Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. Digestion characteristics of forages, including perennial ryegrass at different stages of maturity, and supplementary feeding for dairy cows grazing pasture A thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy In Animal Science Institute of Food Nutrition and Human Health Massey University, Palmerston North New Zealand. Alexandre Vieira Chaves 2003 Abstract This thesis defines digestion kinetics for perennial ryegrass ( Lolium perenne L.), which is the main component of diets fed to dairy cows in New Zealand. Chemical composition and digestion kinetics were measured in fr esh minced ryegrass as it matured and leaf, stem and inflorescence of several grass species. In sacco and in vitro incubations were used to define rates of degradation and nutrient release. Two short-term grazing trials were used to evaluate contrasting silages as supplements for cows fed restricted amounts of summer pasture. The minced preparation of ryegrass resulted in a similar distribution of dry matter (DM) between part icle size fraction and rumen digesta from cows fed pasture. Mincing released 0.46 – 0.80 of crude protein into the soluble fraction, with highest proportions for mature grasses which had low CP concentrations (about 8 g CP/100 g DM). In contrast, the majority of fibre remained in the insoluble fraction but rates of degradation (k) a pproximately halved as grass matured. In vitro yield of VFA was similar for immature and mature minced ryegrass (after 12 hours VFA was equivalent to about 30% of DM), even though ammonia concentration declined to very low values for stem and mature grass. This suggests the rapid initial microbial growth was able to sustain a high level of DM degradation to VFA with mature grass. The summer pasture used for silage supplementation was of uncharacteristically good quality so the expected contrasts between maize, pasture, sulla ( Hedysarum coronarium), lotus ( Lotus corniculatus) and sulla/maize silage mixtures were less than expected. Milk responses to lotus silage supplements were greater than other silages (e.g.: 290 g milksolids from 54 MJ ME by lotus versus 110 g milksolids from about 50 MJ ME supplied by other silages). Pasture substitution was low (0.06 – 0.33). The Cornell Net Carbohydrate and Protein System (CNCPS) was chosen for evaluation of cow trial data because it uses feed degradation parameters as input variables to estimate nutrient supply. Model prediction of milk yield matched observed values when cows maintained liveweight. Milk yield wa s underestimated at low intakes and overestimated at high intakes because no allowance is made for nutrient partitioning between milk production and liveweight change. ii Acknowledgments What does PhD mean? A PhD means inspiration, commitment, responsibility, education, teamwork and fun as well (not writing this thesis though!). Firstly I must thank my parents, particularly my mother who encouraged me to receive all this education, and also my father who challenged me all the time and supported me during my whole life. NZODA for providing the opportunity and financial support for me to make this dream become reality. A dream that I have had since my first year of University, back in 1989. Eleven years later that dream began to become true. Dr. Waghorn, Garry, “The old man”: my mentor and godfather here in New Zealand, a good one. Energetic, hard working, committed with students, family and animals, funny, supportive, and of course, with his good and “bad” days. Garry has always been ready and keen to listen and he lp. I am not sure if I could have made it this far without you. Muito obrigado, Garry . De verdade . I met Colin Holmes in 1998. He helped me with classes, farm trips, trials, and provided me with advice, and friendship. A person that I admire because of the way he is. For me, he is a “legen d” in pasture dairy systems. Dr Warren McNabb and Tom Barry were two important aids when I came to New Zealand the first time to do work experience and helped me discover in which area I would like to do my PhD. I am also grateful to Professor Jo hn Hodgson for his time and help during my first experience in down under. Thank you, guys. My supervisor Dr Ian Brookes that had a hard time trying to understand my Brazilian accent! He was always on time an d prompt with any requests. Also thanks for the fast return of my papers and chapter drafts with useful suggestions and comments. iii I am also grateful to my other supervisor, Sharon Woodward, physically far away from me (in Hamilton), but she gave me an opportunity to work with my favourites animals: dairy cows. In addition, the velocity with which she returned my requests (e- mails; data; files). I really appreciate that! Special thanks also to the staff from Dexcel dairy farm No 5 such as: Pat, Spenser, Roger, Elena, Deanne, etc. and Dr Simon Woodward for assistance with the smoother running of the trials. I had an awesome time over there. I will never forget it. Cheers ☺ ! The continuous support by Sylvia Hook er and people from the international student office at Massey University during my PhD helped to make this thesis possible. Jason Peters became my first real friend here in down under. Really good technician at AgResearch Grasslands, surfer, best friend and flatmate. Friendship forever. Thanks buddy! Also Jennifer, I mean , Dr Burke. We share a lot of knowledge, data, support and friendship. As PhD students we suffered almost on the same degree of stress. I wish you all the best. I would also like to acknowledge all of the Nutrition and Behaviour Platform (Grasslands Research Centre), AgResearch Limited. Special thanks to Dean Corson and Vilma Rodriguez for the NIRS analyses and calibrations; Dr. Graeme Attwood and Dong Li for VFA analyses; Dr. Neville Grace for the “teaching” during lunch breaks; Denise for the bindings and paper work; Tony Dunn for his effort, helping running in vitro incