2.3 Total dry matter disappearance after 4 hour in vivo incubations, of fresh cuttings of two alfalfa cultivars (cv. AC Grazeland and Beaver) relative to maturity of the cultivar 45 2.4 Dry matter disappearance (%) of fresh cuttings of alfalfa (cv. AC Grazeland, Beaver) after 4 hour in vivo incubations relative to the maturity of the standard alfalfa cultivar, Beaver 46 2.5 Dry matter disappearance from fresh cuttings of two immature alfalfa cultivars (early bud, Trial 5; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance . . . 51 2.6 Dry matter disappearance from fresh cuttings of two mature alfalfa cultivars (early blossom, Trial 4; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance . 52 2.7 Dry matter disappearance from freshly masticated, immature alfalfa cultivars (early bud, Trial 1& 2; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance . 53 2.8 Dry matter disappearance from freshly masticated, mature alfalfa (early blossom, Trial 3&4; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance 54 3.1 Rates of gas produced from LIRD parental cultivars during the first 14 h of in vitro incubation compared to estimates predicted by an exponential model of gas production (based on a 48 h fermentation) 73 3.2 Rates of gas produced from LIRD parental cultivars during the first 14 h of in vitro incubation compared to estimates predicted by an exponential model of gas production (based on a 24 h fermentation) 74 3.3 Rates of gas produced from LIRD synthetic cultivars during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 76 3.4 Gas produced from whole, perforated and ground leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 80 3.5 Rates of gas production from whole, perforated and ground leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production 81 3.6 Gas produced from cut and crushed leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 83 ix 3.7 Rates of gas produced from cut and crushed leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production 84 3.8 Gas produced from leaves at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 88 3.9 Rates of gas produced from leaves at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production 89 3.10 Gas produced from stems and petioles at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 92 3.11 Rates of gas produced from stems and petioles at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production 93 4.1 High and low mean maturities (Mean Stage Count, MSC) in regrowth of two alfalfa cultivars (cv. AC Grazeland, Beaver) during four consecutive digestibility trial feeding periods 116 L i s t o f A b b r e v i a t i o n s A c r o n y m D e f i n i t i o n 4hNbDMD Cumulative dry matter disappearance from nylon bags after a 4 h in vivo incubation A Asymptotic total extent of DMD or GP, a parameter in a kinetic model ACG AC Grazeland B r , (LIRD-4), the LIRD alfalfa cultivar from the 4 t h cycle of selection ADF Acid detergent fiber ADL Acid detergent lignin BVR Beaver, a standard alfalfa cultivar and a parent of AC Grazeland B r C m Maximum specific rate of DMD h"1 or GP h 1 , a parameter in a kinetic model C s Average rate of DMD h"1 during the incubation interval 0.25 h to 4 h. D M Dry matter DMD Dry matter disappearance D M D R S % ' Dry matter disappearance as a proportion of the DMD of the standard cultivar GP Gas production IRD Initial rate of digestion LIRD Low initial rate of digestion LIRD-1, 2, or 3 A LIRD alfalfa cultivar from the 1st, 2 n d or 3 rd cycle of selection MSC Mean stage number by stem count, a numerical index of maturity for alfalfa MSW Mean stage number by (dry) stem weight, a numerical index of maturity for alfalfa NDF Neutral detergent fiber OM Organic matter REPS Replicate plant samples SCC Soluble cellular components T 0 Lag time, h, a parameter in a kinetic model xi Acknowledgements Curiosity never killed the cat, it only taught her a lesson about life, that learning is a difficult, lifelong adventure. The lesson for others is that life can be short. Knowledge is hard won, made easier only by the passing on of truths and techniques for the curious by our teachers and mentors. For this pursuit in life we are forever in their debt. My greatest debt is to Dr. R. M . Tait, my graduate supervisor, colleague and friend, who opened the door for my candidacy and extended his scholarship, guidance, understanding and forbearance to me throughout my graduate program. Opportunities sometimes happen by chance but sometimes they are created. Dr. K.-J. Cheng is one who creates opportunity, which is a great talent. I and many of my colleagues have benefited from his endeavours and we are all in awe of his capabilities. For his support and counsel I am most grateful. At various times and under a wide range of circumstances I needed the support of the other members of my graduate committee, Drs. J. A. Shelford, M . Pitt, W. Majak and Dr. J. R. Thompson. Their suggestions were greatly appreciated. I am also grateful for the indispensable advice of others including Drs. S. N . Acharya, Z. Mir, E. Okine, and L. Goonwardene and especially Dr. T. A. McAllister. Technical expertise is one of the two ingredients necessary for good science. The other is money. I owe many donuts to the technical staff in the Ruminant Microbiology and Feed Biotechnology Program, Livestock Sciences Section, at Agriculture and Agri-Food Canada, Lethbridge. When asked they were willing. Financial and logistical support was provided by Alberta Agriculture, Food and Rural Development and grants from the Alberta Agricultural Research Institute and the Alberta Cattle Commission. I have a picture in my office, taken around 1955, of an agrologist, a man in his mid 30's, squinting into the sun, holding a sheaf of fertilized and unfertilized smooth bromegrass in either hand. The man is my father, Eldor. R. Berg, whose pride in his family, his education and his profession set an example to all and nurtured our talents. This work is dedicated to him. Preface Early in 1991, the District Agriculturist at Ponoka, A B , and I sat down with a local cattleman, to work out the details of project proposal. We wanted to demonstrate the productivity, cost and profitability of growing alfalfa for pasture instead of grass. The project was our response to a common question asked by ranchers in the spring, "What can I spray to eliminate the alfalfa in my pasture?" Most of the project costs were for applying herbicides and fertiliser to an old field of alfalfa. The proposal was rejected because, in the view of one critic, we should not be demonstrating cost effective ways of eliminating alfalfa. Unfortunately, we had no other way of demonstrating the value of keeping it. Alfalfa grazing is not a new idea. Many ranchers made that abundantly clear to me at their kitchen tables and out in their pastures. They willingly shared their expertise and opinions on the merits of grazing alfalfa, all the more valuable because they had incorporated their ideas into their management. Several have since become outspoken advocates, professing the profitability and efficiency of grazing alfalfa to all who will invite them to speak, evangelists in a sceptical ranching community. The work we accomplished over the next three years was quite minor, amounting to a few demonstrations that converted few leading edge graziers to alfalfa. Truthfully, some had already adopted an alfalfa grazing system and only needed me to give their neighbours an unbiased confirmation that it was profitable and did not ki l l all their cows. So, the invitation to work on the problem of alfalfa and bloat in a network of highly skilled researchers was as unexpected as it was exciting and flattering. M y greatest regrets are that I had to put family, fishing and the mandolin aside for a few years. A pessimistic economist once said to me, "There are no large, hidden, net present values in agricultural production anymore. The potential to lower costs or increase profits is minimal, amounting to a few percentage points on the positive side of zero". Perhaps, but there is an exception to every rule, including economic ones. Building a better alfalfa takes us one step closer towards harvesting its full potential, a harvest that has been beyond the grasp of generations of graziers. Bjorn P. Berg, PAg xii i Chapter 1 A Review of Alfalfa Grazing and Bloat Theory 1.1 INTRODUCTION Alfalfa is an outstanding crop. It produces exceptional yields of high quality forage under intensive management and it tolerates a wide range of environmental conditions. These attributes make it an ideal choice for feeding ruminant livestock. Unfortunately it deviates from this ideal because its consumption and digestion sometimes interferes with normal rumen function, causing animal distress and even death. The problem is exacerbated under grazing because the hazard is unpredictable and has defied analysis. Ruminants are nutritionally dependent on the digestive efficiency of a continuous-flow, anaerobic, steady-state, fermentation ecosystem in their rumen. Changes in the animal's diet or health may disturb the rumen's functions but it normally adjusts to maintain a stable environment. The system can be destabilized if the mechanisms that compensate for the disturbance are disabled or overwhelmed, as for example when a disease limits an animal's ability to eat or regurgitate. The distress observed in an animal with a distended, bloated rumen is an obvious manifestation of a destabilized rumen ecosystem. Bloat is a serious problem in cattle and sheep (Howarth et al. 1991). Research on bloat is motivated by the need to understand its etiology and test the efficacy of novel ideas for its control. These goals are not mutually exclusive because proving hypotheses about the condition may lead to solutions that prevent it or reduce its incidence and potential for economic loss. For the rancher though, research aimed at preventing bloat is the priority. Effective treatments are needed to avoid economic loss, whatever caused the problem. This review is concerned with the prevention of bloat because its control is an important hurdle to the adoption of alfalfa grazing. The extent to which bloat limits alfalfa's use in western Canadian pastures is summarized by estimating alfalfa's grazing production potential if bloat were not limiting. The diagnosis and treatment of pasture bloat are described in old theories about the cause of bloat, 1 specifically the foam hypothesis (Reid 1960). More recent theories, such as the genetic predisposition hypothesis (Cockrem et al. 1983) and the theory of cell rupture (Howarth et al. 1978a), are considered since they have led to entirely different strategies of prevention, selecting bloat resistant strains of livestock or alfalfa. The review culminates in a brief analysis of the latter strategy, the breeding of alfalfa to reduce bloat incidence, and outlines a series of tests designed to evaluate a new strain selected for this purpose. 1.2 G R A Z I N G A L F A L F A : A N O P P O R T U N I T Y C u r r e n t S t a t u s Alfalfa is the most productive, perennial forage in Western Canada (Walton 1983) as it is in most other regions of North America (Michaud et al. 1988). This does not mean that it is the exclusive forage in the field, or even that it is a prevalent choice for grazing. On the contrary, if graziers are willing to seed pastures with alfalfa at all, it is at low rates, in mixtures designed to limit alfalfa plant densities to less than 20% of the total in an established stand. Furthermore, stand management is not aimed at obtaining a maximum benefit from the alfalfa. The most common grazing strategy for alfalfa is the 'hay-and-graze' management system. Mixed stands of grass and alfalfa are usually cut for hay and stock graze the regrowth, or aftermath, in the fall. As the stands age, winter-kill and over-harvest take their toll, leaving too few alfalfa plants and too many weeds to justify haying. Fields that are still marginally productive as pasture are rendered bloat-safe by removing the remaining alfalfa with a herbicide. Otherwise, these depreciated stands are renovated by cultivating and seeding to an annual cereal. Thus the most common grazing strategy is a strategy of using alfalfa, but not very much. Minimal alfalfa grazing is a creditable management strategy aimed at reducing the incidence of bloat. Pasture bloat can be devastating; it occurs unpredictably, preventive measures are few and not always effective, and severely affected animals may die in less than one hour without immediate surgical 2 intervention (Howarth et al. 1991). Losing animals for any reason can be traumatic; a grazier will never consider the loss as just an operational hazard. Consequently, if the best strategy appears to be avoiding any situation in which bloat might occur, then the simplest strategy is to avoid grazing alfalfa altogether. Unfortunately, bloat occurs anyway. Annual mortality due to bloat is as high as 1.5% of all grazing cattle and sheep. Each year rumours circulate about calamities, where bloat kills more than 5% of an individual herd. These statistics vary little between surveys on several continents, and have not changed in spite of more than 30 years of alfalfa and bloat research (Wolf and Lazenby 1972; Howarth 1975; Howarth et al. 1984). Goplen (1989) valued the annual loss in North America at $125 million. Table 1.1 The potential economic benefit of grazing alfalfa to calf production in Alberta Calf Ranch Alberta 1997 Statistics2 Weaned weight (kg)2 210 Price ($/kg)2 $2.42 Grazing season (days) 120 Breeding herd (cows) 1 250 1,952,000 Weaned calf crop2 88% 220 1,717,760 Gross weight weaned (kg) 185 46,200 360,729,600 Gross return $447.22 $111,804 $872,965,632 Potential Returns Grazing Alfalfa Additional ADG (kg)2 0.15 Seasonal weight gain/calf (kg) 18 Gross weight gain/calf 16 3,960 30,919,680 Gross margin $38.33 $9,583 $74,825,626 Marginal increase in returns 9% 9% 9% Weaned calf equivalent 0.08 19 147,237 2 Burns and Standaert 1985; Dunford and Jewison 1997 Potential for Improvement Operating at low risk, that is not grazing alfalfa to reduce the risk of bloat, means foregoing marginally better returns. Most economic models of grazing systems give a decided advantage to the use of more legumes (Russell et al. 1981). In a review of North American grazing studies, Burns and Standaert (1985) estimated that legumes in pasture increased the daily gains of calves by .15 kg. Over a typical Western Canadian grazing season (120 days) a calf could gain an extra 18 kg. An individual producer, with a breeding herd of 250 cows would gain an additional 9% in annual returns (Table 1.1). This is the production equivalent of weaning 19 more calves. The performance and behaviour of grazing cattle are significantly affected by alfalfa in the pasture. Walton et al. (1981) reported that cattle on pasture with a composition of 12% alfalfa (by dry weight) grazed longer (2.4 h d"') but gained less (0.14 kg d"') than animals on a stand containing 42% alfalfa. Dry matter digestibility on the light alfalfa stand averaged 8.3% lower over the grazing season. While Walton et al. (1981) attributed these differences to rotational grazing management subsequent studies by Popp et al. (1997a) showed that the management system had little impact in any specific year. Thus, the main reason for the improved gains was the amount of alfalfa in the field. When bloat is controlled or eliminated stocking rates and rates of gain from alfalfa can be phenomenal. On dryland pastures containing 50% to 91% alfalfa, Popp et al. (1997b) reported carrying capacities ranging from 103 to 357 steer d ha'1 and rates of gain from .68 to 1.49 kg d"1 over a four year trial period. The pastures were never stocked to obtain maximum gains but the total liveweight gains were quite respectable, ranging from 107 to 462 kg ha"1. Acord (1969) stocked irrigated alfalfa pastures to extremes and obtained carrying capacities of 1488 to 2975 steer d ha"1 and rates of gain ranging from 0.56 to 1.1 kg d"1 over two years. Total liveweight gains were 1506 to 1944 kg ha"1. Clearly, the strategy of limiting alfalfa in pastures carries a significant opportunity cost. The prairie in western Canada has been stocked over its carrying capacity for much of the 20th Century. The current demand for pasture is greater than the supply (MacAlpine et al. 1997). Yet in most regions, alfalfa's high production potential has not been exploited except as a preserved feed; it does not contribute to the supply of pasture to any great extent. In Alberta, hay aftermath provides only 12% of the province's supply of pasture. Alone, alfalfa contributes less than 5% to the province's total carrying 4 capacity (MacAlpine et al. 1997). The wide economic gulf between grazing alfalfa and any other alternative strategy is an incentive to change and an opportunity for technological innovation. Also grazing alfalfa may be the only strategy with the potential to meet our growing demand for pasture. 1.3 B L O A T M A N A G E M E N T Diagnosis A rancher's normal response after a bloat incident is to examine the animals and the conditions they were in, including the plants they were eating, to see if something can be learned that will help predict and prevent another occurrence. If they can find a common factor responsible for bloat perhaps they can diagnose the problem early, treat it before there is a death, or at least make its occurrence more predictable. Our current knowledge of bloat management has been directed by many of these discoveries. CAUSES OF BLOAT. Bloat ensues as a chronic manifestation of disease, a dysfunction of the upper digestive tract, or from the consumption of a bloat-provoking feed (Cole et al. 1945; Johns 1954; Cole et al. 1960; Howarth et al. 1978a; Garry 1990a). The rumen becomes tympanitic when the rate of gaseous discharge is less than the rate of gas produced from fermentation. Bloat is symptomatic of many conditions that interfere with normal eructation and rumen motility including hypocalcaemia, vagal nerve damage, abomasal displacement, thoracic inflammation, ruminal stasis and obstructions of the cardia or reticulo-omasal orifices. Bloat is also a symptom of diseases like pneumonia, tetanus, and reticulo-peritonitis. Plant species known to cause bloat in grazing cattle include legumes such as alfalfa (Medicago spp.), red, alsike, and subterranean clovers (Trifolium spp.) and sweet clover (Melilotus spp.); and vegetative grass forages such as winter wheat, triticale, and the rye grasses (Triticum spp., Triticosecale spp., Secale spp., Lolium spp.). In confined feeding systems, cattle bloat when their diets contain processed cereal grains; preserved feeds such as alfalfa or clover hays, pellets and even corn 5 silage (often associated with increased digestibility from harvesting conditions or subsequent processing); or poorly processed tubers and fruits (animals choke on potatoes, turnips, apples and kiwifruit) (Cole et al. 1945; Ayre-Smith 1971; Howarth 1975; Waghorn, G. 1997 pers. comm.). TYPES OF BLOAT. A distinction is made between frothy bloat and free-gas bloat: amorphous, non-layered rumen contents and the presence of a stable foam are associated with frothy bloat; defined, normal layering of the rumen contents and the absence of a stable foam are associated with free-gas bloat (Cole and Boda 1960; Howarth 1975; Garry 1990b). Both types of bloat can occur simultaneously (Boda et al. 1956). However they could arise from different pathological conditions that require different prophylaxes. For example, an animal that has.been grazing legume pasture and has contracted pneumonia may bloat because the infection affects the animal's ability to eructate. Retention of rumen gas may lead to a free-gas bloat while the digestion of the alfalfa forage may create a non-pathological froth. Other distinctions likely reflect differences in the feed or the by-products of digestion rather than the etiology of the condition. Frothy feedlot and pasture bloat differ in some rumen parameters, viscosity is greater and pH is often lower in feedlot bloats. While these measures may create useful subclassifications of bloat for diagnosis and treatment, they more generally suggest that a range of feeds and rumen conditions can generate stable foams (Clarke and Reid 1974; Cheng et al. 1976) rather than a type of bloat. Similarly, the difference between sub-acute and acute bloat is also one of magnitude. Sub-acute bloat - A state of sub-acute bloat exists when the animal has difficulty discharging gas from the rumen. The condition is asymptomatic, so the animal shows few signs of distress, but it can stimulate behavioural and physiological adaptations in the animal. Eructation and feeding behaviours are modified as the rumen's static pressure increases, to adjust to the new gas dynamics (Cole et al. 1945). Grazing bouts are shorter, rumination times are reduced and ruminal movements increase in frequency (Hancock 1954). Production losses are primarily a result of reduced feed intake (Johns 1954; Reid and Johns 1957; Alder et al. 1967; Hall et al. 1988). In cases of sub-acute, frothy bloat on legume pasture, the rumen has normal motility and low to moderate pressure but may be fully charged with a stable, amorphous foam, containing elevated chlorophyll levels, cation imbalances and an increased capacity to produce gas (Cole and Boda 1960; Reid 1960; Howarth et al. 1977, 1978b; Majak et al. 1980, 1985, 1986a, 1986b; Ledgard et al. 1990; Majak and Hall 1990). The danger for animals with sub-acute, frothy bloat is that they are predisposed to the onset of acute bloat (Majak et al. 1983; Hall et al. 1988). Acute bloat - The development of an acute bloat can be rapid or protracted, with a sub-acute state remaining stable for extended periods (Lindahl et al. 1957). For acute bloat to occur, interactions between the animal and the feed source, the by-products of digestion or the microbial environment must escalate to a breakpoint beyond which fermentation gases begin to accumulate at a rate faster than the existing compensating mechanisms can expel them. If this point is not reached, the bloat may remain sub-acute and even abate without incident. Acute bloat often develops in conjunction with alterations in the forage quality, fluctuations in digestive conditions, when handling stress or a disease affects the animal's physiological status, or during changes in the ambient environment (Hall et al. 1984; Garry 1990b; Waghorn 1991; Hall and Majak 1991, 1995; MacAdam et al. 1995). The additional gas held in the rumen during an acute bloat generates high pressure, leading to severe distension and distress. Therapeutics for the treatment of acute bloat are limited by time, especially if emergency medical intervention is required to prevent asphyxiation or internal haemorrhage and the death of the animal (Garry 1990a, 1990b). Proper diagnosis is a lesser concern when the difference between death and life is a matter of a few minutes. Consequently many experienced ranchers and veterinary practitioners use several remedies (chasing, tubing, drenching with oils, detergents, or pluronics, trocarization, and rumenotomy), chosen sequentially or at random, no matter the cause of the bloat, to relieve the distension. Ranchers need ways to control digestion or detect sub-acute bloats before acute bloats develop and generate serious economic losses (Clarke and Reid 1974; Howarth 1975). Distension is the first clinical symptom used to detect bloat but it is generally insufficient to ascertain the severity of bloat or to verify the onset of acute bloat (Lindahl et al. 1957; Garry 1990b). Other visual symptoms of distress that show severity include panting, frequent urination, stamping the hind feet, kicking at the belly, or an abnormal stance, usually with forequarters and head elevated (Boda et al. 1956; Garry 1990b). However, animals vary in their physical ability to adapt to the pressure and in their individual response to discomfort. A change in girth is not linear with respect to changing ruminal pressure (Waghorn 1991). As the rumen expands, it fills the abdominal cavity, stretching the muscles and exerting pressure on the internal organs. Discomfort will be more severe in animals that have small body cavities, larger internal organs, or layers of non-elastic fat, connective tissue and muscle. Thus the only objective measure of severity is intra-ruminal pressure (Waghorn 1991) and for intact animals the recommended procedure is palpation of the left flank (Lippke et al. 1972). In ruminally cannulated animals the material expelled when the cannula is opened yields ample evidence of bloat pressure and severity. Frothy and free-gas bloat - Visually, the distension resulting from frothy bloat is indistinguishable from free-gas bloat. In the case of frothy bloat on pasture the clinical symptom is the presence of a stable foam that sequesters the gas products of fermentation and retains them in the rumen (Reid 1960; Moate et al. 1997). Free-gas bloat may have a different etiology but on legume pasture, bloats may be a combination of free-gas and froth. Again, for intact animals, the only reliable, external, diagnostic procedure is palpation of the left flank to establish whether the rumen contents are abnormally uniform, due to the presence of foam, or stratified normally as in free-gas bloat (Garry 1990b). A second, more invasive protocol, gastric intubation, can be used to expel gas and some rumen contents to confirm the diagnosis. Free-gas bloat may not be as prevalent in ruminally cannulated animals because gas can be expelled through the fistula. Thus, the severity and the incidence of bloat or the degree of distension may be underestimated by cannulated animals if pasture bloats are normally a mix of free gas and froth. 8 However, the cannula provides a ready means of distinguishing between froth and a rumen distended with forage from a recent meal. Theory and Practice of Bloat Prevention Generally, correct diagnoses of the causes of bloat are made with the expectation that a reliable course of action can be taken to control or eliminate it. This is certainly true of bloat caused by a medical condition (Garry 1990a), but perhaps less so for frothy bloat (Howarth 1975). Popular recipes and rules of thumb on how to suppress frothy bloat on pasture or manage bloated animals, are quite variable although there are striking similarities in publications over a 250-year period (Beddows 1952; Anderson 1997). Turning animals out to pasture late in the day or onto mixtures containing grasses and trefoil were important management rules in 1716, while drenching with some special concoction is still a common remedy even today. The ingredients for the drench have changed considerably from 'terpentine in beer', but soaps, oils and other organic solvents are still recommended. Al l methods fail occasionally, which suggests that there are problems recognizing predisposing conditions that result in acute bloat, selecting an appropriate therapy or determining what caused a bloat (when treatment is given without diagnosis). On the other hand, the consistency of the prophylaxes proves that some treatments have a high degree of efficacy. Numerous theories have been proposed to explain bloat phenomena or the mode of action of preventive measures. Many are obsolete, but a few have maintained their currency for more than fifty years. None could be called a unified theory, which speaks to the obstinacy of the problem (Cole et al. 1945; Waghorn, G. pers. com. 1996). T H E EXCESSIVE CONSUMPTION HYPOTHESIS. Possibly the oldest hypothesis is the theory that excessive consumption causes bloat. It is supported by observations that fasted or hungry animals have high consumption rates during the first few hours after being turned out to pasture. Dougherty et al. (1987) reported that alfalfa consumption rates for the first hour of grazing by fasted beef heifers were 2.4 and 3.0 kg D M h"1 for low and high herbage allowances, respectively. This level of intake may produce gas at a rate exceeding 2 1 min"1 or more than 120 1 h"1. If eructation is suppressed, less than 60 1 of gas may produce a serious bloat (Waghorn 1991). In contrast, the mean intake of beef steers continuously grazing alfalfa was much lower, 1.1 kg OM h"1 (Popp et al. 1997b). Hence the old recommendation to keep animals well fed and move them cautiously, without interrupting their feeding regimen, may be well founded. Rates of consumption or gas production do not have to be extreme for gas to accumulate; the animal just needs to have a problem expelling the amount of gas produced. The threshold level of gas necessary for bloat varies widely between animals. Waghorn (1991) reported that some animals needed to accumulate only 15 1 of gas to obtain an intra-ruminal pressure of 10 cm water (Grade 1 level of bloat) while others required 50 1. Moate et al. (1997) found that non-bloated animals had rumen head-space gas volumes of 4 1 while the volume in bloated animals exceeded 20 1. Lacking the evidence of foam or a high rate of gas production, Moate et al. (1997) were forced to conclude that a major cause of bloat is simply a failure to eructate. However, the reason for the failure may be far from simple (Garry 1990c) because it could include behavioural problems (eg. stress related suppression of rumen motility), biochemical disorders (eg. an impaired satiety function) or pathological conditions (eg. peritonitis). Excessive consumption with an attendant high rate of gas production has not been confirmed as a cause of bloat but, contrary to the opinion of some reviewers (Clark and Reid 1974; Howarth 1975), neither has it been ruled out. Only two studies compared consumption patterns relative to bloat incidence (Hancock 1954; Johns 1954). Both studies used healthy, well-fed animals that were familiar with the feeding regimen and found no bloat associated with changing amount or timing of feedings. Consumption and bloat incidence in stressed, hungry or unhealthy animals has not been examined. THEORY OF ANIMAL SUSCEPTIBILITY. Another old theory holds that bloat is a function of the animal's inherent level of susceptibility (Cole et al. 1945; Johns 1954). Cattle are more disposed to bloat than 10 sheep and the susceptibility of individual animals varies widely (Ayre-Smith 1971; Clarke et al. 1974; Colvin and Backus 1988). Younger animals are more susceptible than older animals (Howarth 1975) suggesting that, with experience, individual animals can cope with bloat-provoking conditions. Learning grazing skills early in life from experienced mothers may have an impact on the off-spring's subsequent bloat susceptibility (Ramos and Tennessen 1992). A corollary, the genetic predisposition hypothesis, has been examined in detail only in cattle. Reports of bloat in ruminants other than sheep or cattle are rare and usually cited in conjunction with a veterinary procedure or a confined feeding system (Clarke and Reid 1974). Bloat incidents in the commercial deer (Cervus spp.) and bison (Bison bison) industries in western Canada have not been documented although anecdotal accounts suggest it is extremely low for deer but may be a growing concern in commercial bison feedlots. Natural selection should eliminate the alleles of bloat-prone animals from herds that regularly encounter bloat-provoking conditions. However, divergent selection for high (HS) and low (LS) susceptibility to bloat in cattle found no specific traits other than that HS animals maintained comparatively high volumes of fluid in the rumen (Cockrem et al. 1987a, 1987b; Carruthers et al. 1988). The condition is heritable and a function of physiological or behavioural attributes, but identifying specific characters has been slow and unrewarding (Cockrem et al. 1983; Howarth et al. 1984). On the one hand high volumes of rumen fluid may restrict the head-space available for gas expansion implying that HS animals will reach stressful intra-ruminal gas pressures earlier than LS. On the other hand, if HS cattle retain more digesta in the rumen for longer periods then perhaps they pass digesta through the rumen at a rate slower than LS non-bloaters (Okine et al. 1989). Breeding a new class of livestock seems prohibitively expensive and inappropriate especially if bloat is an endemic trait. Comparisons between non-bloating and bloating ruminants to learn if specific traits are associated with bloat have not been made and are unlikely, considering the expense needed to identify ruminant species that are truly non-bloating. And if these tests were ever conducted, the preferred outcome would be to exploit the non-bloating species rather than waste the time transferring its 11 genetic capacity to sheep or cattle. Thus, management of the animals and their forage probably holds more promise than selecting non-bloating breeds of livestock (Clarke and Reid 1974; Majak et al. 1995). T H E FOAM HYPOTHESIS. The most influential theory on bloat is not so much one of cause as effect. Combining a normal to excessive rate of gas formation with rumen liquor, foaming agents and foam stabilizers has the undesirable effect of turning the rumen contents into a stable froth (Johns 1954; Reid 1960). The theory explained a great deal of the phenomena that had been observed over the previous 30 years and spawned a 40 year search for foaming agents, foam stabilizers, destabilizers, the source of gas and an examination of the role taken by specific microbes in the development of bloat foams. Before the enunciation of this theory the prevailing philosophy was that an accumulation of gas, mostly in a free-form state (Cole et al. 1945), caused bloat. After the Reid (1960) paper, the study of free-gas bloat was virtually abandoned, perhaps to the detriment of our understanding of rumen function and the etiology of bloat (Howarth 1975; Moate et al. 1997). Corollaries to the foam hypothesis include theories that protein, pectins, saponins, lipids, cations, polysaccharide slimes and cellular fragments of alfalfa stabilize the gas bubbles that are generated during digestion (Johns 1954; Mangan 1959; Pressey et al. 1963; McArthur et al. 1964; Miltmore et al. 1970; Clarke and Hungate 1971; Gutek et al. 1974; Cheng et al. 1976; Howarth et al. 1977, 1978b; Majak et al. 1980; Majak and Hall 1990; MacAdam et al. 1995; Mathison et al. 1999). Axiomatically, lipids, in a dual role, and condensed tannins destabilize foam (Pressey et al. 1963; Cooper et al. 1966; Stifel et al. 1968). Howarth (1975) suggested that the distribution of surface active substances like proteins and lipids in the rumen liquor may affect their ability to stabilize or destabilize the rumen froth. Subsequent investigations (Howarth et al. 1978b) led to the suggestion that the lipid membranes of chloroplasts, fragmented from mastication and bacterial maceration, acted as nucleation sites for bubble formation. Obviously the creation and stabilization of rumen foam are complicated and interactive processes. One major bloat management strategy has emerged as a consequence of the foam hypothesis, 12 using novel prophylactics to prevent bloat foams. The value of this strategy is that it is direct, with a history of traditionally applied remedies behind it, and it is an accepted means for ranchers to control the problem (Cole et al. 1945; Cole and Boda 1960; Howarth 1975). However, these preventatives are not remedies. Although they may affect the rate and stability of foam formation, they do not eliminate bloat. Bloat still may occur because of problems with their administration and they may not be effective in all situations. Complications arise in grazing or feeding systems when the efficacy of the substance is affected by dosage dependancy or microbial adaptation or if it suppresses digestive efficiency. The prophylactic may have a plant origin. Condensed tannins (CT's), also known as proanthocyanidins, are polymeric compounds of flavan-3-ols or flavan-3, 4-diols found in many common forage plants (Jackson et al. 1996). They are produced by the plant as a chemical defence against herbivory since their protein-precipitating ability inhibits enzyme activity and cellulose digestion, and their astringent taste affects their palatability (McMahon et al. 2000). Extractable CT's were identified as foam destabilizers in the non-bloat-causing perennial forage legumes, sainfoin (Onobrychis viciifolia) and birdsfoot trefoil (Lotus corniculatus) (Jones et al. 1973; Gutek et al. 1974). Research teams in Canada and New Zealand are currently trying to reduce bloat incidence and increase protein digestibility by using these naturally occurring substances in grazing and feeding systems (Waghorn and Jones 1989; Waghorn et al. 1990; McMahon et al. 1999, 2000; Barry and McNabb 1999). A common management practice is to treat all animals and all conditions as if they are predisposed to bloat. Animals are provided with a daily dose of a specific prophylactic that the herdsman hopes will prevent bloat. Prophylactics are given in several ways including feeding them in a customized mineral supplement, dissolving them in drinking water, spraying them on pasture, 'drenching' animals before or after feeding or grazing, or by inserting a mechanical, time-release bolus containing the agent into the rumen. Invariably, the efficacy of an agent is dosage dependant so any reduction in the number of incidents and the severity of bloat is a function of the concentration of the agent in the rumen. Concentration, in turn, is dependant on the amount and frequency of administration of the agent, its rate 13 of degradation and passage through the rumen. Thus the ability of an agent to control bloat may be strongly influenced by a stockman's herd management program, by the adaptation of the microbial populations to the agent or by a change in the rumen environment, and by extraneous confounding factors, like weather or palatability, that are beyond of the working range of the prophylactic's management protocol. Many materials have been used to control the foam in bloat (Reid and Johns 1957; Ayre-Smith 1971; Clarke and Reid 1974). The apparent effectiveness of a specific product is directly related to its clinical efficacy weighted by its comparative expense and its ease of administration. Prohibitive costs, administrative difficulties or regulatory barriers for an effective bloat preventive often result in the substitution of less expensive products of low efficacy (Hall and Majak 1992; Hall et al. 1994b). Pluronic detergents such as poloxalene have been proven effective but are too expensive and difficult to administer for general use in North America (Bartley et al. 1965; Acord et al. 1968, 1969; Dougherty et al. 1992; Popp et al. 1997b). Sodium bicarbonate and commercial laundry soaps are substituted in spite of their proven ineffectiveness (Reid and Johns 1957). Rumour supports the practice because ranchers report their subjective observations, attributing low bloat incidents to a difference between using the product and not using it, when the effect could be equally attributed to livestock genetics, behaviour or management (Cole and Boda 1960; Acord et al. 1968; Dougherty et al. 1989a, 1989b; Warner 1997). Alcohol ethoxalates in a pluronic detergent carrier (AEPD), used in New Zealand and Australia for nearly 40 years, have recently been re-examined (Stanford et al. 2000). Yet regulatory restrictions may prevent their use in North America. Similarly, monensin, an ionophore, is currently registered for use in Canada, New Zealand and Australia, but not the United States (Bergen and Bates 1984). The situation with ionophores is not likely to improve because gaining approvals to use antibiotics to enhance feed digestibility or digestive characteristics will be increasingly problematic in the future. Critically, in the context of the foam hypothesis, bloat cannot be eliminated. The causes of frothy bloat are so ubiquitous that therapies have had to focus on treating the symptoms, froth and distension, 14 rather than the cause. Besides, any strategy that promises to eliminate bloat is antithetic because bloat originates from an outwardly normal digestive process; eliminating bloat will inhibit digestion. Regardless of its immediate outcome, eliminating froth will not eliminate bloat; free-gas bloats will still occur. Exclusive pursuit of the foam hypothesis has left us with strategies targeted at reducing the risk of frothy bloat as opposed to managing bloat in all conditions. THE CELL RUPTURE HYPOTHESIS. A relatively new theory considers the development of bloat to be a consequence of the readiness of cells to rupture. The cell rupture theory of bloat proposed by Howarth et al. (1978a; 1982) was actually an advanced theory of forage digestion in the ruminant forestomach. Underlying the theory was the supposition that the initial rate of digestion was limited by the surface area and thickness of the plant cell wall. This was based on evidence that microbial digestion of intact leaf tissues proceeded in steps, each of which could be rate limiting. Bacterial colonization of the leaf surface (around stomata or lesions in the leaf) was followed by their subsequent penetration of the epidermal layer. Proliferation of microbial cells within the inter-cellular spaces macerated the tissues and allowed other bacteria to adhere to the cell walls. Eventual disruption of the cell wall resulted in an invasion of the intra-cellular space and the development of colonies of microbes within the cell wall fragments (Cheng et al. 1980; Howarth et al. 1984). Howarth et al. (1978a) derived their theory on bloat from the differences observed between the cell walls of different plant species, while Goplen et al. (1993) showed that thicker cell walls could affect the rate of digestion within a single species. Clearly the physiognomy of the cell wall had as much potential to influence the rate of digestion as any other rate limiting factor. Had they considered it they probably would have restated the cell rupture theory of bloat as a general theory of forage digestibility. A cell rupture theory of digestibility introduces the idea that the rate of digestion is a function of cell size, shape, volume, surface area, arrangement and wall structure. The conceptual framework includes the supposition that initially, microbial degradation of plant tissues is regulated at the cellular 15 level by a series of barriers. Once the barriers are removed or breached, microbial activity is regulated by competitive interactions and the demand for available nutrients or the accumulation of by-products of digestion and waste. Cells in one tissue type can also be barriers that restrict microbial access to cells in other tissues (Wilson and Mertens 1995). Cell surface area, volume and the potential to rupture have been considered in a few models of digestibility (Fisher et al. 1989; Wilson and Mertens 1995). Ruminants evolved a system to capture some of the energy in the cell wall but this was likely a secondary effect that developed from a need to liberate the nutrients contained within the cell (Van Soest 1994). In the rumen, slower rates of cell lysis would reduce the digestibility of specific plant tissue types forcing the animal to eat less of that plant or develop a capability to break the cell wall. The cell rupture theory also implies that equivalent rates of cell lysis will generate similar limits to digestibility in the ruminant and equivalent bloat potentials. As the plant tissues become macerated and the cell walls are broken apart in the rumen, many forages create a froth of cell constituents. Rumen foam is normal, a result of the mixing of gas and fluid during digestion. If the rate of cell lysis is limited then the volume, stability and rate of foam formation will also be constrained. For example on the plant side, some grass species (eg. perennial ryegrass, Lolium perenne) have leaf cell walls that are thinner than alfalfa but their respective cell sizes, shapes and arrangements constitute a greater barrier to bacterial adherence and invasion (Wilson 1993; Moghaddam and Wilman 1998). The bloating potential of these forages would be determined by the animals' susceptibility, not the plants' digestibility. As plant cells mature their walls develop into a complex external matrix to protect the cell contents and support the structural integrity of the plant. Of all the factors affecting cell wall digestibility and the potential for rupture, the most important is maturity (Buxton and Casler 1993). In forages, the rate of maturity is a critical comparative index that is easily taken for granted. Comparing the extent and rates of dry matter disappearance, gas production, digestibility and nutrient contents will not be meaningful if all the plant material was collected on the same day. Such comparisons assume that all 16 plants and all plant parts develop at the same rate and are equivalent in maturity on the day of collection. No two species of plants are that alike. Yet these types of comparisons are widely accepted as valid indications that one species is nutritionally superior to another (Hoffman et al. 1993). Selecting a character in a bloat-provoking plant that affects the etiology of bloat in a herbivore is as esoteric a strategy as that of selecting non-susceptible strains of livestock, differing only in that breeding plants is less expensive. Picking a character for selection is just as difficult because its expression may not be universal, it may vary with the physiological status of the plant or the environment and, though the trait may be associated with bloat, it may not be directly responsible for the development of bloat in a grazing animal. Canadian researchers have looked at protein fractions, chloroplast membranes, leaf tissue disruption and leaf and stem digestibility (Miltmore et al. 1970; Howarth et al. 1977, 1978a, 1978b, 1979, 1982; Majak et al. 1995). They compared some of these traits with those found in bloat-safe forages and developed a procedure for breeding alfalfa plants to reduce the risk of bloat (Howarth et al. 1979, 1982, 1991). 1.4 ALFALFA FOR GRAZING Since the turn of the 20th century, alfalfa has been bred to improve its quality, productivity and adaptation to the agronomic conditions in North America. It was primarily selected for its value in stored feed and most cultivars were selected under a mechanical harvest protocol. Alfalfa was rarely grazed in Canada, so it was never subjected to grazing pressure to induce grazing tolerance. The demand for grazing types or pasture alfalfas has risen with the demand for improved productivity and reduced cost of pastured forage. Three criteria need to be met before an alfalfa strain can be called a grazing type. First, it must tolerate the environmental conditions. In western Canada, these conditions include harsh limitations like summer drought, severe winters, short growing seasons, and marginally productive soils. Second, it must 17 tolerate grazing, including intermittent but severe defoliation and trampling. Third, and most important, it must possess a nutritional quality that enhances its suitability under grazing and reduces the incidence of bloat. The screening of genotypes for each criterion has been independent, the only area of crossover being that breeders of new strains have relied on previously screened genetic material. Precursors of a Grazing Alfalfa Canadian researchers have successfully met the first criterion, environmental tolerance, releasing several alfalfa cultivars for dryland pasture and rangeland seedings that are industry standards. Some early alfalfa cultivars are called grazing-types although they were never specifically bred for grazing (Heinrichs 1963). Recently released varieties are hardy plants with traits that include persistence, fall-dormancy, drought and winter-hardiness, low-set crowns, creeping roots, and disease and pest resistance. There were tradeoffs, yield in particular, for survival in cold regions (Lorenz et al. 1982; Berdahl et al. 1989; Caddel 1997). Dryland alfalfa cultivars have a degree of grazing tolerance because a few traits, such as low-set crowns and creeping roots, enhance their survival in pasture. Although some cultivars have been tested in grazing trials (Berdahl et al. 1986), they were never selected for grazing tolerance. A grazing-tolerant alfalfa must persist under a regimen of severe defoliation and animal impact. Grazing-tolerant cultivars have decumbent growth habits, more crown buds and greater residual leaf cover after grazing, which may help maintain higher levels of total nonstructural carbohydrates in their root systems (Smith et al. 1989; Brummer and Bouton 1991, 1992). The high feed quality of alfalfa may make it less than ideal for grazing because some factors responsible for its quality, such as digestibility and protein content, are implicated in bloat (Miltmore et al. 1970; Howarth et al. 1977). Breeding for lower quality is antithetical, alfalfa breeding programs rarely maintain lower quality lines except to evaluate traits for selecting lines of higher quality (Allinson et al. 1969; Shenk and Elliot 1970, 1971). However, species such as sainfoin or trefoil {Lotus spp.) seldom 18 cause bloat. These differences were considered for developing a bloat resistant strain of alfalfa (Howarth et al. 1978a, 1979, 1982; Fay et al. 1980, 1981; Goplen et al. 1980, 1985; Kudo et al. 1985). Thick plant cell walls are characteristic of bloat-free legumes. Rumen bacteria take more time to invade and rupture cells in these species than in alfalfa (Howarth et al. 1979; Lees, 1984), so alfalfa strains selected for low leaf tissue disruption could be bloat resistant (Howarth et al. 1982). The breeding program, undertaken by Agriculture and Agri-Food Canada, selected individual alfalfa genotypes for dry matter disappearance (DMD) when fresh clippings from the tops (15 cm) of vegetative leaders were incubated in vivo in nylon bags. Plants with a low initial rate of disappearance (LIRD), defined as a low DMD, were screened and intercrossed through four cycles of selection. The final cultivar, LIRD-4, (released as AC Grazeland) has a DMD 15% lower than a standard cultivar, Beaver (Goplen et al. 1993). Evaluating Low Initial Rates of Digestion in Alfalfa The new LIRD cultivar is unique because selection was limited to genotypes that had a reduced digestibility in fresh vegetative cuttings, during a short in vivo incubation period. Almost all previous selection for quality in alfalfa has been carried out using in vitro techniques on dried samples, with long incubations, and where quality is assumed to be directly related to a few specific plant constituents (Hill et al. 1988). The argument could be made that broadly selecting a group of undefined plant constituents under the label 'digestibility' is risky, the outcome is less certain than the more traditional technique of narrowly selecting for one or two characters that are well correlated with digestibility. However neither a broad nor a narrow technique could be considered exclusively better when the outcome is as uncertain as the cultivar's ability to influence the behaviour or performance of grazing animals. If anything, the new cultivar is arguably closer to the ideal of breeding to influence animal performance. JUSTIFICATION FOR RESEARCH. Every extension of new ideas will generate some controversy. The LIRD alfalfa breeding program was not excepted and has received its share of criticism. One of the first 19 objections concerned the use of the nylon bag digestion technique to screen alfalfa genotypes. Other techniques were available, including in vitro assays such as gas production, that could control some aspects of experimental variation better than the nylon bag in vivo assay. For example dry matter disappearance from a nylon bag may simply be related to the pore size of the bag, the larger the pore, the more material disappears in a specific interval. Another criticism concerned the effect of limiting digestion over a short time; selection might just create an alfalfa cultivar with a low apparent digestibility. A third criticism considered that reducing the capability of leaf cells to rupture in alfalfa would just produce another species of bloat-free legume; further work on improving other species, such as sainfoin, or cicer milkvetch (Astragalus cicer) or even grass would be more profitable. There were other criticisms and some of these were warranted as well. The LIRD cultivars have a potential to improve grazing livestock productivity if only because they will stimulate interest in grazing alfalfa. However they have not been used in any field situation that reflects normal ranching management practice. A host of questions are asked, from the mundane to the profound, whenever the subject is discussed. A sampling includes the following: Is a LIRD alfalfa genetically modified? Is it unique? Did some mutation result during selection? What is the nature of digestion in a LIRD cultivar? Is it nutritionally similar or inferior to other alfalfa cultivars? Can this LIRD trait, whatever it is, be found in a particular part of the plant? Is a LIRD trait confined to leaves or cells of leaves, because animals don't just eat leaves? What limits were placed inadvertently on performance during the selection of the cultivars? Yield? Disease resistance? How does all this relate to the cell rupture hypothesis? Consequently the imperative existed to evaluate the final LIRD cultivar, AC Grazeland and its progenitors in a concerted battery of tests. HYPOTHESIS AND RESEARCH OUTLINE. A thorough analysis of the cultivar's digestion profile may answer some of the criticisms and questions, solicit others and perhaps, in some small measure, advance the science of nutrition and ruminant digestive physiology. The overall hypotheses for this work are: that 20 there are no differences between the LIRD cultivar, AC Grazeland, its progenitors, and its parental standard Beaver, by standard measures of rumen digestibility (the rates and kinetics of digestion); and that a LIRD character or trait in alfalfa is not a distinct entity that makes any difference to the potential to cause or reduce the risk of bloat. The scope of this thesis is broad, making it difficult to test without refinement. So the approach will be to examine the nature of digestion of forage from the synthetic LIRD cultivars using several perspectives and methods. Chapters 2 and 3 are reports on the respective results of experiments designed to measure individual rates and overall kinetics of dry matter disappearance in vivo and gas production in vitro. Together, Chapters 2 and 3 examine the LIRD trait from the perspective of the plant, measuring the result of the application of cell rupture theory on the digestion of individual plant parts at several stages of development. In contrast, the experiments in Chapter 4 examine the LIRD trait from the perspective of the animal, measuring the effect of feeding a LIRD cultivar on the apparent digestibility of alfalfa in sheep. The effect of a LIRD cultivar on bloat incidence is viewed from two perspectives as well, between wethers fed a controlled diet and the unrestricted diets of grazing steers reported in Chapters 4 and 5, respectively. Chapter 6 constitutes a synthesis of what was learned during the study about cell rupture theory, the development of the LIRD cultivars and bloat, proposes a unification of theories, and identifies some ideas for future research. Appendix I is an essay on the use of nonlinear statistical models and Appendix II is a collection of tables that complete the data summary. 21 C h a p t e r 2 R a t e s o f D r y M a t t e r D i s a p p e a r a n c e 2.1 INTRODUCTION The cell rupture theory of legume pasture bloat advanced by Howarth et al. (1978a) proposed that bloat resistance in legumes is a function of the susceptibility of leaf tissues to disruption during digestion. The theory was based on several comparative studies that showed resistance to tissue disruption (Lees et al. 1981, 1982; Lees 1984) and less dry matter disappearance during digestion, were characteristics of the bloat-free legumes cicer milkvetch (Astragalus cicer), birdsfoot trefoil (Lotus corniculatus) and sainfoin (Onobrychis viciifolia) (Cheng et al. 1980; Fay et al. 1980, 1981; Howarth et al. 1982). They further hypothesized that a bloat-free strain of alfalfa (Medicago sativa) would require a 25 to 30% reduction in dry matter disappearance (DMD) after 6-8 hours of digestion in the rumen and used this criterion to devise a selection protocol for a breeding program (Howarth et al. 1982). Since the 8 hour reduction goal was impractical for screening large numbers of plants in a breeding program, and because DMD at 8 h was dependant on the rate of digestion in earlier periods, they targeted a 4 h reduction to achieve their goal. Thus the specific criterion was selection for a low initial rate of digestion (LIRD, IRD) in alfalfa, defined as a comparative reduction in DMD from nylon bags after a 4 hour in vivo digestion interval (4hNbDMD) (Howarth et al. 1984; Kudo et al. 1985). Nurseries were established in 1980 using the parental lines Anchor, Beaver, Kane and Vernal. The 1st two synthetic cultivars LIRD-1, LIRD-2, (called LIRD cycle 1 or Cycle 1, etc.) were selected by retaining 5% of the plants in the nursery with the lowest IRD. (The 2 n d synthetic, LIRD-2 was selected from the LIRD-1 nursery by retaining 5% of the LIRD-1 plants with the lowest IRD.) The LIRD-2 progenies were intercrossed to create the 3 rd synthetic, LIRD-3. Progeny tests identified 20 parental genotypes of the LIRD-3 line with a mean IRD reduction of 15% compared to Beaver, the standard parental check. These genotypes were intercrossed to create the 4 th synthetic, LIRD-4 (Goplen et al. 22 1993) which was subsequently released in 1998 as the cultivar AC Grazeland. Throughout the program LIRD has been called a rate. However as measured and reported it was the total extent of DMD after four hours (Kudo et al. 1985; Goplen et al. 1993; Hall et al. 1994a). The actual 4 th hour rate of DMD, which is a derivation of the total DMD at 4 h, and the effect selection for LIRD may have had on other rates, or on total DMD at 8 h, has not been quantified. This is an important omission. Part of the hypothesis proposed by Howarth et al. (1982) included the supposition that a 25% overall comparative reduction at eight hours would be necessary to achieve bloat free-status. The assumption that a comparative reduction based on the 4hNbDMD would achieve this goal was never tested (Goplen et al. 1993; Hall et al. 1994a). The selection protocols for the LIRD breeding program used only fresh plant material collected daily from individual plants. Material was harvested when the plants were in the 'pasture stage of growth' defined as the pre-bud to mid-bud stages of development (Goplen et al. 1993). Only the upper 15 cm of alfalfa plants were used in the nylon bag digestibility trials. Although an alfalfa field, plot or plant may appear to have an average level of maturity, the stems on individual plants will be in several stages of development. Additionally, problems during seeding, differences in the soil's nutrient status and variation in other environmental conditions could affect the rate at which individual plants mature. Goplen et al. (1993) suggested that the LIRD character was affected by environmental conditions and management, and hypothesized that variations in plant maturity would affect its ability to reduce bloat incidence. Data on partial 4hNbDMD obtained at several sites illustrate the confounding effects of management (cutting and regrowth periods, irrigation and dryland) and environment (soil type, climate and location) (Goplen et al. 1993; Hall et al. 1994a). In 1992, three 4hNbDMD trials were conducted at two locations, Kamloops BC and Lethbridge AB. Alfalfa plots were clipped, irrigated and allowed to regrow for 27 d and 46 d at each respective location, after the first set of trials, and 26 d and 58 d, respectively, after the second trials. The different periods between trials in the same year are a result of differences in the environment and in some management practices (time allocated for rest and regrowth) 23 between the two locations. Overall, 4hNbDMD was lower at Lethbridge (37.4%) compared to Kamloops (48.4%o), perhaps because the management (rest) periods were longer at Lethbridge and the plant material was more mature. Thus, field management, the environment and location may significantly affect the outcome of subsequent bloat trials. Studies of alfalfa leaves and stems show that their digestibility in vitro declines with advancing maturity. The decline is greater in stems than leaves at all stages of development, but even within a particular stage, digestibility declines with age (Kalu and Fick 1983). The in vitro digestibility of individual stems at pre-bud to mid-bud stages of development may be quite variable, ranging between 75% and 55% (Kalu and Fick 1983). Since the tallest stems are generally the most mature, the plant material used by Goplen et al. (1993) may have been slightly older and therefore less digestible than the majority of stems from each plant. On the other hand, the cuttings from the stem tips are the least mature parts of each stem. Thus, maturity of the stem and age of the stem's tip may have confounding effects on the selection and observation of a LIRD character. Further, field observations by technicians at Lethbridge, AB and Kamloops, BC suggested that the LIRD-3 and AC Grazeland cultivars matured at a faster rate than the comparable standard, Beaver. Differences in IRD could be a reflection of the difference in rate of maturity. Consequently a concise assessment of the maturity of the plant material was needed if the effect of a LIRD character was to be isolated from the effect of maturity. Besides plant maturity, digestibility and LIRD may be confounded with seasonal growing conditions, animal behaviour and daily variations in ruminal digestion. Seasonal changes and cycles of growth and regrowth could affect the expression of a LIRD trait. Goplen et al. (1993) observed that the variation in IRD was high between different days so they confined their comparisons to daily differences using chopped (but relatively intact) forage. Mastication by a grazing animal will damage fresh forage more than coarse chopping and affect the process of digestion. Rates of digestion for masticated alfalfa peaked earlier than rates for intact tissues (Fay et al. 1980). If the plant materials are processed in a way that does not mimic the effect of the animals' ingestive behaviour, they may show a different response 24 when an animal consumes it. Further, the LIRD trait should be well expressed under a range of conditions if it is to be reliable in a bloat control pasture management program. With the release of the cultivar AC Grazeland, the breeding program ended. AC Grazeland is the first practical application of the cell rupture hypothesis. The end of the breeding program then is just the beginning of the assessment of DMD and its relationship to bloat. The cultivar's effect on DMD still needs to be quantified in relation to the variability that may be encountered in a field situation. That is the purpose of the following analysis. A series of experiments was conducted to examine IRD differences based on rates and extents of DMD using AC Grazeland and the standard parental cultivar, Beaver. Trials were designed to: • determine the rate of maturation of AC Grazeland and compare it with that of the standard; • evaluate 4hNbDMD in relation to stage of development; • examine rates of DMD h"1 in a 4 h incubation interval, also in relation to maturity; • compare IRD between AC Grazeland, other LIRD cultivars and a standard alfalfa, Beaver, using different collection and processing methods and plant material (fresh cuttings, masticate and dried and ground samples); • analyse kinetic models of extent and rates of DMD during 48 h in vivo digestion intervals; • predict the DMD difference between AC Grazeland and Beaver after 6 to 8 hours incubation from the perspective of achieving a bloat-free status. 2.2 M A T E R I A L S A N D M E T H O D S M o r p h o l o g i c a l S t a g e o f D e v e l o p m e n t The rates of maturity of AC Grazeland, the trial cultivar, and Beaver, a parental standard, were evaluated by comparing their mean stage of development at standard growth intervals. Plants were grown in irrigated fields, to measure maturity in a normal outdoor management setting, and in a growth chamber 25 to confirm differences within a standardized environment. All trials were conducted at the Agriculture and Agri-Food Canada Research Centre near Lethbridge, AB (49°42' N; 110°47' W). MATURITY IN THE FIELD. Stands of AC Grazeland and Beaver were established on adjacent 1 ha fields in the summer of 1993. The soil on the site is a slightly alkaline clay loam(pH 7.0-8.0), a Dark Brown Chernozem, and the area receives about 400 mm of precipitation annually (Appendix III). Fields were seeded at a rate of 11 kg/ha, fertilized with 225 kg/ha of 12-51-0 (N-P-K), flood irrigated as required and harvested for hay twice during the first growing season. The field of AC Grazeland required partial re-seeding (at the same rate) during the establishment year to obtain a pure, continuous stand. Annual weeds were controlled by haying and grasses and other perennial weeds were removed by spot spraying with glyphosate. Procedures for field collection and staging of plant samples followed those of Kalu and Fick (1981). Samples of each alfalfa variety were collected at 3 and 5 wk intervals after growth started in the spring of 1994 and 3 and 8 wk after the first harvest. A set of 12 randomly located, paired 0.1 m 2 microplots were harvested from each field on each sampling date between 0600 and 0630 hrs MDT (Mountain Daylight Time). All alfalfa stems in each microplot were cut to a 3 cm stubble, weeds, litter and old stems were discarded and the sample was bagged separately and stored in a cooler at 4°C. Samples were removed from the cooler one at a time, for processing. Each stem was examined and categorized as to its morphological stage of development. Definitions of each morphological stage followed those of Fick and Mueller (1989): Stage 0 - early vegetative stem, length< 15 cm, no reproductive organs; Stage J - mid-vegetative stem, length 16-30 cm, no reproductive organs; Stage 2 - late vegetative stem, lengths 31 cm, no reproductive organs; Stage 3 - early bud stem, 1 -2 nodes with visible buds, no other reproductive organs; Stage 4 - late bud stem, ^3 nodes with visible buds; 26 Stage 5 - early flower stem, 1 node with 1 open flower, no pods; Stage 6 - late flower stem, 2:2 nodes with open flowers, no pods; Stage 7 - early seed pod stem, 1-3 nodes with green pods. Going beyond seven stages was not necessary because the fields were harvested before the plants had developed pods. Al l stems in each stage were counted for each sample. The stems in each stage of development from one of the paired microplots were then weighed, bagged, dried at 55 °C for 48 h, and re-weighed to determine dry matter. The stems of the other paired microplots were returned to the cooler and processed within 2 h for the nylon bag digestion trials. MATURITY IN A CONTROLLED ENVIRONMENT. Uninoculated seeds from each cultivar were started in sterile growth medium (2:1:1 vermiculite, peat, sand) in transplant cylinders. A set of 15 cm diameter pots were filled with the same growth medium and four plants from each cultivar were transplanted into each pot, 25 pots per cultivar. Pots were fertilized with 250 ml of a 200 ppm aqueous solution of 20-20-20 (N-P-K) at the start and after each 21 d interval. The plants were watered daily and kept in a growth chamber where the temperature was regulated at 22°C (day) / 18°C (night) with 18 h of light. Pots were rotated weekly within the chamber to reduce location effects associated with walls, aisle and doorway. After the plants were fully established (when most of them had blossomed at 50 d) they were cut to a 5 cm height, watered and fertilized and allowed to regrow. Fourteen days after cutting and every 7 d thereafter, all living stems in a random selection of 10 pots were counted and categorized into their current stage of development. Plants grown in the controlled environment were used only to measure stem maturity. Care was taken not to damage stems during staging so that they would continue to grow for the next period. Dry Matter Disappearance ANIMAL MANAGEMENT. A l l animals used in the nylon bag experiments were ruminally cannulated and 27 handled according to the guidelines of the Canadian Council for Animal Care (CCAC 1993). The animals were mature Jersey (521 ±32 kg) or Holstein steers (over 800 kg), housed in individual pens and fed a diet of alfalfa cubes (70% alfalfa : 30% timothy hay) and a supplement (1:1 pelleted mix of canola screenings and barley for the Jerseys, rolled barley only for the Holsteins). Jerseys received a mix of 1:1 cubes and supplement, Holsteins 5:1, in the amount of 1.5% of body weight over two feedings per day (0800 and 1600 hrs). All animals were started on the diet at least 10 d before the commencement of trials. Mature Hereford cows with both esophageal and rumen fistulas were used for the collection of fresh masticate and to incubate the nylon bags while grazing. Before the first nylon bag trial and between trials they were maintained on the same diet as the Holstein steers. The Herefords had no processed feed while on alfalfa pasture. They grazed stands of pure alfalfa for at least 7 d before masticate collection and throughout the nylon bag experiments. Water and trace mineral salt was always available. N Y L O N BAG PROCEDURES. Preparation of nylon bags followed the methods established by Goplen et al. (1993). Replicate 53 fj.m mesh nylon bags (custom sewn, 6x13 cm, with 1 cm edges) were labelled, dried and weighed before each trial. They were then charged with the weight of plant material appropriate for the trial, re-weighed and prepared for incubation. Bags were closed by tying the open ends tightly to a sample retaining string attached to a weighted plastic incubation bottle. The bags and bottles to be incubated were immersed in warm water (40°C) for 10 minutes before insertion into the rumen. As successive groups of bags were removed from the rumen, they were immediately immersed in a pail of cold water to slow fermentation. Changes to the method included increasing the number of replications from duplicate incubations of one plant sample (Goplen et al. 1993) to triplicate or quadruplicate incubations of two replicate plant samples; the preparation of a separate series of non-incubated (0 h), replicated bags for each plant sample in every trial to measure DMD without incubation; and a dry matter recovery technique that used the incubated nylon bag as the filter rather than a 100-mesh screen. Non-incubated bags were not attached to incubation bottles. Al l bags, including the non-28 incubated bags, were washed in warm running water until the water ran clear then dried at 55 °C for 48 h and weighed. FRESH STEM TIP CUTTINGS (EXPERIMENT l). The LIRD breeding program used the unique selection method of incubating fresh stem tip cuttings in nylon bags in vivo. Goplen et al. (1993) commenced nylon bag incubations in the spring of any given year, when plants in the nursery were vegetative or budding. They stopped sampling and harvested the plots when the plants had matured, then recommenced nylon bag trials on the immature regrowth. Individual plants were only sampled on one day. Thus the selection for a low, comparative IRD was carried on throughout the growing season but no plants were sampled on different days during the season. Field conditions and maturity could have affected the outcome. Five nylon bag digestion trials were conducted to compare DMD from fresh cut stem tips of field grown AC Grazeland and Beaver alfalfa at different, early stages of development throughout the growing season. The timing of the trials was structured so that a range of early maturities would be available for analysis with duplicate trials on consecutive days: Trial 1 - early spring growth, 4 h incubations only, Day 1; Trial 2 - early spring growth, 4 h incubations only, Day 2; Trial 3 - late spring growth, 4 h incubations only, Day 1; Trial 4 - late spring growth, 4 h and a kinetic series of incubations, Day 2; Trial 5 - late summer regrowth after hay cut, 4 h and a kinetic series of incubations. A sixth trial, a consecutive day replicate of Trial 5, could not be completed due to labour shortages. The source of fresh plant material for all nylon bag trials was six of the twelve alfalfa samples randomly collected from the fields of AC Grazeland and Beaver to measure the stage of development in each field. (The other six field samples were dried to estimate standing crop dry matter.) This fresh material (kept at 4°C and placed in nylon bags no more than 4 h after the morning collection) was 29 prepared by cutting off the top 15 cm (stem tips) of each set of staged stems. The stem tips of three of the six field samples from one cultivar were then chopped into 1 cm lengths and mixed. The other three field samples were treated similarly, making two replicate plant samples (REPS) for each cultivar. (This procedure was the same as that used by Goplen et al. (1993) to process individual plant stems but they did not measure the stage of development of each stem and they made only one sample for each plant.) Two grab samples from each REPS were weighed and dried at 55°C for 48 h, and re-weighed to estimate the dry matter in the fresh material. For Trials 1, 2 and 3, sub-samples of each REPS, weighing approximately 5 g each, were weighed into quadruplicate nylon bags and incubated into the rumen of a Holstein steer for 4 h. In Trials 4 and 5 additional nylon bags were charged with fresh plant cuttings from each REPS to create a series of incubation intervals that included the 4 h group. In Trial 4, triplicate bags of each REPS were incubated for six time intervals (4, 6, 8, 10, 24, 48 h). In Trial 5, quadruplicate bags of each replicate were incubated for nine time intervals (4, 6, 8, 10, 12, 16, 24, 48, 72 h). Only the Holstein steers on the high forage diet were used for the fresh cuttings trials and all nylon bags were inserted into the rumen, three hours after the morning feeding. The usual procedure for an incubation series, such as those in Trials 4 and 5, is to start the clock and insert each incubation group consecutively beginning with the group requiring the longest incubation. In other words, each group in the series is inserted into the rumen at the appropriate time as the timer counts downward toward 0 h. Al l bags are removed at once at the end of the trial. This procedure is only valid for material that will not change during the period between sample collection and the start time of any respective incubation. Unfortunately this excludes fresh plant material because it would wilt, die and may begin to decay from the outset, resulting in an uncontrolled variation between the material incubated for the longest time and that incubated for the shortest. Separate field collections before each incubation interval in the series is also not an option because maturity would vary between each sample. The only option was to invert the procedure, by placing all bags into the rumen at once, 30 starting the clock at 0 h and removing groups of bags at the appropriate time as the timer counted upwards. Thus for each of the five trials, all bags were inserted at once into the rumen of a Holstein steer, three hours after the morning feeding on the day that plant material was collected, and each group of bags was removed after the required incubation interval. FRESH MASTICATE FROM GRAZING CATTLE (EXPERIMENT 2). Ultimately it is the animal's decision to consume a specific plant part from the array of plant material available in a field. Consequently, any analysis should account for the animal's choice, especially if the treatment is meant to elicit a specific response in the animal. If the LIRD character is to be effective in preventing bloat, then its properties should be detectable in the diet of animals grazing these plants. Accordingly, several nylon bag trials were carried out to examine the kinetics of dry matter disappearance of AC Grazeland and Beaver as selected and masticated by cattle grazing in relatively unrestricted conditions. The two adjacent fields of AC Grazeland and Beaver alfalfa were fenced into two paddocks in each field. Two mature Hereford cows, both fitted with rumen cannulas and plugged esophageal fistulas, were randomly allocated to an ungrazed paddock in the AC Grazeland field, and two others were put in a paddock in the Beaver field (4 animals in total). Each Hereford pair grazed their assigned paddock for 7 d before the start of the first nylon bag trial. Four nylon bag DMD trials were conducted, in pairs on adjacent days, and pairs of trials were separated by a 7 d growth period to allow the forage to reach a later stage of development. The fields were hayed, irrigated and any weeds were spot sprayed with glyphosate. Trial 1 started on Day 26 (days of regrowth after harvest)(09/08/1995), Trial 2 on Day 27 (10/08/1995), Trial 3 on Day 34 (17/08/1995) and Trial 4 on Day 35 (18/08/1995). The pair of Herefords in the AC Grazeland paddock for Trials 1 and 2 was crossed-over to the ungrazed Beaver paddock between Trial 2 and Trial 3. The pair of animals in the grazed Beaver paddock was crossed-over to the ungrazed AC Grazeland paddock simultaneously and all animals grazed their assigned paddock for a 48 h adaptation period before the start of Trial 3. 31 On the first day of each trial, one cow from each paddock (generally the tamest animal) was haltered and fitted with an ostomy bag over her open esophageal fistula. Animals were not fasted overnight but were moved from their beds to a handling chute early in the morning, before grazing commenced. The esophageal plug was removed and a flexible nylon strap was belted around her neck on which the first ostomy bag was clipped. Each animal was allowed to graze while on a tether, in an ungrazed area of her paddock. All masticate was collected for 0.5 h, removing and replacing ostomy bags as they filled. Contents of ostomy bags that appeared contaminated from regurgitation were discarded by the handler and the bag was relocated over the fistula. Few samples were discarded because the animals were hungry and did not regurgitate much during the short collection intervals. After collection the esophageal plug was replaced and the cow was returned to her assigned paddock and allowed to graze while the nylon bags were prepared. Maturity of the area grazed was subjectively estimated because the masticated material selected by the animals was too damaged to make a more precise measurement. The masticate samples were kept in a portable, insulated, iced cooler during collection and in a walk-in cooler at 4°C during processing for nylon bags. The liquid salivary component was separated from the plant masticate by straining the masticate through one layer of cheesecloth, without squeezing, into a plastic container. Triplicate nylon bags sufficient for four consecutive incubation intervals (4, 8, 20, 48 h), in duplicate animals, were charged with approximately 5 g of the masticate from one alfalfa cultivar. A colour-coded tag with a long twine, identifying the trial, incubation interval and animal, was attached to each weighted incubation bottle. All cows were brought into a holding chute next to the field, the timer was started at 0 h and all bags containing AC Grazeland masticate were inserted through the cannula into the rumens of each cow grazing that cultivar. Bags containing Beaver masticate were inserted into the rumens of the pair of cows grazing the Beaver paddocks at the same time. Nylon bags containing masticate were not pre-soaked and insertion into the rumen occurred no later than four hours after the initial collection of the masticated forage. The twine with the colour-coded tags was strung through the rumen cannula plug, leaving the 32 tags outside to identify bag groups for removal. The cows were then returned to their respective paddocks and allowed to graze freely between bag removal times. Approximately 30 min before the end of each incubation interval, the animals were taken out of the field to the holding chute. At the appropriate time, the tagged bottles with attached bags for that incubation interval were removed, dropped into a pail of cool water and the animals were returned to their respective paddocks. DRIED, GROUND STEM TIPS OF LIRD ALFALFA PROGENITORS (EXPERIMENT 3). If the LIRD character is a cellular component then a progressive change in 4hNbDMD of each successive LIRD synthetic should be evident. In addition, the relationship of rates of DMD to time should show a level of reduction. Samples of AC Grazeland (LIRD-4) and the synthetics, LIRD-1, LIRD-2, and LIRD-3 and Beaver, the parental standard alfalfa, were obtained from the breeder plots at the Agriculture and Agri-Food Research Station in Saskatoon, SK. The alfalfa was grown in replicated plots under dryland conditions. Stem tip clippings were taken from the four replicated plots of each cultivar at the 'pasture stage of growth, (vegetative to mid-bud)' as described by Goplen et al. (1993), dried at 55 °C and ground in a Wiley mill to 1 mm. A 100 g composite sample, made up of 25 g sub-samples from each plot replicate, was used for ruminal digestion in nylon bags. Each of three replicate nylon bags was charged with 2 g sub-samples of each composite sample for each of seven incubation intervals (2, 4, 6, 8, 12, 24, 48 h) in two Jersey steers on the mixed forage and concentrate diet. Since the plant material was dried and would not change perceptibly between the beginning and end of the trial, the insertion procedure for nylon bags was staggered (as opposed to the staggered removals for the fresh plant material in Experiments 1 and 2), the timer was started for the longest incubation interval, 48 h, and counted downwards to 0 h. Bags and bottles were presoaked in warm water before insertion. Al l bags requiring 48 h incubations were inserted into the rumen one hour after the animal had been fed on the first day of the trial. Successive insertions of the remaining bags were made at the appropriate times. All bags were removed at 0 h. 33 Statistics MORPHOLOGICAL STAGE OF DEVELOPMENT. The statistics mean stage number by stem count (MSC) and mean stage number by stem dry weight (MSW) were calculated on the data from each field plot and growth chamber pot. General formulas for these frequency indexes are given in Equation 2.1 and 2.2, 7 MSC = Y, (S-Ns)/C (2.1) 5=0 7 MSW = Y^ (S-Ds)/W (2.2) S=0 where N is the number of stems in each stage, S; D is the dry weight of stems in each stage; and C and W are the total count of stems in the sample and the total dry weight of the sample, respectively (Kalu and Fick 1981). The yield statistic, W, was also the total standing crop dry matter for each plot. Means and standard errors and other analyses for the stage of development and standing crops in each collection period were calculated using the STAT and MGLH procedures in SYSTAT® (Wilkinson et al. 1992). For the data collected on field grown alfalfa, one-way analyses of variance were calculated to detect differences in maturity and standing crop between cultivars in each collection period. The MSW score for each cultivar's field plot was also regressed against its corresponding MSC value. For the controlled environment trial, a two-way analysis of variance (the factors being cultivar and week of regrowth) was used to evaluate differences in maturity between cultivars. MSC scores for the pots containing each cultivar in the growth chamber were regressed as a linear function on days of growth. FRESH CUTTINGS (EXPERIMENT l). The mean DMD from the REPS of each cultivar was calculated for every incubation interval, including the 0 h non-incubated samples, in all five trials. The mean stage of 34 development was calculated from the MSC scores of the three field plot samples that went into each REPS. Differences in the extent of DMD between cultivars were evaluated using an analysis of the variation in DMD at each incubation interval. Descriptive statistics and analysis were output from the STAT and MGLH procedures in SYSTAT® (Wilkinson et al. 1992) and the numerical analysis tools in Quattro Pro® (Corel Corp. 1996). To evaluate comparative reductions in 4hNbDMD, Goplen et al. (1993) reported the DMD of each LIRD cultivar as a proportion of the DMD of the parental standard, Beaver (DMD relative to standard, D M D R S %; Equation 2.3). The rate of reduction in IRD was defined as (100 - D M D R S %) but actually this statistic is just a measure of the extent of reduction not the rate. Differences in extent of 4hNbDMD in the five trials were reported here using the same protocol: mean 4hNbDMD of the LIRD cultivar, AC Grazeland was calculated as a percentage of the mean 4hNbDMD of the standard cultivar, , (AhNbDMD of LIRD cultivar) DMD D% =100 * (2.3) R S (AhNbDMD of Beaver) V ' Beaver for each trial (Equation 2.3). The 4hNbDMD in Trials 1, 3 and 4 was also regressed against the MSC of the incubated samples and a third-order polynomial model was fit to the data to illustrate the relationship between an increasing stage of development and 4hNbDMD. If the LIRD trait was designed to slow the rate of degradation then rates need to be compared, as well as extents. A rate of DMD for an extent parameter, such as 4hNbDMD, would be the value divided by the total time, giving an average rate of DMD h"1. Rates can be derived in this manner for any incubation interval, including a 0-time point; mathematically they are the first derivative of the function of cumulative DMD on time. Since a total DMD was obtained by washing bags without incubation (OhNbDMD), two rates were calculated for the 4 h incubations, C„, the rate of initial dry matter loss from 0 h bags, and C s, a secondary net rate composed of dry matter losses through solublization and fermentation and dry matter gains from microbial mass accumulation over the remaining 4 h interval. 35 Since it normally took about 15 min to wash the 0 h nylon bags, Cs was actually calculated as the average rate over the remaining 3.75 h (Equation 2.4). The mean Cs for the REPS of each cultivar was calculated for all five trials. Differences in Cs between cultivars were evaluated using an analysis of variance. n = (OhNbDMD - OhNbDMD) FRESH MASTICATE (EXPERIMENT 2). For the masticated forage selected by the animals, differences in 4hNbDMD between alfalfa cultivars and at each time point in the 48 h incubation series were evaluated using the same statistical procedures as in Experiment 1. A three-way analysis of variance (cultivar, day of collection in paired trials and paired animals as factors) was used to examine differences within trial pairs (Trials 1 and 2; Trials 3 and 4). LIRD ALFALFA PROGENITORS (EXPERIMENT 3). The parameters OhNbDMD and 4hNbDMD were extracted for each animal and cultivar and analysed using the statistical procedures of Experiment 1. Differences between animals and cultivars were evaluated using a two-way analysis of variance. KINETIC ANALYSIS. Rumen microbial growth and fermentation are described mathematically by kinetic models (Hobson and Jouany 1988). The patterns of growth and decay exhibit common behaviours that are classically defined into four consecutive phases (Stanier et al. 1970), a lag phase, a phase of exponential growth or decay, a stationary phase and a death phase. Conceptually, a model of the first three phases can be inverted to so that it can describe phenomena that are either increasing or decreasing. Parameters derived from each phase can then be used to describe or compare different fermentations or microbial activities. A regression curve was fitted to the time-course nylon bag data in Experiments 2 and 3, by a 36 nonlinear, least squares, iterative algorithm (Wilkinson et al. 1992) using the Gompertz growth model as modified by Zwietering et al. (1990). This is a three-parameter, first-order exponential model regressing the dependent variable against time (Equation 2.5). For the dependant variable, YDMD, the proportional yDMD=A^xp(-cxp[^(X-t) + l]) (2.5) dry matter disappearance, the parameters are / i m , the maximum rate of disappearance, X, the lag time, and A, the asymptote. The model is a mathematical analogy of a sigmoidal growth or fermentation process. Lag time and the maximum rate of DMD are highly correlated because X is defined as the x-intercept of the tangent to the curve at the inflection point, with the slope of the line being fim. (A detailed review and justification for this model are included in Appendix I.) Iterative regression solutions on multi-parameter models overestimate the coefficient of determination (all regressions had raw R2>.99) because the residual sum of squares is reduced as more parameters (and degrees of freedom) are added to the regression model. The NONLIN algorithm in SYSTAT® calculates a standardized coefficient, R2, corrected for this overestimate (Wilkinson et al. 1992). This is the coefficient reported here. Asymptotic standard errors of the estimated parameters were also calculated so that differences between estimated parameters could be tested with a t-test for independent sample means. Factoring the first derivative C"7A) of each model yielded curves that illustrated the variation in the rate of DMD and directly compared the rates of DMD of the cultivars. 2.3 R E S U L T S A N D D I S C U S S I O N M o r p h o l o g i c a l S t a g e o f D e v e l o p m e n t Two experiments were designed to compare the maturity of AC Grazeland with that of the standard cultivar, Beaver. In the first experiment, maturity, as measured by the mean morphological stage of 37 development (Kalu and Fick 1981), was monitored in the field under normal growing conditions. In the second experiment, the rate of development was measured in a strictly controlled growth environment. IN THE FIELD. Partial reseeding appeared to correct the poor establishment and low plant density in the stand of AC Grazeland observed in the fall of the establishment year. The standing crops of AC Grazeland and Beaver were similar through the next growing season (Table 2.1). Differences were slight and favoured Beaver. After the first cutting, AC Grazeland regrew well because standing crops in both fields were nearly equivalent. Table 2.1 Standing crop dry matter and morphological stage of development of alfalfa (cv. AC Grazeland and Beaver) in the spring and after harvest Standing Crop (kg/ha) MSC Z Growth Period AC Grazeland Beaver AC Grazeland Beaver Spring Growth 3 weeks 2,409 ±277 y 2,823 ±334 1.92a ±0.060 1.70b ±0.082 5 weeks 4,460 ±691 4,698 ±526 2.43a ±0.058 2.07b ±0.072 Post-Harvest Regrowth 3 weeks 2,229±285 2,233 ±437 1.37a±0.098 1.37a±0.149 8 weeks 4,823 ±596 5,056 ±553 2.60a ±0.155 1.65b ±0.110 2 Mean stage number by Stem Count y Mean ± Standard Error; spring n=60, 36 for 3 and 5 wk; regrowth, n= 24, 12 for 3 and 8 wk. a ' b Means followed by the same or no letter under the same heading are not different (P>.05) The field grown AC Grazeland matured at a rate different from Beaver (Table 2.1). Stems elongated faster and AC Grazeland plants were generally taller than Beaver. By mid-June AC Grazeland was in early bloom (about 10% of the stems in Stage 5) whereas Beaver had only a few blossoms by that time. Maturity differences were more noticeable after the first hay cut with regrowth of AC Grazeland reaching MSC = 2.0 and bud stage at least one week earlier than Beaver. 38 Time (days of growth) Figure 2.1 Change in maturity (Mean Stage number by stem Count, MSC) of two alfalfa cultivars (cv. AC Grazeland, Beaver) grown in a controlled environment over a 35 day period IN A CONTROLLED ENVIRONMENT. Early maturity was also observed in plants grown in growth chambers with standardized day lengths, temperatures, soil and nutrients. AC Grazeland matured at a rate faster than its parental cultivar, Beaver (Figure 2.1) confirming anecdotal accounts and our field observations that AC Grazeland 'flowers early'. Beaver matured at a rate of 0.06 MSC d"1 (R2=.89, n=40, SEM=.004) compared to 0.09 MSC d"1 for AC Grazeland (R2=.84, n=40, SEM=.006) in a controlled environment (P< .05). Consequently, in the field, AC Grazeland may be as much as half a stage of development ahead after three weeks growth (Table 2.1). Fick and Onstad (1988) reported that measures of alfalfa quality were highly correlated with MSW and confirmed the predictive capabilities of the index for the continental United States. Kalu and Fick (1983) had previously reported that the index MSW was highly correlated with in vitro quality measurements associated with the development of cell walls in alfalfa, such as neutral detergent fibre 39 (NDF), acid detergent fibre (ADF) and acid detergent lignin (ADL). Mueller and Fick (1989) subsequently regressed MSW on MSC so that MSC data could be used to simplify data collection for forage quality investigations. Their regression was linear if they used data from alfalfa stands that were less than eight weeks old or had a MSC <, 3.5. The double sampling procedure used to collect MSC data at Lethbridge also provided material to evaluate MSW. The regression of MSW on MSC had a high degree of linearity for the irrigated fields at Lethbridge (Figure 2.2). The parameters in the model were also very similar to those reported by Mueller and Fick (1989) suggesting that the equations relating forage quality to MSW derived by Kalu and Fick (1983) could be used to predict differences in forage quality between the cultivars with an established measure of confidence. MSC Figure 2.2 Relationship between MSC (Mean Stage number by stem Count) and MSW (Mean Stage number by stem Weight) of two alfalfa cultivars (cv. AC Grazeland, Beaver) grown under irrigation 40 A confidence interval on the means, which explained 95% (CI9 5) of the variation in MSC, was calculated for each cultivar (from the means and standard errors of the data in Figure 2.1) and converted into MSW (using the linear relationship in Figure 2.2). These values were then expressed as a normalized range of MSW for 95% of the variation in maturity. For example, the CI 9 5 for maturity in AC Grazeland was 2.2<MSW<2.6 and for Beaver, 2.0<MSW<2.2 after 21 d growth. Using the predictive equations of Kalu and Fick (1983) and the CI 9 5 for MSW, the proportion of cell wall in AC Grazeland after 21 d growth, estimated by NDF, would fall in the range 35.7<NDF%<37.4. The C I 9 5 for Beaver was less (P<.05), 34.2^NDF%<;35.2. Similarly, the CI 9 5 for ADF was a range of 25.7<ADF%<27.0 for AC Grazeland, which was greater (P<.05) than that for Beaver 24.7<ADF%<25.4. The amount of lignin was also greater (P<;.05) in AC Grazeland, 5.0sADF%<5.4, than in Beaver, 4.8<ADF%<5.0. Differences in maturity between the cultivars in the field (Table 2.1) were actually greater and probably account for most of variation in the LIRD cultivars reported by Goplen et al. (1993). Dry Matter Disappearance Several experiments and individual trials within them were conducted to examine the differences in DMD between LIRD cultivars and the Beaver standard. In the process, the trials had to more precisely define the standard alfalfa, in this case Beaver at a stage of development that could cause bloat. Further, the experiments needed to account for the effects of maturity and the processing of plant materials, including animal mastication, to develop a more thorough understanding of DMD in relation to the LIRD cultivar, AC Grazeland. The experiments were classified by the type of plant material. The trials in Experiment 1 used fresh cuttings from plant stem tips; in Experiment 2 the substrate was freshly masticated plant material selected by esophageally fistulated cattle; and in Experiment 3 the substrate was the dried and ground stem tips collected from the plants in the original LIRD breeder plots using the procedures of Goplen et al. (1993). (Tabular results for Experiments 1 and 2 are listed in Appendix II.) 41 FRESH STEM TIP CUTTINGS (EXPERIMENT l). In fresh cut alfalfa, which was similar to the material collected by Goplen et al. (1993) during the LIRD breeding program, D M D R S % (dry matter disappearance relative to the standard) over a 4 h incubation interval appeared to be less, suggesting a greater difference, when AC Grazeland was in earlier stages of development (Table 2.2). Goplen et al. (1993) had also observed this during LIRD-3 trials although they had no replication within a location nor any trials that measured maturity. The DMD R S % of AC Grazeland (94.2%) was not significant (Table 2.2). In other words there was no difference between the disappearance of dry matter from AC Grazeland or Beaver stem tips. Further, none of the D M D R ^ in any of the five trials in Experiment 1 was reduced enough to be attributable to a low 4hNbDMD in the LIRD cultivar. Overall variability was high, since the pooled standard error of the mean (4.13 %) was almost double that reported by Goplen et al. (1993) during the LIRD-4 progeny tests. However greater variation was expected in Experiment 1 because the plant material used in the five trials came from every stage of development and several plants within each 0.1 m 2 sample plot. That was one reason for conducting five trials. Goplen et al. (1993) had used only the tops of individual plants, likely the tallest, most mature stems, which would reduce the variability. Goplen et al. (1993) had also reported no DMD R S % differences for all 36 progenies of AC Grazeland that they attributed to weather and harvesting difficulties. Weather and harvest were not factors at Lethbridge during Experiment 1, but the results confirm those of Goplen et al. (1993) that there was no difference in 4hNbDMD between the LIRD cultivar and the standard. Goplen et al. (1993) defined the IRD in their breeding program as the accumulated DMD from nylon bags over the first four hours of incubation. In fact this is an extent parameter, defined as an accumulation over a fixed interval (Albrecht et al. 1989) rather than a rate. Actual rates of fermentation and DMD may vary during this interval, with changes in the solubility, hydration and fermentation characteristics of the substrate and the activity levels of the microbial community. 42 Table 2.2 Rates of dry matter disappearance after a 4 hour digestion in vivo of fresh cuttings, masticate and genetic precursors of the cultivar AC Grazeland, and maturity relative to the standard alfalfa, Beaver Trial Group DMD R S (%)z C y (DMD% h"') MSC* AC Grazeland Beaver AC Grazeland Beaver Experiment 1 Fresh cuttings Trial 1 87.3 6.70 8.13 1.80a ±0.066 1.83a ±0.147 Trial 2 94.4 1.63 2.73 2.02a ±0.198 1.74a ±0.167 Trial 3 95.4 9.11 9.99 2.43a ±0.107 1.94a ±0.187 Trial 4 104.1 4.92 4.20 2.52a ±0.101 2.27a ±0.056 Trial 5 89.8 4.75 5.56 1.36a ±0.158 1.34a ±0.175 Mean±SE w 94.2ns±4.13 5.42" ±0.880 6.12a ±0.884 2.03a ±0.097 1.82b ±0.085 Experiment 2 Fresh masticate Trial 1 & 2W 100.1ns ±2.55 2.72a ±0.105 1.76a ±0.024 early bud Trial 3 & 4 103.4ns±1.99 2.86a ±0.062 3.12a ±0.231 early bloom Experiment 3 Dried and ground Beaver 100.0 3.05 ±0.010 nd LIRD-1W 92.5ns ±0.61 1.96* ±0.072 nd LIRD-2 95.0ns±0.57 2.05* ±0.135 nd LIRD-3 95.4ns±3.31 1.86* ±0.329 nd AC Grazeland 92.4ns ±2.56 1.80* ±0.367 nd 2 D M D R S = (Mean 4 hr DMD% of LIRD cultivar / Mean 4 hr DMD% of Beaver)* 100 y C s = Rate s e c o n d a r y of DMD = (4hr DMD% - Ohr DMD%) / 3.75 h x Mean Stage number by stem Count; nd=not determined w Mean ±Standard Error; Expt 1 n=10 for DMD R S ; n=20 for C s; n=60 for MSC. Expt 2 n=8 for DMDRS; n=6 for C s in each trial group. Expt 3 n=6 for D M D R S and C s for each cultivar. a ' b Means in adjacent columns under one heading followed by the same letter are the same (P>.05) n s Mean LIRD cultivar DMD compared to the standard, Beaver are not different; * different (P=.05) 43 Another extent parameter is the soluble and presumably fermentable substrate that is available immediately after the start of an incubation. For nylon bags, these initial 'rates' of DMD are approximated by hydrating and washing samples of the substrate without subsequent incubation. The mean total (0 h) DMD of all trials in Experiment 1 for AC Grazeland and Beaver was 8.07% (SE=.707%) and 7.58% (SE=.525%) respectively, which were not different (P>.05). This soluble fraction is not instantaneously removed at the 0 h time point. The process of solublization takes time, approximated by washing which normally took about 15 min of running warm water through the 0 h set of nylon bags. Recalculated, these initial (0 h) rates of DMD, C0, were approximately 30% h"1. This is a 'best' approximation of the maximum rate of DMD that might occur during a 4 h incubation interval when measurements are made at only two time points, the beginning and the end. A series of incubations would be required to derive any other rates of DMD. The mean 4hNbDMD of the fresh cuttings was 28.5% (SE=1.72%) and 30.9% (SE=1.76%) for AC Grazeland and Beaver, respectively. The average hourly DMD rates may be approximated by subtracting the mean total 0 h DMD from these values and dividing by the remaining time (3.75 h). These secondary rates, C s, for each trial are shown in Table 2.2. The mean rates of secondary DMD for the fresh cuttings trials, were 5.42% h"1 (SE=0.880%) and 6.12% h"1 (SE=0.884%) for the respective cultivars. In other words, the C s of AC Grazeland was 88.6% of Beaver but the difference was not significant (P>.05). The maturity of AC Grazeland did not differ (P>.05) from that of Beaver within each trial in Experiment 1 but over all trials AC Grazeland was more mature (P^.05). Trials were not timed to coincide with any particular date or maturity except to ensure that they occurred in early stages of development, that is, earlier than early bloom (Stage 5). The relationship between maturity and 4hNbDMD was not linear at least over the range of maturities found in these trials (Figure 2.3). The best fitting (P<.05) third-order polynomial regression of 4hNbDMD on MSC produced negative coefficients indicating that DMD declined with maturity in both cultivars. However the relationship between DMD 44 6 ^ Q 50% -, 45% 40% 35% 30% 25% 20% 15% A A Beaver o A C Grazeland I Polynomial Regression, Beaver Polynomial Regression, A C Grazeland » o y = 7.7882 - 12.295x + 6.8517x2 -1.2739x3 R 2 = 0.5894 A y = 14.663 - 20.790X + 9.9498x2 -1.5684x3 R 2 = 0.3942 1.5 1.75 2 2.25 Maturity (MSC) 2.5 2.75 Figure 2.3 Total dry matter disappearance after 4 hour in vivo incubations, of fresh cuttings of two alfalfa cultivars (cv. AC Grazeland and Beaver) relative to maturity of the cultivar and MSC was not strong (low R2) because of the variation in DMD between trials. Howarth et al. (1982) had also observed this variation and suggested that the only way to reduce it was to limit comparisons to incubations carried out on a single day. There are many sources of variation in nylon bag digestion trials but they can be broadly categorized as originating from one of four sources, the plant, the animal, the environment or the experiment. In field trials some sources of variation can be controlled and a few other variables can be measured, but it is impossible to control or measure all of them. The variation in DMD between trials may be caused by unmeasured sources such as wilting in plants experiencing drought stress, or an animal that consumes a great amount of cold water, or a different technician handling the samples on the day of the trial. Selecting for a LIRD character in stem tips appears to have affected 4hNbDMD in two ways 45 (Figure 2.3). First, the DMD of AC Grazeland was reduced in comparison to Beaver at the earliest stages of development. Second, the normal decline in DMD with advancing maturity was delayed in AC Grazeland to a later stage of development than its occurrence in Beaver. The decline in DMD with maturity then parallels the DMD for Beaver. 50% -| 45% -40% -2 35% -Q 2 30% -Q 25% -20% -15% -A Beaver o AC Grazeland | Polynomial regression, Beaver Polynomial regression, A C Grazeland y = 7.7882 - 12.295x + 6.8517x2 -1.2739x3 R2 = 0.5894 A y = - 9.2061 + 13.43x - 6.0775X2 + 0.8824x3 R 2 = 0.5185 1.5 1.75 2 2.25 Maturity of Beaver (MSC) 2.5 Figure 2.4 Dry matter disappearance (%) of fresh cuttings of alfalfa (cv. AC Grazeland, Beaver) after 4 hour in vivo incubations relative to the maturity of the standard alfalfa cultivar, Beaver Comparisons are always made against some standard, in this case the cultivar, Beaver. During the LIRD breeding program, Goplen et al. (1993) compared the 4hNbDMD of each alfalfa plant in the nursery against every other plant and finally against the 4hNbDMD of Beaver. The cuttings were taken when Beaver was in the correct stage for harvest, so all comparisons were made against this standard at its stage of development. Figure 2.4 illustrates this comparison in the present study. The DMD of AC Grazeland was lower than Beaver when Beaver was immature (MSC< 1.75). As the standard cultivar matured, the difference in DMD between the two cultivars declined and effectively disappeared. This 46 comparison assumes that both cultivars age at the same rate, and that the rate of change in DMD is the same as each cultivar matures. However, apparently these assumptions are not valid (Figures 2.1, 2.3). Maturity plays a substantial role in the DMD of the standard, Beaver, and its effect is similar but delayed in the LIRD cultivar, AC Grazeland. FRESH MASTICATE FROM GRAZING CATTLE (EXPERIMENT 2). When the animals selected and masticated the forage there were no DMD R S % differences (P>.05) between the two cultivars, at two different, though early, stages of development (Table 2.2). There were also no differences in the masticate trials between animals and trial series in the same stage of development. Consequently the four trials in Experiment 2 were pooled into two larger series, designated Trials 1&2, and Trials 3&4. For masticated forage the mean 4hNbDMD was 87.1% (SE=0.006%) and 85.3% (SE=0.006) for AC Grazeland and Beaver, respectively, which was more than twice the extent of DMD from stem tip cuttings in Experiment 1. The mean C s was 2.79% h"1 (SE=0.064%) and 2.44% h"1 (SE=0.404%) for AC Grazeland and Beaver, respectively which was lower than the rates for fresh cuttings. The extent and rates of DMD differ from Experiment 1 because mastication liberated more of the available dry matter in the 0 h pool. Animal preference and ingestive behaviour affect the outcome of digestion (Allison, 1985; Van Soest 1994), a fact that is also illustrated by the differences in C s between the individual trial groups in Experiment 2 on masticated forage (Table 2.2). While there was little difference in the C s of AC Grazeland between trials, the C s of Beaver almost doubled and exhibited greater variability. DRIED, GROUND STEM TIPS OF LIRD ALFALFA PROGENITORS (EXPERIMENT 3). As in Experiments 1 and 2, there were no differences (P>.05) in DMD R S % between the dried and ground samples of Beaver and the LIRD synthetic cultivars obtained from breeder plots. The mean 4hNbDMD was 60.7% (SE=0.013%) and 65.7% (SE=0.017%) for AC Grazeland and Beaver, respectively. Curiously, the LIRD precursors of AC Grazeland did not show a progressive decline in DMD R S % with each cycle of selection. Only the 47 products of the first (LIRD-1) and the last (LIRD-4, AC Grazeland) cycles of selection had lower D M D R S % in relation to Beaver. The C s of each LIRD synthetic was lower (P<.05) than Beaver (Table 2.2) but not consecutively, that is, LIRD-2 did not follow LIRD-1 with a lower C s. The lower rates are a result of a greater cumulative DMD from Beaver after accounting for the 0 h soluble losses. Technical staff at the Agriculture and Agri-Food Research Station in Saskatoon used the experimental collection technique instituted by Goplen et al. (1993) for the present study. Consequently, Beaver was probably in an earlier stage of development than AC Grazeland when the samples were cut. Interestingly, all LIRD cultivars exhibited the same comparative reduction in the secondary rate of DMD, Cs. Thus, most of the difference between Beaver and the LIRD synthetics appears to have occurred in the first cycle of selection for LIRD-1. The variability increased with each subsequent cycle of selection so that overall, the difference in C s between the LIRD cultivars was not significant (P>.05). Kinetic Models The time course data collected in Experiments 1, 2 and 3 were entered in a nonlinear model of the kinetics of DMD regressed against time to facilitate comparisons about DMD between the cultivars. (Time course data were collected in two trials in Experiment 1, Trials 4 and 5 on fresh cut alfalfa; in all four trials in Experiment 2, the pooled groups Trials 1&2, and 3&4 on fresh, masticated alfalfa; and in Experiment 3 using the ground samples of the LIRD cultivars obtained from Saskatoon, SK.) The maturity of the fresh plant material was known, because it was measured while collecting the fresh cut samples in Experiment 1 and subjectively estimated in Experiment 2 while tending the animals during fresh masticate collections (Table 2.2). Consequently it was easier and more meaningful to refer to the trials in each of these experiments as belonging to one of two states of maturity, immature (early bud; Trial 5, Experiment 1 and Trial 1&2, Experiment 2) and mature (early blossom; Trial 4, Experiment 1 and Trial 3&4, Experiment 2). 48 Table 2.3 Parameter estimates of in vivo dry matter disappearance from fresh cuttings and masticate of the LIRD cultivar AC Grazeland and Beaver alfalfa AC Grazeland Beaver A 2 c m T 0 R2(n) A c m T 0 R2(n) Experiment 1 Fresh cuttings Early bud 89.2 y 6.67 0.43 0.978 89.9 6.27 -0.23 0.973 (Trial 5) ±0.160 ±0.052 ±0.053 (63) ±0.169 ±0.050 ±0.057 (61) Early blossom 88.8 4.11 -1.94 0.969 90.3 3.69 -2.71 0.982 (Trial 4) ±0.278 ±0.043 ±0.095 (35) ±0.354 ±0.039 ±.098 (32) Experiment 2 Fresh masticate Early bud 94.2 7.03 -12.64 0.938 95.2 6.20 -15.37 0.790 (Trials 1&2) ±0.186 ±0.268 ±0.531 (50) ±0.218 ±0.258 ±.068 (41) Early blossom 94.0 5.72 -16.33 0.787 90.3 9.81 -8.66 0.911 (Trials 3&4) ±0.212 ±0.319 ±0.970 (47) ±0.191 ±0.416 ±0.388 (38) 2 Parameters estimated fory=a*exp(-exp(((cm*e)/a)*(t0-x)+l)), the kinetic growth (modified Gompertz) model: where a= the asymptotic value, cm = the maximum specific rate of disappearance, t0 = lag time and e=exp(l). A (DMD%), C m (DMD%) h"1 and T 0 (h) are parameters in a specific trial. y Estimate ± Asymptotic Standard Error of the Estimate The asymptotic estimates of the total DMD after 48 h digestion, A, (Table 2.3, 2.4) were highest in freshly masticated forage, lower (P<.05) in fresh cut samples and lowest (P<.05) in dried material. Fay et al. (1980) observed a similar ranking of treatments using homogenized, masticated and whole fresh plant material while measuring gas production but they did not dry and grind any samples to learn if standard sample processing could also affect the results. Clearly, the sample treatment has a significant impact on DMD of fresh plant material, that could either obscure or confound results of breeding trials designed to select forage strains for grazing. The maximum rate of DMD, C m in the kinetic model, is the peak rate of DMD and occurs at the point of inflection of the curvilinear relationship between total DMD and time. Within Experiment 1, immature (early bud) stem tips had higher peak rates of DMD (P^.05) than mature (early blossom) stem tips. Also, the fresh cuttings from AC Grazeland had a higher peak rate of DMD (P<.05) than Beaver. However, within Experiment 2, there were no differences in C m because the variability was much greater 49 Table 2.4 Parameter estimates of in vivo dry matter disappearance of AC Grazeland, its precursors (LIRD-1, LIRD-2, LIRD-3), and the standard alfalfa, Beaver Parameters2 Cultivar A y c T 0 R 2(n) Beaver 84.4 ±0.352* 2.55 ±0.067 -22.2 ±0.648 0.878 (46) LIRD-1 82.9 ±0.407 2.26 ±0.073 -25.0 ±0.889 0.833 (45) LIRD-2 82.4 ±0.301 2.59 ±0.068 -22.1 ±0.646 0.876 (45) LIRD-3 82.8 ±0.331 2.59 ±0.070 -22.2 ±0.660 0.847 (48) AC Grazeland 79.4 ±0.322 2.39 ±0.071 -23.8 ±0.783 0.886 (48) z Estimates of DMD from ground and dried plant samples, Experiment 3. y Parameters estimated for y=a*exp(-exp(((cm*e)/a)*(t0-x)+l)), the kinetic growth (modified Gompertz) model: where a= the asymptotic value, cm = the maximum specific rate of disappearance, t0 = lag time and e=exp(l). A (DMD%), C m (DMD%) h"1 and T 0 (h) are parameters in a specific trial. x Estimate ± Asymptotic Standard Error of the Estimate in the masticated samples. This lack of differences was probably an effect of animal selectivity and feeding behaviour during collection of the samples. The tethered animals were very selective about what they wanted to consume when they appeared to be bothered by the ostomy apparatus, the proximity of the handler, or the desire to be with the other cattle in the area. The C m in Experiment 1 coincided somewhat with our expectations about rates of DMD declining with increasing plant maturity. Both cultivars had lower C m in more mature plant material (Table 2.3). However, the higher peak rate of DMD for AC Grazeland regardless of maturity, was unexpected. The implication that selecting for a low cumulative amount of DMD would result in an increase in the maximum rate of DMD is difficult to reconcile. This effect was not evident in the dried and ground samples of the LIRD cultivar series (Table 2.4). Although none of the parameter estimates of either A or C m were significantly lower in the LIRD cultivars than Beaver, lower cumulative amounts of DMD were accompanied by lower maximum rates of DMD. However, the reason for the discrepancy is rational. The asymptotic total DMD of AC Grazeland and Beaver were equivalent (P>.05) in Experiment 1 because they were at similar (P>.05) stages of development within each trial. Since AC Grazeland was 50 -8 -4 0 4 8 12 16 20 24 28 32 36 40 44 48 TIME (h) Figure 2.5 Dry matter disappearance from fresh cuttings of two immature alfalfa cultivars (early bud, Trial 5; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance selected for a reduced accumulating total DMD early in the digestive process, the maximum rate of DMD would have to increase to result in an equivalent total DMD. The third parameter estimated by the kinetic model is the lag time, T 0 (Table 2.3). In fermentation cultures T 0 is defined as the point in time marking the start of the phase of exponential decay or disappearance. The interval between the time that the first derivative of the model equals zero and the lag time (T0 - T d y / d t _0) is called the lag phase. So T 0 is also the end of the lag phase. T 0 is not a rate but it is a covariate of the maximum rate, C m . If C m is low compared with its time of occurrence, or if the fermentation curve is displaced (to the left), T 0 can be a negative number. Most lag times and all lag phases predicted by the model were theoretical, that is, the value is negative and is extrapolated from the model because no data was collected at that time. The effect is caused by sample processing, where damage to the plant tissues before insertion into the fermentation 51 -12 -8 -4 0 4 8 12 16 20 24 28 32 36 40 44 48 TIME (h) Figure 2.6 Dry matter disappearance from fresh cuttings of two mature alfalfa cultivars (early blossom, Trial 4; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance medium shortens or eliminates the time otherwise required for hydration and invasion of those tissues. Thus the dried and ground plant material produced extremely negative lag times because these tissue samples required essentially no time to reach an exponential phase of disappearance from a nylon bag. The fresh cut plant material produced slightly negative (or positive) lag times because the intact cuticle and epidermal tissues in these samples slowed the initial rate of disappearance. The difference in lag times between masticated and chopped plant materials is actually an estimate of how well the animal's prehension prepares the food for subsequent digestion by microbes. Up to this point comparisons between the cultivars have used standard points within the function of DMD on time. The curves (Figures 2.5, 2.6, 2.7, and 2.8) illustrate several things. First, the selection pressure for LIRD targeted the lower third of the exponential phase when rates of DMD are increasing rapidly. Selection pressure at this point appears to have had little impact. 52 100, r 1 0 90. - 9 80. Y - 8 Model, AC Grazeland - 7 0 s o Mean, A C Grazeland — —Model , Beaver - 6 (%/h) DMD { // 'h-/ / / / 40\. • Mean, Beaver Model, Rate, AC Grazeland (dy/dt) Model, Rate, Beaver (dy/dt) . 5 - 4 MD Rate / / i i • I I I ! v / / 7 / 30V / / / / / / / / 20 -'VI 10" \ ^ 1 1 1 n \ \ \ \ . ^ \ \ \ \ \ \ \ \ \ \.\ - 3 . 2 . 1 n a 1 -20 1 1 1 1 -16 -12 -8 -4 1 0 4 1 8 12 16 20 24 28 32 36 40 44 — 1 — « 48 TIME (h) Figure 2.7 Dry matter disappearance from freshly masticated, immature alfalfa cultivars (early bud, Trial 1&2; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance At most it may have forced the point of inflection, C,„, higher and earlier (P<.05) in AC Grazeland than in Beaver (Figures 2.5 and 2.7). This is the exact opposite of what was intended (Howarth et al. 1982; Goplen 1993). To obtain a measure of bloat resistance, the curve for LIRD cultivars was hypothesized to have a lower and later point of inflection. The curves for DMD reveal that differences in DMD are a function of the rates of DMD. However they also illustrate that a lower and later point of inflection is only a function of plant maturity (Figure 2.5 and 2.6). To obtain a level of bloat resistance comparable to birdsfoot trefoil or cicer milkvetch, Howarth et al. (1982) postulated that a LIRD alfalfa cultivar should have a difference of 25 - 30 % DMD less than the standard after 6-8 hour incubations. The standard was not defined but for the sake of argument let it be Beaver alfalfa at an immature stage of development. Using values estimated by the Gompertz model, immature AC Grazeland had a DMD 4.9% to 2.1% lower than the standard alfalfa, immature Beaver 53 6 ^ Q — Model, AC Grazeland o Mean, AC Grazeland — —Model, Beaver • Mean, Beaver Model, Rate, AC Grazeland (dy/dt) — Model, Rate, Beaver (dy/dt) n — i — i — i — i — y - r 10 9 - 8 - 7 6 5 4-5 | on 4-4 I Q -- 3 .. 2 1 0 -24-20-16-12 -8 -4 0 4 8 12 16 20 24 28 32 36 40 44 TIME (h) 48 Figure 2.8 Dry matter disappearance from freshly masticated, mature alfalfa cultivars (early blossom, Trial 3&4; cv. AC Grazeland, Beaver) compared to a model of total extent and rate of disappearance (Figure 2.5), after 6-8 hours incubation. This level of difference is not statistically significant. The separation is greater if the effect of plant maturity is included. Mature AC Grazeland (Figure 2.6) had a DMD 17.7% to 22.2% lower than the immature Beaver standard (Figure 2.5). Since AC Grazeland matures at faster rate than Beaver (Figure 2.1) this difference should begin to develop in the field after 28 d of growth or more. Second, some exponential models use a parameter to estimate the 'very rapidly degradable component' (McDonald 1981). When describing DMD from nylon bags, this fraction is the proportion of the substrate that immediately disappears at r=0, the start of the fermentation process (these values are represented physically by the DMD from washed but non-incubated nylon bags). In the Gompertz model this fraction is calculated as the y-intercept of the function. Selection for LIRD appears to have driven this intercept downwards in AC Grazeland in comparison to Beaver alfalfa at the same stage of 54 development (Figures 2.5, 2.6). Third, for masticated forage, the inflection point ( C J occurs to the left of the function's y-intercept (Figures 2.7 and 2.8). Thus any comparison of DMD at 4 h must be related to the stationary phase of DMD when the actual rate of DMD is declining. It is interesting to speculate whether selection for LIRD would have had an effect if it had actually been applied at this point. Fourth, any 'processing' or tissue damage to the forage will affect 4hNbDMD. Even the minimal processing caused by cutting the samples appears to have moved the lag times of the functions into the negative, 'theoretical' realm, left of the y-axis. 2.4 CONCLUSIONS An analysis of the time course data shows that selection for a low initial rate of digestion, a LIRD trait in alfalfa, has affected the kinetics of dry matter disappearance (DMD in vivo). Targeting selection to reduce total DMD at a single time point induced compensatory effects in adjacent intervals and in related parameters. Selection pressure was placed on the lower third of the exponential phase of DMD (at 4 h) and although there was no evident change in the selected rate, the result was an unintentional selection for greater maximum rates, C m , of DMD. The use of in vivo techniques complicates the mathematical analysis of rates of DMD because continuous-flow fermentations like those in the rumen do not start at time 0, the origin in Euclidean geometry. These fermentations are on going and can be in one of several dynamic states (phases in the jargon of microbiology) before the introduction of any substrate for testing. Consequently the most unusual condition during testing would be one where the rumen state would not affect the outcome of the test and no impact on the mathematical derivation of extent or rate parameters. Similarly, completely intact substrates are an unusual condition, they are always processed to some degree. The more the substrate is processed, the more the mathematically derived in vivo fermentation curve will be displaced 55 to the left; lag times (T0) will be increasingly negative and the y-intercept of the function will rise. In the continuous first order model used in this study, T0 and C m are covariates by mathematical convention and the natural extension of their linear relationship runs into the negative region in a geometric plane. Negative numbers for kinetic parameters are difficult to reconcile with a biological interpretation. However, a negative lag time (T0) has two immediate interpretations. First, the substrate may be of such poor quality that the peak rate of DMD is low relative to the time it takes for substrate fermentation to reach a stationary phase of disappearance. This is probably not a correct interpretation for alfalfa forage because breeding has turned it into one of the highest quality forages available. The second interpretation is more likely, negative lag times are measures of the effect that processing the substrate (chopping, grinding, masticating) has on the rates and extent of DMD. As a corollary, only if the fermentation process starts after the origin, will T0 be a positive number. This can only occur if the substrate is undamaged and the rumen fermentation state is in a stationary phase. Comparisons using the dried and ground samples of the LIRD synthetic cultivars indicate that as a group, the rates of DMD are all similar, but lower than the standard, Beaver. Each successive cycle of selection was not a sequential, progressive reduction in 4hNbDMD, most of the change appears to have occurred in the first cycle of selection that produced the LIRD-1 synthetic. Minor morphological differences were neither measured nor controlled during the breeding program and were not used in the collection procedure for the LIRD synthetics during this study. However the results suggest that the LIRD synthetics display reduced rates of DMD only when compared against a later maturing standard, Beaver, and no differences when compared within the LIRD group. To summarize, selection for reduced 4hNbDMD has not been without consequence. Dried and ground samples of the LIRD cultivars, including AC Grazeland, had reduced average rates of DMD (Cs) during the first four hours of incubation. The plant material was harvested using the maturity of Beaver, a later maturing cultivar, to coordinate the timing of collections. When maturities were similar, and fresh cuttings were incubated in nylon bags, the maximum rate of DMD of AC Grazeland was greater than the 56 standard cultivar, Beaver. Otherwise, there were no differences. These results are inconsequential considering that the goal was to improve alfalfa's bloat resistance by reducing the initial rate of digestion. During the first four hours of incubation DMD was quite variable and reflected the interactions of DMD with in situ rumen conditions and maturity or processing of the substrate before incubation. Neither the extent nor the rates of DMD were sufficiently low that the difference between the LIRD cultivar and the standard alfalfa was unequivocal. The differences between trials on fresh cuttings and those on fresh masticate also demonstrated that animal ingestive behaviour or the adaptation of the rumen ecosystem to the forage could at least partly nullify any ameliorative effect of a LIRD alfalfa on DMD. Given the battery of replicated tests applied, including several different processing methods, incubation techniques and animal feeding regimens, there was little evidence that AC Grazeland dry matter disappeared to any extent less than that of the standard, Beaver when they were at similar stages of development. When rates of dry matter disappearance were compared there was either no difference or the standard, Beaver was the lesser, implying that the standard would be the more effective at reducing bloat incidence. Since this is antithetical, the null hypothesis (no difference) is therefore accepted. The only consequence of the LIRD breeding program that appears to transcend all others is the development of early maturity. In fact this is probably the most significant result; selection has increased the rate at which the LIRD cultivar AC Grazeland matures in comparison to the standard, Beaver. Early maturity is an important agronomic characteristic. The ability to mature earlier and regrow sooner is an advantage for grazing and hay production. It could be used to advantage by graziers because the risk of bloat may decline sooner with the more rapid cell wall development that accompanies the maturing process in AC Grazeland. 57 Chapter 3 Gas Production Kinetics 3.1 INTRODUCTION When whole leaves of legumes like cicer milkvetch (Astragalus cicer) or sainfoin (Onobrychis viciifolia) are digested by rumen microbes, they produce less gas than leaves from bloat-causing alfalfa (Medicago sativa) (Fay et al. 1980). Two things are implied by a fermentation that produces comparatively low amounts of gas, first that the barriers to microbial invasion are more resistant (less digestible) or second that the plant materials differ and produce different fermentation by-products. Leaves have only a few barriers that can resist microbial attack, externally the outer cuticle and internally the vascular tissue and to a much lesser extent, the cell walls. Digestion proceeds in a characteristic way starting with the breaching of a barrier by rumen microbes followed by invasion, disruption and finally degradation of the tissues behind the barrier (Cheng et al. 1980; Fay et al. 1981). Leaves from cicer milkvetch and sainfoin have thicker epidermal and mesophyll cell walls (Sant and Wilson 1982; Lees 1984) making the cells in these leaves more resistant to lysis (Howarth et al. 1978a). Digestibility of plant tissues is negatively correlated with maturity (Buxton 1996). The amount of cell wall naturally increases in alfalfa leaves and stems, and digestibility declines, with advancing age and stage of development (Kalu and Fick 1983; Albrecht et al. 1987). The amount of cell wall and its structure are limits to microbial digestion (Wilson and Mertens 1995). The discovery of thicker cell walls in non-bloating legumes led Howarth et al. (1982) to propose the cell rupture hypothesis of bloat and undertake a breeding program for bloat-resistance in alfalfa by selecting for increased thickness of leaf epidermal and mesophyll cell walls. Howarth et al. (1982) measured the dry matter disappearance (DMD) from leaves and proposed a theoretical time course for the fermentation of a bloat-safe alfalfa. A breeding program was initiated and Goplen et al. (1993) reported on its success that culminated in the release of the final cultivar, AC Grazeland in 1998. The 58 criterion for selection was a comparative reduction in DMD after a 4 h incubation, which was referred to as a low initial rate of digestion (LIRD, IRD). Within alfalfa clones, DMD varied between years, but the epidermal and mesophyll cell walls of AC Grazeland leaves were more than 11% thicker than those of the comparable cultivar, Beaver (Goplen et al. 1993). Although the target of selection was the cell wall in leaf tissue, the material used for the in vivo incubations was the combined pre-bud (ie. vegetative) to mid-bud stem and leaf cuttings. Using this procedure, it seems likely that selection would have affected stem tissues. It also seems likely that variations in maturity would have a confounding effect because cell walls thicken with maturity and resistance to microbial invasion is affected by cell wall thickness (Wilson and Mertens 1995). Gas production (GP) in vitro and DMD in vivo are highly correlated (Bliimmel and 0rskov 1993). A LIRD character should be detectable by either technique although only GP was used to support the original hypothesis on cell rupture. However, except for the initial experiments that showed the differences between leaves of bloat-free and bloat-causing legumes (Fay et al. 1980), GP was not used to evaluate any of the changes during the LIRD selection program (Kudo et al. 1985; Hall et al. 1994a). Bloat is also affected by the rate and extent of GP (Waghorn 1991; Moate et al. 1997). The effect of a LIRD character on GP is unknown. A true time course of digestion, compared with the theoretical one suggested by Howarth et al. (1982), may vary with incubation technology, the treatment of the leaves (Fay et al. 1980), the inclusion of stems and petioles or differences in maturity. Thus, the objectives of this study were to evaluate the kinetics of GP in the progenitors of AC Grazeland and compare the GP profile of AC Grazeland, with a parental standard, Beaver, using different GP technologies and a range of plant materials. Experiments were conducted to achieve these objectives and examine the rates and extents of GP. Trials were designed to: • compare the extent and rates of in vitro GP between the LIRD synthetic alfalfa cultivar, AC Grazeland and all parental lines used in the LIRD program, Anchor, Beaver, Kane and Vernal; 59 • compare the extent and rates of GP between AC Grazeland, the last cultivar in the LIRD series, LIRD-1, the first synthetic, and the parental standard Beaver; • compare rates of GP from leaflets of AC Grazeland and Beaver, using different processing methods (fresh intact, cut or crushed, and dried, ground samples); • compare GP from leaflets and stems and petioles of AC Grazeland and Beaver in relation to specific, early stages of development; • analyse kinetic models of extent and rates of GP; and • identify the source of any GP differences between AC Grazeland and Beaver that may suggest bloat-resistance. 3.2 MATERIALS AND METHODS The rate and extent of GP from in vitro fermentations of AC Grazeland tissues were evaluated in three experiments using two methods. Some plant materials or standards were common to each experiment to facilitate comparisons. Gas Production Experiments (Alfalfa Comparisons) EFFECT OF CULTIVAR (EXPERIMENT l). Two trials were conducted to examine the extent and rates of GP from cultivars used in the LIRD alfalfa breeding program. In Trial 1, the rates of GP of the LIRD parental cultivars were compared with those of the final cultivar in the line, AC Grazeland. In Trial 2 the rates of GP of the first and last LIRD synthetic cultivars, LIRD-1 and LIRD-4 (AC Grazeland) were compared against the standard parental strain, Beaver. Samples of the four LIRD parental lines, cv. Anchor, Beaver, Kane and Vernal, and the two LIRD synthetic cultivars, LIRD-1 and LIRD-4 (AC Grazeland) were obtained from the breeder plots at the Agriculture and Agri-Food Research Station in Saskatoon, SK. The alfalfa was grown in well-60 managed, replicated plots under dryland conditions. Stem tip clippings (top 15 cm of individual plants) were taken from the four replicated plots of each cultivar at the 'pasture stage of growth' as described by Goplen et al. (1993), dried at 55°C and ground in a Wiley mill to 1 mm. A 100 g composite sample, made up of 25 g sub-samples from each plot replicate, was used as the GP substrate. For Trial 1, 500 mg sub-samples of the dried material from each parental cultivar and A C Grazeland were weighed into one of 20 replicated incubation bottles. GP was measured using the manometric method (Waghorn and Stafford 1993). The gas volume was recorded at intervals over 48 hours (1 , 2, 3, 4, 5, 6, 8, 12, 24 and 48 h). Changes in pH were monitored with a Fischer Accumet® Model 825 pH meter by removing three replicates from each cultivar set (at 2, 4, 5, 6, 8, 12 and 24 h). For Trial 2, dried samples of each cultivar were funnelled into the bottom of one of 4 replicated, incubation syringes. The Hoenheim method (Menke et al. 1979) was used to inoculate the samples and measure GP in each syringe. The fermentation gas volume was recorded for 48 hours (0, .5, 2,4, 8, 12, 24 and 48 h). EFFECT OF LEAF TREATMENT (EXPERIMENT 2). Fay et al. (1980) reported that the handling and processing treatments applied to samples of fresh leaves affected the rates and cumulative amounts of GP. The purpose of Experiment 2 was to evaluate the effect of increasing levels of tissue damage on in vitro GP from leaves of A C Grazeland and Beaver at early stages of development. Stands of A C Grazeland and Beaver were established on adjacent 1 ha fields at the Agriculture and Agri-Food Canada Research Centre near Lethbridge, A B (49°42' N ; 110°47' W) as described in Chapter 2. Procedures for field collection and determining the morphological stage of development of plant samples followed those of Kalu and Fick (1981). Four randomly located, 0.1 m 2 microplots were harvested from each field between 0600 and 0630 h M D T . A l l alfalfa stems in each microplot were cut to a 3 cm stubble, weeds, litter and old stems were discarded and the sample was bagged separately and 61 stored in a cooler at 4°C for no more than 2 h. Samples were removed from the 4°C cooler one at a time, for processing. Each stem was examined and separated into a group based on its morphological stage of development. Definitions of each morphological stage followed those of Fick and Mueller (1989): Stage 2 - late vegetative stem, lengths 31 cm, no reproductive organs; Stage 3 - early bud stem, 1-2 nodes with visible buds, no other reproductive organs; Stage 4 - late bud stem, 2:3 nodes with visible buds. Intact fresh leaflets without petioles were clipped from the top 15 cm of the late vegetative (stage 2) and early bud (stage 3) stems and returned to the cooler. Al l stems in other stages and the remaining Stage 2 and 3 stems and lower leaflets were discarded. Five treatments were applied to the leaves before in vitro incubation: 1) fresh whole leaflets served as the undamaged control also used by Fay et al. (1980); 2) leaflets were crushed in a mortar and pestle using a 'A twist and moderate pressure to mimic damage from mastication (leaflets were flattened but intact, often with the cuticle displaced); 3) leaflets were cut into transverse sections of 1 cm or less with a sharp blade, a treatment similar to that used in the nylon bag trials for selecting LIRD cultivars (Goplen et al. 1993); 4) leaflets were perforated eight times with a sharp needle simulating another intermediate, artificial level of damage; and, 5) dried, ground leaflets (dried at 55 °C, 48 h; 1 mm screen) collected 2 days ahead of the fresh leaflets, using the same collection procedure. Three replicated syringes were charged with sub-samples of leaflets weighing approximately 1 g (fresh weight) or 250 mg (dried weight) for each treatment on each cultivar. The remaining fresh leaf material was placed in small brown, paper bags and dried at 55°C for 48 h to determine the dry matter. The samples were inoculated using the Hoenheim method and GP was measured over 72 h (at 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 16, 20, 24, 32, 48 and 72 h). 62 EFFECT OF MATURITY (EXPERIMENT 3). The digestibility of stems and leaves in the same stage of development are quite different, and the digestibility of stems or leaves decline with increasing maturity (Kalu and Fick 1983). GP may differ in the same manner but has not been evaluated in fresh alfalfa. Further, GP could differ between a LIRD cultivar and a standard. Experiment 3 was designed to measure in vitro GP from leaflets or stems and petioles of AC Grazeland and Beaver at three stages of development. Alfalfa stems were collected and classified into morphological stages of development using the same procedures as Experiment 2. The stem tips (top 15 cm) of the late vegetative (stage 2), early bud (stage 3) and late bud (stage 4) stems for each cultivar were cut off, placed in individual open trays labelled with the sample type and returned to the cooler. All stems in other stages and the remaining Stage 2, 3 and 4 lower stems and leaflets were discarded. Intact fresh leaflets without petioles were clipped from the stem tips of each cultivar and placed in separate set of trays from the stems and petioles. The leaflets were cut into 1 cm transverse sections (leaf treatment 3, Experiment 2). Stems and petioles were cut into transverse sections, approximately 2 mm in length, with a sharp blade. Three replicated syringes were charged with sub-samples for each cultivar, stage of development and leaflet or stem and petiole combination. Each sub-sample weighed approximately 1 g (fresh weight). The remaining fresh leaf or stem and petiole material was dried in paper bags at 55 °C for 48 h to determine the dry matter. The samples were inoculated using the Hoenheim method and GP was measured over 72 h (intervals 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 16, 20, 24, 32, 48 and 72 h). G a s P r o d u c t i o n M e t h o d s Two methods were used to measure the in vitro gas production from the plant materials collected for each experiment. The first method, referred to as the Hoenheim method, was developed by Menke et al. (1979). The apparatus consists of individual, glass syringe, incubation cylinders supported horizontally 63 by a rotating rack housed in an incubation oven set at a constant 39°C. The rack can hold only 53 syringes, which limits the number of treatments and replications that can be handled. The second method, referred to as the manometric method, consisted of a set of closed incubation bottles linked to a manometric apparatus that measures the gas volume from fermentation at the prevailing atmospheric pressure (Waghorn and Stafford 1993). About 120 bottles can be handled by a skilled technician (2-4 gas measurements every minute) allowing more treatments or replicates than the Hoenheim cylinders. The method chosen for a trial within an experiment depended on the number of treatments and replicates for that specific trial and the limitations of the method. The manometric method used 50 ml, sealed incubation bottles that had a limited volume for gas expansion. Trials using this method required a controlled fermentation so the gas pressure would not rupture the seals, blow out the manometric measuring device or adversely affect the rate of fermentation. In previous GP trials, Fay et al. (1980) had reported reduced rates of gas production using strained rumen fluid instead of a 1:1 blend of rumen fluid and rumen solids. (Both media were mixed with a neutral buffer at a ratio of 5:1, vol/vol buffer to rumen inoculum.) Straining and siphoning off the particulate from the rumen fluid samples would reduce the amount of fermentable substrate in the rumen liquor and increase the lag phase associated with the start of the fermentation process. Volume limitations were not a handicap for the Hoenheim method so the inoculum did not require as much filtration or buffering. Further, the plant sample size used in each incubation syringe was half that of the bottles and the volume of buffered rumen fluid inoculum used was greater. These factors combine to reduce the required ratio of buffer to rumen fluid in the inoculant, an important consideration especially because over half the gas produced in these in vitro experiments originates from the buffer (Getachew et al. 1998). R U M E N F L U I D C O L L E C T I O N . On the day of each trial a sample of rumen contents including fluid and particulate, was siphoned into a pre-warmed thermos from two steers one hour before their morning feeding. Al l animals used in the experiments were ruminally cannulated and handled according to the 64 guidelines of the Canadian Council for Animal Care (CCAC 1993). The animals were Holstein steers (over 1600 kg), housed in individual pens and fed a diet of alfalfa cubes (70% alfalfa : 30% timothy hay) and a rolled barley supplement mixed at a ratio of 4:1, in the amount of 1.5% of body weight over two feedings per day (0800 and 1600 hrs). Steers were started on the diet at least 10 days before the trials. MANOMETRIC METHOD. The rumen contents from the thermos were strained through six layers of cheesecloth into a large Erlenmeyer flask. The flask was covered and placed in an incubator at 39°C for 30 minutes to allow the particulate fraction to separate from the rumen fluid. A continuous flow of C 0 2 was channelled into the flask and the particulate layer was siphoned off leaving the rumen fluid in the flask. The rumen fluid was mixed 1:4, under C0 2 , with a neutral (pH=7.0) artificial saliva buffer (McDougall 1948), capped with a butyl rubber stopper and stored at 39°C in an incubation oven. Replicated 500 mg samples of dried and ground plant material were weighed into a series of 50 ml incubation bottles. A thermometer, magnetic stir rod and a continuous flow of C 0 2 were placed in the flask containing the buffered rumen fluid and the flask was placed on a magnetic hot plate set to stir at a gentle rate and maintain a temperature of 39°C. The initial pH of the buffered rumen fluid was measured with a Fischer Accumet® Model 825 pH meter. A flow of C 0 2 was started to expel the air in each incubation bottle and 20 ml of the buffered rumen fluid was pipetted into each bottle as the inoculum. Bottles were immediately sealed with a rubber stopper and crimped metal cap, and placed in a tray on a shaker in an incubation oven at 39°C. A replicate set of buffered rumen fluid-only bottles was prepared to correct for the amount of gas produced from inoculum alone. GP was measured by the displacement of a column of distilled water in a sealed graduated cylinder. A hollow surgical needle, attached to the cylinder by a small surgical tube, was inserted through the self-sealing rubber cap on each bottle, releasing the gas and depressing the water column. The gas volume produced in each interval was measured to the nearest 1 ml, by recording the location of the bottom of the water column meniscus in relation to the marks (ml) on the graduated cylinder. 65 HOENHEIM METHOD (MENKE'S APPARATUS). The rumen contents from the thermos were strained into a large Erlenmeyer flask through two layers of cheesecloth and the strained rumen fluid was mixed 1:2, with the neutral (pH=7.0) artificial saliva buffer (McDougall 1948) under C0 2 . The initial pH of the buffered rumen fluid was measured with Fischer Accumet® Model 825 pH meter, the flask was sealed with a butyl rubber stopper and placed in an incubation oven set at 39°C. The 100 ml incubation syringes were labelled and the syringe plungers were greased with petroleum jelly to create a seal between the ground glass syringe and the inner wall of the cylinder barrel and to lubricate the plunger. Replicated plant samples of approximately 1 g, fresh or 250 mg, dried and ground, were weighed and funnelled into the barrel of each syringe. A lubricated plunger was inserted to expel the air. A siphon, thermometer, magnetic stir rod and a continuous flow of C 0 2 were placed in the flask containing the buffered rumen fluid and the flask was put on a magnetic hot plate set to stir at a gentle rate and maintain a temperature of 39°C. Approximately 30 ml of buffered rumen fluid inoculum was injected into the base of each syringe. The syringe was inverted, the remaining air was expelled and the end of the syringe was sealed with surgical tubing and a clamp. A replicated set of buffered rumen fluid-only syringes were also prepared to correct for the gas produced from the inoculum alone. Each syringe was held vertically and a record was made of the volume of the contents (0 h reading). Syringes were inserted horizontally into a motorized rack inside an incubation oven at 39°C. The motor was set to rotate the rack gently to agitate the contents of the syringes. The gas produced in each syringe was measured at regular intervals to the nearest 0.5 ml, by recording the location of the base of the plunger in each syringe in relation to the gradations (ml) marked on the outer barrel of the cylinder. When the plunger had moved beyond the 60 ml mark, the clamp was removed on the surgical tubing and gas was gently discharged from the syringe to the 40 ml mark. This averted damage to the apparatus if the plungers were expelled too far from the barrel and prevented leakage of fluid and gas. After the last GP reading, the content of each syringe was emptied into a 50 ml beaker and a pH reading was taken with the Fischer Accumet® Model 825 pH meter. 66 Statistics DESCRIPTIVE STATISTICS. The accumulating total GP at each time for each replicate (bottle or syringe) was corrected for the sample dry matter weight and the residual gas produced from the buffered inoculum and expressed as ml (100 mg DM)" 1. The mean GP of each cultivar or treatment was calculated for every incubation interval, in all three experiments. The specific rate of GP h"1 for each replicate during a specific interval was calculated by dividing the corrected total GP in the specific interval by the number of hours in the interval. Differences in the extent of GP and the specific rates of GP between cultivars or treatments were evaluated using an analysis of the variance in GP for each incubation time. Comparisons were made using a t-test for differences between means of independent samples or Tukey's post-hoc HSD comparison of groups of means. Calculations, descriptive statistics and analysis were output from the STAT and M G L H procedures in SYSTAT® (Wilkinson et al. 1992) and the numerical analysis tools in Quattro Pro® (Corel Corp. 1996). KINETIC ANALYSIS. Rumen microbial growth and fermentation are described mathematically by kinetic models (Hobson and Jouany 1988). The patterns of growth and decay exhibit common behaviours that are classically defined into four consecutive phases (Stanier et al. 1970), a lag phase, a phase of exponential growth or decay, a stationary phase and a death phase. Conceptually, a model of the first three phases can be inverted to so that it can describe phenomena that are either increasing or decreasing. Parameters derived from each phase can then be used to describe or compare different fermentations or microbial activities. A regression curve was fitted to all time-course GP data by a nonlinear, least squares, iterative algorithm (Wilkinson et al. 1992) using the Gompertz growth model as modified by Zwietering et al. (1990). This is a three-parameter, first-order exponential model regressing the dependent variable against r G J ) M - e x p ( - e x p [ ^ - 0 + l ] ) (3.1) 67 time (Equation 3.1). For the dependant variable, YGP, gas production, the parameters are jum, the maximum specific rate of GP, X, the lag time, and^4, the asymptote, when the rate of GP (or the first derivative, dy/dt) approaches zero. Lag time and the maximum specific rate of GP are highly correlated because X is defined as the x-intercept of the line tangent to the curve at the inflection point, with the slope of the line being jum. (A detailed review and justification for this model are included in Appendix I.) The model gives solutions analogous to a sigmoidal growth or fermentation curve. The estimated specific rate of GP per hour (GP h"1), the models' first derivative (fr/dl), was calculated from the parameter estimates for each solution. Factoring the first derivative yielded curves that illustrated the variation in the rate of GP and gave a direct comparison of rates of GP between two cultivars at all time points of interest. The advantage of the modified Gompertz model is that the dependant variable can be used to analyse many different components and products in fermentation systems (Zwietering et al. 1990; Beuvink and Kogut 1993). The three parameters provide numerical estimates of three periods common to most fermentations, the lag phase, the exponential phase and the stationary phase for a particular data set. Iterative regression solutions on multi-parameter models overestimate the coefficient of determination (R2>.99) because the residual sum of squares is reduced as more parameters (and degrees of freedom) are added to the model. The NONLIN algorithm in SYSTAT® calculates a standardized coefficient, R2, corrected for this overestimate (Wilkinson et al. 1992). This is the coefficient reported here although the differences between the corrected and uncorrected coefficients were minor because the degrees of freedom often exceeded 50. Asymptotic standard errors of the estimated parameters were calculated and differences were evaluated with an un-paired (independent samples) t-test. 68 3.3 RESULTS AND DISCUSSION Effect of Cultivar on Gas Production (Experiment 1) Two trials were conducted to examine the rates and extent of GP from cultivars in the LIRD alfalfa breeding program. Two techniques, the Manometric Method and the Hoenheim Method were used in the first and second trials, respectively, to find out if one or the other was better at detecting LIRD trait. Cultivars in the first trial were the parental strains, Anchor, Beaver, Kane and Vernal, and the final cultivar from the breeding program, AC Grazeland. The second trial used two cultivars produced during the LIRD breeding program, the synthetic from the 1st cycle of selection, LIRD-1, the synthetic from the 4 th and final selection cycle, LIRD-4 or AC Grazeland, and the parental standard, Beaver. PARENTAL CULTIVARS (TRIAL 1). Generally the total and rate of gas produced by AC Grazeland were the same as that of the LIRD parental strains. There was no discernable pattern of reduced GP suggesting a LIRD character. The only significant differences in cumulative GP were between Anchor and Vernal during the first hour and AC Grazeland and Vernal over the first 3 hours of incubation but the rate of GP differed only in the first hour (Table 3.1). Reliance on these data would have erroneously picked Vernal as a LIRD cultivar. (Detailed tabular results of Experiment 1, 2 and 3 are presented in Appendix II.) The lack of differences was, to a degree, expected. Fay et al (1980) suggested that the high rates of gas produced from homogenized fresh leaf tissues obscured the GP differences between species of bloat-free and bloat-inducing legumes. Grinding the plant samples might also obscure differences in a manner equivalent to the homogenization treatment of Fay et al. (1980). High rates of GP during short lag phases would make detection of a LIRD trait difficult. The experimental procedures needed to be modified to improve the likelihood of detection because the difference between AC Grazeland and its parents was not likely to be as pronounced as the difference between species of legumes. Fay et al. (1980) mixed solid and liquid rumen contents 1:1 for their buffered rumen fluid 69 Table 3.1 Rates of gas production during in vitro digestion of the LIRD cultivar, AC Grazeland and its parental lines, cv. Anchor, Beaver, Kane and Vernal Cultivar Time GO AC Grazeland Anchor Beaver Kane Vernal S E 2 Cumulative Gas Production (ml 100 mg DM"1) l 0.37a 0.37a 0.28a'b 0.31ab 0.22b 0.035 2 0.95a 0.87a'b 0.88ab 0.83a-b 0.74b 0.048 3 1.60a 1.54a-b 1.57a'b 1.47a'b 1.38" 0.057 4 2.22 2.18 2.16 2.11 1.96 0.069 5 2.77 2.72 2.63 2.61 2.48 0.093 6 3.32 3.33 3.19 3.16 3.14 0.121 8 4.98 4.75 4.72 4.64 4.75 0.151 12 7.55 7.66 7.42 7.36 7.59 0.198 24 12.39 13.34 12.67 12.95 12.18 0.286 48 15.54 15.87 15.76 15.89 15.36 0.131 Rate of Gas Production (ml 100 mg DM" h"1) 1 0.37a 0.37a 0.28a'b 0.3 l a b 0.22b 0.035 2 0.58 0.51 0.60 0.52 0.51 0.031 3 0.64 0.66 0.69 0.63 0.65 0.033 4 0.61 0.64 0.59 0.62 0.58 0.022 5 0.50 0.49 0.48 0.46 0.49 0.019 6 0.61 0.61 0.56 0.56 0.66 0.041 8 0.79 0.75 0.76 0.74 0.80 0.025 12 0.64a 0.67a-b 0.66ab 0.67a-b 0.71b 0.016 24 0.39 0.45 0.43 0.46 0.38 0.024 48 0.13 0.13 0.13 0.13 0.14 0.004 2 Pooled Standard Error; n=140 for each cultivar a , b Tukey's HSD pairwise comparison. Means in the same row followed by the same or no letter are not significantly different (P^.05) 70 inoculum, to increase the rate of gas produced from whole alfalfa leaves. They felt that the complement of bacteria used for>m vitro studies should be standardized but similar to that found in the rumen (Cheng et al. 1979). We reduced the solids in the inoculum by filtration, aspirated off the particulates that passed through the filter expecting that this would lengthen the fermentation process and help reveal any subtle differences between cultivars. This strategy was successful because the lag times for the LIRD parentals were around 2 hours (Table 3.2), values similar to those reported by Fay et al. (1980) for fresh homogenized leaves using liquid rumen fluid alone. The rates of GP from the ground LIRD parental samples were 1 to 2 ml (100 mg DM)"' lower after an 8 h incubation than those reported by Fay et al. (1980) for fresh, whole alfalfa leaves. The LIRD parental samples were mixed clippings of stem and leaf material ranging in maturity from the vegetative Table 3.2 Parameter estimates of gas production from 48 hour in vitro incubations of the LIRD cultivar AC Grazeland and its parental lines, cv. Anchor, Beaver, Kane and Vernal Parameters A 2 P£_ Cultivar Beaver 15.4 ±0.301* 0.73 ±0.023 1.82 ±0.177 0.990 Anchor 15.6 ±0.316 0.77 ±0.026 1.93 ±0.171 0.986 Kane 15.6 ±0.303 0.74 ±0.024 2.00 ±0.178 0.990 Vernal 14.6 ±0.288 0.76 ±0.026 2.05 ±0.174 0.979 AC Grazeland 14.8 ±0.318 0.73 ±0.025 1.58 ±0.180 0.981 All Cultivars 48h(n=688) 15.2 ±0.137 0.75 ±0.011 1.87 ±0.079 0.984 24h(n=665) 13.9 ±0.197 0.77 ±0.012 1.87 ±0.078 0.980 z Parameters estimated fory=a*exp(-exp(((cm*e)/a)*(t0-x)+l)), the kinetic growth (modified Gompertz) model: where a= the asymptotic value, cm = the maximum rate of gas production, t0 = lag time and e=exp(l). A [ml (100 mg DM)"1], C m [(ml (100 mg DM)"1) h"1] and T 0 (h) are estimated parameters in a specific trial. y Estimate ± Asymptotic Standard Error of the Estimate 71 to late bud stages of development. Including stems should improve the likelihood of detecting differences between cultivars because the LIRD character is not likely confined to leaves alone. As alfalfa stems age their digestibility declines more than leaves (Kalu and Fick 1983). Thus the lower rates are partly due to the highly filtered inoculum and partly a result of plant maturity and the inclusion of stems. Examination of the nonlinear regression parameters also revealed no divergent pattern of GP associated with a LIRD cultivar (Table 3.2). The overall estimate of total GP, A, for the 48 h in vitro incubation, was 15.2 ml (100 mg DM)"1. Vernal and AC Grazeland had the lowest estimates of the group. The estimated maximum rate of GP, C m was 0.75 ml (100 mg DM)"1 h"1 and occurred 9.4 hours after the start of incubation. AC Grazeland, Beaver and Kane had the lower estimates in this grouping. The lack of differences between cultivars gave us an opportunity to examine the Gompertz model of exponential growth on GP (Zwietering et al. 1990) using a large data set. We were particularly interested in the ability of the model to mimic the lag phase and early exponential phase of fermentation. The information in these phases is important because selection for LIRD was made by amplifying the DMD differences at one isolated time point early in the fermentation process. The model was needed to compare the information between trials that measured DMD from nylon bags, between other GP trials, and for its potential as a selection criterion for further plant breeding. General parameters for the Gompertz model were estimated using all the GP data collected during a 48 h fermentation interval (Table 3.2). Figure 3.1 illustrates the fit of these parameters to the mean total and rates of gas produced during the first 14 hours of incubation. Cumulative GP was slightly overestimated by the model during the lag phase. The effect of this overestimation was apparent when the first derivative of the Gompertz model, dy/dt, was compared with the observed mean gas rates (Figure 3.1). The 48 h model predicted a lower maximum rate of GP than the actual maximum mean gas rate observed. Further, the model estimated that the maximum rate would be reached one hour later than it was recorded. The discrepancy is partly explained by the different calculations that created the two 72 data sets. In the model the rate of GP (dy/dt) is a continuous variable, a rate occurs instantaneously at each time point. It is not a mean, although it has the same units as the mean. In contrast, the mean rate of GP h"1 is an average of the rates of GP recorded over a set of adjacent intervals. — I 1 1 1 1 1 1 1 1 1 1 1 1 1 1- 0 0 2 4 6 8 10 12 14 Time (h) Figure 3.1 Rates of gas produced from LIRD parental cultivars during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production (based on a 48 hour fermentation) The mean rate of GP was bimodal, with the first peak occurring around 3 hours after the start of fermentation and the second at 8 hours. In a heterogeneous substrate such as alfalfa leaves and stems, observed rates are a summation of the gas produced from several simultaneous fermentation processes (Getachew et al. 1998). A multi-modal pattern of peaks and valleys will occur because these processes advance along different time-lines. For example, the first peak could occur if some bacteria reached an early exponential growth phase by digesting easily fermentable materials during the first few hours of incubation. The second peak would occur when other bacteria reached exponential growth while digesting less soluble or more complex substrates. 73 9 -| r 0.9 8 - - 0.8 7 - ^A' • - 0.7 _ Gas Production l/100mgDM) 6 -5 -4 -/ A A ' \ P / \ / / / * / / / / / A"' • / "A - 0.6 1 - 0 . 5 ° - 0.4 1. Gas Production E 3 - / • A/ -0 .3rS Gas Production / —*— Model, 24 h Gas Production (A CO 2 - - K A Mean, Gas Production - 0.2 O 1 -- - A - - Model, 24 h Gas Rate (dy/dt) — A — Mean, Gas Rate - 0.1 o " A l l l 1 • i • i i • i i n 0 1 1 1 2 4 1 6 1 1 1 1 1 1 1 8 10 12 1 14 - \j Time (h) Figure 3.2 Rates of gas produced from LIRD parental cultivars during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production (based on a 24 hour fermentation) The Gompertz model does not mimic multi-modality so it must be used with actual data and only to illustrate general trends. Otherwise a LIRD character might not be detected because its effects may be confounded by the model's inability to discriminate between processes. Abandoning the model for another was possible but not desirable because the Gompertz model has two important advantages: it fits nonlinear fermentation processes better than most other models and its parameters can be given a biological interpretation (Appendix I; Zwietering et al. 1990). Limiting the model to consecutive data of 24 hours duration affected the general parameters slightly, (the asymptote was lower, Table 3.2) and coincidentally the predicted C m (Figure 3.2) matched the actual maximum rate. Imposing the limit was justified because a 48 hour (or longer) rumen cycle time is well beyond the period of interest for a LIRD alfalfa, while a shorter limit (12 hours or less) does not give the model enough information. The calculation of parameters requires that the process reach an asymptote so that the asymptotic standard 74 errors are low, conferring interpretable estimates to the lag, exponential and stationary phases of GP. LIRD SYNTHETIC CULTIVARS (TRIAL 2). Analysis of the rates of gas produced from the LIRD synthetics, LIRD-1 and AC Grazeland, also revealed no clear pattern of low, or high, production that might indicate Table 3.3 Rates of gas production during in vitro digestion of the alfalfa cultivars, LIRD-1 and A C Grazeland (LIRD-4) compared to the standard alfalfa, Beaver Cultivar Time (h) AC Grazeland LIRD-1 Beaver S E Z Cumulative Gas Production [ml (100 mg DM)"1] 0.5 1.90 1.89 1.78 0.114 2 5.21 5.60 5.30 0.216 4 9.88 10.71 10.29 0.282 8 15.61 16.51 15.88 0.341 12 17.61 18.49 17.96 0.491 24 20.39 21.15 20.54 0.543 48 21.70 Interval Rate of Gas Production [ml (100 mg DM)"1] h"1 0-0.5 3.71 3.78 3.56 0.228 0.5-2 2.24 2.47 2.34 0.088 2-4 2.33a 2.56b 2.50a'b 0.056 4-8 1.43 1.45 1.40 0.047 8-12 0.50 0.50 0.52 0.023 12-24 0.23 0.22 0.22 0.005 24-48 0.05 z Pooled Standard Error; n=28 for each cultivar; n=32 for Beaver a b Tukey's HSD pairwise comparison. Means in the same row followed by the same or no letter are not significantly different (P< .05) 75 a LIRD character (Table 3.3). The cumulative gas produced by the final LIRD synthetic, AC Grazeland was lower than either of the other cultivars but the differences were not significant (P>.05). The only significant difference in the rates of GP occurred after 4 hours incubation between AC Grazeland and LIRD-1. Interestingly, the rate of gas produced from the LIRD synthetics also exhibited a bimodal pattern (Figure 3.3). A high peak rate occurred during the first hour of fermentation followed by a second, much lower peak rate at 4 hours. To put these rates in perspective, after 0.5 h in vitro incubation, AC Grazeland leaves and stems were producing gas at an average rate of 0.62 1 (kg DM)"1 min"1. In vivo gas production between 1.6 and 3.3 1 min'1 is sufficient to produce bloat (Waghorn 1991; Moate et al 1997). Mean intake rates of alfalfa were 3 kg D M h"1 for the first hour of grazing reported by Dougherty et al. (1987). Time (h) Figure 3.3 Rates of gas produced from LIRD synthetic cultivars during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 76 The parameters estimated by the Gompertz model (Table 3.4) approximated the amount of gas produced during the first 12 hours except in the lag phase (Figure 3.3). The model's first derivative, dy/dt missed predicting the high peak rate during the lag phase and slightly underestimated the rates of gas produced in the exponential and stationary phases. T a b l e 3.4 P a r a m e t e r es t ima tes o f gas p r o d u c t i o n f r o m 24 h o u r in vitro i n c u b a t i o n s o f t h e L I R D c u l t i v a r s AC G r a z e l a n d , L I R D - 1 , a n d the s t a n d a r d a l f a l f a , B e a v e r Parameters A z C m T 0 R 2 Cultivar Beaver 20.6±0.248y 2.31 ±0.105 -0.054 ±0.148 0.980 LIRD-1 20.2 ±0.306 2.61 ±0.129 0.065 ±0.139 0.981 AC Grazeland 19.5 ±0.320 2.36 ±0.118 0.022 ±0.149 0.989 All Cultivars 24h(n=84) 19.8 ±0.180 2.49 ±0.072 0.054 ±0.083 0.983 z Parameters estimated for y=a*exp(-exp(((cm*e)/a)*(t0-x)+l)), the kinetic growth (modified Gompertz) model: where a- the asymptotic value, cm = the maximum specific rate of gas production, t0 = lag time and e=exp(l). A [ml (100 mg DM)"1], C m [(ml (100 mg DM)"1) h"1] and T 0 (h) are estimated parameters in a specific trial. y Estimate ± Asymptotic Standard Error of the Estimate Although the samples of the LIRD synthetics were ground and dried like the LIRD parental strains, the cumulative GP was higher, similar to those values reported by Fay et al. (1980). The difference between the LIRD parent and synthetic trials was attributed to the mixed particulate and fluid inoculum used in the LIRD synthetics trial. Partial inclusion of the particulate in the buffered rumen fluid inoculum decreased the lag time (Tables 3.2, 3.4), and moved the maximum rate of GP from the 8th to the 1st hour of incubation (Tables 3.1, 3.3), increasing it by nearly five times over the rate for rumen fluid with low amounts of particulate. 77 Table 3 .5 Parameter estimates of in vitro gas production from fresh leaves and stems of the LIRD cultivar A C Grazeland and Beaver alfalfa, and the effect of maturity and treatment AC Grazeland Beaver Treatment A z c m T 0 R 2 A c m T 0 R2 Experiment 2 Leaflets (Growth Stage 2&3) Whole 46.6 y ±12.99 0.97 ±0.121 7.62 ±1.796 0.981 34.8 ±4.541 1.03 ±0.040 6.10 ±0.512 0.9746 Cut 21.7 ±0.664 1.29 ±0.041 2.41 ±0.212 0.995 23.2 ±0.489 1.51 ±0.045 2.46 ±0.190 0.9764 Perforated 24.4 ±0.379 1.87 ±0.051 1.98 ±0.157 0.995 20.6 ±0.330 1.80 ±0.059 2.35 ±0.172 0.9953 Crushed 24.9 ±0.347 1.96 ±0.051 1.14 ±0.148 0.983 21.9 ±0.297 1.96 ±0.059 0.87 ±.154 0.9941 Ground 24.0 ±0.274 2.26 ±0.063 0.35 ±0.137 0.991 23.0 ±0.268 2.23 ±0.065 0.21 ±0.141 0.9916 Experiment 3 Leaflets (Cut) Stage 2 25.5 ±0.594 1.49 ±0.040 2.19 ±0.175 0.991 16.8 ±0.704 0.93 ±0.037 2.82 ±.272 0.993 Stage 3 19.1 ±0.646 1.18 ±0.49 2.03 ±0.270 0.986 18.5 ±0.597 1.10 ±0.040 2.57 ±0.236 0.981 Stage 4 14.9 ±0.525 0.94 ±0.042 2.27 ±0.286 0.995 20.3 ±0.625 1.15 ±0.038 2.22 ±0.223 0.993 Stems and Petioles (Cut) Stage 2 19.8 ±0.303 1.86 ±0.062 1.73 ±0.159 0.986 19.6 ±0.312 1.78 ±0.059 1.79 ±.164 0.993 Stage 3 21.0 ±0.299 2.00 ±0.064 1.48 ±0.151 0.990 20.5 ±0.295 1.94 ±0.062 1.34 ±0.154 0.993 Stage 4 11.1 ±0.326 0.95 ±0.055 1.19 ±0.303 0.985 19.6 ±0.310 1.76 ±0.058 1.34 ±0.165 0.992 z Parameters estimated fory=a*exp(-exp(((cm*e)/a)*(t0-x)+l)), the kinetic growth (modified Gompertz) model: where a= the asymptotic value, cm = the maximum specific rate of gas production, t0 = lag time and e=exp(l). A [ml (100 mg DM)"'], C m [(ml (100 mg DM)"') h"1] and T 0 (h) are estimated parameters in a specific trial. y Estimate ± Asymptotic Standard Error of the Estimate. n=51 for each regression. 78 Effect of Leaf Treatment on Gas Production (Experiment 2) Fay et al. (1980) determined that different processing techniques change the rate and amount of GP from fresh legume leaves. The LIRD breeding program used fresh plant stem tip samples, cut into 1 cm lengths but all subsequent evaluations of LIRD cultivars focussed only on the leaves (Goplen et al. 1993). No evaluation was available to show which processing technique, if any, was more effective at detecting GP differences associated with a LIRD alfalfa cultivar. Consequently in Experiment 2, GP was measured from fresh alfalfa leaflets that had been physically damaged in controlled ways. Leaflets from stage 2 and 3 stem tips of AC Grazeland and Beaver were left whole, cut, perforated, crushed, or dried and ground as in Experiment 1. Besides comparing GP from several processing techniques, the purpose of Experiment 2 included a comparison of GP from leaflets of AC Grazeland and Beaver. GROUND OR WHOLE LEAVES. Rates and cumulative extent of GP from AC Grazeland and Beaver leaves, collected from the tips of stems in growth stages 2 or 3, were not different whether they were fresh, whole and intact or dried and ground (Figure 3.4; Appendix Tables II.5,11.9). Lag times for whole leaves were long, more than 6 hours for AC Grazeland and 7 hours for Beaver (Table 3.5), which is a testament to the resilience of the barriers in plant leaves to invasion by microbes. Lag time is correlated with the maximum rate of GP, so the long lag times for whole leaves meant that maximum rates of GP were not reached during the first 12 hours of incubation (Figure 3.5). In contrast, lag times for ground leaves were less than 1 hour and maximum rates occurred within the first 4 hours of incubation. Fay et al. (1980) found differences in GP between bloat-causing and bloat-free species of legumes after 4 hour in vitro incubations. However, the differences between species within each group were subtle and obvious differences did not occur until after 8 hours incubation. We were looking for differences between cultivars within one species, so any dissimilarities over the first 12 hours of incubation were expected to be, at most, subtle. During the first 6 hours, whole leaves of AC Grazeland produced an average of 11.5% more gas than Beaver (Figure 3.4), but by the 12* hour they had produced 79 11.8% less gas. However, none of the cumulative hourly means were significantly different (P>.05). Rates of GP varied widely (Figure 3.5), differing (P<.05) at 5 hours, by 71.8% h"1 [Beaver= 0.25 ml (100 mg D M 1 ) h"'; AC Grazeland= 0.43] and at 7 hours, by 40.2% h"1 (Beaver= 0.58; AC Grazeland= 0.35). Rates of GP from whole AC Grazeland leaves averaged 13.8% h"1 more than Beaver during the first 6 hours of incubation, and 22.7% h"1 less than Beaver over the next 6 hours (Figure 3.5). Dried and ground leaf differences were trivial (Figure 3.4). AC Grazeland produced an average of 1.5% more gas from ground leaves than Beaver during the first 6 hours of incubation, and 3.0% more for the next 6 hours. Q D) E o o c o (0 25 -r 201 151 O 10 TS O 51 Model • Mean, • — Model o Mean, Model x Mean, Model + Mean, Model © Mean, Model A Mean, Ground, Beaver Ground, Beaver , Ground, A C Grazeland Ground, AC Grazeland , Perforated, Beaver Perforated, Beaver Perforated, AC Grazeland Perforated, AC Grazeland Whole, Beaver Whole, Beaver , Whole, AC Grazeland Whole, AC Grazeland -i 1 1 1 1 1 1 i 6 8 Time (h) \ 1 1 h 10 12 H 1 14 Figure 3.4 Gas produced from whole, perforated and ground leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production 80 Thus, the whole or ground leaf treatments did not clearly segregate the cultivars based on the presence or absence of a LIRD character. CRUSHED, PERFORATED OR CUT LEAVES. Mastication by a grazing animal damages fresh forage to varying degrees. Some of it will be left whole, some will be cut or abraded and some will be crushed. Bacteria will invade and digest this material and produce by-products, including gas, at rates that depend on the initial amount of structural damage. Thus the prehensile behaviour of the animal could have an O U) E o o CD & CO re O 3.0 -p* 2.51 2.01 1.5 1 1.0 1 0.5 1 0.0 • Model, Ground, Beaver (dy/dt) o Mean Rate, Ground, Beaver - * — Model, Ground, A C Grazeland (dy/dt) A Mean Rate, Ground, AC Grazeland — * — Model, Perforated, Beaver (dy/dt) x Mean Rate, Perforated, Beaver - - - - - Model, Perforated, AC Grazeland (dy/dt) + Mean Rate, Perforated, AC Grazeland — • — Model, Whole, Beaver (dy/dt) Mean Rate, Whole, Beaver - Model, Whole, A C Grazeland (dy/dt) Mean Rate, Whole, AC Grazeland + + + + 6 8 Time (h) 10 12 14 Figure 3.5 Rates of gas production from whole, perforated and ground leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production 81 effect greater than a cultivar's in vitro resistance to cell rupture. Fay et al. (1980) collected masticate from cattle to measure in vitro GP. The LIRD character was selected by cutting alfalfa stem tips into 1 cm lengths before in vivo digestion (Goplen et al. 1993). These methods were standardized in the present study into three treatments constituting a midrange of artificial leaf damage, crushing, cutting and perforating. The only leaf treatment that consistently and significantly showed differences in GP between cultivars was the most artificial one, perforation of leaves with a sharp needle. Perforation gave an intermediate level of damage, since its parameters fell between those of cut and crushed leaves (Table 3.5). Total cumulative GP and the rates of gas produced by each alfalfa cultivar were significantly (P<.05) segregated at the beginning of incubation and after 10 hours, but the separation was clear throughout the range (Figure 3.4; Figure 3.5; Appendix Table II.7). The mean gas produced from crushed AC Grazeland leaves was only 6.8% less than Beaver after a 6 hour incubation and none of the differences up to 12 hours after the start of incubation were significant (Figure 3.6; Appendix Table II.8). The hourly mean rates of gas produced by AC Grazeland were much higher (P<.05) than Beaver in the 7th, 10th, 11th, and 14th hours after the start of incubation (Figure 3.7). Models of GP for crushed AC Grazeland and Beaver leaflets produced curves with similar maximum rates and lag times but different asymptotes (Table 3.5; Figure 3.6). After 6 hour incubations, cut AC Grazeland leaves produced 10.4% less gas than Beaver, a value in contrast with the other treatments. However, neither the cumulative gas produced (Figure 3.6) nor the rate of production during each interval (Figure 3.7) were significantly different (Appendix Table II.6) although AC Grazeland was. consistently lower than Beaver. Cut leaves produced fermentations with similar lag times, but AC Grazeland had a lower maximum rate and asymptote (Table 3.5). Maximum rates of GP increased in relation to the degree of tissue damage caused by the treatment of the leaves (Table 3.5). Lag times were inversely related, decreasing with greater damage. There were few differences between cultivars; none could be attributed to the effect of a LIRD character. 82 25,. Q U) E o o c o o 3 o (0 re O 201 • Model, Crushed, Beaver Mean, Crushed, Beaver Model, Crushed, AC Grazeland Mean, Crushed, AC Grazeland Model, Cut, Beaver Mean, Cut, Beaver Model, Cut, AC Grazeland A Mean, Cut, AC Grazeland o Time (h) Figure 3.6 Gas produced from cut and crushed leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production The rates of gas produced from cutting leaves agreed with the theory behind the expression of a LIRD character, given that cutting was part of the selection process for LIRD. Theoretically, a LIRD trait should show a reduction in cumulative DMD during the lag or early exponential phases of fermentation. Although there were no significant differences, the mean cumulative gas produced from cut leaves of the LIRD cultivar diverged slightly from the standard at the start of incubation and ended by producing slightly less gas overall (Figure 3.6). The mean rate of GP was slightly lower than the standard after the lag and through the first 14 hours of incubation (Figure 3.7). Variation was high (mean C V . >10%) and 83 3.0.,-2.51 Q 2.0 D) E o o 3^ 1.51 I 1.0 w O 0.5 1 o • Model, Crushed, Beaver (dy/dt) o Mean Rate, C rushed , Beaver Model, Crushed, AC Grazeland (dy/dt) A Mean Rate, Crushed, AC Grazeland - • — Model, Cut, Beaver (dy/dt) • Mean Rate, Cut, Beaver • - - Model, Cut, A C Grazeland (dy/dt) o Mean Rate, Cut, AC Grazeland 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 2 4 6 8 10 12 14 Time (h) Figure 3.7 Rates of gas produced from cut and crushed leaves of the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production may have been reduced by increasing the number of replications. However, the space in the gas syringe apparatus limited the replication for each treatment. Perforation of leaves gave an unusual result (Appendix II, Table II.7). A strong manifestation of a LIRD character was expected to show a pattern where less gas would be produced by the LIRD cultivar at the start of incubation after which it would either catch up to the standard in a later time, or produce lower gas overall. This was the pattern in all the other leaf treatments, whole, ground, cut and crushed, although in each case the divergence was not statistically significant (P>.05). If the needle perforations were consistently applied to leaves of each cultivar, then the results should have been similar to the other 84 trials, with at least one intersection and a divergence. The results were dissimilar so the most likely cause was an experimental error, either too few holes were made in the Beaver leaves, or too many in the AC Grazeland samples, resulting in greater GP from perforated AC Grazeland leaves. Mastication was mimicked by crushing leaves in a mortar. Gas produced from this level of damage showed an interesting pattern, in spite of the lack of significant differences. The gas produced by crushed LIRD cultivar leaves was slightly reduced after a short lag because the rate of GP was slightly lower than the standard cultivar (Figure 3.6, Figure 3.7). However, the rate of acceleration of the fermentation was higher in the LIRD replicates. Consequently, the accumulating volume of gas and the rates of production caught up to and exceeded those of the standard after 8 hours incubation. The result was a small displacement of the derived curve to the right (Figure 3.7). This was also the pattern observed for the ground leaf samples (Figure 3.5). Effect of Maturity on Gas Production (Experiment 3) Maturity has a profound effect on the digestibility of forage (Kalu and Fick 1983). Individual plant parts respond differently as they age. Stem tissues lignify to a greater extent than leaf tissues because the demand for structural support is greater in stem than in leaf tissues (Albrecht et al. 1987). The selection procedure for a LIRD trait was a mass selection technique in that an amorphous mixture of stem and leaf tissues was placed under selection pressure. Goplen et al. (1993) did not examine stem tissues and considered maturity only as a condition affecting the start of the selection process in any particular year. Experiment 3 was designed to examine GP in leaflets separate from stems and petioles, and further, to evaluate the effect of maturity on those tissues. LEAFLETS. The amount and rates of GP were measured on leaves at three different stages of development. Field collections were made just after the first Beaver alfalfa stems entered the late bud stage. A few stems in the AC Grazeland were in early flower by that time, so the stage 4 stems collected 85 from this field had matured sometime during the previous seven days. Kalu and Fick (1983) showed that digestibility declined as leaves at a particular stage, aged. In other words, the digestibility of young leaves on mature stems could be the same as older leaves on immature stems. Leaves on stem tips are not always young or newly emerged; stems can stop maturing for a number of reasons so age and maturity are not always covariates (Fick et al. 1988). GP from Beaver and AC Grazeland leaves also varied with maturity and age. For example, leaves from late vegetative stems of Beaver were chronologically older and produced less gas (P<.05) than the newly emerged late bud leaves (Table 3.6). The relationships (or lack thereof) between the numbers in Table 3.6 are illustrated by the curves in Figure 3.8 where the means in Table 3.6 were plotted against the best-fitting nonlinear curve estimated by the Gompertz model. Estimated parameters for each regression are given in Table 3.5. The first derivate of each model, J Y/ d l, the specific rate of GP, is illustrated by the curves in Figure 3.9. The rates of GP (Appendix II) for the two cultivars had similar patterns (Figure 3.9). As a group the GP differences from leaves within or between cultivars were minor. The amount of gas produced from Beaver leaves in either the late vegetative or early bud stages of development were similar (P>.05) for the first 11 hours in vitro and the amount of GP from early bud leaves did not differ (P>.05) from that of late bud leaves. Similarly, the gas produced from leaves of either late vegetative or early bud AC Grazeland stems did not differ (P>.05) during the first 8 hours but after 8 hours incubation the cumulative amount of gas from leaves of late vegetative stems was greater. Leaves from late bud AC Grazeland stems produced less gas (P<.05) than leaves at earlier maturities in this cultivar (Table 3.6). Consequently, except for the 2 n d and 3 r d hours after the start of incubation, overall GP from leaves did not differ between the cultivars (P>.05) (Table 3.6, all stages) confirming the results of the earlier trials on composite leaf samples. Lag times, T0, were between 2 and 3 hours (Table 3.5). Estimated maximum rates of GP during the exponential phase, C m , were similar across maturities for Beaver but declined from least mature to most mature for AC Grazeland. Total GP for Beaver leaves 86 Table 3.6 Cumulative gas production from leaflets in 3 stages of development during in vitro digestion of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Beaver AC Grazeland Time(h) Leaf2z Leaf3 Leaf4 Leaf2 Leaf3 Leaf4 SE y 1 -0.06x -0.05 0.06 0.28 0.09 0.09 0.107 2 0.27a 0.44a 0.76ab 0.98" 0.75ab 0.61ab 0.107 3 0.92ab 1.18abc 1.69bcd 2.16cd 1.73bcd 1.26abc 0.124 4 1.70ab 2.23abc 2.68bcd 3.41cd 2 g4bcd 2.07abc 0.184 5 2.48ab 3.09abc 3.61bc 4.73cd 3.9\bc 2.80abc 0.222 6 3.33abc 4 I4a^c<*e 4.66bcde 5.99ef 5.03cdef 3.62abcd 0.239 7 4.24abc ^ J Qabcde 5.88bcde 7.45ef 6.16cdef 4.56abcd 0.281 8 5.03abc 6.18abcde g gybcde 8.84ef 7.26cdef 5 4 7 abcd 0.325 9 5.94ab 7 34abc 8.09bc 10.44d 8.38bc 6.46abc 0.368 10 6.66abc 8.21abcde O, Q^ bcde 11.84f 9.37cde 7 23 0.383 11 7.57abc 9.26abcde 10.07bcde 13.16f 10.43cde 8.23abcd 0.391 12 8.29ab 10.13bcde 11.01cde All Stag 14.28f es 11.32cde 8.86abc 0.367 1 -0.02 0.16 0.063 2 0.49a 0.78b 0.089 3 1.26a 1.71" 0.149 4 2.20 2.76 0.209 5 3.06 3.80 0.276 6 4.04 4.86 0.330 7 5.10 6.04 0.400 8 6.03 7.18 0.463 9 7.13 8.44 0.543 10 7.96 9.49 0.615 11 8.97 10.63 0.654 12 9.81 11.51 0.709 2Leaf2=leaf in late vegetative growth (stage 2); Leaf3=early bud (stage 3); Leaf4=late bud (stage 4) y Pooled standard error; n=36 for each stage of leaf; n=108 for all stages " Mean accumulated gas produced per unit of DM = ml (100 mg DM)"1 a b Tukey's pairwise comparison. Means in the same row followed by the same or no letter are not significantly different (P>.05) estimated by the asymptote, A, were also similar but declined with maturity for AC Grazeland leaves. The curves (Figure 3.8 and 3.9) show that there is a general pattern of GP from leaflets of both cultivars and that no single maturity or cultivar has an accumulation or rate of GP that is lower than any other. 87 17.5 15.0 1 Q |> 12.5 i o o H 10.0 + c .2 <~ 7.5 4-3 "D O £ 5.0 + </) re O Model, Leaf2, Beaver • Mean, Leaf2, Beaver -Model, Leaf2, AC Grazeland o Mean, Leaf2, AC Grazeland Model, Leaf3, Beaver x Mean, Leaf3, Beaver Model, Leaf3, AC Grazeland + Mean, Leaf3, AC Grazeland Model, Leaf4, Beaver o Mean, Leaf4, Beaver Model, Leaf4, AC Grazeland A Mean, Leaf4, AC Grazeland • o Time (h) Figure 3.8 Gas produced from leaves at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production STEMS AND PETIOLES. With few exceptions, the amount of gas produced from stems at different stages of development did not differ within a cultivar. GP from stems of Beaver in late bud and early bud stages did not differ (P>.05) (Figure 3.10; Table 3.7) and the amount of gas from incubating late vegetative and late bud stems did not differ either. Gas from early vegetative and early bud AC Grazeland stems differed only in the 6th and 7th hours of incubation (P .^05). However, the cumulative GP from AC Grazeland's late bud stems was lower than either early vegetative or early bud stems (P<.05) (Figure 3.10; Table 3.7). 88 2.5 T Q D) E o o 2.01 • Model o Mean • A — Model A Mean - * — M o d e l x Mean • + - - Model + Mean • — M o d e l • Mean * - - Model o Mean Leaf2, Beaver (dy/dt) Rate, Leaf2, Beaver , Leaf2, AC Grazeland (dy/dt) Rate, Leaf2, AC Grazeland Leaf3, Beaver (dy/dt) Rate, Leaf3, Beaver , Leaf3, AC Grazeland (dy/dt) Rate, Leaf3, AC Grazeland Leaf4, Beaver (dy/dt) Rate, Leaf4, Beaver Leaf4, AC Grazeland (dy/dt) Rate, Leaf4, AC Grazeland Figure 3.9 Rates of gas produced from leaves at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production Rates of GP for stems peaked in the 4 t h or 5* hour after the start of incubation and showed greater variability than the rates estimated from the model's first derivative (Figure 3.11). The late bud stems of AC Grazeland produced gas at a slower rate than those of any other maturity. As a group, GP from AC Grazeland stems in all stages of development did not differ (P>.05) from Beaver stems (Table 3.7, all stages). The gas accumulated from AC Grazeland stems was slightly lower than Beaver and coincidentally, lower (P<.05) only during the 4 th hour of incubation. 89 Table 3.7 Cumulative gas production from stems and petioles in 3 stages of development during in vitro digestion of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Time(h) Beaver AC Grazeland SE y Stem2 2 Stem3 Stem4 Stem2 Stem3 Stem4 1 -0.28ax -0.02ab -0.08ab -0.1 l a b 0.16b 0.0 l a b 0.088 2 0.62 1.21 1.12 0.72 1.22 0.69 0.153 3 2.13ab 3.22" 2.88ab 2.18ab 3.09b 1.72a 0.268 4 4.33ab 5.55b 5.34ab 4.6 l a b 4.30b 3.06a 0.332 5 6.32bc 7.83cd 7 1 6 b c d 6.78bcd 8.19cd 4.03a 0.305 6 8.18bc g y y c d e 8.76bcd 8.66bcd 10.10de 4.95a 0.267 7 9 7 6 b c d 11.30cdef 10.31bcdef 10.21bcde 11.54def 5.69a 0.261 8 11.07bc 12.58cd 11.58bcd 11.59bcd 12.81cd 6.37a 0.277 9 12.32bc 13.80cd 12.85bcd 12.85bcd 14.09cd 6.95a 0.275 10 13.34bcd 14.86cdef 13.76bcde 13.94bcdef 15.16def 7.62a 0.262 11 14.45bc 16.00cd 14.80bcd 15.03bcd 16.07cd 8.23a 0.270 12 15.41bcd 16.74cdef 15.58bcde 1 5 g 9 b c d e f 16.86def 8.68a 0.265 All Stages 1 -0.13 0.02 0.059 2 0.98 0.88 0.117 3 2.75 2.33 0.226 4 5.07" 3.99a 0.270 5 7.11 6.33 0.484 6 8.91 7.91 0.582 7 10.46 9.15 0.659 8 11.74 10.25 0.730 9 12.99 11.30 0.806 10 13.99 12.24 0.852 11 15.08 13.11 0.895 12 15.91 13.81 0.934 z Stem2=stems in late vegetative growth (stage 2); Stem3=early bud (stage 3); Stem4=late bud (stage 4) y Pooled standard error; n=36 for each stem stage; n=108 for all stages " Mean accumulated gas produced per unit of DM = ml (100 mg DM)"' a b Tukey's HSD pairwise comparison. Means in the same row followed by the same or no letter are not significantly different (P>.05) 90 Kinetic Models The purpose behind developing a LIRD alfalfa was to take the cell rupture theory of bloat to its logical conclusion and develop a bloat-safe alfalfa. A theoretical kinetic description of a bloat-safe alfalfa was proposed based on the dry matter disappearance (DMD) of other bloat-safe legumes. For gas production, a LIRD character would likely be expressed in the same way as DMD. The kinetics of ruminal digestion of a LIRD alfalfa should show less gas produced in the early phases of fermentation in comparison to some standard. Generally the kinetic parameters of digestion would either show an extended lag phase (later lag time), a reduced maximum rate of GP or low cumulative amount of GP in the early phases. Table 3.5 is an array of the estimated parameters for kinetic models of GP on all leaf and stem treatments and maturities. These are estimates of the gas produced or the time of occurrence for each phase of fermentation. Theoretically, a LIRD character should consistently show up as some reduced level or extended time in one or more of these parameters, preferably associated with the LIRD cultivar, AC Grazeland. For example, whole AC Grazeland leaves qualify in this regard, since the estimated maximum rate of GP is less and the lag time is later than the corresponding values for Beaver. However the asymptotic standard errors are so large that there is virtually no difference between these estimates. There was one trend associated with the treatment applied to the leaflets. As the leaflets of either cultivar became more damaged or fragmented, the maximum specific rate of GP increased and the lag time decreased (Table 3.5; Figure 3.5; Figure 3.7). Only for cut leaflets was the maximum specific rate reduced (P .^05) in the LIRD cultivar. When maturity was considered in Experiment 3, C m decreased (P<.05) with advancing maturity in the LIRD cultivar, but remained static (P>.05) in the standard (Table 3.5; Figure 3.9). Only the earliest maturity was different, and in that case the C m was greater in the LIRD cultivar. Thus, there is no evidence of a reduced or low rate of GP in leaflets associated with the LIRD cultivar. Insofar as there were few rate differences there were also few cumulative differences in GP from leaflets except at the asymptote (Table 3.5; Figure 3.4; Figure 3.6; Figure 3.8). 91 20.0 T Q O) E o o c o •o o (0 o 15.01 10.01 • Model Mean, Model Mean, Model Mean, — Model + Mean, Model o Mean, — -Model A Mean , Stem 2 Stem 2, , Stem 2 Stem 2, Stem 3 Stem3, , Stem 3 Stem3, Stem4 Stem4, , Stem4 Stem4 , Beaver Beaver , AC Grazeland AC Grazeland , Beaver Beaver , AC Grazeland AC Grazeland , Beaver Beaver , AC Grazeland AC Grazeland Time (h) Figure 3.10 Gas produced from stems and petioles at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by an exponential model of gas production Certainly none of the cumulative GP functions displayed any great tendency that would suggest the existence of a reduced IRD. The original design for Experiment 3 did.not include a procedure for measuring GP from stems and petioles. The focus of the LIRD breeding program and all previous assessments was on the leaves (Goplen et al. 1993). The procedure was modified when the methods used by Goplen et al. (1993) were reviewed. They used pre-bud to mid-bud stem tips, not just leaves, to measure in vivo DMD and select plants for the next breeding cycle. Thus, stems and petioles were subjected to the same selective pressure as leaves during the LIRD alfalfa breeding program. The range of maturity was relatively narrow 92 Q O E o o 0) 8. (0 o 2.50 2.25 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00 2 x o X • © - Model, Stem2, Beaver (dy/dt) Mean Rate, Stem2, Beaver Model, Stem2, A C Grazeland (dy/dt) Mean Rate, Stem2, AC Grazeland • Model, Stem3, Beaver (dy/dt) Mean Rate, Stem3, Beaver • Model, Stem3, AC Grazeland (dy/dt) Mean Rate, Stem3, AC Grazeland •Model, Stem4, Beaver (dy/dt) Mean Rate, Stem4, Beaver Model, Stem4, AC Grazeland (dy/dt) Mean Rate, Stem4, AC Grazeland I » I h H h + H H H 1 1 1 1 1 6 8 Time (h) 10 12 14 Figure 3.11 Rates of gas produced from stems and petioles at different stages of development from the alfalfa cultivars AC Grazeland and Beaver during the first 14 hours of in vitro incubation compared to estimates predicted by the first derivative of an exponential model of gas production given the annual cycle of development in alfalfa. However, as plants mature through these early stages, the ratio of stem to leaf material rises substantially, with the major increase occurring within 2 weeks of the early bud stage (Albrecht et al. 1987). Thus the selection pressure for LIRD would have been greater on stems than leaves, and greater on stems at later stages of development. Consequently, a LIRD character would be likely to show an effect greater in stems than leaves associated with advancing maturity. In vitro incubations of stems and petioles from the LIRD cultivar show clear evidence of a reduction in GP at the late bud stage of development. During the earlier stages, lag times and maximum 93 rates of GP from stems were similar in both cultivars (Table 3.5; Figure 3.10 and 3.11). Resistance to invasion and digestion by rumen microbes was evident in the increased variability of the lag time, the reduced maximum rate of GP, C m , and the lower rates of GP from stems of the LIRD cultivar. Thus, the LIRD character is resident in and expressed by in the stems and petioles and is confounded with maturity. Since the proportion of cell wall increases with maturity (Albrecht et al. 1987), and the LIRD character is, fundamentally, an increase in cell wall (Goplen et al. 1993), it is reasonable to conclude that the LIRD trait and maturity are the same effect. 3.4 CONCLUSIONS Extended lag times or reduced rates of gas production were not conspicuous in the LIRD parental and synthetic cultivars when compared against a standard, using two different GP methodologies. Each GP trial was also comparative in that the LIRD cultivar, AC Grazeland was compared against the same standard, Beaver, in effect providing two standards within each trial. There was no evidence of any significant barrier in the dried, ground samples that would delay or reduce the initial rates of GP or create a low cumulative extent of GP during the first 14 hours of incubation. The lack of any pattern could be attributed to the design of the experiment, since the use of finely processed samples or the source of the gas (the buffer reacting with the production of short chain fatty acids) could affect the outcome. However, if the barrier is cell walls thick enough to delay microbial degradation to prevent bloat, then the difference should have been more obvious. The LIRD breeding program was initiated because the differences that were observed in the rates of DMD of alfalfa leaf tissues were amenable to change with selective pressure. The rates of GP also varied in the experiment on fresh leaf tissues, but reflect the level of damage to which the leaves were subjected rather than differences between the cultivars. This experiment was also comparative with the trials on the LIRD parent and synthetic cultivars because two standards, AC Grazeland and Beaver, and 94 one treatment, dried, ground samples, were common. The results show that there was little difference in GP between fresh leaflets of the LIRD cultivar, AC Grazeland and the standard, Beaver. Certainly there was no evident decrease in extent or reduction in rates of GP during the initial 14 hours of incubation showing that a delay in digestion or a significant barrier existed within LIRD alfalfa leaflets. In the final analysis, fresh leaflets from stem tips at 3 early stages of development were compared because maturity has a considerable effect on the digestibility of fresh feed sources. This experiment also had a set of comparable standards with the other experiment, the same two alfalfa cultivars, one common leaf treatment, cutting, and two similar stages of development. The only differences in rates and extent of GP from leaflets were those associated with maturity and aging. The selective pressure for a LIRD trait was not exclusively applied to leaves but was exerted on both leaves and stems. Because of the physiological and morphological changes that occur as plants mature, the greatest selection pressure was likely applied to the stems and petioles. This possibility was confirmed by the GP results when the stems and petioles, collected for the leaflet trial, were subjected to the same in vitro incubations. The rates of gas production were lower, the maximum rate was reduced and the lag time exhibited greater variability in the most mature stem material of the LIRD cultivar, AC Grazeland. Cell walls are one of the few barriers in plants that will delay bacterial digestion and reduce DMD and GP. As alfalfa matures the amount of cell wall increases markedly in stems and to a lesser extent in leaves. Considering that the LIRD trait and maturity affect early cell wall development, it seems reasonable to conclude that they are one and the same. 95 Chapter 4 Feed Composition, Digestibility and Bloat 4.1 INTRODUCTION The genetic history of alfalfa in North America has been one of breeding for improved yield, disease and pest resistance, tolerance to environmental conditions and nutritional quality (Hill et al. 1988). Only a few cultivars have been selected for their use in pasture. Only one, AC Grazeland, has been selected to reduce bloat. (Goplen et al. 1993; Caddel 1997). AC Grazeland comes out of a conceptual and scientific analysis of the plant characteristics that cause bloat in ruminants (Howarth 1975; Howarth et al. 1978a, 1982). Theoretically, bloat was considered a consequence of the ease with which cell walls rupture during digestion. The theory was supported by a thorough scientific survey of bloat resistant and bloat causing legumes which determined that thicker cell walls were a barrier to bacterial invasion in the bloat resistant species (Fay et al. 1980; Cheng et al. 1980; Lees 1984). When the alfalfa cultivar AC Grazeland was compared with a standard, Beaver, known to cause bloat, it was found to have thicker cell walls (Goplen et al. 1993). If cell walls are thick and a barrier to digestion then there is a better than average possibility that the digestibility of the forage has changed. Increased amounts of cell wall, and the lignification of that wall, is a normal developmental characteristic in herbaceous plants. Maturity in alfalfa is closely linked to changes in forage quality especially in the early stages of development (Kalu and Fick 1983; Albrecht et al. 1987). Cell wall development is associated with declining digestibility. In herbaceous plants like alfalfa, maturity continuously advances through the growing season. This is a severe obstacle when comparisons are needed between forages in grazing and fresh feeding situations because the digestibility of the feed source is in a state of change. Animals and rumen microbes respond readily by adapting to changes in maturity. All of this variation makes comparative experiments difficult to control and evaluate. 96 The digestibility of alfalfa cultivars has been compared but few tests have been conducted in vivo. Allinson et al. (1969) and Shenk and Elliot (1970; 1971) tested a number of varieties by in vitro assay to help them select traits that were directly associated with digestibility. Albrecht et al. (1987) used two cultivars, one thick-stemmed and tap-rooted, the other thin and creeping-rooted, but found no in vitro digestibility differences between them. Thomas et al. (1968) also used two cultivars and found a difference sufficient to recommend cross-breeding and selection based on in vitro quality characteristics. Kudo et al. (1985) conducted an in vivo feeding trial using sheep that compared the first synthetic cultivar with a low initial rate of digestion, LIRD-1, against Beaver. In designing their experiments, they assumed that maturity would be controlled by timing feed collections in standardized hay cuttings and regrowth periods (1st, 2n d, and 3 rd cuts in one season) (Kudo et al. 1985). The forage grown prior to the 1st cut and regrown after the 2 n d and 3 rd cuts was fed when it was in a 'vegetative to mid-bud' stage of development. The area seeded was small (<0.5 ha) and plots of each cultivar were contiguous so they assumed that variations in maturity, caused by environmental or edaphic factors, were minor. They also assumed that the LIRD and standard cultivars matured at the same rate after cutting so any difference in maturity between cultivars at the beginning and the end of each trial would also be minor. If the assumption that both cultivars matured at the same rate was violated, then comparisons between strains would tend to show significant differences favouring the early maturing strain. Kudo et al. (1985) did not report any results for organic matter or cell wall digestibility. Instead, they found highly significant differences between cultivars, between 'cuts' and in some interactions between alfalfa strains within cutting periods for every rumen parameter measured (chlorophyll, protein and carbohydrate content and pH). They attributed these differences to the effect of the LIRD trait. However, they left the LIRD trait undefined; it was anything that could result in a reduced amount of cell rupture. If the LIRD cultivar matured at a faster rate than the standard, then they may have actually been measuring the effects of forage maturity on these rumen parameters. Instead of showing how rumen conditions are influenced by a specific LIRD character, the results would show the effect of variation in 97 the timing of the harvest and the maturity of the forage crop regrowth at different points in the growing season. If a LIRD trait was present, its expression would have been confounded with a trait for early maturity. The effects of maturity can also be confounded in a feeding trial within a short period. The nutritional quality and digestibility of most feed fractions, such as neutral detergent fibre (NDF) or crude protein (CP), change substantially through the early stages of plant development (Kalu and Fick 1983). In the Kudo et al. (1985) investigation, each feeding trial had a duration of two weeks and would not have begun until sufficient forage was available to carry out a full trial. To have attained a 'vegetative to mid-bud' level of maturity, the alfalfa would have had to be between 21 and 35 days old at the start of a trial and consequently, 35 to 49 days old at the end. Although Kudo et al. (1985) did not report the nutritional quality or the digestibility of various feed fractions, in vitro digestibility would have fallen and cell wall content would have risen perhaps as much as 20% during each trial. (The range was determined by the difference in values predicted from the equations of Fick and Mueller 1989, relating nutritional quality to maturity.) Increasing maturity during a trial would therefore mask any differences between cultivars at the beginning of that trial or anytime within the trial. The previous digestibility trial showed that a LIRD cultivar could induce changes in rumen conditions but the reason for those changes were not examined (Kudo et al. 1985). Goplen et al. (1993) measured the width of the cuticle and wall of epidermal and mesophyll leaf cells and determined that the total cell wall in the LIRD cultivars was thicker than that found in a standard alfalfa. If the LIRD cultivars have a thicker cell wall, then they may have an increased amount of cell wall. Further, if the wall is a barrier to microbial invasion then it may affect the relative digestibility of the LIRD cultivars. An in vivo assessment of the amount and digestibility of the cell wall or its nutritional analog, fibre, was missing. The research needed to be re-directed toward the cell wall and the effect of maturity on the digestibility of cell walls. Moreover, bloat incidence needed to be evaluated in relation to the amount of cell wall rather than an effect of cell wall development, such as reduced dry matter 98 disappearance (DMD) or gas production (GP). To evaluate the digestibility of cell walls of LIRD cultivars, an alfalfa strain selected for use in a grazing context, digestibility trials had to be conducted in vivo, the cultivars would have to be compared against a standard and they had to use fresh forage cuttings as the only feed source. Researchers would need to control variations in age or rates of maturity between the subject and the standard cultivar. Otherwise, differences in digestibility associated with the degradation rate of cell wall would be confounded between a difference in cell wall composition and its rate of development. Accordingly the objectives of this investigation were to: • evaluate the composition of fresh cut LIRD cultivars at several stages of development and in comparison to a standard; • evaluate the effect of maturity on cell wall development and account for changes associated with advancing stages of development in the LIRD cultivar; • determine the digestibility of feed fractions in the LIRD strain associated with the cell wall and compare them with a standard; and • assess the potential of the LIRD cultivar to reduce bloat when fed to sheep. 4.2 MATERIALS AND METHODS Two alfalfa cultivars, Beaver and the synthetic, LIRD-3, were used as progenitors in the development of the final LIRD cultivar, AC Grazeland. Beaver was used as one of the parental lines and as the standard for all comparisons of DMD; the progeny of LIRD-3 was used to synthesize AC Grazeland. The comparative digestibility of these cultivars was not known so it was evaluated in two experiments using Beaver as the common standard. 99 Digestibility Experiments F I E L D A N D SITE DESCRIPTION. Al l trials were conducted at the Agriculture and Agri-Food Canada Research Centre near Lethbridge, AB (49°42' N; 110°47' W). The soil on the site is a slightly alkaline clay loam (pH 7.0-8.0), a Dark Brown Chernozem that receives about 400 mm of precipitation annually (Appendix III). Stands of Beaver and LIRD-3 were established in adjacent 9 x 90 m plots in the spring of 1992. Plots were seeded at a rate of 11 kg/ha, fertilized with 225 kg/ha of 12-51-0 (N-P-K), sprinkler irrigated as required and harvested for hay twice during the initial growing season. Annual weeds were controlled by haying and grasses and other perennial weeds were spot sprayed with glyphosate. D I G E S T I B I L I T Y OF LIRD PROGENITORS ( E X P E R I M E N T i). The two progenitors of AC Grazeland, LIRD-3 and Beaver were fed to sheep in an experiment designed to compare the relative, apparent digestibility of a LIRD alfalfa strain against a standard. Some methods used by Kudo et al. (1985) were replicated so that parallels could be drawn between that trial and this subsequent one. Trial feeding periods- Two trials were conducted during 1993, conforming to the early hay cutting regimen used by Kudo et al. (1985). The first trial, Cut 1, had a 10 d feeding period that began June 10, which was between 28 and 35 d after growth started in the spring. The plots were cut to a 10 cm harvest height at the end of the trial period, sprinkler irrigated and allowed to regrow for 30 d before the start of the second trial. Cut 2, the second trial, had an 8 d feeding period that began July 24. The herbage in the plots was in the vegetative to mid-bud growth stages (Kudo et al. 1985) at the start of each trial. Animal husbandry and feed allowance- Twelve Suffolk wethers (wt 81.6 ±9 kg) were blocked by weight and randomly assigned to feedings of either LIRD-3 or Beaver alfalfa (6 per cultivar) for the first cross-over interval in the first trial. Animals were ruminally fistulated and fitted with a butyl rubber cannula. Al l animals were handled according to the guidelines of the Canadian Council for Animal Care (CCAC 100 1993). The wethers had been previously treated for internal and external parasites and were provided with free access to water and trace-mineral salt throughout the experiment. During each trial they were housed in individual metabolic crates (0.6 x 1 m) and between trials in individual pens (1 x 3 m) with wood shavings for bedding. Wethers were adapted to their respective diets, crates and a fecal collection bag harness for 10 d before the start of each trial. Fresh alfalfa feed was cut daily at 0800 MDT using a Swift Current flail mower set at a 10 cm cutting height. During the adaptation period for the Cut 1 trial wethers received year-old, LIRD-3 or Beaver hay ad libitum for the first 6 d and fresh cut alfalfa for the final 4 d. Fresh cut alfalfa was fed for the entire adaptation period of the Cut 2 trial. The diet in each experimental collection period consisted of fresh alfalfa (daily DM feed allowance 1.7% of body weight) fed twice daily at 0900 and 1600 MDT. Diets were crossed-over between animals by feeding 6 wethers LIRD-3 for half the trial period and Beaver for the other half. The fresh feed allowance was adjusted daily depending on the moisture content of the feed over the previous 48 h. This was the same allowance used by Kudo et al. (1985) in the LIRD-1 trial although they restricted feeding to once per day and feed was available for 2 h only. Sample collection and processing- A daily grab sample of the fresh alfalfa feed was dried in a forced air drying oven at 55 °C for 48 h to determine the moisture content and adjust the feed allowance. The sample was then placed in a covered plastic bin containing a composite of the daily feed samples for each cultivar during the collection period. Fecal collection bags were closed at the start of each trial period and emptied once daily between 0800 and 0900 MDT. The daily fecal output wet weight was recorded and a sub-sample (approximately 10% of the fresh weight) was dried at 55 °C for 48 h and composited in a bin for each wether. Composite samples of feed from each cultivar and feces from each wether for each cross-over interval were stored separately during each trial period. Thus, two composite feed samples for each cultivar and two composite fecal samples for each wether were collected during each trial period. After each trial the dried samples were ground in a Wiley mill to 1 mm. 101 DIGESTIBILITY OF AC GRAZELAND (EXPERIMENT 2). The alfalfa cultivars AC Grazeland and the standard parental cultivar, Beaver, were fed to sheep to compare their apparent digestibilities while maintaining a measured, low differential between their respective maturities and to increase the risk of bloat. Alfalfa was grown and the digestibility trial was conducted at the same site as the previous trials. AC Grazeland and Beaver had been established the previous year in adjacent 1 ha fields (described in Chapter 2). The fields were harvested for hay and flood irrigated three weeks before the start of the experiment. Trial feeding periods and maturity- The digestibility trial required fresh-cut alfalfa as the only feed, with a minimal maturity differential between cultivars. Ideally the forage needed to be immature (vegetative to late bud stages of development) (Kalu and Fick 1983) and remain in that stage long enough to complete the feeding period. Realistically the fields and cultivars could not be expected to remain in one stage of development for periods much longer than one week or in these immature stages for longer than 3 weeks. Even if some way could be found to slow the rate of maturity, the age of the plant material might be a confounding factor. Two procedures were considered to address these concerns. The first procedure was designed to manage the length of the regrowth period and the age of the stand. Four weeks after a hay harvest, seven large macro-plots (approximately 20 m2) within each cultivar/field were to be cut weekly, beginning five weeks before the start of the digestibility trial and continuing through the trial. Theoretically this series of plots would regrow into matched feed sources of each cultivar in equivalent stages of development. This procedure would compensate for a cultivar with a faster rate of maturity and correct for stand age over the 35 day feeding trial period. However, it was inflexible, requiring strict timing for at least two months. It was abandoned when the start of the trial had to be delayed and water delivery problems to parts of the field reduced the size and number of macroplots to four. The second procedure, which was used, was designed to find areas within each field with the least mature herbage as measured by the mean stage of development by stem count (MSC) index of Kalu 102 and Fick (1981). A procedure of stratified, systematic sampling, using a combination of subjective and objective techniques was adapted from forest inventory and survey methodology to reduce the variation between daily harvested samples (Mueller-Dombois and Ellenberg 1974). Seven days before the start of the feeding trial each field was subjectively surveyed for areas that contained the least mature alfalfa. (Three areas were known because they had been cut for'use in the other procedure). The largest of these areas was further examined for morphological similarities within the site. Six 0.1 m 2 microplots, placed systematically every 5 m along each of two parallel transects, 1 m apart (twelve plots per cultivar) were harvested from each field. Al l alfalfa stems in each microplot were cut to a 3-cm stubble, weeds, litter and old stems were discarded and the sample was bagged separately (Kalu and Fick 1981). Each stem from a microplot was examined and separated into a group based on its morphological stage of development following the definitions of Fick and Mueller (1989). The MSC scores for the twelve microplots in each cultivar were placed in rank order and split into two groups, those with a high or low mean MSC. The cultivar means of the low MSC groups were compared (using a t-test for differences between unpaired samples) and microplot scores were moved between the high and low groups until the test showed no significant difference or the lowest difference possible given the variation between cultivars. Generally this stratified procedure left six or more of the twelve microplots in a low mean MSC group, and identified transect sections with high MSC microplot scores. In the field these more mature sites and some of the immediately adjacent areas were flagged so that they could be avoided during daily feed collections. The procedure was repeated four times, at the start of the feeding trial and at 7 days, 18 days (the half-way point) and again at the end of the trial. Thus the stratified MSC procedure also divided the digestibility trial into four feeding periods of 7 to 10 days each (Figure 4.1) based on the availability of alfalfa in an early stage of development. Animal husbandry and feed allowance- Ten Romanov x Suffolk wethers (wt 48.6 ±2 kg) were blocked by weight and eight (plus the two spares) were randomly assigned to feedings of either AC Grazeland or 103 Beaver alfalfa (4 per cultivar). Animals were ruminally fistulated and fitted with a butyl rubber cannula. All animals were handled according to the guidelines of the Canadian Council for Animal Care (CCAC 1993). The wethers had been previously treated for internal and external parasites and were provided with free access to water and trace-mineral salt throughout the experiment. During most of the trial they were housed in individual metabolic crates (0.6 x 1 m) allowing for separate collection of feces and urine. Because of the length of the experiment (nearly 7 weeks), the animals were given scheduled exercise periods (as required by CCAC) in a communal pen (5 x 20 m) for 4 h every 7 d, and in individual pens (1 x 3 m) for 7 d every 3 wk. All wethers were weighed during each short exercise period. To continue fecal collections during the long exercise period, after 18 days the wethers were fitted with harness and open canvas fecal collection bags. On Day 21 the bags were closed and the wethers were moved to individual pens. On Day 29 the bags were removed and they were returned to the crates where they remained until Day 36, the end of the trial. Fresh alfalfa from a pre-selected area was cut daily between 0630 and 0730 MDT using a Swift Current flail mower set at a 10 cm cutting height. For the first five days of a 10 day feed adaptation period the wethers received an increasing amount (approximately 10% of daily feed allowance) of fresh alfalfa in their feed bunks in the morning and an unrestricted feeding of alfalfa cubes in an afternoon feeding. During the final five days of the adaptation period, they were given only fresh alfalfa ad libitum in the morning feeding. Uneaten alfalfa from the previous morning feeding was discarded. Animals were fed the same cultivar throughout the experiment. Over the 36 day collection period the diet was fresh alfalfa (daily feed allowance 1.7% of body weight on a dry matter basis) fed once daily at 0800 MDT. Bloat - Bloat was a prospect and expectation of Experiment 2, even though Experiment 1 and a previous digestibility trial with a LIRD cultivar (Kudo et al. 1985) had not reported any bloat. Other work had found that early morning feedings, cooler temperatures and autumn forage increase the number of bloat incidents in cattle (Hall et al. 1984; Majak and Hall 1990; Hall and Majak 1991, 1995). Feeding twice 104 per day also increases bloat incidence in cattle (Majak and Hall 1990) but this has not been shown for sheep (Colvin and Backus 1988) and no bloat was reported in Experiment 1 on two feedings per day. So wethers were fed once daily and feed was left in the bins for 24 h, in contrast to Kudo et al. (1985). Wethers were carefully monitored for signs of distension and distress. Ruminal tympanites were diagnosed and bloat was confirmed by palpating the left flank (Garry 1990b). Ruminal distensions resulting from bloat were relieved by removing the central plug in the cannula. Bloat incidence was recorded as the number of distensions in 24 h since the previous feeding. A set of rumen fluid samples was collected in the event that no bloats occurred. Rumen fluid samples were taken on four days (Day 6, 13, 28 and 35) from three sheep in each group. Collections were made immediately before the morning feeding and at 1, 2, 4, and 6 hours after feeding. Rumen fluid pH was measured immediately after collection and the samples were stored at -40°C. Sample collection and processing- Grab samples of feed (approximately 500 g wet weight) were collected prior to the daily feeding and dried at 55°C for 48 h in a forced-air drying oven to estimate the dry matter (DM) content. The values over a two day period were averaged and used to correct the amount of dry matter fed to the wethers. The fresh weights of the orts and feces from the previous 24 hours were measured before the daily feeding. Contamination of these samples from leaking ruminal cannulas and rumen fluid losses due to bloat were noted as they occurred. Daily sub-samples of the orts and feces for each wether were dried in the same manner as the feed samples. All daily samples were stored separately in labelled Whirl-pac® bags. After the experiment each sample was separately ground through a 1 mm screen using a Wiley mill. Laboratory Methods EXPERIMENT 1. The feed and fecal samples collected in Experiment 1 were divided into sets for analysis. A single analytic set consisted of (5) replicate sub-samples of each composite feed sample collected from 105 the (2) cultivars during the (2) cross-over periods in the (2) trials (n=40) plus (5) replicate sub-samples of the fecal composite sample collected for the (6) wethers during the (2) cross-over periods in the 2 trials (n=240). Sub-samples weighed approximately 2 g each. One analytic set was weighed into crucibles, dried in a forced air drying oven at 110°C for 48 h and re-weighed to determine the dry matter. Nitrogen in the feed and feces was measured in a second set of sub-samples using the Kjeldahl procedure (method 7.025; AO AC 1990). Neutral detergent fibre (NDF), acid detergent fibre (ADF), and lignin (ADL) were consecutively analysed on a third set of sub-samples using the methods of Van Soest et al. (1991). EXPERIMENT 2. An analytic set from Experiment 2 consisted of 2 g replicates (2) from the daily fecal samples (36) from each wether (8) (n=576) and the daily feed sample from each cultivar (2) (n=144). Dry matter, nitrogen, NDF and ADF were measured with the same procedures used in Experiment 1 (Van Soest and Robertson 1980). Nitrogen was measured in the feed only. The ash content was measured on the set of sub-samples used in the dry matter determination by transferring the samples immediately after the final weighing to a muffle furnace, heating them to 550°C for 5 h and re-weighing the crucibles. Statistics All feed and fecal measurements and DM intake of the wethers were converted to a DM basis and summarized by cultivar and feeding period. Calculation of the soluble cellular component (SCC) and the cell wall constituents, cellulose and hemicellulose, followed the treatment used by Van Soest (1994). The organic matter (OM) content of feed and feces of Experiment 2 was calculated by subtracting the respective ash value from the DM content. The maturity of each cultivar in each experiment was estimated by back-solving for the index mean stage of development by stem weight (MSW) on NDF content using the equations of Kalu and Fick (1983). The apparent digestibility of a feed component was calculated as the proportion lost from the total consumed during each feeding period. Daily records of any animal in Experiment 2 that had been relieved of bloat during the previous 24 hours were deleted. 106 The experimental design required that the animals be considered as replicates within treatments. In Experiment 1 the animals fed LIRD-3 were crossed over to the Beaver treatment (and vice versa) at the mid-point during each feeding period. In Experiment 2 no animals were crossed over between treatment groups (between cultivars within feeding periods) because an interaction with maturity in a feeding period would be confounding; variations in digestibility due to variations in maturity between feeding periods could be misinterpreted as an interaction between animals and the other treatment group (cultivar), or purely as an animal effect in a cross-over design. Experiment 1 was a 2 x 2 factorial design and Experiment 2 was a 2 x 4 factorial using cultivars and trial periods as the treatments or factors. The statistical model for both experiments was Y =u +C +P +C P +e (d I) which estimated the population mean (u) and the effects of cultivars (C,) (i=2), feeding periods (P,) (j—2 or 4), the interaction of cultivars within feeding periods (C/j) and residual error (e,-,) on the apparent digestibility of each feed component (Y^. Analysis of variance in the model was computed using the G L M procedure in SAS® for Experiment 1 and the MGLH Fully Factorial ANOVA statistical program in SYSTAT® (Wilkinson et al. 1992) for Experiment 2. Where effects were significant (P<.05), the means were compared using Tukey's HSD test. The frequency of bloat incidents within cultivars was compared against a random Poisson frequency distribution using a X 2 goodness of fit test (Sokal and Rolf 1995). The comparison was made to learn whether the incidence of bloat was a random phenomenon or occurred with greater frequency in wethers fed one of the cultivars. 107 4.3 RESULTS AND DISCUSSION Goplen et al. (1993) suggested that an increase in the amount of cell wall was the main result of the LIRD breeding program. Increasing amounts of cell wall are associated with reductions in digestibility and advancing maturity in forages. Two experiments were conducted to examine the apparent digestibility of LIRD cultivars in comparison to a standard. Digestibility of LIRD Progenitors (Experiment 1) The immediate progenitor of AC Grazeland, the final cultivar in the LIRD breeding program is LIRD-3. Progeny testing showed the LIRD-3 plants had consistently lower IRD than Beaver alfalfa (Goplen et al. 1993). Little is known about the dietary consequences of a low IRD. The feed composition or quality of LIRD-3 has been compared with the Beaver (limited to nitrogen and insoluble fibre content) but only in relation to the incidence of bloat (Hall et al. 1994a). A low IRD may have affected the comparative digestibility of the strain when consumed by an animal. The objective of Experiment 1 was to examine the feed composition and digestibility of LIRD-3 by feeding fresh herbage to sheep. FEED COMPOSITION. A plant's dry matter (DM) content is important because low values, especially those below 25%, have been implicated as a causal factor of bloat in cattle (Hall and Majak 1991). At any particular stage of development, D M is mostly influenced by the environment. Dry conditions will increase DM, and moist conditions will reduce it (Smith 1970). Thus the period between harvest and the most recent rainfall or irrigation will affect the amount of DM in a fresh feed sample. The D M content of the fresh feed in the first trial, Cut 1, was lower (P^.05) than the D M during the second trial, Cut 2 (Table 4.1). During Cut 1, the DM of LIRD-3 was less (P^.05) than Beaver. Consequently, the overall D M of LIRD-3 was less than Beaver for both trials (P<.05). The only other feed composition difference was NDF. The NDF content of the feed in the first 108 Table 4.1 Feed fractions of LIRD-3 and Beaver alfalfa, during two feeding periods coinciding with the seasonal maturity during a first (spring) and second (regrowth) hay cut Feed Fraction2 D M % % of D M 2 N % NDF% ADF% ADL% Cut Ii Cut 2 LIRD-3 Beaver LIRD-3 Beaver Cut 1 Cut 2 LIRD-3 Beaver SEM (pooled) 23.0a 28.2" 29.0" 30.5b 25.6a 29.8b 26.0a 29.4b 1.63 3.4 3.8 3.5 3.6 39.4a 43.6a'b 42.8a-b 44.1" Between Periods 29.4 31.3 30.6 29.8 3.6 3.6 41.5a 43.5b Between Cultivars 30.4 30.2 3.5 3.7 0.07 41. l a 43.9b 1.06 30.0 30.6 0.42 5.0 5.2 5.8 5.6 5.1 5.7 5.4 5.4 0.18 2 D M % = dry matter percentage as fed; all other fractions are percentage of DM. N = Nitrogen ; NDF = Neutral Detergent Fibre; ADF = Acid Detergent Fibre; ADL = Acid Detergent Lignin. y Feeding periods: Cut 1 = 10 days starting June 10; Cut 2 = 8 days starting July 24. "•b Means, in the same column under the same heading, followed by the same letter or no letter are not different (P>.05). For each period or cultivar, n=20. trial was lower (P^.05) than in the second, and overall, the NDF in LIRD-3 was lower (P<.05) than Beaver (Table 4.1). This fraction is commonly used as an analog of the amount of cell wall since the extraction process recovers most of the major cell wall components, including hemicellulose, cellulose and lignin (Van Soest 1994). The proportion of NDF in alfalfa increases predictably (R2>.95) as the plant matures (Kalu and Fick 1983). This predictability made it possible to estimate the stage of development of the two cultivars in each trial by back-solving for MSW from NDF using the quadratic equation developed by the Kalu and Fick (1983). Back-solving for each cultivar's stage of development confirmed 109 Table 4.2 Calculated feed fractions of LIRD-3 and Beaver alfalfa, during two feeding periods coinciding with the seasonal maturity during the first and second hay cuts Feed Fractionz SCC% Hemicellulose % Cellulose % Cutn LIRD-3 60.6 10.0 24.4 Beaver 56.4 12.3 26.1 Cut 2 LIRD-3 57.2 12.2 24.8 Beaver 55.9 14.3 24.2 Between Periods Cut 1 58.5 11.2 25.3 Cut 2 56.6 13.2 24.5 Between Cultivars LIRD-3 58.9 11.1 24.6 Beaver 56.2 13.3 25.2 2 A l l fractions are a percentage of DM. OM= Organic Matter= 100 - Ash %; SCC= Soluble cellular component 100 - NDF%; Hemicellulose= NDF% - ADF%; Cellulose= ADF% - ADL% (Van Soest 1994). y Feeding periods: Cut 1 = 10 days starting June 10; Cut 2 = 8 days starting July 24. a ' b Means, in the same column under the same heading, followed by the same letter or no letter are not different (P>.05). For each period or cultivar, n=20. that LIRD-3 was less mature than Beaver in Cut 1 and at an equivalent maturity to Beaver in Cut 2. During Cut 1, the estimated maturity of LIRD-3 was an early bud stage of development (MSW = 3.0) while that of Beaver was a late bud stage (MSW = 4.1). During the second trial, Cut 2, the cultivars were in similar late bud stages of development (MSW = 3.9 and 4.3, for LIRD-3 and Beaver, respectively). Environmental variations where the plots were established could account for the differences. The site is on a south-facing slope next to a large grove of trees. The soil is saline (e. c. > 8 mmhos/cm) and the LIRD-3 plot was partially shaded for 2 h each morning (R. MacKenzie, pers com. 1995). The other feed fractions, N , ADF and ADL did not differ between the two feeding periods or between the cultivars (Table 4.1). 110 Table 4.3 Apparent two feeding periods digestibility of LIRD-3 and Beaver alfalfa feed fractions in wethers, during coinciding with the seasonal maturity of the first and second hay cuts Apparent Feed Fraction Digestibility2 D M % NDF% ADF% ADL% Cut 1 y Cut 2 LIRD-3 Beaver LIRD-3 Beaver S E M X pw Cut 1 Cut 2 SEM P LIRD-3 Beaver SEM P 60.7a 64.2" 58.7a 66.2b 0.897 0.043 62.5 62.5 0.634 0.983 59.7a 65.2" 0.634 0.000 36.8 47.4 37.3 49.4 1.771 0.666 38.3 47.7 34.3 47.2 1.561 0.278 Between Periods 42.1 43.3 1.252 0.501 43.0 40.7 1.104 0.171 Between Cultivars 37.1a 48.4b 1.252 0.000 36.3a 47.5" 1.104 0.000 6.5 4.1 4.9 6.2 1.111 0.121 5.3 5.5 0.786 0.862 5.7 5.2 0.786 0.641 z Al l fractions are on a D M basis. DM%= Dry matter digestibility; NDF = Neutral Detergent Fibre; ADF = Acid Detergent Fibre; ADL= Acid Detergent Lignin. y Feeding periods: Cut 1 = 10 days starting June 10; Cut 2 = 8 days starting July 24. x Standard Error of the Mean w Probability of a difference (P<.05) between groups for that feed fraction. a , b Means, in the same column under the same heading, followed by the same letter or no letter are not different (P>.05). For each period or cultivar n=20. The OM, SCC, hemicellulose and cellulose values (Table 4.2) were derived from the values for ash, DM, NDF, ADF and ADL in Table 4.1. They provide another perspective on the feed components and cell wall composition although their statistical differences are the same as those of the parameters 111 Table 4 .4 Apparent digestibility of calculated LIRD-3 and Beaver alfalfa feed fractions in wethers, during two feeding periods coinciding with the seasonal maturity of the first and second hay cuts Feed Fractionz OM% SCC% Hemicellulose % Cellulose % Cutl* LIRD-3 67.8 76.2 32.4 44.8 Beaver 68.8 77.2 46.6 56.4 Cut 2 LIRD-3 65.8 74.7 44.8 41.2 Beaver 70.1 79.5 54.0 56.7 Between Periods Cut 1 68.3 77.0 39.7 50.6 Cut 2 67.9 77.3 49.2 48.9 Between Cultivars LIRD-3 66.8 75.5 39.3 43.0 Beaver 69.4 78.3 50.5 56.6 z A l l fractions are on a D M basis. OM= Organic Matter= 100 - Ash %; SCC= Soluble cellular components= 100 - NDF%; Hemicellulose= NDF% - ADF%; Cellulose= ADF% - ADL% (Van Soest 1994). y Feeding periods: Cut 1 = 10 days starting June 10; Cut 2 = 8 days starting July 24. For each period or cultivar, n=20. from which they were derived. For example, SCC is the material lost during the extraction of NDF. It comprises much of the soluble intra-cellular material as well as pectins, which are a cell wall component not recovered during the extraction of NDF. FEED DIGESTIBILITY. The apparent digestibility of LIRD-3 alfalfa D M was less (P<.05) than that of Beaver (Table 4.3). The cell wall digestibility reflected the pattern evident in D M digestibility, that is overall, NDF and ADF of LIRD-3 were less digestible (P<.05) than Beaver (Table 4.3) although the differences were not significant within each trial. The digestibility of lignin was low and did not differ (P>.05) between cultivars. 112 The derived feed fractions, OM, hemicellulose and cellulose, reflected the digestibility differences of their originating values, differing between cultivars (P<.05), but not between feeding periods or in the interaction between feeding periods and cultivars (Table 4.4). Digestibility of SCC is a compound mathematical derivation of D M digestibility and NDF digestibility and reflects the variation and interaction between these variables. Comparing the SCC digestibility between the two cultivars while considering the predicted stages of development in each trial was interesting and instructional. The digestibility of the SCC was similar (P>.05) between the two cultivars when Beaver was more mature than LIRD-3 (in Cut 1) and dissimilar (LIRD-3 less, P<.05) when they were at the similar stages of development. Goplen et al. (1993) suggested that the main result of the LIRD selection pressure was an increase in leaf cell wall thickness and implied that the amount of cell wall was greater in a LIRD cultivar. When the proportional cell wall fractions from mixed leaf and stem herbage of LIRD-3 were compared with Beaver, none were greater in the LIRD cultivar (Table 4.1, 4.2). This negation introduces a contentious argument: if a LIRD-3 cell wall contains less material and its cell wall thickness is greater (Goplen et al. 1993) then it must occupy more volume. Further, since the digestibility of the LIRD-3 cell wall (namely the NDF, ADF and ADL feed fractions, and their derivatives, hemicellulose and cellulose) was similar to or less than the standard, the LIRD-3 cell walls must be more resistant to microbial activity. So in summary the tenets of this hypothesis are that not only is a LIRD-3 cell wall thicker, its volume is occupied by less material that is a greater barrier to microbial invasion. Support for this theory may be found in an evident biochemical change, perhaps an increase in the less digestible materials within the cell wall. However the evidence shows otherwise, there is no increase in hemicellulose or lignin, the least digestible components of the cell wall, in LIRD-3. Another explanation may be that the pattern of construction of the LIRD cell wall has changed, creating a thatch or interleaved arrangement between adjacent cell walls, similar to those patterns found in grasses, that resist microbial adhesion. This hypothesis was not investigated but evidence of a different cell wall would have been noticed and 113 reported during the cell wall thickness measurements conducted by Goplen et al.(1993). There is another highly plausible explanation of the observations: that the greater cell wall thickness in the LIRD cultivars, as measured by Goplen et al. (1993), was actually a consequence of a difference in maturity. This suggests that cell walls of the more rapidly maturing LIRD cultivars would be thicker than the standard when both were cut at the same time. The implications are that cell walls of the cultivars at equivalent stages of development, or at an earlier stage for the LIRD cultivar, are of similar thickness, occupy the same volume and are equal in amount, and therefore are equivalent barriers to microbial invasion and digestion. This explanation rescinds the previous hypothesis and replaces it with another, that LIRD-3 is an early maturing strain of alfalfa with a normal, though rapidly developing, cell wall. Howarth et al (1982) assumed that a low rate of cell rupture would reduce the IRD but probably would not affect overall digestibility. The difference in digestibility in the LIRD-3 cultivar in Experiment 1 shows that this assumption is not correct. The digestibility of most LIRD-3 feed fractions was the same or lower than the standard cultivar (Tables 4.3, 4.4). Thus, in comparison, the overall digestibility of NDF in LIRD-3 was less than that of Beaver and the NDF content was lower. Digestibility of AC Grazeland (Experiment 2) AC Grazeland was the final alfalfa strain selected in the LIRD breeding program. Goplen et al. (1993) found the results of comparative trials for dry matter disappearance equivocal and suggested that maturity was a factor affecting the expression of the LIRD character. Hall et al. (1994a) also suggested that the cultivar's effect on bloat incidence in cattle declined with increasing maturity. Differences in maturity had occurred in previous digestibility trials (LIRD-1, Kudo et al. 1985; LIRD-3, in Experiment 1 of the present study) but with no bloat reported, its effect on bloat incidence could not be deduced. The plan for Experiment 2 was to compare the feed composition and digestibility of AC Grazeland against its parental standard, Beaver, at an early stage of development and a high potential for bloat. 114 ALFALFA MATURITY. The stratified, systematic sampling procedure was relatively rapid and initially assisted in delineating areas with equivalent maturities for harvest and feeding. However, the MSC index became increasingly unreliable forjudging stand maturity as the feeding trial progressed. Average MSC did not relate well to the maturity or age of the stand after the regrowth period exceeded 35 days. Cooler temperatures and shorter day lengths in late August and September may have delayed the onset of flowering in both cultivars and caused MSC values to decline. Static or declining MSC values after 5 weeks of regrowth was originally identified by Kalu and Fick (1981). Alfalfa plants in older canopies initiate new stem buds as they mature, resulting in increased numbers of stems in the early developmental stages. These new stems skewed the morphology-based MSC index downwards. Kalu and Fick (1981) resolved the problem by using an index based on weight, MSW (mean stage by stem weight), because the contribution of the new immature stems to the total dry matter was low. The MSW index was considered for this trial but rejected because the drying interval (48 h) was too long. In spite of its limitations, the MSC values were still useful. Dividing the MSC scores into sub-groups based on high and low means helped identify the mature sites to be avoided during subsequent harvesting. The means of these sub-groups also diverged creating a wide maturity difference between AC Grazeland and Beaver (Figure 4.1). Beaver's high mean MSC sub-group declined after only 35 days of regrowth while the same sub-group for AC Grazeland continued to rise. The difference was likely a result of the response of each cultivar to the light frosts that occurred on several mornings during the 2 n d feeding period rather than a rapid development of new stems. The flowers, buds and stem tips damaged by frost dried quickly and were abraded off by the wind. Their loss meant that other vegetative characteristics associated with maturity, such as elongation of the nodes and branching from leaf axils, had to be used to identify the stage of development of a stem. The damage seemed greater in the Beaver field. Although both cultivars were selected from cold hardy stock, alfalfa cultivars vary in their tolerance to cold. Beaver, the older variety (released in the late 1950's) is not as frost tolerant as newer strains (Fairey and Fairey 1993). 115 Digest ib i l i ty Tr ial P e r i o d s (days) 6 9 12 15 18 21 24 27 30 33 36 + -t- + +• + -+-O CO 3 (0 3.5 4-3 4, 2.5 4-42 2 1.5 4-H i 21 • AC Grazeland, High Mean & SE • AC Grazeland, Low Mean & SE • Beaver, High Mean & SE o Beaver, Low Mean & SE -+- •+- H — 56 28 35 42 49 Alfalfa Regrowth (days after harvest) Figure 4.1 High and low mean maturities (Mean Stage Count, MSC) in regrowth of two alfalfa cultivars (cv. AC Grazeland, Beaver) during four consecutive digestibility trial feeding periods Examination of the MSC scores at the start of the 3 rd feeding period showed a wide separation had occurred between mature and immature sites and between cultivars (Figure 4.1). This difference was expected although it was early, coming after only 35 days of regrowth. No attempt was made to match maturities between cultivars; the only requirement of the digestibility trial was to harvest the least mature forage from each field consistently. In hindsight this strategy was appropriate because it helped reduce the variation in maturity of the feed. The low mean MSC scores differed by values of less than 1.0 while the high mean MSC scores differed by values exceeding 1.5. FEED COMPOSITION. The DM increased from Period 1 to Period 4 during the experiment, most likely because of increasing soil moisture deficits and the time from the last irrigation (Table 4.5). Dry matter 116 content is the proportion of the total weight of the fresh plant remaining after the water and a few volatile chemicals have evaporated. Periods of moisture deficits reduce the moisture in the plant tissues and concentrate DM. The fields received no rain and were not irrigated during the experiment. The amount of D M may also be evidence of maturity and age. Older, mature herbage is less succulent because it contains a greater proportion of structural material (Smith 1970). The range of morphological stages in the forage harvested from the AC Grazeland stand was wider and there were more samples collected from later stages of development than Beaver (Figure 4.1). Consequently, maturity was likely responsible for the higher D M content in AC Grazeland throughout the trial (P<.05). The amount of cell wall, estimated by the amount of NDF (Table 4.5), did not differ between cultivars in any period. Both cultivars had more cell wall (P<.05) in the late compared with the early feeding periods. Predicted stages of development for Periods 1, 2, 3, and 4, were MSW scores of 1.1, 1.0, 1.6 and 1.5 respectively (using the mean NDF values for each period and the equations of Kalu and Fick 1983). These results contrast with those of Goplen et al. (1993) who reported that leaf cell walls of all LIRD synthetic cultivars were thicker than the Beaver standard and that the differences occurred in the early stages of vegetative growth. In fairness, Goplen et al. (1993) had measured thickness using electro-micrographs of leaflet cell walls and the increase may not have been sufficient to raise the total amount of cell wall. They suggested that the reduction in IRD in the LIRD cultivars was related to the increasing thickness of cell walls in each selection cycle. However their data actually show that the IRD did not correlate well with cell wall thickness; the relationship varied inconsistently between selection cycles and between years. The greatest increase in cell wall appeared early in the selection process, in the Cycle 1 and 2 synthetic LIRD cultivars. The abaxial epidermal and mesophyll cell walls increased only for the first two cycles of selection then declined while the adaxial cell wall continued to increase in size through all four cycles. The IRD of the Cycle 4 LIRD synthetic, AC Grazeland, was 'unusually high' (that is not low compared with Beaver), and yet the total thickness of the epidermal cell walls was the highest reported for the 117 Table 4.5 Feed fractions of AC Grazeland and Beaver alfalfa during four consecutive feeding periods beginning 26 days after a hay cut Feed Fractionz % of D M 2 D M % N % NDF% ADF% Mean SE ; Mean SE Mean SE Mean SE Period 1y AC Grazeland 17.0 0.870 3.48a 0.132 29.3 0.992 25.9 0.874 Beaver 15.3 0.870 3.75a 0.147 27.8 0.992 25.2 0.874 Period 2 AC Grazeland 16.3 0.872 3.52a 0.120 28.5 0.990 25.3 0.870 Beaver 15.9 0.872 3.38a 0.120 28.2 0.990 25.7 0.870 Period 3 AC Grazeland 20.9 0.696 3.10a 0.104 32.1 0.790 27.8 0.700 Beaver 19.2 0.696 3.23a 0.098 32.2 0.790 28.3 0.700 Period 4 AC Grazeland 23.8 0.760 2.58b 0.120 32.3 0.870 28.8 0.770 Beaver 20.7 0.760 3.27a 0.145 30.8 0.870 26.9 0.770 pw 0.45 0.02 0.70 0.38 Between Periods Period 1 16.2a 0.617 3.61a 0.099 28.6a 0.700 25.6a 0.620 Period 2 16.1a 0.617 3.45a-b 0.085 28.3a 0.700 25.5a 0.620 Period 3 20.1b 0.492 3.16bc 0.072 32.1" 0.560 28.0b 0.490 Period 4 22.2C 0.544 2.93° 0.095 31.5b 0.620 27.9" 0.550 P 0.00 0.00 0.00 0.00 Between Cultivars AC Grazeland 19.5" 0.403 3.17a 0.060 30.5 0.460 27.0 0.400 Beaver 17.8" 0.403 3.41b 0.065 29.8 0.460 26.5 0.400 P 0.00 0.01 0.23 0.43 z D M % = dry matter percentage of feed as fed; other fractions are percentage of DM: N= Nitrogen; NDF= Neutral Detergent Fibre; ADF = Acid Detergent Fibre. SE = Standard Error of Mean. y Feeding periods begin 26 days after hay cutting. Period 1= 7 days (ends on Day 33); Period 2= next 7 days (to Day 40); Period 3= next 11 days (to Day 51); Period 4= last 9 days (to Day 60). a , b Means, in the same column under the same heading, followed by the same letter or no letter are not different (P>.05). For Periods 1 to 4, n=14, 14, 22, and 18, respectively; n= 34 for each cultivar. w Probability of a difference (P^.05) between groups for that feed fraction 118 whole breeding program (Goplen et al. 1993). In other words thick cell walls did not reduce the IRD of AC Grazeland. A subsequent test of the Cycle 4 progeny corroborated this inconsistency but the investigators discounted it on the basis that sampling had occurred in wet and cool conditions at a 'later bud stage'. During the tests on the Cycle 3 progeny, thickness measurements were taken on epidermal and mesophyll cell walls at three different stages of development, defined by Goplen et al. (1993) as pre-bud, bud and flowering. They omitted the data summary from their report because the only difference they had observed was between the Beaver and Cycle 3 cultivars in the cell walls of the pre-bud tissues. This observation was important because it led them to suggest that the LIRD trait was only expressed at the early stages of vegetative growth. On the other hand, if the LIRD trait was actually an early maturity characteristic, it would be most prevalent in comparisons between slower maturing strains at their early stages. The difference between the two strains would decline whenever the slow developmental strain was given an opportunity to mature. Delayed harvesting because the fields were too wet, a hot dry spell during the early period of growth, or a late season harvest when light and temperature slow the rates of growth and maturity (typical of 3 rd cuts in northern Saskatchewan) would reduce the differences between the two strains. Table 4.5 shows that, if the maturity differential is minimal, as it was in each trial period of the present study, the amount of cell wall (estimated by NDF) in the LIRD cultivar, AC Grazeland is similar to the standard, Beaver. When both cultivars are in the same stage of development, whether an early or late stage, the amount of cell wall is also similar between cultivars. The only difference occurs when the two cultivars are in different stages of development a result which supports the theory that selection for a low IRD was selection for a different rate of development. These results raise one other concern. If a LIRD alfalfa cultivar is not just an early maturing strain, then the barrier to microbial digestion responsible for reducing the IRD must be in the amount of 119 Table 4.6 Calculated feed fractions of AC Grazeland and Beaver alfalfa during four consecutive feeding periods beginning 26 days after a hay cut Feed Fractionz OM% SCC% Hemicellulose % Period 1y AC Grazeland Beaver Period 2 AC Grazeland Beaver Period 3 AC Grazeland Beaver Period 4 AC Grazeland Beaver Period 1 Period 2 Period 3 Period 4 AC Grazeland Beaver 85.1 84.7 85.4 85.0 86.7 86.5 86.5 86.6 84.9 85.2 86.6 86.5 85.9 85.7 70.7 72.2 71.5 71.8 67.9 67.8 67.7 69.2 Between Periods 71.5 71.7 67.9 68.5 Between Cultivars 69.5 70.3 3.3 2.7 3.2 2.5 4.3 4.0 3.5 3.8 3.0 2.8 4.1 3.7 3.6 3.2 z Al l fractions are a percentage of DM. OM= Organic Matter= 100 - Ash %; SCC= Soluble cellular components= 100 - NDF%; Hemicellulose= NDF% - ADF%. y Feeding periods begin 26 days after a hay cut. Period 1= 7 days (ends on Day 33); Period 2= next 7 days (to Day 40); Period 3= next 11 days (to Day 51); Period 4= last 9 days (to Day 60). For each period, n=14, 14, 22, and 18, respectively; for each cultivar, n= 34. lignified, indigestible tissue. ADF is a measure of the insoluble cellulose, lignin and cutin in the cell wall. There were no differences, except those associated with maturity, between the ADF fractions of AC Grazeland and Beaver (Table 4.5). The feed fractions OM, SCC, and hemicellulose, derived from the analytic information, reflect 120 the lack of differences found in those parameters and corroborate those attributable to maturity (Table 4.6). For example, during the first two feeding periods the OM content of the two cultivars was similar (P>.05) but lower (P<.05) than in the 3 rd and 4 th feeding periods. Alfalfa tissues have a greater OM content in the stems than the leaves and the differences increase as the tissues mature (Smith 1964, 1970). Maturity was apparently controlled by the selective harvest procedure and effectively limited the cultivars and the forage fed to sheep to only two distinguishable stages of development. These stages of development could be simply described as early vegetative (Periods 1 and 2) and late vegetative (Periods 3 and 4). The proportion of SCC contained in these immature leaf tissues was greater than those found in the more mature early bud and late bud tissues in Experiment 1 (Table 4.2). Similarly hemicellulose increased with maturity, but varied little between cultivars with the same maturities (Table 4.2, 4.6). FEED DIGESTIBILITY. The harvesting procedure was designed to select the least mature forage available. In effect this placed an upper limit on the diversity of morphological stages that would be collected in the daily cuttings from each cultivar. This also reduced the maturity differential between cultivars in any particular period. Any digestibility difference would measure the difference between cultivars with similar (or the least variable) morphologies. The apparent D M digestibility declined from Period 1 to Period 4 (P^.05) as the plants matured. However, over the whole trial and within each trial period, there were no differences in D M digestibility between the two cultivars (P>.05) (Table 4.7). Evidently the maturity differential within each feeding period between the cultivars was low enough that its effect was minimized. The apparent digestibility of the cell walls in the LIRD cultivar was greater (P<.05) than the standard cultivar. This result was not expected. The digestibility of NDF and ADF declined for both cultivars during the trial although the differences were minor (P>.05) within each period. So the difference between cultivars probably reflects the relatively high NDF digestibility in AC Grazeland during Period 1 and the relatively low NDF digestibility in the Beaver forage in Period 4 (Table 4.7). 121 Table 4.7 Apparent digestibility of fresh feed fractions of AC Grazeland and Beaver alfalfa when fed to wethers during four consecutive feeding periods Feed Fraction Digestibilityz D M % NDF% ADF% Period 1y AC Grazeland 72.5 53.6 55.2 Beaver 70.0 45.6 50.2 Period 2 AC Grazeland 69.6 50.4 52.4 Beaver 69.0 48.5 52.5 Period 3 AC Grazeland 67.1 46.2 48.5 Beaver 67.0 44.9 48.4 Period 4 AC Grazeland 66.6 43.3 47.8 Beaver 66.2 39.0 43.4 SEM* 0.916 2.197 2.146 p w 0.57 0.43 0.50 Between Periods Period 1 71.3a 49.6a 52.7a Period 2 69.3ab 49.4ab 52.5ab Period 3 67.0bc 45.5abc 48.5abc Period 4 66.4C 41. l c 45.6C SEM 0.648 1.554 1.518 P 0.00 0.00 0.00 Between Cultivars AC Grazeland 68.9 48.4a 51.0 Beaver 68.0 44.5b 48.6 SEM 0.458 1.099 1.073 P 0.17 0.02 0.14 Al l fractions are on a D M basis. DM%= Dry Matter digestibility; NDF= Neutral Detergent Fibre; ADF= Acid Detergent Fibre y Consecutive feeding periods: Period 1= 7 d; Period 2= 7 d; Period 3= 11 d; Period 4= 9 d. * Standard Error of the Pooled Mean. w Probability of a significant difference (P< .05) between groups for that feed fraction. a ' b Means, in the same column under the same heading, followed by the same or no letter are not different (P^.05).For Periods 1 to 4, n=14, 14, 22, and 18, respectively; for each cultivar, n=34. 122 T a b l e 4 . 8 A p p a r e n t d i g e s t i b i l i t y o f c a l c u l a t e d feed f r a c t i o n s o f A C G r a z e l a n d a n d B e a v e r a l f a l f a w h e n f e d to w e t h e r s d u r i n g f o u r c o n s e c u t i v e f e e d i n g p e r i o d s Feed Fraction Digestibilityz O M % SCC % Hemicellulose % Period 1 y AC Grazeland Beaver Period 2 AC Grazeland Beaver Period 3 AC Grazeland Beaver Period 4 AC Grazeland Beaver Period 1 Period 2 Period 3 Period 4 AC Grazeland Beaver 85.3 82.6 81.4 80.9 77.3 77.5 77.0 76.4 84.0 81.2 77.4 76.7 80.3 79.4 80.2 78.8 77.2 76.9 76.9 77.4 77.5 78.5 Between Periods 79.5 77.0 77.2 78.0 Between Cultivars 41.5 -7.4 33.4 13.6 30.4 19.5 7.0 6.3 17.1 23.5 24.9 6.6 77.9 77.9 28.1 8.0 2 All fractions are on a D M basis. OM%= Organic Matter= 100- Ash%; SCC = Soluble cellular components= 100 - NDF%; Hemicellulose %= NDF% - ADF% (Van Soest 1994). y Consecutive feeding periods: Period 1 = 7 d; Period 2 = 7 d; Period 3 = 11 d; Period 4 = 9 d; n=14, 14, 22, and 18, respectively; overall n=68. The difference may be a cumulative one that, although significant statistically, is not a major biological effect. However, if the LIRD cultivar is early maturing, it may lay down cellulose fibrils and pectic substances quickly, perhaps in a looser matrix (Terashima et al. 1993). With little lignification, these tissues would allow cellulolytic microbes more opportunity to degrade the walls (Dehority 1993). 123 The digestibility of fresh alfalfa OM decreased (P<.05) between the early vegetative and late vegetative stages of development (Table 4.8). However there were no differences (P>.05) between the two cultivars within any feeding period. Decreasing OM digestibility is associated with increasing amounts of fibre or cell wall in maturing forage (Kalu and Fick 1983; Tamminga et al. 1993; Van Soest 1993). The digestibility of SCC was high, more than 75%, and similar (P>.05) between periods and cultivars. It was also similar to the SCC digestibility reported in Experiment 1, in more late bud alfalfa, perhaps suggesting that, once liberated from the surrounding walls, the intra-cellular materials are digested by rumen microbes with some degree of constancy. The variation in hemicellulose digestibility was high, reflecting the relatively high variation in NDF and ADF digestibility (Table 4.7). Hemicellulose digestibility also appeared to be affected by the high number of bloat incidents in Period 1. Losses of rumen fluid from bloated animals occurred when the rumen cannula was opened to relieve the pressure and distension. The proportional losses would be greater from feed fractions with lower digestibility because these fractions would accumulate in the rumen and be retained for longer periods. BLOAT INCIDENCE. Wethers bloated 23 times during the trial. The number of bloats between animals fed each cultivar did not differ (P>.05) (11 bloats on Beaver and 12 on AC Grazeland). However, the distribution of bloats was considerably more frequent (P<.05) in Period 1 (12 bloats) than in Period 2 (5 bloats), Period 3 (2 bloats) or Period 4 (4 bloats). The frequency distribution of bloats was not random (Table 4.9). Overall incidents of one bloat per day occurred less frequently than expected, while incidents of 2 or more bloats per day occurred with greater frequency. The frequency of bloat in animals fed the standard cultivar, Beaver, was clustered. In other words, if one animal bloated there was a greater than expected probability that other animals would also bloat at the same time. This tendency was not evident in the pattern of bloats of the AC Grazeland fed animals. Bloat frequency in these animals was randomly distributed. 124 Table 4.9 Observed and expected (Poisson) frequency distributions of bloat in sheep fed two alfalfa cultivars (cv. A C Grazeland and Beaver) AC Grazeland Beaver Overall Frequencyz Observedy Expected Observed Expected Observed Expected 0 133 132.49 136 133.41 270 265.89 1 10 11.04 5 10.19 15 21.23 2 1 0.46 3 0.39 2 0.85 3 0 0.00 0 0.00 0 0.02 4 0 0.00 0 0.00 1 0.00 5 0 0.00 0 0.00 0 0.00 Statistic Mean 0.083 0.083 0.076 0.076 0.080 0.080 SE" 0.025 0.024 0.028 0.023 0.021 0.017 CD W 1.090 1.007 1.479 1.006 1.621 1.003 X 2 NS ** *** z Bloats / day of feeding y Cumulative days of feeding; for AC Grazeland and Beaver, n=144; Overall n=288 " Standard error of the mean w Coefficient of dispersion = s2/mean. CD=1 indicates random distribution of incidents, CD>1 indicates clumping, CD<1 indicates dispersion (Sokal and Rohlf 1995). X 2 Goodness of Fit test, ** significant at 0.01 probability;*** = 0.001 probability level A clustered frequency distribution means that a unique set of conditions is required to foster the expression of a phenomenon. Generally, if an event is rare but occurs at random then the probability of a single occurrence is low. The expectation of several events occurring at one time is therefore very low. When a rare phenomenon exhibits clustering at higher than expected frequencies it suggests that a unique set of conditions have developed or a threshold has been breached. For bloat, the animal's normal behavioural and physiological mechanisms compensate for conditions that otherwise might destabilize the rumen system. The occurrence of a single random bloat is therefore much less frequent than expected because the animal's coping mechanisms help maintain an equilibrium, even a state of sub-acute bloat. When conditions destabilize the system, the frequency of bloat rises well above the numbers expected if it was just a random event. 125 4.4 C O N C L U S I O N S The amount of cell wall in the standard alfalfa cultivar, Beaver, varied considerably within the relatively narrow range of maturity between the two experiments. The NDF content in Beaver increased between early vegetative and late bud stages from a mean of 29.8% to 43.9%. The amount of NDF in two related LIRD cultivars was respectively, 30.5% and 41.1%. At the same time, the digestibility of the cell wall in Beaver increased, from 44.5 % to 48.4% between early vegetative and late bud stages. The mean digestibilities of the two LIRD cultivars were 48.4% and 37.1% respectively. Apparently the digestibility of cell walls declines at a faster rate in the LIRD cultivars than in the standard. The relative decline in digestibility that followed a slight increase in maturity in AC Grazeland during Experiment 2 supports this conclusion. The digestibility of AC Grazeland cell walls declined from a mean of 52.0 % during the first half of the experiment (early vegetative) to 44.8% in the second half (late vegetative). The mean NDF digestibility of Beaver declined from 47.0% to 41.5%, for the same respective periods. Clearly fibre assays from whole plant feeding trials show that something is happening to the cell walls of the LIRD cultivar as it matures that is different from the standard alfalfa. Perhaps it is a thickening process, although this implies that the amount of cell wall has increased, for which there was little evidence in these experiments. The cell rupture hypothesis as a causal theory of bloat is supported by these results. The decline in bloats between early and late vegetative alfalfa was accompanied by a decline in D M and OM digestibility, an increase in cell wall content and a decline in cell wall digestibility. The greatest number of bloats occurred when early vegetative alfalfa was fed, fewer bloats occurred during the feeding of late vegetative alfalfa and no bloats occurred when the alfalfa was in either early bud or late bud stages of development. Thus an increase in the barrier to microbial degradation, associated with maturity of the feed was accompanied by a drop in bloat incidents. The results suggest that the animals could maintain a relatively stable digestive environment if the cells did not break apart too easily. 126 The greatest number of bloats occurred during the first two feeding periods of Experiment 2. A cluster with the highly unusual frequency of four bloats per day was recorded during Period 1. Coincidentally the digestibility of the OM, NDF and ADF feed fractions was the highest in the trial during these two periods. Since the wethers had no other obvious sources of stress that might have induced bloat it appears that feed digestibility was the malefactor in this study. The range of digestibility in this trial was relatively narrow, less than 10 percentage points separated the highest and the lowest mean OM digestibility over a 36 day period. The transition point between normal digestion with an occasional bloat incident and serious bloat outbreaks appears to be an apparent OM digestibility above 80%. Similar thresholds could be suggested for the other feed components. For example, a forage cell wall content (NDF value) below 29% or a NDF digestibility above 49% was also associated with bloat outbreaks. Fresh alfalfa dry matter below 25% may be another threshold, perhaps one that is easier to evaluate than digestibility under field conditions. 127 C h a p t e r 5 B l o a t I n c i d e n c e i n G r a z i n g S t e e r s 5.1 INTRODUCTION Bloat in Feeding Trials Some legumes that do not cause bloat produce less gas and have relatively lower rates of in vivo dry matter disappearance (DMD) in comparison to alfalfa (Medicago sativa L) (Howarth et al. 1979; Fay et al. 1980). To determine whether DMD was correlated with bloat in alfalfa, a series of preliminary trials was conducted using ruminally carmulated cattle fed diets of either fresh cut alfalfa or alfalfa hay, once each day. The in vivo DMD of fresh alfalfa was greater (P<.05) on days when one or more bloat incidents were reported (DMD—65.7%) than when there were no bloats (DMD=60.2%) (Hall et al. 1994a). These results supported the theory that the rate of tissue disruption was a cause of bloat, and further, that alfalfa strains with a low rate of DMD could be less bloat-inducing. Breeding an alfalfa with a low initial rate of digestion (LIRD; defined as a low DMD during 4 h nylon bag incubations, or 4hNbDMD) was an arduous process that took more than 10 years (Howarth et al. 1979; Goplen et al. 1993). Except for the preliminary feeding studies, no other test on bloat incidence was conducted until sufficient seed of the strain from the 3 rd breeding cycle, LIRD-3, was available for small field plantings. Irrigated stands were established at Kamloops, BC and three feeding trials were conducted to measure the effect of the LIRD-3 cultivar on bloat incidence (Hall et al. 1994a). Comparisons were made against one of the LIRD cultivar's parental strains, the standard, Beaver. Bloat incidents occurred on 46.5% of 372 animal-days on feed, an incidence that would be considered catastrophic in a normal ranching enterprise. Overall, the incidence of bloat was 20% lower in cattle fed the LIRD-3 cultivar but this was an equivocal result because LIRD-3 was associated with low bloat incidence in only one trial. 128 Bloat in Grazing Trials of LIRD Alfalfa The first grazing trials to evaluate the effect of any LIRD alfalfa strain on bloat were conducted on the small LIRD-3 fields at Kamloops (Hall et al. 1994a). Occasionally during bloat studies, the number of distensions or their severity is insufficient to show treatment differences (Ashford and Heinrichs 1967; Wolf and Lazenby 1972) so a restricted grazing regimen was used to mimic the feed availability of an experiment using penned animals. Ruminally fistulated steers were limited to 6 h grazing periods and fasted for the remainder of the day. Bloat occurred on 60.6% of the 330 animal-days of grazing, an incidence much greater than in the previous feedlot trials. Under these circumstances, the overall bloat incidence was only 2% less in cattle grazing the LIRD-3 cultivar and the LIRD cultivar did not affect bloat incidence in any particular trial. Creating the appropriate bloat inducing conditions to make meaningful comparisons is a difficult proposition. Most known bloat management techniques are directed toward reducing the potential for bloat. Sustaining a high level of bloat incidents is a self-taught skill that engenders no respect from members of the ranching community even if it is necessary for a statistical test. There are no guides on how much or how little bloat is needed or what forage or animal factors create more or less bloat. The restricted grazing and feedlot trials used to measure the effect of LIRD-3 alfalfa on bloat are an example of this problem. Restricted feeding regimens encourage binge feeding behaviour in animals. Although previously discounted as a factor in bloat (Johns 1954; Howarth 1975) binge feeding followed by overnight fasting increases bloat susceptibility in individual animals. Majak et al. (1983) called this the animals' predisposition to bloat. To increase the potential for bloat in experiments using penned animals, alfalfa must be harvested at an early stage of development or cut high to reduce the amount of stem in the feed. Early or 'high cut' harvests create another set of problems, including one of having insufficient feed to complete the trial. As a compromise, some less digestible plant material is fed which reduces the bloat challenge (that is, the probability that an individual animal may bloat). In grazing situations where feed availability is not limited, the animal has more choice and eats the more digestible, 129 bloat provoking plant parts (Dougherty et al. 1989b). Restricted grazing would therefore be expected to create a greater bloat challenge than in a comparable penned animal feeding trial using the same alfalfa. The definition of a bloat challenge can be a rating based on a subjective four point scale: low (0-10% bloat frequency), moderate (11-25%), high (26-50%) or extreme (>50%). The range within each interval is a logical division rather than an arithmetic progression: bloat incidence in most ranching situations occasionally exceeds a frequency of 5%, but rarely reaches 10%, and a bloat incidence above 25% would be considered catastrophic. High or extreme bloat challenges are necessary to improve the probability that a statistical test for treatment differences will not conclude in an error (type I). However there is a risk in an extreme bloat challenge that bloat incidence may be too high and no test will detect a true difference (type II error). If the conditions in the LIRD-3 feedlot trial created a high bloat challenge then the conditions in the restricted grazing trials were extreme. All feedlot trials were conducted during periods when the measured 4hNbDMD differences between LIRD-3 and Beaver were less than 5.7% (actual 4hNbDMD values were 57.8% and 61.3% respectively, or a DMDRS%>94.3%; Hall et al. 1994a). The restricted grazing trials were conducted when 4hNbDMD differences were less than 11.0% (actual values were 53.6% and 60.2% for each respective cultivar; DMDRS%>89.0%; Hall et al. 1994a). The bloat challenge in these trials may have been too high or at least high enough to mask the bloat-reducing effects of LIRD-3, especially if the amount of bloat reduction was predicated on small variations in initial digestion kinetics. The inconclusive bloat results from the LIRD-3 cultivar were discouraging. The program had suffered from a lack of sufficient seed for bloat trials so when a limited amount of seed of AC Grazeland, the product of the fourth, and last, cycle of selection, became available, the research team decided to conduct a full series of grazing trials in several locations in western Canada (Berg et al. 1996, 1997). The objective was to measure bloat occurrence and severity in steers grazing AC Grazeland compared with Beaver under a moderate to high bloat challenge and an unrestricted grazing regimen. 130 5.2 MATERIALS AND METHODS The experiment was conducted in a series of annual grazing trials at three locations (stations) over a three year period. This is the minimum standard number of station years required for registration of a cultivar. The number of annual grazing trials varied at each location depending on growing conditions including the purity and amount of regrowth in the alfalfa stands and the availability of labour and livestock. Comparisons were made difficult because no alfalfa cultivar has a known bloat potential with which to compare, even among the older, dryland alfalfa varieties (Ashford and Heinrichs, 1967). Thus the grazing trials would be valuable if only because the tests could establish a standard for future comparisons of bloat reduction. A parental cultivar of the LIRD line, Beaver, is widely grown in western Canada, is used as a standard in comparative variety trials and has a record of inducing bloat. It was an obvious candidate for use as a standard in bloat trials although it is not regarded as a pasture or grazing-type alfalfa. Field Conditions and Management SITE DESCRIPTIONS. Project managers at each location met annually ahead of each growing season to coordinate trials, discuss methods and report on interim results. At each location they adapted to differences in agronomy and livestock handling but replicated the trial procedures as closely as possible. Trials were conducted at three sites in western Canada, the Agriculture and Agri-Food Canada Research Centres near Lethbridge, AB (49°42' N; 110°47' W), Kamloops, BC, (50°42' N; 120°24' W) and Melfort, SK (52°65' N; 104°45' W). This report is limited to the methods and results at the Lethbridge site. F I E L D M A N A G E M E N T . Stands of AC Grazeland and Beaver were established on adjacent 1 ha fields in 1993 as described in Chapter 2. Each field was divided by electric fencing into two paddocks of equal 131 area, to facilitate a crossover of animals between grazing periods within each grazing trial. Animal handling lanes were fenced, and lanes and fence lines were mowed immediately before each grazing trial. Pastures were irrigated as required, while unstocked. Grazing aftermath was removed by mowing or by intensively grazing the area with another group of cattle after each grazing period. This eliminated any mature alfalfa stems that might be carried over to be consumed in a subsequent trial and maintained the fields and crossover paddocks in early stages of development for the next grazing period. ALFALFA MATURITY. Maturity of the fields was recorded subjectively before each grazing trial. During the first year the three grazing trials were coincident with the late vegetative (stage 1 to stage 3, mid-vegetative to early bud stems), early bud (stage 2 to stage 4, late vegetative to late bud stems) and late bud (stage 3 to stage 5, early bud to early flower stems) stages of development in AC Grazeland. (Alfalfa plants exhibit a range of maturity as each stem develops. Specific stages follow the definitions of Fick and Mueller, 1989). In the second grazing season AC Grazeland was in a late vegetative to early bud stage for all three grazing trials (Berg et al. 1996). Animal Management All animals were ruminally fistulated, fitted with a ruminal cannula and managed according to the guidelines of the Canadian Council for Animal Care (CCAC 1993). The herd was predominantly Jersey steers (body weight 480 ± 30 kg) ranging from 1 to 3 years of age. At the commencement of the study, all animals were familiarized with the handling facilities to expedite their retrieval from pasture. Stock waterers and mineral salt boxes were accessible from all four paddocks. Paddocks were purposely understocked to permit a high degree of diet selection by the cattle. Cattle were fed alfalfa hay for at least ten days before each trial. 132 G R A Z I N G P E R I O D S . Before each grazing trial, either 8 or 10 steers, depending on animal health (no infection, disease or other problem that might suppress appetite) and forage availability, were randomly allocated to one of two grazing groups. The 4 or 5 replicate grazers in each group were assigned to graze either AC Grazeland or Beaver for the first of two periods in each grazing trial (Berg et al. 1996). When either the number of cumulative bloats exceeded 24 or reduced forage availability in the assigned paddock ended the first grazing period, each group was crossed-over to the ungrazed paddock of the other cultivar for the second grazing period (Hall et al. 1994a). These limits had been previously validated statistically by Hall et al. (1994a). Thus the length of each grazing period was not fixed except that after 7 to 10 days of grazing, forage availability often declined to such a degree that no more bloats could be expected. B L O A T M E A S U R E M E N T . N O bloat preventatives were used during or between grazing trials. Grazing cattle were observed every 1 to 2 h, from 0530 to 2300 daily and scored for distension against a 5-point index (Majak et al. 1983; Berg et al. 1996). A subjective rating of distension is generally insufficient for assessing bloat because froth and pressure must be gauged to evaluate bloat severity. Thus bloat indexes often include other observed conditions (Lindahl et al. 1957; Lippke et al. 1972; Majak et al. 1980). Sometimes a separate scale is used to score specific attributes, such as rumen volume or pressure (Cockrem et al. 1987b; Waghorn 1991). The scales employed in this study are shown in Table 5.1. Obvious distension in the field was the first condition observed and used to report bloat. Animals scoring 3, 4 or 5 on any distension scale were removed from the pasture and vented by opening the cannula to release the ruminal pressure. Any extrusa was cleaned away and the animal was returned to the paddock immediately. Bloat severity can be scored in ruminally cannulated animals by noting the presence of a stable foam and estimating the ruminal pressure. 133 Table 5.1 Distension and rumen condition indexes used to estimate bloat incidence and severity in studies of alfalfa grazing Rumen Conditiony Score Distensionz Frothx Pressure 1 None No None 2 Slight, left side No Normal 3 Left side Yes Normal 4 Left and right sides Yes Moderate 5 Severe Yes Severe zMajaket al. 1980, 1983 yBergetal. 1996 "Majak etal. 1983 w Lindahl et al. 1957, adapted by Berg et al. 1996 Two indexes were used, a distension rating for field scoring (Majak et al. 1980, 1983) and a separate index to describe ruminal pressure (Lindahl et al. 1957) and frothiness (Jacobson et al. 1957). The rumen condition index was recorded only when the cannula was opened, whether to relieve distension pressure or during routine cleaning and examinations (Berg et al. 1996). Pressure was scored by observing the distance that the rumen contents were expelled upon opening the cannula. In a case where there was no expulsion, the rumen contents were usually below the lip of the cannula, a rumen contraction would not expel any fluid and the rumen pressure could be characterized as 'none' due to low volume. In a rumen with 'normal' pressure (10 mm Hg measured by Lindahl et al. 1957), the contents usually spilled out of the cannula during a rumen contraction, but would fall directly down the side of the animal. In a 'moderately' pressured rumen (20-30 mm Hg measured by Lindahl et al. 1957) the contents would be expelled a distance of up to 1 m. In a 'severely' pressured rumen (>40 mm Hg measured by Lindahl et al. 1957) the contents were expelled beyond the 1 m mark (the record was 12 m). The pressure index was combined with the observation of a stable froth (Jacobson et al 1957; Majak et al 1983) into a 5 point scalar. 134 The use of the rumen condition index allowed moderate free-gas bloats, which may score 2 or 3 as distension but only 1 or 2 as rumen condition, to be deleted from the analysis. One case of bloat was recorded each time an animal scored 2 or higher on the distention index , once during a grazing day (ie. one calendar day in a grazing period, the established procedure of Majak et al. 1983). Trial periods were ended when the combined number of bloat cases for all treatments exceeded 24 (Hall et al. 1994a). Total animal-days of grazing were calculated as the sum of the days each animal was on pasture. The total animal-days of grazing on bloat-positive days was the sum of the days each animal was on pasture during days when one or more cases of bloat were recorded. Analysis was restricted to the first case of bloat reported for each animal on any particular grazing day. Second, third and occasionally fourth cases of bloat in the same animal on the same grazing day were recorded as multiple bloats but were otherwise omitted in the analysis, as were all data from the first day of grazing in a new paddock. Multiple bloats by an animal on one grazing day were considered as evidence of bloat storms. An animal with distension and rumen condition scores of 3 or greater was recorded as one animal-day of severe (acute) bloat. Statistical Analysis Bloat incidence (number of bloats per animal-day of grazing) was treated as a binomial variable. Bloat frequency (number of bloats per grazing day) was compared to a Poisson distribution (Sokal and Rohlf 1995) to find out whether bloat was a random phenomenon during the trials. The incidence and severity of pasture bloat on the two cultivars were evaluated by comparing the proportion of bloats in each treatment in each grazing trial using a Cochran Q test adapted for crossover experiments (Cochran 1950, Hall et al. 1994a). 135 5.3 RESULTS AND DISCUSSION Bloat Frequency Distribution Six grazing trials were conducted at Lethbridge over two years, 1995 and 1996 (Berg et al. 1997). No distension warranted intervention (that is, no animal had a distension score of 3 or greater) during the three trials in 1996, although conditions were subjectively similar to those 1995 trials where bloat did occur (Appendix III). For example, there were no distensions above a score of 2 in the 1996 trials although rumen condition scores of 3, a state of sub-acute bloat, were regularly observed during routine management of the animals (weekly care and cleaning of the fistula and cannula). The lack of bloats is consistent with other bloat studies in Saskatchewan on dryland alfalfa (Ashford and Heinrichs, 1967). Table 5.2 Bloat incidents in cattle grazing two alfalfa cultivars (cv. AC Grazeland and Beaver) Grazing Period2 AC Grazeland Beaver Reduction Trial (%) Spring Trial 1 June 21 - July 7 1995 6 4 -50.0% Summer Trial 2 Aug 8 - 24 1995 18 20 10.0% Fall Trial 3 Sept 15 - Oct 1 1995 Qa 20b 55.0% Total 33 44 25.0% 2 No bloats during trials 4, 5 and 6 from May 21 - June 10 1996, July 7- July 30 1996 and Aug 17 -Sept 5 1996. a , b One tail probability test of AC Grazeland<Beaver at P=.05. Numbers in the same row followed by the same or no letters do not differ (P>.05). Bloat occurred throughout the 1995 grazing season at Lethbridge (Table 5.2). No bloats were recorded in the spring 1996 trial, and the incidence of bloat was too low in the 1995 spring trial to evaluate treatment differences. The late summer trial of 1995 had the highest bloat incidence but the late summer, 1996 trial had no bloats. A scheduled irrigation between grazing periods within the 1996 late 136 summer trial did not affect bloat incidence. The 1996 autumn trial was also bloat-free. However, the fall is a period of high bloat risk (Hall and Majak, 1991) and bloat incidents were high during the autumn grazing in 1995. Table 5.3 Observed and expected (Poisson) frequency distributions of bloat in cattle grazing two alfalfa cultivars (cv. AC Grazeland and Beaver) at Lethbridge, AB AC Grazeland Beaver Overall Frequencyz Observed y Expected Observed Expected Observed Expected 0 285 257.74 282 248.07 588 505.72 1 13 50.89 13 58.46 12 109.52 2 5 5.02 3 6.89 6 11.86 3 6 0.33 10 0.54 5 0.86 4 4 0.02 5 0.03 6 0.05 5 1 0.00 1 0.00 0 0.00 6 0 0.00 0 0.00 2 0.00 7 0 0.00 0 0.00 4 0.00 8 0 0.00 0 0.00 1 0.00 9 0 0.00 0 0.00 0 0.00 10 0 0.00 0 0.00 1 0.00 Statistic Mean 0.197 0.197 0.236 0.236 0.217 0.217 SEX 0.041 0.025 0.045 0.027 0.041 0.019 CD W 2.650 1.003 2.746 1.003 4.821 1.002 X 2 * * * *** * * * z Bloats / day of grazing y Cumulative days of grazing; for each cultivar n=314, overall n=628 x Standard error of the mean w Coefficient of dispersion = s2/mean. CD=1 indicates random distribution of incidents, CD>1 indicates clumping, CD<1 indicates dispersion (Sokal and Rohlf 1981). *** X 2 Goodness of Fit test between observed and expected, significant at the 0.001 probability level. One common assumption made by ranchers is that the risk of bloat is greater in the spring. The Lethbridge results show otherwise, confirming earlier reports that late summer or autumn is the more dangerous period for bloat in western Canada (Hall and Majak 1991, 1995). Phenological development and maturity in alfalfa occur at different rates during the growing season primarily in response to 137 photoperiod and temperature (Bidwell 1974; Fick et al. 1988). Alfalfa plants respond to increasing daylight hours during the spring growth period with rapid morphological change. Development is less rapid during the late summer and fall as the day length and temperatures decline (Fick et al. 1988). These characteristics limit the time that immature forage is available to the animal during the spring in comparison to autumn. The overall bloat challenge during the trials was moderate. The bloat frequency in the first grazing season at Lethbridge was 21.7% (136 cases of distension recorded in 628 days of grazing during the first grazing season) (Table 5.3). Furthermore, over half the cases of distension were classified as severe bloat (Berg et al. 1996). Of these severe bloats, 3.5% had progressed to the point where death was potentially imminent (rumen condition scores of 5). In spite of the popular opinion that bloat occurs frequently when given an opportunity, bloat qualifies statistically as a rare event (Sokal and Rohlf 1995) because its mean frequency is typically low (Table 5.3). The incidence of solitary bloats (one animal bloat per day) had a lower frequency than expected (12 observed versus 109 expected), and clustered bloats (two or more animals bloating per day) occurred more frequently than expected. When compared against a Poisson distribution, bloat incidents did not occur randomly and were not independent of one another. Coefficients of dispersion suggest clumping (CD> 1) (Table 5.3), corroborating ranchers' observations that bloat occurs in 'bloat storms', where little or no bloat occurs most of the time, but under certain conditions the incidence will be high and can be extreme. In a bloat storm many animals may bloat at once, and if given the opportunity, as they were in these trials, a few have multiple bloats in a 24-hour period. On 12 different days at Lethbridge, clustered bloat incidents exceeded 4 and on one well-remembered day there were 10 (Table 5.3). The group of animals grazing AC Grazeland had 3 multiple bloat cases on 2 separate days. In contrast, during the same grazing periods there were 17 cases of multiple bloats on 10 different days on Beaver. Thus, the incidence of multiple bloats was 82% lower in cattle grazing AC Grazeland. 138 All cattle bloated at least once and 11 of the 17 head used in these trials bloated more than 5 times during the first grazing season. None of the animals used in the trials could be considered a chronic bloater. Two animals showed a propensity (P^.05) to become distended more frequently on the standard cultivar, Beaver, than on AC Grazeland. The overall coefficient of dispersion on the frequency of bloats in individual animals was 1.9, suggesting a degree of non-random clustering of bloat incidents (Berg et al. 1996). In other words, if one animal bloated, others were likely to be found in the same condition. Effect of Cultivar and Plant Maturity on Bloat Grazing AC Grazeland had an effect (Ps.05) on bloat incidence in only one grazing trial (Table 5.2). Overall there were 25.0% fewer bloat incidents on AC Grazeland than Beaver over 2 station-years of grazing trials (Berg et al. 1997). However the wide range, from no reduction to 50%, reflects the variation inherent in the occurrence of bloat at different times of the year and in different years. There were 456 animal-days of grazing on bloat days (those days when one or more animals in any field had a distension score ^3) in the first grazing season at Lethbridge (Table 5.4). Although the overall bloat incidence in cattle grazing AC Grazeland during bloat days was 16% lower than the incidence grazing Beaver, it was not statistically significant (P>.05). On the other hand, the incidence of severe bloat (rumen condition ^ 3) on bloat days, occurred 25% less often (P<.05) on AC Grazeland than on Beaver. Largely this was due to the highly significant (P<.01) reduction in bloats in animals that grazed AC Grazeland at a late vegetative stage of development. Bloat incidence on bloat days when early or late bud alfalfa was grazed did not differ (P>.05) between the two cultivars. Alfalfa is considered more bloat provocative when it is immature. During this study most severe bloats occurred on the lush, immature forage in the autumn trials (Berg et al. 1996). Distensions happened 29.6% less often (P^.05) on AC Grazeland than on Beaver, and the occurrence of severe bloat was 55.2% lower (P<.05) when AC Grazeland was in a late vegetative stage of growth (Table 5.4). These differences disappeared with increasing plant maturity and the total number of bloats declined as well. 139 Table 5.4 Effect of alfalfa maturity on bloat (total incidents / days of grazing) in cattle grazing two alfalfa cultivars (cv. AC Grazeland and Beaver) at Lethbridge, AB Alfalfa Maturity AC Grazeland Beaver P 2 All Distension*1 Late Vegetative x 0.171 0.243 0.027 Early Bud 0.117 0.133 0.306 Late Bud 0.125 0.117 0.676 Overall 0.136 0.162 0.069 Severe Bloatw Late Vegetative 0.064 0.143 0.003 Early Bud 0.092 0.102 0.372 Late Bud 0.050 0.033 0.893 Overall 0.072 0.096 0.044 2 One tail probability test for AC Grazeland < Beaver (after Hall et al. 1994; Cochrane 1950) y Includes all sub-acute and acute bloats; all distension scores >2, rumen scores ^3 " Cumulative days of grazing on which bloats occurred: late vegetative, 140 days; early bud, 196 days; late bud, 120 days; overall, 456 days. w Includes acute bloats only; distension scores ^3, rumen scores ^3. The greatest reduction in bloat was achieved by grazing AC Grazeland when the cattle were most at risk but the bloat challenge was reduced with increasing alfalfa maturity no matter which cultivar was grazed. Thus maturity is more important than the cultivar for preventing or reducing the incidence of bloat. Determining the actual maturity of alfalfa forage consumed by a grazing animal is extremely difficult. An attempt during a previous experiment, using masticate from esophageally fistulated cows (Chapter 2) was unsuccessful. The material was too damaged to make accurate assessments of stem development, and rarely were whole stems eaten. However, in another experiment (Chapter 4), the relative change in maturity between AC Grazeland and Beaver was measured during a 36 day regrowth period in the early fall. The results showed that plants in the AC Grazeland field entered late bud and early bloom stages more rapidly than plants in the Beaver field and that this created a maturity differential between the two cultivars (Figure 4.1). This maturity differential was observed in several other experiments and within a controlled environment (Chapter 2, 3). Four of the six grazing trials in the current experiment had a low bloat challenge (<10%). In all 140 but one of the six trials, AC Grazeland was in an early bud or later stage of development when the animals were turned into the fields (the stage of development in the autumn trial at Lethbridge in 1995 was late vegetative). During each of these trials, which were generally 14 days or more including crossovers, this cultivar would have moved through the late bud to early blossom stages or later. Beaver would have been in a correspondingly similar or earlier stage of development throughout. Thus the low bloat incidence and the lack of difference between the two cultivars are likely a consequence of maturity rather than any other inherited characteristic. As previously mentioned, four of the six grazing trials conducted during two different years had a low bloat challenge (<10% incidence). No difference in bloat incidence was found between AC Grazeland and Beaver in these four trials. During the other two grazing trials, the bloat challenge was moderate or high but AC Grazeland had fewer bloat incidents (P^.05) only in the moderate challenge trial. Interestingly, this latter trial was conducted late in the growing season on autumn regrowth. 5.4 CONCLUSIONS AC Grazeland, an alfalfa cultivar with a low initial rate of digestion, has shown an ability to reduce the incidence and severity of bloat in grazing cattle in one of six trials. The effect was apparent mostly during a high risk period in conditions that created a moderate bloat challenge. Early stages of development in either the LIRD cultivar or the standard, create a greater bloat challenge and will result in more bloat incidents. Late season grazing apparently has the same effect. Under these circumstances, the LIRD cultivars' ability to advance morphologically and physiologically may reduce the risk of bloat sooner than the current standard. AC Grazeland and the other LIRD cultivars do not eliminate the development of the stable foams associated with sub-acute or acute frothy bloat. Thus, they cannot be considered 'bloat-safe' or 'bloat-free'. Although there is strong evidence that foam will accumulate, perhaps the rate of accumulation is 141 sufficiently slow that it reduces the rate of distension, allowing the normal coping mechanisms of the animal to limit the development of severe tympanites. AC Grazeland and its precursors may add a degree of safety in bloat storms by decreasing the frequency and severity of acute bloats, giving the grazier time to gain control of the situation. In this sense AC Grazeland may be considered a 'bloat-reduced' alfalfa. 142 Chapter 6 Analyses and Implications 6.1 SYNTHESIS In Summary Myriads of procedures exist for testing the efficacy of different feeds and their comparative digestibility in ruminants. The method chosen must always be adapted to fit the situation, so the lack of fit and the results will always be the subject of much critical thought. Such is the nature of scientific progress: cycles of change, adaptation and criticism take us from where we have been to where we are today and where we will be tomorrow. It is a bit audacious to consider that a small digestibility trial is a model of scientific advancement, but nevertheless it is true. We are where we are today because a search for a solution to a problem, in this case bloat in ruminants, directed us to look in a particular way, effect a change, and appraise the result. So where are we and what was learned? Probably the most important result is the confirmation that the cultivar, AC Grazeland differs from the standard Beaver using standard measures of rumen digestibility. Thus the original hypotheses are rejected but as with many others, the rejections are qualified. When AC Grazeland and Beaver are compared at dissimilar stages of development, a reduction in tissue digestion is a characteristic of the LIRD cultivar. The reduction is not consistent, linear with respect to maturity in either cultivar, or unaffected by the local environment. The differences in rates or extent of digestion between AC Grazeland and Beaver are not radical and are not exclusively acquired from the original expected source, leaf tissues; instead they are minor (bordering on statistical insignificance) and their primary locus is the stem tissue. In the field, the differences are a result of an earlier or faster rate of maturity in the LIRD cultivar, AC Grazeland. Confirmation of differences between AC Grazeland and its progenitors were not as forthcoming. Consequently the original hypotheses of no difference are accepted but again some qualification is 143 needed. The only tests that were conducted used dried and ground plant tissue samples collected from each cultivar without correspondingly detailed information on the relative stages of maturity of each cultivar. The acceptance of the hypotheses must be considered in the context that the testing and measurement procedures may have been insufficient. The most important hypothesis required a test of the bloating inducing properties of AC Grazeland against a previously undefined standard. Previous tests of this hypothesis, using a progenitor of AC Grazeland, had been equivocal, favouring rejection when whole plant material was fed to penned animals and acceptance in restricted grazing trials. In other words, the LIRD cultivar may have induced fewer bloat incidents but the conditions during the testing procedures were confounding. Subtle differences appeared to be overwhelmed in extreme bloat challenges (also undefined). The alfalfa standard within the current study was defined as the cultivar Beaver, a parent of the LIRD strain, in an early bud stage of development and the bloat challenge was scaled so that comparisons could be made between experiments, should they occur. During four grazing, in conditions of low bloat challenge, there were no bloat incident differences between AC Grazeland and Beaver, leading to acceptance of the hypothesis. The bloat challenge was moderate to high during two other grazing trials, in one of which there were significantly fewer bloat incidents on AC Grazeland, leading to rejection of the hypothesis. Ranchers would rarely manage herds or forage resources to maintain moderate or high bloat challenges. Consequently, in practice, no differences in bloat incidence will likely accrue to ranch operations by grazing AC Grazeland. Our current understanding of bloat points toward one conclusion that it is an endemic trait of domestic ruminants and cannot be eliminated. Consequently, there will be a continuing need to control it and reduce its severity. AC Grazeland may be an important tool in this respect: the severity and multiple bloat incidents characteristic of 'bloat storms' were reduced in cattle grazing this cultivar during both low and moderate bloat challenges. Again this is probably a result of a maturity differential between it and the standard cultivar within comparative trials. A rancher trying desperately to prevent livestock 144 deaths during a bloat storm would neither admit nor understand that AC Grazeland was helping in the situation. However if the decision has been made to graze alfalfa for the profit it can give the operation, a risk appraisal will rank the choice of AC Grazeland better than others. On a L I R D Alfalfa When Howarth et al. (1978a) formulated the hypothesis that the rate of cell rupture might be a factor in bloat, it was based on the difference between alfalfa leaf cell fractions and those of bloat-free legumes, such as sainfoin. The leaves of all species were collected at weekly intervals but all plant parts were sampled, older leaves were mixed with recently emerged leaves and they summed the analytic results over all periods. Therefore they must have assumed that leaf age did not affect cell wall development or digestibility. Procedures for collection included a broad range for maturity ('bud to early bloom stages' in preliminary trials; 'prebud to seedset stages' in the full data set; Howarth et al. 1978a) and specific differences were not reported. So they must have assumed that all species matured at the same rate, or that the differences were so minor that they did not need to be evaluated. (Knowing this, it has been interesting to watch sainfoin and alfalfa growing in the same irrigated field at Lethbridge AB. Sainfoin invariably blossoms before the alfalfa buds.) While subsequent in vitro and in vivo digestion trials considered a narrower range of maturity ('pre-bud or bud stages', Fay et al. 1980; 'prior to onset of flowering', Sant and Wilson 1982; 'early to mid-bud stage', Howarth et al. 1982) differences in rates of development between species were not measured. Furthermore, by collecting plant material from different species or cultivars on one day, or by summarizing over time intervals, these investigators held on to the assumption that cell wall digestibility was unaffected by specific rates of development or age. The effects of age and maturity on cell wall thickness and digestibility can be easily overlooked, confounded and misinterpreted in comparative studies. The distribution of thick and thin walled cells in various plant tissues were examined for mid-season (56 to 61 days growth) or older (87 to 121 days) 145 plant material (Moghaddam and Wilman 1998; Wilman and Moghaddam, 1998). The investigators reported thicker epidermal cell walls in alfalfa than those observed by Lees (1984) but the leaves used in the latter study were no more than 35 days old. Sainfoin was grown to the same age as alfalfa but for some reason the sainfoin stem buds failed to initiate and the plants did not mature normally (Wilman and Ahmad 1999). They concluded that sainfoin and alfalfa leaf cell walls were similar in thickness, in contrast to Lees (1984) measurements, and implied that leaf tissue disruption was not as great a factor in bloat as surmised by Howarth et al. (1982). An alternate explanation is that the sainfoin in Lees' (1984) study had matured normally. At the start of selection trials for a LIRD trait, Howarth et al. (1982) observed that the day of harvest affected in vivo DMD. The effect was confounded because DMD of leaf tissue differed between the same plants harvested on consecutive days and between cutting dates within the growing season. Consequently they proposed that the procedure to select for a low IRD would need to compare groups of plants collected on the same day. The in vivo digestion method used by Goplen et al. (1993) limited selection pressure to alfalfa stem tips from plants in the 'pasture stage of growth (pre-bud to mid-bud)'. Maturity was defined as a range because logistically determining when each plant in the nursery attained a more specific level of maturity would have been impossible. Thus maturity was used only as a limit to the timing of digestion trials (start, end and cutting schedules between trials). Goplen et al. (1993) also accepted the counsel of Howarth et al (1982) and tested the digestibility of each plant only once on one day. This condition had an unexpected side effect. Plants with faster rates of maturity would develop more cell wall before their evaluation date so the genes from these plants would be more likely to remain in the pool for the next cycle of selection. Because each test period took more than 10 days to complete, plants with faster rates of maturity within the sampling period were also more likely to make the final selection. When intercrossed, the new synthetic cultivars had a lower digestibility and more cell wall but they also reached later stages of development at earlier ages than the 146 standard cultivar. Thus, screening alfalfa clones for low initial rates of digestion was a screen of cell wall against age favouring the selection of early maturing genomes. In its current guise, the definition of a LIRD alfalfa is any cultivar that matures earlier than an existing standard. Comparisons of rates or extent of digestion between the LIRD synthetics, AC Grazeland or its progenitors, LIRD-1, LIRD-2 or LIRD-3, and any other cultivar using collection methods based on a common event, such as a harvest cutting or a regrowth period, will likely confirm the original differences. The implications of this are great. Most alfalfa cultivars registered in Canada are evaluated in comparative trials against the same standard, Beaver. Yield and quality are evaluated on a single day of harvest, also in relation to the same standard. Thus most cultivars will appear to be of higher quality and will be slower maturing than AC Grazeland because the current test procedure confounds quality and maturity. Instead, all cultivars should be tested for yield and quality at the same stage of development not the same age. On Bloat Bloat is not accepted as a tolerable risk in ruminant livestock production, but few hazards are. Death from old age might be one. However, if the demand for greater feed conversion efficiencies continues, acute bloat will rank with other hazards associated with an increased risk of loss in intensive livestock production. We will continue to see acute bloat in circumstances where the digestibility of the feed is well beyond any range that would have been encountered in the ruminant's evolutionary history. For cattle or sheep, pure stands of vegetative alfalfa are certainly in this category but so are feedlot and dairy rations. The amount of dry matter disappearance hypothesized by Howarth et al. (1982) to bring bloat to an acceptable or nonexistent level was 25% lower than a standard cultivar, eight hours after ingestion. This goal was determined by comparing the DMD of alfalfa with non-bloating legumes. Significant reductions in the risk of acute bloat may be accomplished earlier and with marginally smaller changes in 147 digestibility. The major reduction of bloat incidence in sheep, in a feeding regimen designed to induce bloat, occurred after the mean apparent dry matter digestibility of the two alfalfa cultivars fell from 71.3% to 67.5 % (ie. 3.8%). Sheep are less prone to bloat than cattle (Colvin and Backus 1988) but if the principle is the same, the frequency of boat incidents will hinge on relatively minor changes in alfalfa digestibility. On a Unified Theory CELL RUPTURE HYPOTHESIS. For more than 50 years, distinct kinds of bloat have been described because the origins appeared to differ by location (pasture or feedlot), feed source (legume or grain) or type (free-gas or froth). The cell rupture theory eliminates the need for these categories because it implies that all locations and feed sources have a potential for any of several types of bloat (including sub-acute and acute bloat). Thus the theory is inclusive of those circumstances where bloat has been reported but regarded as an exception to some unwritten rule about the cause of bloat. For example the 'rule' that grass does not cause bloat is antithetic to the fact that bloat does occur on annual grass forages such as winter wheat and lush ryegrass pastures (Cole et al 1945, Ayre-Smith 1971). As an application of cell rupture theory, these exceptions are worth examining because the theory suggests that there is some circumstance where grass cells may rupture at a rate sufficient to cause bloat. The cell rupture hypothesis helps our understanding of bloat by providing a more integrated model of bloat and the digestive processes in the rumen. It is an important step toward a unified theory of bloat occurrence and digestion but alone it is a segregated theory (Waghorn, G. pers. comm. 1996). The theory that the rate of cell rupture is a cause of bloat is still a functional division of digestion because it regards bloat as having either an exclusive plant or an animal origin and it offers no consideration for animal susceptibility. In a unified theory the distension known as bloat is only one clinical symptom in a progressive dysfunction of digestion that has both plant and animal origins. In other words a unified hypothesis is 148 that bloat occurrence, or even a predisposition to bloat, is caused by an imbalance between the rate of cell rupture and the rates of substrate intake and discharge of by-products. So at least two congruent arguments must be integrated in a unified theory: 1) a theory about the rate limiting steps in the ruminal digestion of plant materials, filled adequately by the cell rupture hypothesis, and 2) a hypothesis about ruminal rates of passage. PASSAGE RATE THEORY. The hypothesis that bloat is one consequence of changes in ruminal rates of passage is a generalization that uses the old theories of excessive consumption and animal susceptibility as corollaries. Excessive consumption is an immoderate change in the rate of intake. Animals may have a greater or lesser susceptibility to bloat because of their inherent rate of passage, or a congenital or pathological predisposing condition that changes the rates of passage. The passage rate theory purports that the cause of bloat is directly related to the rates of food intake, rates of bacterial invasion, maceration and fermentation and the rates of output of food and fermentation products from the rumen. An acute bloat may be one result of a radical change in the rates of ruminal digestion and passage. The theory is supported by the observation that: 1) the difference, in bloat incidence and severity, between a strain of dairy cattle selected for high susceptibility to bloat and a low susceptibility strain is inversely related to the rate of passage of food through the rumen (Cockrem et al. 1987b; Carruthers et al. 1988; Okine et al. 1989; Okine and Mathison 1991); 2) the rate of gas expulsion may be limited by the animal's failure to eructate, and that free-gas bloat is an important cause of distension that may be as prevalent as frothy bloat (Cole et al. 1945; Moate etal. 1997); 3) bloat can arise from pathological conditions, stress, motility disorders and other ailments that obstruct or change the dynamics of rumen flow (Garry 1990a, 1990b, 1990c); 149 4) animals adapt behaviourally to changing flow rates and the accumulation of gas and foam by reducing intake, curtailing rumination and increasing rumen contractions (Hancock 1954; Alder etal. 1967; Waghorn 1991); 5) a state of sub-acute bloat can be sustained for an indefinite period, this state being dependant on a steady input of the same feed ingredients offset by countervailing outputs of gas and other by-products of fermentation, and that the balance changes when either the input or output streams falter, a circumstance which warrants that it is the change between flow rate states rather than the existence of a particular state that induces an acute bloat (Lindahl et al.1957); and, 6) microbial invasion and digestion of leaf and stem tissues proceed sequentially in a series of rate limiting steps that vary with the plant tissue type (Cheng et al. 1980) and that rates of intake and passage through the rumen can differ between feeds of equivalent overall digestibility (Troelsen and Campbell 1969). The rate of passage theory, like the cell rupture theory, also includes the exceptions often excluded from consideration in foam theories. Bloat incidents have been reported on an amazingly wide variety of feed sources, including those considered bloat-safe, such as grass pastures, trefoil, hay and corn silage (Cole et al. 1945). Most legume species including the non-bloat-causing legumes like sainfoin and cicer milkvetch, develop stable foams in vitro (Cooper et al. 1966). Under the foam hypothesis these exceptions could be ignored only if the kind of bloat being examined was predefined, say as in an alfalfa bloat. However, if bloat arises because a threshold limit to passage has changed then there is little reason to make distinctions except to define the context of the changed rate. In practice, manipulating rates of intake, passage and fermentation in the rumen are traditional ranching activities in another guise. Input and output from the rumen can be indirectly affected by adopting grazing and feeding regimens that increase or decrease rumen clearance, or reduce gas and foam production. The decision to leave animals in one paddock to graze the remaining less palatable forage will reduce the rate of intake. Moving them from a lush vegetative paddock to a mature, senescent 150 stand will have a similar effect. Making these decisions in the context of promoting rapid clearance rates has only been suggested but has not been delineated (Majak et al. 1995). So for the present, control will continue to depend on the rancher's skill at balancing the availability and digestibility of pastured forage with the grazing behaviour of the animals (Hancock 1954; Allison 1985; Hall et al. 1994a; Majak et al. 1995). 6.2 FUTURE DIRECTIONS When Howarth et al (1978a) looked for common characteristics responsible for bloat resistance in sainfoin and birdsfoot trefoil, two traits were identified, leaf cell resistance to rupture and the presence of condensed tannins. The results led to the establishment of a team tasked with the development of bloat resistance in alfalfa. Individuals in the team took one of two research directions. The first direction, since abandoned, required the identification, isolation and transfer of the genes responsible for the production of condensed tannins using sainfoin as the donor. Perhaps the traces of this research will be picked up after the alfalfa genome is mapped. The success of the second direction came early during the LIRD breeding program, a net reduction in IRD. However, this program has also been abandoned. To make alfalfa bloat resistant will still require either a major reduction in digestibility or the acquisition of other factors such as condensed tannins. The development of new cultivars is unlikely because of the substantial commitment of resources that were required in the LIRD breeding program. Besides, there are other more economical ways of obtaining early maturity, by screening for early development of inflorescence, for example, that do not require livestock. The Cycle 1 and Cycle 2 LIRD cultivars should be tested for release and registration. Most of the genetic change occurred during these selection cycles. Since no other cultivars are available with a quantified increase in the rate of development, they may be of some value alongside AC Grazeland. 151 There are significant advantages to early maturity in prairie agriculture. In some regions the standard alfalfa maturation period of 35 to 45 days causes harvest to occur during inclement weather periods. Just a 5 day grace period might be enough time to take up quality hay instead of rain-damaged feed. For graziers, early turn-outs to pasture would extend the grazing season or improve pasture productivity during the spring. Techniques other than digestion in vivo, such as gas production, will be used to assess food value and will be used more frequently to select traits for breeding programs. However, ultimately the animal must be used to test forage more often than in the past. The indiscriminate substitution of in vitro techniques has left western Canada with a spate of grasses and legumes lacking grazing tolerance and many have low nutrient levels in periods outside normal hay cutting intervals. Resources for plant breeding programs are limited and pursue the commercial crops sometimes exclusively. It is a circular argument; new money for the non-bloat legumes such as sainfoin will not be forthcoming until they prove themselves capable of competing for the resources devoted to alfalfa, but they cannot compete without new investment. Instead there may be a hidden opportunity in mixing legumes such as sainfoin, with alfalfa and developing grazing and feeding strategies to use them to advantage over alfalfa. The condensed tannins in some of these plants are untested, natural, potentially useful feed additives. 152 References Acord, C. E., Butcher, J . and Thomas, D . W. 1968. Bloat on dry alfalfa and legume pastures. J. Anim. Sci. Abstr. 27: 1105. Acord, C. E. 1969. Beef production on irrigated pastures. Utah Sci. 30: 7-9. Albrecht, K. A., Wedin, W. F. and Buxton, D . R . 1987. Cell-wall composition and digestibility of alfalfa stems and leaves. Crop Sci. 27: 735-741. Alder, F. E. S., Cowlishaw, S. J . , Newton, J . E. and Chambers, D . T. 1967. The effect of level of nitrogen fertilizer on beef production from grazed perennial ryegrass/white clover pastures. 2. Animal production from irrigated swards. J. Brit. Grassland Soc. 22: 230-238. Allison, C. D . 1985. Factors affecting forage intake by range ruminants: a review. J. Range Manage. 38: 305-311. Allinson, D . W., Elliot, F. C. and Tesar, M. B. 1969. Variations in nutritive value among species of the Medicago genus as measured by laboratory techniques. Crop Sci. 9: 634-637. Anderson, B. 1997. Dealing with bloat in grazing alfalfa. Pages 45-52 in Johnson, J., ed. Proceedings of the 2nd National Alfalfa Grazing Conference, Omaha, NE, Aug 16. Certified Alfalfa Seed Council, Davis, CA. AOAC. 1990. Official methods of analysis. (15th Ed.). Assoc. Official Analytical Chemists, Arlington, VA. Ashford, R . and Heinrichs, D . H. 1967. Grazing of alfalfa varieties and observations on bloat. J. Range Manage. 20: 152-153. Ayre-Smith, R . A. 1971. Pasture bloat in cattle. Aust. Vet. J. 47: 162-169. Barry, T. N. and McNabb, W. C. 1999. The implications of condensed tannins on the nutritive value of temperate forages fed to ruminants. Brit. J. Nutr. 81: 263-272. 153 B a r t l e y , E . E . , L i p p k e , H . , P f o s t , H . B . , N i j w e i d e , R . J . , J a c o b s o n , N . L . a n d M e y e r , R . M . 1 9 6 5 . Bloat in cattle. 10. Efficacy of poloxalene in controlling alfalfa bloat in dairy steers and lactating cows in commercial dairy herds. J. Dairy Sci. 4 8 : 1657-1662. B e d d o w s , A . R . 1952 . Observation on hoove, hoven, blown, or bloat taken from agricultural writings, 1716-1827. J. Agric. Soc. Univ. Coll. Wales 3 3 : 12-15. Reprint. B e r d a h l , J . D . , W i l t o n , A . C , L o r e n z , R . J . a n d F r a n k , A . B . 1986 . Alfalfa survival and vigor in rangeland grazed by sheep. J. Range Manage. 3 9 : 59-62. B e r d a h l , J . D . , W i l t o n , A . C . a n d F r a n k , A . B . 1989 . Survival and agronomic performance of 25 alfalfa cultivars and strains interseeded into rangeland. J. Range Manage. 4 2 : 312-316. B e r g , B . P . , T a i t , R . M . , M c A l l i s t e r , T . A . , M a j a k , W . , A c h a r y a , S . a n d C h e n g , K . - J . 1996 . Bloat in cattle grazing an alfalfa cultivar selected for reduced digestibility and low bloat incidence. Pages 111-115 in Williams, M . J. ed. Proceedings of the Forage and Grassland Conference, Vancouver, BC, June 13-15. American Forage and Grassland Council, Georgetown, TX. B e r g , B . P . , C o u l m a n , B . , G o p l e n , B . , H a l l , J . , M a j a k , W . , M c A l l i s t e r , T . A . , M c C a r t n e y , D . , T a i t , R . M . a n d C h e n g , K . - J . 1997 . An alfalfa cultivar selected for low initial rate of digestion reduces bloat in grazing steers. Page 73 in Johnson, J., ed. Proceedings of the 2nd National Alfalfa Grazing Conference, Omaha, NE, Aug 16. Certified Alfalfa Seed Council, Davis, CA. B e r g e n W . G . a n d B a t e s , D . B . 1984. Ionophores: their effect on production efficiency and mode of action. J. Anim. Sci. 58 : 1465-1483. B e u v i n k , J . M . W . a n d K o g u t , J . 1993 . Modeling gas production kinetics of grass silages incubated with buffered ruminal fluid. J. Anim. Sci. 71:1041-1046. B i d w e l l , R . G . S . 1974 . Plant physiology. Macmillan Publ. Co. N. Y . B l u m m e l , M . a n d 0 r s k o v , E . R . 1993 . Comparison of in vitro gas production and nylon bag degradability of roughages in predicting feed intake in cattle. Anim. Feed. Sci. Technol. 4 0 : 109-119. 154 Boda, J. M., Cupps, P. T., Colvin, H., Jr., Cole, H. H. 1956. The sequence of events preceding death of a cow in acute experimental bloat on fresh alfalfa tops. J. Amer. Vet. Med. Assoc. 128: 532-535. Brummer, E. C. and Bouton, J. H. 1991. Plant traits associated with grazing-tolerant alfalfa. Agron. J. 83: 996-1000. Brummer, E. C. and Bouton, J. H. 1992. Physiological traits associated with grazing-tolerant alfalfa. Agron. J. 84: 138-143. Burns, J. C. and. Standaert, J. E. 1985. Productivity and economics of legume-based vs nitrogen-fertilized grass-based pasture in the United States. Pages 56-71 in Barnes, R. F., Ball, P. R., Brougham, R. W., Marten, G. C. and Minson, D. J. Proceedings of a tri-lateral workshop. Forage legumes for energy efficient animal production. Palmerston North, N . Z., April 30, 1984. USDA. Agric. Res. Serv., Springfield, VA. Buxton, D. R. 1996. Quality-related characteristics of forages as influenced by plant environment and agronomic factors. Anim. Feed. Sci. Technol. 59: 37-49. Buxton, D. R. and Casler, M. D. 1993. Environmental and genetic effects on cell wall composition and digestibility. Pages 685-714 in Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J. (eds). Forage cell wall structure and digestibility. Amer. Soc. Agron. Madison, WI. Caddel, J. 1997. Alfalfa grazing varieties - past, present, future. Pages 9-19 in Johnson, J., ed. Proceedings of the 2nd National Alfalfa Grazing Conference, Omaha, NE, Aug 16. Certified Alfalfa Seed Council, Davis, CA. Carruthers, V. R., Bryant, A. M., Cockrem, F. R. M. and McLeay, L. M. 1988. Quantity of digesta in the reticulo-rumen of cows differing in susceptibility to bloat. N . Z. J. Agric. Res. 31:111-119. CCAC. 1993. Guide to the care and use of experimental animals. Can. Council on Animal Care. Ottawa ON. Cheng, K.-J., Fay, J. P., Howarth, R. E. and Costerton, J. W. 1980. Sequence of events in the digestion of fresh legume leaves by rumen bacteria. Appl. Environ. Microbiol. 40: 613-625. 155 Cheng, K.-J., Hironaka, R., Jones, G. A., Nicas, T. and Costerton, J. W. 1976. Frothy feedlot bloat in cattle: production of extracellular polysaccharides and development of viscosity in cultures of Streptococcus bovis. Can. J. Microbiol. 22: 450-459. Clarke, R. T. J. and Hungate, R. E. 1971. Bloat in cattle. X X X V . Microbial activities in the reticulo-rumens of cows differently susceptible to legume bloat. N. Z. J. Agr. Res. 14: 108-121. Clarke, R. T. J. and Reid, C. S. W. 1974. Foamy bloat of cattle. A review. J. Dairy Sci. 57: 753-785. Clarke, R. T. J., Jones, W. T. and Reid, C. S. W. 1974. Bloat in cattle: XLI. Proteins of saliva as possible genetic markers in breeding programmes to reduce bloat susceptibility in cattle. N. Z. J. Agric. Res. 17:411-415. Cochran, W. G. 1950. The comparison of percentages in matched samples. Biometrika 37: 256-266. Cockrem, F. R. M., Mcintosh, J. T. and McLaren, R. D. 1983. Selection for and against susceptibility to bloat in dairy cows - a review. Proc. N. Z. Soc. Anim. Prod. 43: 101-106. Cockrem, F. R. M., Mcintosh, J. T., McLaren, R. D. and Morris, C. A. 1987a. In vitro gas and foam production by rumen fluid from cows of genetically high or low susceptibility to pasture bloat. Anim. Prod. 45: 37-42. Cockrem, F. R. M., Mcintosh, J. T., McLaren, R. D. and Morris, C. A. 1987b. The relationship between volume of rumen contents and genetic susceptibility to pasture bloat in cattle. Anim. Prod. 45: 43-47. Cole, H. H., Huffman, C. F., Kleiber, M., Olson, T. M. and Schalk, A. F. 1945. A review of bloat in ruminants. J. Anim. Sci. 4: 183-236. Cole, H. H. and Boda, J. M. 1960. Continued progress toward controlling bloat. A review. J. Dairy Sci. 43: 1585-1614. Colvin, H. W., Jr. and Backus, R. C. 1988. Bloat in sheep (Ovis aries). Comp. Biochem. Physiol. 91A: 635-644. 156 Cooper, C. S., Eslick, R. S. and McDonald, P. W. 1966. Foam formation from extracts of 27 legume species in vitro. Crop Sci. 6:215-216. Dehority, B. A. 1993. Microbial ecology of cell wall fermentation. Pages 425-453 in Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J. (eds). Forage cell wall structure and digestibility. Amer. Soc. Agron. Madison, WI. Dougherty, C. T., Bradley, N. W., Cornelius, P. L. and Lauriault, L. M. 1987. Herbage intake rates of beef cattle grazing alfalfa. Agron. J. 79: 1003-1008. Dougherty, C. T., Bradley, N. W., Cornelius, P. L. and Lauriault, L. M. 1989a. Short-term fasts and the ingestive behaviour of grazing cattle. Grass and Forage Sci. 44: 295-302. Dougherty, C. T., Bradley, N. W., Cornelius, P. L. and Lauriault, L. M. 1989b. Ingestive behaviour of beef cattle offered different forms of lucerne (Medicago sativa L.) Grass and Forage Sci. 44: 335-342. Dougherty, C. T., Collins, M., Bradley, N. W., Lauriault, L. M. and Cornelius, P. L. 1992. The effects of poloxalene on ingestion by cattle grazing lucerne. Grass and Forage Sci. 47: 180-188. Dunford, A. and Jewison, G. 1997. Canfax Annual Report. Canfax Research Services, Can. Cattlemen's Assoc. Fairey, N. A. and Fairey, D. T. 1993. Forage cultivar trials: 1993 bulletin. Agric. Agri-Food Can. Bulletin NRG 93-11. Fay, J. P., Cheng, K.-J., Hanna, M. R., Howarth, R. E. and Costerton, J. W. 1980. In vitro digestion of bloat-safe and bloat-causing legumes by rumen microorganisms: gas and foam production. J. Dairy Sci. 63: 1273-1281. Fay, J. P., Cheng, K.-J., Hanna, M. R., Howarth, R. E. and Costerton, J. W. 1981. A scanning electron microscopy study of invasion of leaflets of a bloat-safe and a bloat-causing legume by rumen microorganisms. Can. J. Microbiol. 27: 390-399. Fick, G. W., Holt, D. A. and Lugg, D. G. 1988. Environmental physiology and crop growth. Pages 163-194 in Hanson, A. A., Barnes, D. K., and. Hill, R. R., Jr. eds. Alfalfa and alfalfa improvement. 157 Agronomy Mono. No. 29. Amer. Soc. Agron., Madison, WI. F i c k , G . W . a n d M u e l l e r , S . C . 1989 . Alfalfa quality, maturity, and mean stage of development. Page 1-13 in Information Bulletin 217, College of Agriculture and Life Sciences, Cornell University, Ithaca, NY. F i c k , G . W . a n d O n s t a d , D . W . 1988 . Statistical models for predicting alfalfa herbage quality from morphological or weather data. J. Prod. Agric. 1:160-166. F i s h e r , D . S . , B u r n s , J . C . a n d P o n d , K . R . 1989 . Kinetics of in vitro cell-wall disappearance and in vivo digestion. Agron. J. 8 1 : 25-33. G a r r y , F . 1 9 9 0 a . Evaluating motility disorders of the bovine forestomach. Vet. Med. June, 634-642. G a r r y , F . 1 9 9 0 b . Managing bloat in cattle. Vet. Med. June, 643-650. G a r r y , F . 1 9 9 0 c . Managing forestomach problems caused by motor disease. Vet. Med. June, 651-659. G e t a c h e w , G . , B l u m m e l , M . , M a k k a r , H . P . S . a n d B e c k e r , K . 1998 . In vitro gas measuring techniques for assessment of nutritional quality of feeds: a review. Anim. Feed Sci Technol. 72:261-281. G o p l e n , B . P . 1 9 8 9 . Breeding bloat-safe alfalfa. Alfalfa Seed News 8:10-11. G o p l e n , B . P . , H o w a r t h , R . E . , S a r k a r , S . K . a n d L e s i n s , K . 1980 . A search for tannins in annual and perennial species of Medicago, Trigonella, and Onobrychis. Crop Sci. 20 : 801-804. G o p l e n , B . P . , H o w a r t h , R . E . , L e e s , G . L . , C h e n g , K . - J . a n d M a j a k , W . 1 9 8 5 . An interdisciplinary approach to breeding bloat-safe forage legumes. Pages 228-229 in Obuko, T. and Shiyomi, M. , eds. Proceedings of the 15th International Grassland Congress, Science Council of Japan, Kyoto, Japan. G o p l e n , B . P . , H o w a r t h , R . E . a n d L e e s , G . L . 1993 . Selection of an alfalfa for a lower initial rate of digestion and corresponding changes in epidermal and mesophyll cell wall thickness. Can. J. Plant Sci. 7 3 : 111-122. G u t e k , L . H . , G o p l e n , B . P . , H o w a r t h , R . E . a n d M c A r t h u r , J . M . 1974. Variation of soluble proteins in alfalfa, sainfoin, and birdsfoot trefoil. Crop Sci. 14: 495-499. 158 Hall, J. W. and Majak, W. 1991. Relationship of weather and plant factors to alfalfa bloat in autumn. Can. J. Anim. Sci. 71: 861-866. Hall, J. W. and Majak, W. 1992. Rapid screening of feed supplements for the prevention of legume bloat. Can. J. Anim. Sci. 72: 613-617. Hall, J. W. and Majak, W. 1995. Effect of time of grazing or cutting and feeding on the incidence of alfalfa bloat in cattle. Can. J. Anim. Sci. 75: 271-273. Hall, J. W., Majak, W., Williams, R. J. and Howarth, R. E. 1984. Effect of daily weather conditions on bloat in cattle fed fresh alfalfa. Can. J. Anim. Sci. 64: 943-950. Hall, J. W., Majak, W., Van-Ryswyk, A. L., Howarth, R. E. and Kalnin, C. M. 1988. The relationship of rumen cations and soluble protein with predisposition of cattle to alfalfa bloat. Can. J. Anim. Sci. 68: 431-437. Hall, J. W., Majak, W., Stout, D. G., Cheng, K.-J., Goplen, B. P. and Howarth, R. E. 1994a. Bloat in cattle fed alfalfa selected for a low initial rate of digestion. Can J. Plant Sci. 74: 451-456. Hall, J. W., Walker, I. and Majak, W. 1994b. Evaluation of two supplements for the prevention of alfalfa bloat. Can. Vet. J. 35: 702. Hancock, J. 1954. Studies in grazing behaviour of dairy cattle. II. Bloat in relation to grazing behaviour. J. Agric. Sci. (Camb.) 45: 80-95. Heinrichs, D. H. 1963. Creeping alfalfas. Adv. Agronomy. 15: 317-337. Hill, R. R. Jr., Shenk, J. S., and Barnes, R. F. 1988. Breeding for yield and quality. Pages 809-825 in Hanson, A. A., Barnes, D. K., and. Hill, R. R., Jr. eds. Alfalfa and alfalfa improvement. Agronomy Mono. No. 29. Amer. Soc. Agron., Madison, WI. Hobson, P. N . and Jouany, J.-P. 1988. Models, mathematical and biological, of the rumen function. Chapter 15, Pages 461-511 in Hobson, P. N . ed. The rumen microbial ecosystem. Elsevier Science Publ. Ltd. NY. 159 Hoffman, P. C, Sievert, S. J., Shaver, R. D., Welch, D. A. and Combs, D. K. 1993. In situ dry matter, protein and fiber degradation of perennial forages. J. Dairy Sci. 76: 2632-2643. Howarth, R. E. 1975. A review of bloat in cattle. Can Vet. J. 16: 281-294. Howarth, R. E., Majak, W., Waldern, D. E., Brandt, S. A., Fesser, A. C, Goplen, B. P. and Spurr, D. T. 1977. Relationships between ruminant bloat and the chemical composition of alfalfa herbage. I. Nitrogen and protein fractions. Can. J. Anim. Sci. 57: 345-357. Howarth, R. E., Goplen, B. P., Fesser, A. C. and Brandt, S. A. 1978a. A possible role for leaf cell rupture in legume pasture bloat. Crop Sci. 18: 129-132. Howarth, R. E., Goplen, B. P., Majak, W., Waldern, D. E. and Chaplin, R. K. 1978b. Chloroplast membranes in the etiology of pasture bloat. Can. J. Anim. Sci., Abstracts. 58: 820. Howarth, R. E., Goplen, B. P., Fay, J. P. and K.-J. Cheng. 1979. Digestion of bloat-causing and bloat-safe legumes. Ann. Rech. Vet. 10: 332-334. Howarth, R. E., Goplen, B. P., Brandt, S. A. and Cheng, K.-J. 1982. Disruption of leaf tissues by rumen microorganisms: an approach to breeding bloat-safe forage legumes. Crop Sci. 22: 564-568. Howarth, R. E., Cheng, K.-J., Majak, W. and Costerton, J. W. 1984. Ruminant bloat. Pages 516-539 in Milligan, L. P., Grovum, W. L. and Dobson, A. eds. Control of Digestion and Metabolism in Ruminants. Proceedings of the 6th International Symposium in Ruminant Physiology, September 10-14, Banff, AB. Prentice-Hall, Inc. Englewood Cliffs, NJ. Howarth, R. E., Chaplin, R. K., Cheng, K.-J., Goplen, B. P., Hall, J. W., Hironaka, R., Majak, W. and Radostits, 0.1991. Bloat in cattle. Pages 1-34 in Agric. Canada Publication 1858/E. Ottawa, ON. Jackson, F. S., McNabb, W. C , Barry, T. N., Foo, Y. L. and Peters, J. S. 1996. The condensed tannin content of a range of subtropical and temperate forages and the reactivity of condenses tannin with ribulose-1,5-bis-phosphate carboxylase (rubisco) protein. J. Sci. Food Agric. 72: 483-492. 160 Jacobson, D. R., Lindahl, I. L., McNeill, J. J., Shaw, J. C, Doetsch, R. N. and Davis, R. E. 1957. Feedlot bloat studies. II. Physical factors involved in the etiology of frothy bloat. J. Anim. Sci. 16: 515-524. Johns A. T. 1954. Bloat in cattle on red clover. I. N. Z. J. Sci. Tech. 36A: 289-320. Jones, W. T., Anderson, L. B. and Ross, M. D. 1973. Bloat in cattle. XXXIX. Detection of protein precipitants (flavolans) in legumes. N. Z. J. Agric. Res. 16: 441-446. Kalu, B. A. and Fick, G. W. 1981. Quantifying morphological development of alfalfa for studies of herbage quality. Crop Science. 21: 267-271. Kalu, B .A. and Fick, G. W. 1983. Morphological stage of development as a predictor of alfalfa herbage quality. Crop Sci. 23: 1167-1172. Kudo, H., Cheng, K.-J., Hanna, M. R., Howarth, R. E., Goplen, B. P. and Costerton, J. W. 1985. Ruminal digestion of alfalfa strains selected for slow and fast initial rates of digestion. Can J. Anim. Sci. 65: 157-161. Ledgard, S. F., O'Connor, M. B., Carruthers, V. R. and Brier, G. J. 1990. Variability in pasture nitrogen fractions and the relationship to bloat. N. Z. J. Agric. Res. 33: 237-242. Lees, G. W. 1984. Cuticle and cell wall thickness: relation to mechanical strength of whole leaves and isolated cells from some forage legumes. Crop Sci. 24:1077-1081. Lees, G. L., Howarth, R. E., Goplen, B. P. and Fesser, A. C. 1981. Mechanical disruption of leaf tissues and cells in some bloat-causing and bloat-safe forage legumes. Crop Sci. 21: 444-448. Lees, G. L., Howarth, R. E., and Goplen, B. P. 1982. Morphological characteristics of leaves from some legume forages: relation to digestion and mechanical strength. Can. J. Bot. 60:2126-2132. Lindahl, I. L., Davis, R. E., Jacobson, D. R. and Shaw, J. C. 1957. Feedlot bloat studies. I. Animal and dietary factors. J. Anim. Sci. 16: 165-178. Lippke, H., Reaves, J. L. and Jacobson, N. L. 1972. Rumen pressures associated with the scores of a bloat severity scale. J. Anim. Sci. 34: 171-175. 161 Lorenz, R.'J. 1982. Alfalfa in western grazing management systems. Pages 1-3 in Lorenz, R. J., Ries, R. E., Cooper, C. S., Townsend, C. E. and Rumbaugh, M . D. eds. Alfalfa for dryland grazing. Information Bulletin No. 444. Agric. Res. Serv., USDA, Washington, DC. MacAdam, J. W., Hill, J. and Whitesides, R. E. 1995. Frost induced movement of potassium to the apoplast may increase bloat in ruminants grazing alfalfa. Crop Sci. 35: 1133-1136. MacAlpine, N. D., Engstrom, D. F., Kirtz, J., and Cooke, S. 1997. Resources for beef industry expansion in Alberta. Alberta Agric, Food and Rural Develop. Animal Industry Division Report. Edmonton, AB. Majak, W. and Hall, J. W. 1990. Sodium and potassium concentrations in ruminal contents after feeding bloat-inducing alfalfa to cattle. Can. J. Anim. Sci. 70: 235-241. Majak, W., Howarth, R. E., Fesser, A. C, Goplen, B. P. and Pedersen, M. W. 1980. Relationships between ruminant bloat and the composition of alfalfa herbage: II. Saponins. Can. J. Anim. Sci. 60: 699-708. Majak, W., Howarth, R. E., Cheng, K.-J. and Hall, J. W. 1983. Rumen conditions that predispose cattle to pasture bloat. J. Dairy Sci. 66: 1683-1688. Majak, W., Howarth, R. E. and Narasimhalu, P. 1985. Chlorophyll and protein levels in bovine rumen fluid in relation to alfalfa pasture bloat. Can. J. Anim. Sci. 65: 147-156. Majak, W., Hall, J. W., Rode, L. M. and Kalnin, C. M. 1986a. Rumen clearance rates in relation to the occurrence of alfalfa bloat in cattle. 1. Passage of water-soluble markers. J. Dairy Sci. 69: 1560-1567. Majak, W., Hall, J. W. and Howarth, R. E. 1986b. The distribution of chlorophyll in rumen contents and the onset of bloat in cattle. Can J. Anim. Sci. 66: 97-102. Majak, W., Hall, J. W. and McCaughey, W. P. 1995. Pasture management strategies for reducing the risk of legume bloat in cattle. J. Anim. Sci. 73: 1493-1498. Mangan, J. L. 1959. Bloat in cattle. XI. The foaming properties of proteins, saponins and rumen liquor. N . Z. J. Agr. Res. 2:47-61. 162 Mathison, G. W., Soofi-Siawash, R., Klits, P. T., Okine, E. K. and Sedgwick, G. 1999. Degradability of alfalfa saponins in the digestive tract of sheep and their rate of accumulation in rumen fluid. Can. J. Anim. Sci. 79:315-319. McArthur, J. M., Miltimore, J. E. and Pratt, M. J. 1964. Bloat investigations. The foam stabilizing protein of alfalfa. Can. J. Anim. Sci. 44: 200-206. McDonald, 1.1981. A revised model for the estimation of protein degradability in the rumen. J. Agric. Sci. (Camb.) 96:251-252. McDougall, E. 1.1948. Studies on ruminant saliva. 1. The composition and output of sheep's saliva. Biochem J. 43: 99-109. McMahon, L. R., Majak, W., McAllister, T. A., Hall, J. W., Jones, G. A., Popp, J. D. and Cheng, K.-J. 1999. Effect of sainfoin on in vitro digestion of fresh alfalfa and bloat in steers. Can. J. Anim. Sci. 79: 203-212. McMahon, L. R., McAllister, T. A., Berg, B. P., Majak, W., Acharya, S. N. , Popp, J. D., Coulman, B. E., Wang, Y. and Cheng, K.-J. 2000. A review of the effects of forage condensed tannins on ruminal fermentation and bloat in grazing cattle. Can. J. Plant Sci. Accepted. Menke, K. H., Raab, L., Salewski, A., Steingass, H., Fritz, D. and Schneider, W. 1979. The estimation of the digestibility and metabolizable energy content of ruminant feedingstuffs from the gas production when they are incubated with rumen liquor in vitro. J. Agric. Sci. (Camb.) 93: 217-222. Michaud, R., Lehman, W. F., and Rumbaugh, M. D. 1988. World distribution and historical development. Page 25-91 in Hanson, A.A.,Barnes, D. K., and. Hill, R. R., Jr. eds. Alfalfa and alfalfa improvement. Agronomy Mono. No. 29. Amer. Soc. Agron., Madison, WI. Miltimore, J. E., McArthur, J. M., Mason, J. L. and Ashby, D. L. 1970. Bloat investigations. The threshold fraction 1 (18s) protein concentration for bloat and relationships between bloat and lipid, tannin, Ca, Mg, Ni, and Zn concentrations in alfalfa. Can J. Anim. Sci. 50: 61-68. 163 Moate, P. J., Clarke, T., Davis, L. H. and Laby, R. H. 1997. Rumen gases and bloat in grazing dairy cows. J. Agric. Sci. (Camb.) 129: 459-469. Moghaddam, P. R. and Wilman, D. 1998. Cell wall thickness and cell dimensions in plant parts of eight forage species. J. Agric. Sci. (Camb.) 131: 59-67. Mueller-Dombois, D. and Ellenberg, H. 1974. Aims and methods of vegetation ecology. John Wiley and Sons. Mueller, S. C, and Fick, G. W. 1989. Converting alfalfa development measurements from mean stage by count to mean stage by weight. Crop Sci. 29:821-823. Okine, E. K., Mathison, G. W. and Hardin, R. T. 1989. Relations between passage rates of rumen fluid and particulate matter and foam production in rumen contents of cattle fed on different diets ad lib. Br. J.Nutr. 61:387-395. Okine, E. K. and Mathison, G. W. 1991. Effects of feed intake on particle distribution, passage of digesta and extent of digestion in the gastrointestinal tract of cattle. J. Anim. Sci. 69: 3435-3445. Popp, J. D., McCaughey, W. P. and Cohen, R. D. H. 1997a. Effect of grazing system and stocking rate on the productivity, botanical composition and soil surface characteristics of alfalfa-grass pastures. Can. J. Anim. Sci. 77: 669-676. Popp, J. D., McCaughey, W. P. and Cohen, R. D. H. 1997b. Effect of grazing system, stocking rate and season of use on herbage intake and grazing behaviour of stacker cattle grazing alfalfa-grass pastures. Can. J. Anim. Sci. 77: 677-682. Pressey, R., Synhorst, S. H., Bertram, J., Allen, R. S. and Jacobson, N. L. 1963. Foaming properties of alfalfa and their relationship to bloat. J. Anim. Sci. 22: 970-978. Ramos, A. and Tennessen, T. 1992. Effect of previous grazing experience on the grazing behaviour of lambs. Appl. Anim. Behav. Sci. 33: 43-52. Reid, C. S. W. 1960. Bloat: the foam hypothesis. Paper 10B/4. Pages 668-671 in Proceedings of the 8th International Grassland Congress. Univ. of Reading, Berkshire, Engl. 164 Reid, C. S. W. and Johns, A. T. 1957. Bloat in cattle on red clover. III. Treatment and prevention with anti-foaming agents. N . Z. J. Sci. Tech. 38A: 908-927. Russell, K. D., Hironaka, R. and Wilson, D. B. 1981. An economic analysis of alternative management techniques for beef production on irrigated pastures in Alberta. Can. Farm Econ. 16: 1-7. Sant, F. I. and Wilson, D. 1982. Use of a cellulolytic enzyme digestion technique to distinguish bloat-causing from non-bloat-causing legumes and to select for speed of mesophyll cell wall disintegration in red clover. J. Agric. Sci. (Camb.) 98: 99-102. Shenk, J. S. and Elliot, F. C. 1970. Two cycles of directional selection for improved nutritive value of alfalfa. Crop Sci. 10: 710-712. Shenk, J. S. and Elliott, F. C. 1971. Plant compositional changes resulting from two cycles of directional selection for nutritive value in alfalfa. Crop Sci. 11: 521-524. Smith, D. 1964. Chemical composition of herbage with advance in maturity of alfalfa, medium red clover, ladino clover and birdsfoot trefoil. College of Agric, Univ. of Wise Expt. Stat. Res. Rpt. 16. Smith, D. 1970. Yield and chemical composition of leaves and stems of alfalfa at intervals up the shoots. J. Agr. Food Chem. 18: 652-656. Smith, S. R., Jr., Bouton, J. H. and Hoveland, CS. 1989. Alfalfa persistence and regrowth potential under continuous grazing. Agron. J. 81: 960-965. Sokal, R. R. and Rohlf, F. J. 1995. Biometry. (3rd Ed.). 887 pp. W. H. Freeman & Co., New York. Stanford, K., Wang, Y., Berg, B. P., Majak, W., McCartney, D. H., Baron, V. and McAllister, T. A. 2000. Effects of alcohol ethoxylate andpluronic detergents on the development of pasture bloat in cattle and sheep. J. Dairy Sci. Submitted. Stanier, R. Y., Doudoroff, M. and Adelberg, E. A. 1970. Chapter 9: Microbial growth. The microbial world. 3 rd Ed. Prentice Hall, Inc. NJ. Stifel, F. B., Vetter, R. L., Allen, R. S. and Horner, H. T. 1968. Chemical and ultrastructural relationships between alfalfa leaf chloroplasts and bloat. Phytochem. 7: 355-364. 165 Taminga, S. 1993. Influence of feeding management on ruminant fiber digestibility. Pages 571-602 in Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J. (eds). Forage cell wall structure and digestibility. Amer. Soc. Agron. Madison, WI. Terashima, N., Fukushima, L-F. He and Takabe, K. 1993. Comprehensive model of the lignified plant cell wall. Pages 247-270 in Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J. (eds). Forage cell wall structure and digestibility. Amer. Soc. Agron. Madison, WI. Thomas, J. W., McCampbell, J. L., Tesar, M. B. and Elliot, F. C. 1968. Improved nutritive value of alfalfa by crossing and selection based on an in vitro method. Abstr. J. Anim. Sci. 27: 1783-1784. Troelsen, J. E. and Campbell, J. B. 1969. The effect of maturity and leafiness on the intake and digestibility of alfalfas and grasses fed to sheep. J. Agric. Sci. (Camb.) 73: 145-154. Van Soest, P. J. 1993. Cell wall matrix interactions and degradation. Pages 377-395 in Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J. (eds). Forage cell wall structure and digestibility. Amer. Soc. Agron. Madison, WI. Van Soest, P. J. 1994. Nutritional ecology of the ruminant. 2 n d Ed. Cornell Univ. Press. NY. Van Soest, P. J. and Robertson, J. B. 1980. Systems of analysis for evaluating fibrous feeds. Pages 49-60 in Pidgen, W. J., Balch, C. C. and Graham, M . eds. Standardization of analytical methodology for feeds. IDRC-134e, Int. Dev. Res. Centre, Ottawa, ON. Van Soest, P. J. and Robertson, J. B. and Lewis, B. A. 1991. Methods for dietary fiber; neutral detergent fiber, and non-starch polysaccarides in relation to animal nutrition. J. Dairy Sci. 74:3583-3597. Waghorn, G. C. and Jones, W. T. 1989. Bloat in cattle. 46. Potential of dock (Rumex obtusifolius) as an antibloat agent for cattle. N . Z. J. Agric. Res. 32: 227-235. Waghorn, G. C. 1991. Bloat in cattle. 47. Relationships between intra-ruminal pressure, distension and the volume of gas used to simulate bloat in cows. N . Z. J. Agric. Res. 34: 213-220. Waghorn, G. C , Jones, W. T., Shelton, I. D. and McNabb, W. C. 1990. Condensed tannins and the nutritive value of herbage. Proc. N. Z. Grass. Assoc. 51: 171-176. 166 Waghorn, G. C. and Stafford, K. J. 1993. Gas production and nitrogen digestion by rumen microbes from deer and sheep. N. Z. J. Agric. Res. 36: 493-497. Walton, P. D. 1983. Production and management of cultivated forages. Reston Publ. Co. Reston, VA. Walton, P. D., Martinez, R. and Bailey, A. W. 1981. A comparison of continuous and rotational grazing. J. Range Manage. 34: 19-21. Warner, G. 1997. Grazing beef cattle on irrigated alfalfa. Page 65-68 in Johnson, J., ed. Proceedings of the 2nd National Alfalfa Grazing Conference, Omaha, NE, Aug 16. Certified Alfalfa Seed Council, Davis, CA. Wilkinson, L., Hill, M. and Vang, E. 1992. SYSTAT, statistics. Version 5.2 Edition. Evanston, II. SYSTAT, Inc. 724 p. Wilman, D. and Ahmad, N . 1999. In vitro digestibility, neutral detergent fibre, lignin and cell wall thickness in plant parts of three forage species. J. Agric. Sci. (Camb.) 133: 103-108. Wilman, D. and Moghaddam, P. R. 1998. In vitro digestibility and neutral detergent fibre and lignin contents of plant parts of nine forage species. J. Agric. Sci. (Camb.) 131: 51-58. Wilson, J. R. 1993. Organization of forage plant tissues. Pages 1-32 in Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J. (eds). Forage cell wall structure and digestibility. Amer. Soc. Agron. Madison, WI. Wilson, J. R. and Mertens, D. R. 1995. Cell wall accessibility and cell structure limitations to microbial digestion. Crop Sci. 35: 251-259. Wolf, E. C. and Lazenby, A. 1972. Bloat incidence and liveweight gain in beef cattle on pastures containing different proportions of white clover (Trifolium repens). Aust. J. Exptl. Agric. Anim. Husb. 12:119-125. Zwietering, M. H., Jongenburger, I., Rombouts, F. M. and van't Reit, K. 1990. Modeling of the bacterial growth curve. Appl. Envir. Microb. 56:1875-1881. 167 Appendix I EXPONENTIAL MODELS FOR KINETIC ANALYSIS Introduction A mathematical model is a useful tool for analysing rates of growth or decay because only a few parameters are needed to adequately describe the kinetics of fermentation or to compare treatments and yields with other fermentations. One of the simplest models is a first-order, linear model. Volumes have been written on ways to analyse or transform data into a linear conformation (Neter and Wasserman 1974). However, simple nonlinear models appear to give more elegant and intuitive solutions for many phenomena in natural systems (May 1976). Analysis and Interpretation Nonlinear exponential models are frequently used in the analysis of bacterial cultures because they simulate the sigmoidal pattern typically observed during the growth of a culture or the degradation of a medium (Beuvink and Kogut 1993). The output of by-products, the disappearance of a substrate or the yield of microbial cells are described as functions of time, the independent variable, and rates of change as shaping parameters. Because of their complexity, the shaping parameters in some models are difficult to reconcile with biological phenomena and tedious to compute (Zwietering et al. 1990). Modelers using more than four parameters are often greeted with 'stack overflow' or 'failure to converge' errors when attempting to fit complicated models. For example, Fisher et al. (1989) opined that the difference between a 'rate' parameter and an 'extent' parameter could lead to erroneous conclusions during trait selection for plant breeding. Van Milgen et al. (1993a) stacked a three-level compartmental model of rumen dynamics over models of substrate degradation and microbial mass accumulation but were forced to conclude that still 168 more complex statistical methods were needed. They also encountered the proverbial 'failure to converge' error and the effect of gaining an excellent but erroneous fit of the model to the data set by adding too many parameters to the model (van Milgen et al. 1993b). Groot et al. (1996) took a different approach. Instead of employing the exponent as a rate parameter, it was used to vary a simple, linear rate equation from a first-order model to a third-order one. Each increment in the exponent was defined as a phase and the phases were stacked by summation. The process was similar to the creation of a polynomial. (The phases of the Groot model are full models over all time periods which is a different definition than the usual definition of a phase in fermentation, a narrow time period within one model.) The first two parameters of the Groot et al. (1996) model estimate the asymptote and the half-life of the function. Rates of change were calculated by differentiating the first or second derivatives of each phase in the model. The third parameter, the exponent of the model, was used to shape the mathematical function. Subsequent tests of the Groot model have not been positive (Getachew et al. 1998) probably because at its core it is a first order, linear rate equation. Linear functions do not provide an adequate level of fit for most non-linear data sets (Zwietering et al. 1990). Independent tests of first order models have proven that the best fits of mathematical equations to data from bacterial fermentations are modifications of the Gompertz growth function with 3 parameters (Zwietering et al. 1990; Beuvink and Kogut 1993). The Gompertz models have been tested and are generally advocated for both in vivo dry matter disappearance from nylon bags and in vitro gas production (Schofield et al. 1994; Van Milgen and Baumont 1995; Lavrenfiic et al. 1997; Susmel et al. 1999). Generic Equations A model is just a yardstick for measuring functional response. Each model has a level of utility, an ease of use, some historical precedent and elements in common with other models, all of which justifies their 169 continued relevancy. This is the case for the exponential model (Equation 1), which has been used to describe dry matter disappearance and gas production curves for 20 years (0rskov and McDonald 1979; y=a+(b—); Ct e y=a- ,(A-c<) (1.1) Krishnamoorthy et al.1991; Blummel and 0rskov 1993; Dhanoa et al. 1995) and it is also true for the multitude of linear models (Equation 2) that have assisted in the interpretation of research over many y=a+bc-bt (1.2) more years (Scales et al. 1974; Nocek and English 1986; Fisher et al. 1989; Groot et al. 1996). These two equations are not among the best available for modelling non-linear relationships (Zwietering et al.1990; Beuvink 1993; Beuvink and Kogut 1993) but that does not mean that they should be discarded. They have been modified, though not without criticism, and are used in some laboratories as standard practice (Dhanoa et al. 1995; Sandoval-Castro 1997; Orskov 1997; France et al. 1997). The Gompertz Growth Equation The problem with the logistic equation (Equation 3) and the original Gompertz growth equation a (Equation 4) is a problem common to all the other equations, the parameters need to be interpreted. This is no handicap for a mathematician because the three parameters a, b, and c, are just abstract constants or shape variables used to define the function of the dependent variable, y, on the independent variable, t. 170 (1.4) Interpretation is a concern for animal scientists because a shape variable gives them no real information about the biological phenomena it may represent. So the models needed 'reparameterization' to create parameters with an easily interpretable biological meaning. Zwietering et al. (1990) accomplished this by defining each parameter as a relevant point in three of the phases of bacterial growth, the lag, the exponential growth phase and the stationary phase. Equation 5 is their modified Gompertz model. r=/jeXp(-exp[^l(X-o+i]) - Y=—r.— n , . The three parameters are: A, the asymptote, which occurs during the stationary phase; fim, the maximum specific rate, which occurs during the exponential growth phase; and X, the lag time, which occurs at the end of the lag phase. If the exponential expression is substituted with c H ^ O O + l] (1.6) A then Y=A(e V ) (1-7) and the first derivative becomes ^ =<!!,.)(«-')(."> (1-8) 171 We now have two equations (Equations 7 and 8) that can be used to calculate the extent and rate of growth or degradation for any time point or over any period. When the equations are used as non-linear regression models, the parameters will give relevant biologically interpretable information about a growth or degradation process. Further, instead of adding parameters to the model to determine the intercepts, as was necessary in some exponential and linear models (0rskov and McDonald 1979; McDonald 1981; Krishnamoorthy et al.1991; Bliimmel and 0rskov 1993; Dhanoa et al. 1995), they can be calculated. What remains is a required substitution to put the equations into a format that is easily handled by a statistical or spreadsheet program. Substituting a for A, c for fim, and / for X, Equation (5) becomes Y=a *exp( -exp(((c *exp(l ))/a) *(/ -/) +1)) (1.9) and Equation (8) expands to — =(c *exp(l)) *(exp( -exp(((c *exp(l))/a) *(/-/)+1))) *(exp(((c *exp(l))/a) +1)) (1.10) dt The models in Chapters 2 and 3 are based on Equations (9) and (10). References Beuvink, J. M. W. 1993. Measuring and modelling in-vitro gas production kinetics to evaluate ruminal fermentation of feedstuffs. PhD Thesis. Landbouwuniversiteit te Wageningen, The Netherlands. 172 Beuvink, J. M. W. and Kogut, J. 1993. Modelling gas production kinetics of grass silages incubated with buffered ruminal fluid. J. Anim. Sci. 71:1041-1046. Bliimmel, M. and 0rskov, E. R. 1993. Comparison of in vitro gas production and nylon bag degradability of roughages in predicting feed intake in cattle. Anim. Feed. Sci. Technol. 40: 109-119. Dhanoa, M. S., France, J., Siddons, R. C , Lopez, S. and Buchanan-Smith, J. G. 1995. A non-linear compartmental model to describe forage degradation kinetics during incubation in polyester bags in the rumen. Brit. J. Nutr. 73: 3-15. Fisher, D. S., Burns, J. C. and Pond, K. R. 1989. Kinetics of in vitro cell-wall disappearance and in vivo digestion. Agron. J. 81: 25-33. France, J., Lopez, S., Dijkstra, J. and Dhanoa, M. S. 1997. Particulate matter loss and the polyester-bag method-reply by France. Brit. J. Nutr. 78:1033-1037. Getachew, G., Bliimmel, M., Makkar, H. P. S. and Becker, K. 1998. In vitro gas measuring techniques for assessment of nutritional quality of feeds: a review. Anim. Feed Sci. Technol. 72:261-281. Groot, J. C. J., Cone, J. W., Williams, B. A., Debersaques, F. M. A. and Lantinga, E. A. 1996. Multiphasic analysis of gas production kinetics for in vitro fermentation of ruminant feeds. Anim. Feed Sci. Technol. 64:77-89. Krishnamoorthy, U., Soller, H., Steingass, H. and Menke, K. H. 1991. A comparative study on rumen fermentation of energy supplements in vitro. J. Anim. Physiol. Anim. Nutr. 65: 28-35. Lavrencic, A., Stephanon, B. and Susmel, P. 1997. An evaluation of the Gompertz model in degradability studies of forage chemical components. Anim. Sci. 64: 423-431. May, R. M. 1976. Simple mathematical models with very complicated dynamics. Nature 261: 459-467. McDonald, 1.1981. A revised model for the estimation of protein degradability in the rumen. J. Agric. Sci. (Camb.) 96: 251-252. Neter, J. and Wasserman, W. 1974. Applied linear statistical models. Irwin-Dorsey Ltd., ON 173 Nocek, J. E. and English, J. E. 1986. In situ digestion kinetics: evaluation of rate determination procedure. J. Dairy Sci. 69: 77-87. 0rskov, E. R. 1997. Particulate matter loss and the polyester-bag method-reply by 0rskov. Brit. J. Nutr. 78: 1032-1033. 0rskov, E. R. and McDonald, 1.1979. The estimation of protein degradability in the rumen from incubation measurements weighted according to rate of passage. J. Agric. Sci. (Camb.) 92:499-503 Sandoval-Castro, C. A. 1997. Particulate matter loss and the polyester-bag method. Brit. J. Nutr. 78: 1031-1032. Scales, G. H., Streeter, C. L., Denham, A. H. and Ward, G. M. 1974. A comparison of indirect methods of predicting in vivo digestibility of grazed forage. J. Anim. Sci. 38: 192-199. Schofield, P., Pitt, R. E. and Pell, A. N. 1994. Kinetics of fibre digestion from in vitro gas production. J. Anim. Sci. 72:2980-2991. Susmel, P., Spanghero, M. and Stephon, B. 1999. Interpretation of rumen degradability of concentrate feeds with a Gompertz model. Anim. Feed Sci. Technol. 79: 223-237. Van Milgen, J. and Baumont, R. 1995. Models based on variable fractional digestion rates to describe ruminal in situ digestion. Brit. J. Nutr. 73: 793-807. Van Milgen, J., Berger, L. and Murphy, M. R. 1993a. Digestion kinetics of alfalfa and wheat straw assuming heterogeneity of the potentially digestible fraction. J. Anim. Sci. 71: 1917-1923. Van Milgen, J., Berger, L. and Murphy, M. R. 1993b. An integrated, dynamic model of feed hydration and digestion, and subsequent bacterial mass accumulation in the rumen. Brit. J. Nutr. 70:471-483. Zwietering, M. H., Jongenburger, I., Rombouts, F. M. and van't Reit, K. 1990. Modelling of the bacterial growth curve. Appl. Envir. Microb. 56:1875-1881. 174 II DMD AND GP TABLES List of Tables II. 1 Dry matter disappearance of fresh cuttings during in vivo digestion of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 176 11.2 Dry matter disappearance of fresh masticate during in vivo digestion of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 177 11.3 Rates of gas production during in vitro digestion of the LIRD cultivar, AC Grazeland and its precursors, Anchor, Beaver, Kane and Vernal 178 11.4 Rates of gas production during in vitro digestion of the alfalfa cultivars, LIRD-1 and AC Grazeland (LIRD-4) compared to the standard alfalfa, Beaver 179 11.5 Rates of gas production during in vitro digestion of fresh, whole leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 180 11.6 Rates of gas production during in vitro digestion of fresh, cut leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 181 11.7 Rates of gas production during in vitro digestion of fresh, perforated leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 182 11.8 Rates of gas production during in vitro digestion of fresh, crushed leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 183 11.9 Rates of gas production during in vitro digestion of dried, ground leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 184 11.10 Rates of gas production during in vitro digestion of leaves from late vegetative stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 185 II. 11 Rates of gas production during in vitro digestion of leaves from early bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 186 11.12 Rates of gas production during in vitro digestion of leaves from late bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 187 11.13 Rates of gas production during in vitro digestion of late vegetative stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 188 11.14 Rates of gas production during in vitro digestion of early bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 189 11.15 Rates of gas production during in vitro digestion of late bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver 190 175 Table II.1 Dry matter disappearance of fresh cuttings during in vivo digestion of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Trial 4 2 Trial 5 y Beaver AC Grazeland Beaver AC Grazeland Time(h) D M D % " SE DMD% SE DMD% SE DMD% SE 0 10.86 1.071 8.26 2.233 6.18 0.209 6.33 0.259 4 26.60 1.469 27.65 1.264 27.03 1.321 24.15 1.939 6 34.86 2.177 35.52 1.925 38.91" 1.382 36.03" 1.766 8 36.16" 2.117 41.73" 1.679 46.48 0.921 45.99 1.343 10 40.07 0.555 42.56 2.158 65.92 2.400 65.07 0.666 12 71.80 1.204 72.09 1.108 16 67.88 1.548 70.06 1.710 24 80.33 0.927 81.18 2.272 87.57 1.009 87.84 0.676 48 84.86 3.387 88.02 1.556 90.17 0.956 88.83 0.993 72 90.90 0.359 90.21 0.472 2 Trial 4; fresh cuttings, early blossom stage of development; n=42 for each cultivar y Trial 5; fresh cuttings, early bud stage of development; n=70 for each cultivar " Mean Dry Matter Disappearance as a proportion of total dry matter "•" t-test of independent sample means. Trial means in the same row followed by the same or no letter are not significantly different (P>.05) 176 Table IL2 Dry matter disappearance of fresh masticate during in vivo digestion of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Trials 1&2 2 Trials 3&4 y Beaver AC Grazeland Beaver AC Grazeland Time(h) DMD% x SE DMD% SE DMD% SE DMD% SE 0 79.78 1.624 76.09 0.689 7 2 . 9 9 a 0.768 77.23" 0.540 4 86.72 1.099 86.28 0.449 85.08" 0.325 88.02" 0.326 8 91.60 0.835 90.00 0.506 87.59 0.665 88.44 1.058 16 94.03 0.458 93.16 0.190 20 89.75a 0.782 92.21" 0.687 24 93.65 0.384 48 95.15 0.385 95.11 0.434 91.91" 0.826 95.60" 0.184 z Trials 1&2; fresh masticate, early bud stage of development; n=52 for each cultivar y Trials 3&4; fresh masticate, early bloom stage of development; n=40 for BVR, n=54 for ACG x Mean Dry Matter Disappearance as a proportion of total dry matter a'" t-test of independent sample means. Trial means in the same row followed by the same or no letter are not significantly different (P>.05) 177 Table II.3 Rates of gas production during in vitro digestion of the LIRD cultivar, AC Grazeland and its precursors, Anchor, Beaver, Kane and Vernal Cultivar Time (h) AC Grazeland Anchor Beaver Kane Vernal Cumulative Gas Production2 1 0.37^0.027 y 0.37a ±0.038 0.28ab ±0.028 0.3 l a b ±0.029 0.22" ±0.050 2 0.95a ±0.047 0.87ab ±0.054 0.88ab ±0.034 0.83ab ±0.048 0.74b ±0.055 3 1.60a ±0.061 1.54ab ±0.062 1.57ab ±0.049 1.47ab ±0.042 1.38" ±0.065 4 2.22 ±0.076 2.18 ±0.076 2.16 ±0.067 2.11 ±0.053 1.96 ±0.069 5 2.77 ±0.102 2.72 ±0.104 2.63 ±0.095 2.61 ±0.067 2.48 ±0.091 6 3.32 ±0.124 3.33 ±0.146 3.19 ±0.102 3.16 ±0.095 3.14±0.136 8 4.98 ±0.167 4.75 ±0.146 4.72 ±0.127 4.64 ±0.153 4.75 ±0.170 12 7.55±0.215 7.66 ±0.250 7.42 ±0.127 7.36 ±0.204 7.59 ±0.200 24 12.39 ±0.354 13.34 ±0.288 12.67 ±0.063 12.95 ±0.165 12.18 ±0.416 48 15.54 ±0.043 15.87 ±0.019 15.76 ±0.176 15.89 ±0.069 15.36 ±0.200 Rate of Gas Production* 1 0.37a ±0.027 0.37a ±0.038 0.28ab ±0.028 0.3 l a b ±0.029 0.22b±0.049 2 0.58 ±0.035 0.51 ±0.031 0.60 ±0.028 0.52 ±0.029 0.51 ±0.030 3 0.64 ±0.025 0.66 ±0.024 0.69 ±0.022 0.63 ±0.053 0.65 ±0.030 4 0.61 ±0.021 0.64 ±0.024 0.59 ±0.021 0.62 ±0.021 0.58 ±0.022 5 0.50 ±0.023 0.49 ±0.020 0.48 ±0.017 0.46 ±0.016 0.49 ±0.019 6 0.61 ±0.039 0.61 ±0.051 0.56 ±0.029 0.56 ±0.033 0.66 ±0.054 8 0.79 ±0.015 0.75 ±0.040 0.76 ±0.018 0.74 ±0.030 0.80 ±0.021 12 0.64a ±0.007 0.67ab ±0.020 0.66ab ±0.013 0.67ab ±0.006 0.71" ±0.025 24 0.39 ±0.031 0.45 ±0.012 0.43 ±0.015 0.46 ±0.011 0.38 ±0.031 48 0.128 ±0.002 0.126 ±0.002 0.127 ±0.007 0.133 ±0.001 0.136 ±0.001 z Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1 y Mean ± Standard Error; n=140 for each cultivar * Mean rate of GP hour"1 for that time interval = (ml 100 mg DM"1) hour'1 a b t-test of independent sample means. Means in the same row followed by the same or no letter are not significantly different (P>.05) 178 Table II.4 Rates of gas production during in vitro digestion of the alfalfa cultivars, LIRD-1 and AC Grazeland (LIRD-4) compared to the standard alfalfa, Beaver Cultivar Time(h) AC Grazeland LIRD-1 Beaver Cumulative Gas Production2 0.5 1.86 ±0.042 y 1.89 ±0.134 1.78 ±0.139 2 5.21 ±0.050 5.60 ±0.289 5.30 ±0.232 4 9.88 ±0.189 10.71 ±0.319 10.29 ±0.319 8 15.61 ±0.150 16.51 ±0.500 15.88 ±0.504 12 17.61 ±0.208 18.49 ±0.618 17.96 ±0.544 24 20.39 ±0.229 21.15 ±0.698 20.54 ±0.588 48 Rate of Gas Production" 21.70 ±0.521 0.5 3.71 ±0.084 3.78 ±0.267 3.56 ±0.278 2 2.24 ±0.020 2.47 ±0.127 2.34 ±0.081 4 2.33a ±0.073 2.56" ±0.027 2.50ab ±0.059 8 1.43 ±0.022 1.45 ±0.047 1.40 ±0.062 12 0.50 ±0.019 0.50 ±0.030 0.52 ±0.016 24 0.23 ±0.003 0.22 ±0.008 0.22 ±0.005 48 0.048 ±0.004 z Mean accumulated gas produced per unit of DM = ml 100 mg DM" 1 y Mean ± Standard Error; n=28 for ACG and LIRD-1; n=32 for BVR x Mean rate of GP hour"1 for that time interval = (ml 100 mg DM' 1) hour"1 a , b t-test of independent sample means. Means in the same row followed by the same or no letter are not significantly different (P>.05) 179 Table II.5 Rates of gas production during in vitro digestion of fresh, whole leaves of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Cumulative Gas Production2 Rate of Gas Production y Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 0.74 0.003 0.93 0.110 0.74 0.003 0.93 0.110 2 1.08 0.086 1.35 0.133 0.33 0.084 0.43 0.028 3 1.49 0.164 1.56 0.087 0.41 0.142 0.20 0.053 4 1.66 0.171 1.77 0.129 0.17 0.166 0.21 0.067 5 1.90 0.172 2.19 0.139 0.25a 0.001 0.43b 0.028 6 2.15 0.227 2.40 0.190 0.25 0.084 0.21 0.067 7 2.73 0.229 2.75 0.180 0.58b 0.003 0.35a 0.044 8 3.15 0.310 3.17 0.242 0.41 0.084 0.42 0.063 9 4.06 0.377 3.87 0.238 0.91 0.086 0.70 0.021 10 4.64 0.449 4.29 0.242 0.58 0.085 0.42 0.039 11 5.30 0.525 4.85 0.302 0.66 0.085 0.56 0.062 12 6.21 0.568 5.48 0.376 0.91 0.167 0.63 0.085 14 8.28 0.740 6.95 0.431 1.04 0.112 0.74 0.028 16 10.60 1.110 8.56 0.628 1.16 0.251 0.80 0.098 20 14.66 1.227 12.12 0.783 1.01 0.059 0.89 0.067 24 18.13 0.636 15.79 0.781 0.87 0.152 0.92 0.021 32 22.01 0.202 21.94 0.816 0.49a 0.076 0.77" 0.020 48 24.25 0.331 26.33 0.976 0.14 0.018 0.27 0.076 72 25.57a 0.346 28.95" 0.530 0.06a 0.003 0.11b 0.019 2 Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour"1 for that time interval = (ml 100 mg DM"1) hour"1 a,bt-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 180 Table II.6 Rates of gas production during in vitro digestion of fresh, cut leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 0.98 0.118 0.76 0.182 0.98 0.118 0.76 0.182 2 1.49 0.027 1.31 0.181 0.52 0.139 0.55 0.067 3 2.20 0.049 1.93 0.247 0.71 0.042 0.62 0.068 4 3.17 0.132 2.68 0.313 0.97 0.089 0.76 0.068 5 4.42 0.268 3.86 0.363 1.25 0.148 1.17 0.072 6 5.30 0.310 4.75 0.363 0.88 0.053 0.90 0.003 7 6.54 0.397 5.78 0.378 1.24 0.096 1.03 0.065 8 7.87 0.545 6.88 0.377 1.33 0.153 1.10 0.004 9 9.64 0.690 8.47 0.299 1.76 0.174 1.58 0.134 10 11.07 0.873 9.84 0.296 1.44 0.269 1.38 0.005 11 12.68 1.109 11.15 0.272 1.61 0.249 1.31 0.066 12 14.29 1.349 12.39 0.324 1.61 0.249 1.24 0.115 14 16.93 1.343 14.39 0.178 1.32 0.015 1.00 0.194 16 18.42 1.223 15.91 0.332 0.74 0.170 0.76 0.122 20 20.19 0.565 18.41 0.113 0.44 0.168 0.54 0.029 24 21.68 0.547 20.07 0.123 0.37 0.030 0.41 0.003 32 23.21" 0.199 22.17* 0.240 0.19 0.070 0.26 0.015 48 24.36 0.622 23.79 0.353 0.07 0.035 0.10 0.007 72 25.50 0.633 24.55 0.358 0.05" 0.001 0.03a 0.000 z Mean accumulated gas produced per unit of DM = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour"1 for that time period = (ml 100 mg DM"1) hour"1 a'"t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 181 Table II.7 Rates of gas production during in vitro digestion of fresh, perforated leaves of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 049 a 0.014 0.90b 0.062 0.49" 0.014 0.90" 0.062 2 1.13a 0.097 1.74" 0.171 0.65 0.089 0.84 0.144 3 2.02 0.029 2.65 0.258 0.88 0.124 0.91 0.086 4 3.22 0.208 4.04 0.347 1.20 0.185 1.39 0.112 5 4.83 0.244 5.64 0.398 1.61 0.039 1.60 0.098 6 6.44 0.363 7.52 0.504 1.61 0.122 1.88 0.110 7 8.37 0.524 9.32 0.502 1.93 0.165 1.80 0.056 8 10.15 0.591 11.26 0.439 1.77 0.107 1.94 0.074 9 12.01 0.504 13.21 0.475 1.86 0.111 1.94 0.066 10 13.47 0.381 14.87 0.473 1.46 0.124 1.67 0.056 11 14.52" 0.410 16.18" 0.372 1.05 0.065 1.31 0.105 12 15.65" 0.320 17.36b 0.267 1.14 0.103 1.17 0.158 14 17.19" 0.197 19.23b 0.275 0.77 0.063 0.94 0.041 16 18.24" 0.249 20.82b 0.333 0.53" 0.068 0.80" 0.032 20 19.62" 0.267 22.76" 0.416 0.34" 0.025 0.48" 0.038 24 20.59" 0.126 24.15" 0.415 0.24" 0.036 0.35" 0.008 32 21.66" 0.283 25.95" 0.475 0.13 0.048 0.23 0.011 48 22.79" 0.306 27.54" 0.490 0.07" 0.005 0.10" 0.006 72 24.00" 0.130 28.58" 0.551 0.05 0.008 0.04 0.003 z Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour"1 for that time period = (ml 100 mg DM"1) hour"1 "•"t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 182 Table II.8 Rates of gas production during in vitro digestion of fresh, crushed leaves of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Production y Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE , Mean SE Mean SE 1 1.19 0.117 1.15 0.067 1.19 0.117 1.15 0.067 2 2.52 0.188 2.53 0.237 1.34 0.083 1.38 0.221 3 4.19 0.225 3.60 0.085 1.66 0.059 1.07 0.242 4 6.39 0.332 5.93 0.529 2.20 0.118 2.33 0.561 5 8.43 0.294 7.96 0.762 2.04 0.088 2.03 0.234 6 10.55 0.278 9.84 0.872 2.12 0.083 1.88 0.111 7 12.29 0.283 11.93 0.987 1.73" 0.022 2.09" 0.116 8 13.47 0.245 13.30 0.897 1.19 0.044 1.37 0.100 9 15.13 0.259 15.03 0.932 1.65 0.027 1.73 0.035 10 16.15 0.234 16.47 0.961 1.03" 0.044 1.44" 0.029 11 16.94 0.265 17.69 0.914 0.79" 0.038 1.22" 0.049 12 17.81 0.253 18.84 0.792 0.87 0.046 1.15 0.124 14 19.22 0.287 20.57 0.827 0.71" 0.021 0.87" 0.018 16 20.24 0.349 22.08 0.667 0.51 0.035 0.75 0.081 20 21.50a 0.393 23.73b 0.512 0.31 0.026 0.41 0.040 24 22.36" 0.407 24.95b 0.463 0.22" 0.011 0.31" 0.012 32 23.61" 0.536 26.61b 0.386 0.16 0.017 0.21 0.012 48 24.80" 0.536 2797b 0.312 0.07 0.003 0.09 0.006 72 25.67" 0.479 28.54" 0.328 0.04 0.003 0.02 0.009 z Mean accumulated gas produced per unit of D M = ml 100 mg D M 1 ; n=60 for each cultivar y Mean rate of GP hour"1 for that time period = (ml 100 mg DM"1) hour"1 "•"t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 183 Table II.9 Rates of gas production during in vitro digestion of dried, ground leaves of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 1.70 0.061 1.69 0.168 1.70 0.061 1.69 0.168 2 4.24 0.118 4.18 0.149 2.55 0.139 2.48 0.181 3 6.65 0.022 6.33 0.249 2.41 0.098 2.15 0.102 4 8.93 0.022 8.68 0.367 2.28 0.036 2.35 0.126 5 11.09 0.021 10.70 0.356 2.15b 0.027 2.02a 0.012 6 13.18 0.105 12.98 0.351 2.09 0.084 2.28 0.009 7 14.94 0.151 15.01 0.403 1.76a 0.061 2.02" 0.054 8 16.17 0.199 16.37 0.403 1.24 0.100 1.37 0.063 9 17.41 0.258 17.67 0.383 1.24 0.076 1.30 0.246 10 18.27 0.217 18.85 0.469 0.85 0.241 1.18 0.285 11 19.05 0.283 19.57 0.459 0.78 0.193 0.72 0.057 12 19.77 0.320 20.35 0.479 0.72 0.074 0.78 0.021 14 20.94 0.330 21.59 0.448 0.59 0.008 0.62 0.015 16 21.85 0.302 22.57 0.442 0.46 0.027 0.49 0.031 20 23.09 0.393 23.80 0.412 0.31 0.032 0.31 0.008 24 24.00 0.420 24.78 0.438 0.23 0.019 0.24 0.007 32 25.18 0.404 26.08 0.473 0.15 0.006 0.16 0.004 48 26.09 0.474 27.00 0.527 0.06 0.005 0.06 0.005 72 26.88 0.544 27.39 0.580 0.03 0.003 0.02 0.015 z Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour"1 for that time period = (ml 100 mg DM' 1) hour"1 abt-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 184 Table 11.10 Rates of gas production during in vitro digestion of leaves from late vegetative stem tips of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 -0.06 0.194 0.28 0.071 -0.06 0.194 0.28 0.071 2 0.27" 0.197 0.98b 0.058 0.33a 0.008 0.70" 0.059 3 0.92a 0.140 2.16b 0.182 0.65a 0.060 1.18" 0.124 4 1.70a 0.215 3.41" 0.315 0.79 0.075 1.25 0.164 5 2.48a 0.248 4.73" 0.346 0.78a 0.054 1.32" 0.048 6 3.33a 0.193 5.99b 0.326 0.85a 0.060 1.26" 0.023 7 4.24a 0.166 7.45b 0.422 0.91a 0.044 1.46" 0.096 8 5.03a 0.238 8.84" 0.465 0.79a 0.075 1.39" 0.055 9 5.94a 0.215 10.44" 0.427 0.91a 0.044 1.60" 0.132 10 6.66a 0.194 11.84" 0.466 0.72a 0.077 1.39" 0.046 11 7.57a 0.164 13.16" 0.456 0.91a 0.044 1.32" 0.066 12 8.29a 0.109 14.28" 0.435 0.72a 0.060 1.12" 0.020 14 9.85a 0.086 16.44" 0.489 0.78a 0.014 1.08" 0.048 16 11.23a 0.195 18.45" 0.628 0.69a 0.055 1.01" 0.079 20 13.51a 0.226 21.05" 0.451 0.57 0.016 0.65 0.193 24 15.33a 0.220 24.26" 0.184 0.46 0.022 0.80 0.153 32 17.42a 0.373 27.40" 0.227 0.26a 0.020 0.39" 0.015 48 19.06a 0.499 30.06" 0.358 0.10a 0.013 0.17" 0.009 72 20.11a 0.531 31.58" 0.460 0.04 0.004 0.06 0.010 z Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour"1 for that time period = (ml 100 mg DM"1) hour"1 a , b t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 185 Table 11.11 Rates of gas production during in vitro digestion of leaves from early bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Cumulative Gas Production2 Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 -0.05 0.108 0.09 0.005 -0.05 0.108 0.09 0.005 2 0.44 0.126 0.75 0.040 0.50" 0.020 0.66b 0.035 3 1.18" 0.046 1.73b 0.050 0.74 0.084 0.98 0.089 4 2.23a 0.042 2.84" 0.062 1.05 0.088 1.10 0.012 5 3.09 0.232 3.91 0.146 0.86 0.192 1.07 0.084 6 4.14 0.254 5.03 0.228 1.05 0.024 1.12 0.082 7 5.19 0.352 6.16 0.168 1.05 0.130 1.13 0.059 8 6.18 0.422 7.26 0.181 0.99 0.080 1.10 0.012 9 7.34 0.611 8.38 0.262 1.17 0.189 1.12 0.082 10 8.21 0.619 9.37 0.210 0.87 0.105 0.99 0.052 11 9.26 0.643 10.43 0.225 1.05 0.024 1.05 0.015 12 10.13 0.637 11.32 0.178 0.87 0.049 0.89 0.047 14 11.86 0.671 12.85 0.168 0.87 0.032 0.76 0.005 16 13.22 0.747 14.32 0.161 0.68 0.038 0.74 0.004 20 15.53 0.669 16.61 0.030 0.58 0.051 0.57 0.048 24 17.52 0.707 18.30 0.119 0.50 0.028 0.42 0.022 32 19.70 0.657 20.62 0.523 0.27 0.023 0.29 0.051 48 21.58 0.480 22.78 0.988 0.12 0.015 0.13 0.029 72 22.82 0.491 24.28 1.419 0.05 0.001 0.06 0.018 2 Mean accumulated gas produced per unit of D M = ml TOO mg DM" 1; n=60 for each cultivar y Mean rate of GP hour'1 for that time period = (ml 100 mg DM"1) hour"1 "•bt-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 186 Table 11.12 Rates of gas production during in vitro digestion of leaves from late bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 0.06 0.058 0.09 0.076 0.06 0.058 0.09 0.076 2 0.76 0.008 0.61 0.067 0.70 0.053 0.52 0.083 3 1.69 0.076 1.26 0.152 0.93 0.068 0.65 0.089 4 2.68" 0.072 2.07" 0.182 0.99 0.051 0.82 0.048 5 3.61" 0.096 2.80b 0.141 0.93 0.065 0.73 0.043 6 4.66a 0.199 3.62" 0.222 1.05 0.110 0.82 0.083 7 5.88a • 0.211 4.56b 0.229 1.22" 0.013 0.95" 0.010 8 6.87a • 0.278 5.47b 0.204 0.99 0.069 0.91 0.043 9 8.09a 0.291 6.46b 0.189 1.22" 0.013 0.99" 0.046 10 9.03a 0.364 7.23b 0.148 0.93 0.123 0.77 0.041 11 10.07a 0.375 8.23b 0.186 1.05 0.011 1.00 0.117 12 11.01' 0.385 8.86b 0.113 0.93 0.010 0.63 0.118 14 12.69a 0.387 10.03b 0.111 0.84" 0.028 0.58" 0.059 16 14.32a 0.397 11.23" 0.161 0.81" 0.028 0.60" 0.038 20 16.88a 0.322 12.96b 0.284 0.64" 0.019 0.43" 0.032 24 18.92a 0.257 14.29b 0.287 0.51" 0.026 0.33" 0.012 32 21.71a 0.166 15.94" 0.449 0.35" 0.023 0.21" 0.023 48 24.39" 0.280 17.83" 0.723 0.17 0.011 0.12 0.035 72 26.02a 0.640 18.29" 0.603 0.07 0.016 0.02 0.010 z Mean accumulated gas produced per unit of DM = ml 100 mg DM' 1 ; n=60 for each cultivar y Mean rate of GP hour1 for that time period = (ml 100 mg DM"1) hour"1 "•"t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 187 Table 11.13 Rates of gas production during in vitro digestion of late vegetative stem tips of the LIRD cultivar, A C Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 -0.28 0.124 -0.11 0.049 -0.28 0.124 -0.11 0.049 2 0.62 0.135 0.72 0.285 0.90 0.187 0.83 0.243 3 2.13 0.200 2.18 0.577 1.51 0.067 1.46 0.308 4 4.33 0.193 4.61 0.718 2.20 0.148 2.43 0.142 5 6.32 0.169 6.78 0.615 2.00 0.034 2.17 0.143 6 8.18 0.150 8.66 0.518 1.86 0.031 1.88 0.099 7 9.76 0.190 10.21 0.499 1.58 0.045 1.55 0.019 8 11.07 0.174 11.59 0.499 1.31 0.141 1.38 0.052 9 12.32 0.151 12.85 0.439 1.24 0.136 1.26 0.076 10 13.34 0.065 13.94 0.458 1.03 0.103 1.09 0.025 11 14.45 0.130 15.03 0.477 1.10 0.066 1.09 0.025 12 15.41 0.195 15.89 0.458 0.96 0.066 0.86 0.035 14 16.64 0.115 17.08 0.365 0.62 0.061 0.60 0.069 16 17.74 0.102 18.11 0.325 0.55 0.028 0.51 0.022 20 19.25 0.165 19.66 0.269 0.38 0.030 0.39 0.018 24 20.08 0.271 20.46 0.245 0.21 0.027 0.20 0.009 32 21.17 0.373 21.55 0.205 0.14 0.013 0.14 0.005 48 22.68 0.409 22.82 0.293 0.09 0.003 0.08 0.006 72 23.72 0.392 23.74 0.405 0.04 0.001 0.04 0.005 z Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour'1 for that time period = (ml 100 mg DM 1 ) hour"1 a , b t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 188 Table 11.14 Rates of gas production during in vitro digestion of early bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 -0.02 0.061 0.16 0.056 -0.02 0.061 0.16 0.056 2 1.21 0.021 1.22 0.084 1.22 0.077 1.06 0.049 3 3.22 0.107 3.09 0.068 2.02 1 0.114 1.86 0.030 4 ' 5.55" 0.015 4.30" 0.230 2.33" 0.118 1.21" 0.175 5 7.83 0.167 8.19 0.153 2.28" 0.174 3.89" 0.077 6 9.77 0.081 10.10 0.201 1.94 0.091 1.91 0.074 7 11.30 0.122 11.54 0.197 1.53 0.063 1.44 0.065 8 12.58 0.176 12.81 0.243 1.28 0.071 1.27 0.049 9 13.80 0.125 14.09 0.269 1.22 0.094 1.28 0.065 10 14.86 0.152 15.16 0.245 1.06 0.062 1.07 0.042 11 16.00 0.141 16.07 0.241 1.14" 0.016 0.91" 0.060 12 16.74 0.144 16.86 0.224 0.74 0.024 0.80 0.041 14 17.80 0.107 18.31 0.209 0.53" 0.032 0.72" 0.037 16 18.94 0.094 19.37 0.176 0.57 0.008 0.53 0.021 20 20.40 0.140 20.97 0.167 0.36 0.017 0.40 0.006 24 21.33" 0.132 21.88" 0.141 0.23 0.005 0.23 0.010 32 22.38" 0.151 23.10" 0.143 0.13 0.008 0.15 0.002 48 23.94 0.161 24.43 0.182 0.10 0.006 0.08 0.006 72 24.82" 0.149 25.39" 0.140 0.04 0.002 0.04 0.002 z Mean accumulated gas produced per unit of D M = ml 100 mg DM" 1; n=60 for each cultivar y Mean rate of GP hour"1 for that time period = (ml 100 mg DM"1) hour'1 "•" t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 189 Table 11.15 Rates of gas production during in vitro digestion of late bud stem tips of the LIRD cultivar, AC Grazeland and a standard alfalfa, Beaver Cumulative Gas Productionz Rate of Gas Productiony Beaver AC Grazeland Beaver AC Grazeland Time(h) Mean SE Mean SE Mean SE Mean SE 1 -0.08 0.132 0.01 0.062 -0.08 0.132 0.01 0.062 2 1.12 0.091 0.69 0.156 1.20" 0.084 0.68a 0.094 3 2.88b 0.074 1.72a 0.191 1.76" 0.084 1.03a 0.072 4 5.34" 0.127 3.06a 0.200 2.46" 0.054 1.35a 0.029 5 7.16b 0.219 4.03a 0.227 1.83" 0.097 0.96a 0.040 6 8.76b 0.189 4.95a 0.234 1.60" 0.082 0.93a 0.044 7 10.31b 0.172 5.69a 0.203 1.55" 0.072 0.74a 0.031 8 11.58b 0.200 6.37a 0.224 1.26" 0.029 0.67a 0.021 9 12.85b 0.261 6.95a 0.290 1.27" 0.089 0.58a 0.074 10 13.76" 0.247 7.62a 0.234 0.91" 0.045 0.67a 0.063 11 14.80b 0.194 8.23a 0.277 1.04" 0.053 0.61a 0.046 12 15.58" 0.178 8.68a 0.263 0.77" 0.044 0.45a 0.014 14 16.76" 0.150 9.41a 0.240 0.59" 0.038 0.37a 0.023 16 18.01" 0.283 9.99a 0.240 0.62" 0.102 0.29a 0.013 20 19.33" 0.090 10.95a 0.230 0.33 0.050 0.24 0.009 24 20.25" 0.153 11.40a 0.249 0.23" 0.020 0.11" 0.005 32 21.36" 0.135 11.98a 0.269 0.14" 0.006 0.07a 0.003 48 22.75" 0.248 12.81a 0.322 0.09" 0.009 0.05a 0.003 72 23.73" 0.212 13.42a 0.359 0.04" 0.004 0.03a 0.002 z Mean accumulated gas produced per unit of D M = ml 100 mg D M 1 ; n=60 for each cultivar y Mean rate of GP hour1 for that time period = (ml 100 mg DM"1) hour1 a,"t-test of independent sample means. Production or rate means in the same row followed by the same or no letter are not significantly different (P>.05) 190 vo CS vo H H O r~ cs r~ in m ! S3 > I o 00 r~ co H co CS rH O cs CO CO H I H H H CS co *f *r ^ * cs H H O i Ei <* i < i EH — 1 M B i 0. B co o cs i 1 CO * o in in CO o CS CS 1 M ^ ON i u ON VO CO in r~ vo o vo ON o I w H H H H cs vo r- H H CO I 0j H H in i o. O o o o H ON CO H in in o O i § O o o H Vo Vo CO o O VD 1 M CO CO CS ON CO ON I EH ^ H H CS CS H H rH 1 ^ S H i 2 e ro O — ON O H ON CS ON rH VO 1 CM CM s E E E E I < H ON CO CO vo VO in O I > Ol H CO vo *f co in CO I W H CS rH f-t rH rH rH rH H W >H D 5 Bj O H X Q W CO E ON W > S3 ~ < H 0) a n to 3 B O co ON CO — ON I H I s co I CN u — i H W W E-i M ei w sc 5 IB H 5 2 iJ co W ON PH I ON 2 I H U H I EH 0) EH H 05 X W X O M CO 5C ON ON i n o - i n L n H c o o c o o r ^ o c o Hcs cscscscocscocscs OOOOOOOONHCSCOH LnininoininininoNr^^poN C S C S C S C O C S C S C S O H ^ C S H CSCSCSiHCSCSCSCSCOCSCSCS omoNOOHincovovocovo c o c s r ^ i n c A i n L n v o o o N ^ r * voininininvovor->p~ooN ocscovDHco^coinHr-vo H O ON O CS CS H CS ONr^in^voooH H i - t r H i - t H C S C S C S OCOCOCSHHVOCSin^ ONH vDror~in invocscsco*roNco H r H V O r ^ C S f - H C O ^ H ^ C O O f O c r t i n ^ O N O N r H O O H H ^ i n H O N C S V D O i n c O ^ O H r ^ H O N > H H C S C S C S C 0 C 0 C S H H CO i n c s c s o N c o c o c o c o m m C T i c ^ v o T j i ^ C S H C O V O C O H C O i n t - i C O C S ^ C O V O C S C n V D C S O N H O N C S C O H H H i - I C S C S C S C S C S C S r H r H vo*j*voinr~ONOooocooN c o v o i - i i n o * r c o c ~ c s c ~ o i n i n r ^ r H r - O N c o ^ o i f o i n r ^ c o OVOCSVDCSrO'^'^Hr-HO ocscocscofoomvo^voco CO CS f~ r- CS CO TJ ^ n n H i I I I I I cs m vo in cs I H CS CO ^ in vo cs o CO cs cs rH^ooror-ONcsoo cs in H H co o l H CS CS coin^ONinvoHcsHoro^ r^cs^coroinoNr^voocoON H C S C S C 0 C O C O C O C 0 C O C O C S H fO**inococoincooNr~coo C S i n f - O N O N C O O N C O C O ^ H C O H H H H C S C S C S C S C S C S H H w oi a m PJ w w EH s fa m m D EH O W a J o a. EH > u D D H U O W ft <n o z a O X J BJ U H (B CO Bi OJ >< S " w a a rf 5 S < 2 b V0 (N vo H rH O t- CN t- in in > i o < 00 I- ro H CO CN H o (N co 00 rH 1 M rH rH OJ ro r- 1" CN H rH o i H 1" 1 < 1 frl — I |_j e I 0. B ro r- 00 H •d» CO *T O in in 1 H — oi I u cn VO CO o (N CO m o CO o CN CN  W H M H H 00 r» CN CN rH in CN H r~ i OS m i a O o o o O H CM ro H in in O O i S > i o < o o o rH VO VD ro *r t- O o vo 1 H CN 00 ro CN cn co cn i E-i ~ H CN (  H rH rH ! £ i H • MH p I O — ^< r~ rH O rH r- o CN l i cn 2 s 2 S 2 1 < Cn t- 00 VO CN in rH ro 1 > H oo rH r~ ro CN in CO I w rH CN CN CN CN CN H in rH Ed i J Q 5 os O H 55 Q g | s _ * a > z ~ < H m a: n 00 0 a o f 5 J3 m co, — cn 1 S s i cn — i EH 03 W H H w K H H H ON H H (N o \ 0 ( N ( N r - c o i n t N i n r - o o o r l r l P l l N I / l O H f O l N t H l N P l CS(N(NCMrO(N(N(NCNH(N(N m c o c o ^ i / i h c o o j L n u i a j H otNmvoHco'^coinHr'VD Hoaioc r i r-Ln^^ooH ( N M H l N H r l r l r l H ( N l N ( N 0 3 O C N t ^ ^ c r \ u n r ~ - r 0 L 0 r - c n r o r ^ o i n i n ^ C N H r o m o r -H H VO f> CM H criLn^cr\criHoo a\csvooinco^o r-iHWCNCNromtN ro ^ rH ^ m H r> H a\ r -f O ^ V O M O ^ ' J O O H v C h f S O H ( N ( N ( N r n n C N O J i - I H H l i l v C - J v O l / l h f f i O O O O f f l O ^ CN o o v D H t n o ^ Q o r - M r - o i n in H - J H H O C l C D C O U ) l / l v D O o ( N o o t N c o r o o L n \ D T r v r > o o ro CN r> r- ro ^ ^ n M H t >* t- CO in cn VO CN cn o H t- CN o cn in in rH cn rH O ro rH O rH CN CO rH 1 i ' 1 t CN CN H CO in cn in VD H CN rH o ro ^ > CO H CN <* ro ro in cn t- VD o CO cn X H CN CN ro ro ro ro ro ro CN H ro CO CN CN H CN o o o CN CN cn cn ro cn o in 00 cn •* vo H o O H H CN CN CN CN ro ro ro CN H H a !« Id Oi « pa a. H U H a CQ < D X a cn u m CU u H w >H o W w m 04 o CU E-< > u a 5 5 ta o o pq tu E. < 2 B cn o Q cu • H U ' CD 01 •H cu "P ft e 192 a 55 > O < H < EH I—I 0. e m M ON u Ol w H OJ 0. o s > o 1—1 a E 2 e o in o< ON cn > H u w EH >H < d Q g 04 O H « Q 1 1 £? u > a — < H 01 K to CO 3 B O in S3 Oi CO ~ o\ I v D ( N v o i - i H O T t « r - ^ ( N r > i n i n H H ( N r O l O r - ' * - « J ' - c r C N H H O c o c s u ) ^ c o c D i n i n c o i / ) ( n c o u } r o o r o v x > ^ H f O H r - H O \ H H t ooooHcnroHLninoo^r o o o H v o u j r o ^ r ^ r - o o v D M m ro o) cn co o\ H i-H CN CN H H rH H o\ ^ oo m o\ h ^ u> s s s s s • r>o%rHf* i inr - r - ro U) m H m ro w in rH H CN CN CN H rH CO ajTjimHHVDr^roHcnajog rH CN CN CN CN rH rH r - * u > r H a v o o a \ r - i n v x > o H o r - c n H ^ r o o v D ^ i n o c o C N C N O C N O O \ f O ( N C N C N r H O CN CN CN CN CO CN CN CN CN CN CN I f i U J t O t O O m O r l ' J U J H f N > j > a » r ^ r o r o i n r N o o i n o o i n v o c o c o r ^ v D v o r - o m c o o r ^ 0 ( N M l O r i C O \ f ( O i n H > « ) i - i o o \ o o N r - i n ^ v o o o H C N C N H C N H H H H H C N C N C N r - c o v o m ^ r o c r i o r N V o c o c N m v D i n r r r o r o c N ^ H o o i n c N fflin^CnmHOOH H H C N C N C N m r O C N H < J in H c n r - c N ^ i n c o r - o r - H C N r H r H • ^ r o H o ^ v o o c N r o r - H r o a * H r H C N C N C N C N m m C N H H H ' 3 ' v o ^ v o u i r - c ^ o o o o o o c / * C O V D H L O O - ^ C O r - C N r - O l i l c s r ^ r o m H U ) r n o o i n O r H C T » u —- I EH o C S CO C S CO C O o in vo VO CO CO m EH M co C S p- t- C S C O o cs in vo m C S pi C O X co cs H i i H cs CO W i 1 i 1 1 1 1 • 1 H 3 O 2 in r» CO CO p- co CO ON cs P- p- H ON H ON PH i cn p- vo vo p- cs H vo CO in cs C S O E i H C S C S cs i i i H C S C O w i 1 1 1 1 1 EH J i EH CO m ON in VO rH cs rH O CO I CO I EH H p- (N T C O cn m ON p- vo o CO I C O X H CS C S co CO C O CO CO CO CO cs H I u 1 X \ o H in O CO cs ON vo ON O cs P- rH rH I K ON I ON rH C O ON CO CO p- ON o 0N ON CO in Hi-iHt-iCSCSCSCOCSrHH P Oi 2 m < W D X J ' U rH OJ Oi § OH 5 D OJ ia m OJ EH s Cxi CO w m H o O OI EH W U CO O <u Dl •H u Cn s 0) o > u o w a Q o *o P. B 5 s Dl a VO CN VO rH rH O *f r- CN t> in in ! a > i o CO r- ro rH ro CN H O CN CO 00 H 1 H H rH CM ro in r- CN H H o I H i < i Fn — 1 rH G i o. E vo CO O r- O r~ in rH CO O ro in ro 1 H — cn I u o\ VO rH H ro CO o VO vo cn i w H ro CN CN in rH r~ CN CN CN i a; ro I Oi cs o o o O rH cn ro rH in in o O <tf ! a > i o <: o o o H VO vo ro r- o O vo 1 M CN CO ro CN CO 00 CO i EH —- H rH CN CN rH H rH i 3 i H • CS C I o — vo cn r- cn rH CN o CN i 0< OS 2 s 2 2 2 i < cn CN r~ o H CO CN o i > rH o CO CN ro CO in o CO i w CN rH ro ro rH H CN in H 1 I H CO EH H CO W W X o H VO X cn cn to H cn in cn ro H r- H p- m H H H CN CN CN ro H CN CN rH Q p- rH ro 00 CN CN ro o in H O cn H O CN in i-H CO cn H CO CN CN o H H O o rH H p* p- CN CO CN VO 00 a Q CN CN CN CN CN CN CN CN CN CN CN CN VD 00 CN O 00 ^ r- H H P- rH H X cn H H O 00 r- vo ro ro in CO rH I-I cn 00 00 G\ r- p- r- r> CO r- cn O C N r 0 V 0 H 0 0 T j t 0 0 i n r H P - V 0 H O c n o m r - i n ^ v o o o H C N C N r H C N r H H H r H r H C N C N C N r o o c N r o v o H ^ j ' " i , P * i n o c o i n c r l r o o » i n p * r o ^ i , ^ , c o i n v D H n v o P - c N H r o ^ r H ^ r o o r o 1 > cn in cn cn rH o o rH rH <* in H S . . < cn CN VD o in CO *r o rH r~ rH cn H CO H H CN CN CN ro ro CN rH rH ro X u CN CQ 0 vo *c VD CN CN cn 2 2 2 2 2 2 2 2 2 W cn ro VO CN CN l rH cn in VO rH 1 rH H 1 vo VO in [•*• cn o o o o 00 CO [ 1 00 VO rH in o 00 r- CN r~ o in j 1 i ' 1 H H rH rH H 1 1 J 2 vo in cn 00 r~ Ul vo H ro rH 00 i 1 cn 1 1 cn ro ro 00 in CO CO rH vo r» o ! 1 rH rH 1 i ' H H rH H I H 1 u 1 EH o CN CO CN CO ro O in VO vo 00 CO w EH M ro CN r» CN ro O CN in VO in CN &• CO X ro CN H i 1 rH IN ro W i 1 i i 1 1 i i 1 EH O 2 VO ro O CN vo VO ro rH VD ro ro i n o H v o r o r o i n v o c N r H r H r o i H ro ro ooin^cninvoHCNHOro'^ C N ^ t | r o r o i n c T > p ~ v o o r o c n C N C N r o r o r o r o r o r o r o c N H C N ^ C N C N C O C n i n V O V O P - P - i n H C N C N r o r o r o c N C N •a D X J K U H pa ot « w § CK (xi 2 < m oi oi B o4 w w EH 2 U « -CO W CQ 2 _ » EH o w ta O CU EH > CJ . , D w u o w o ta < CO O a Q B 2 EH 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089771/manifest

Comment

Related Items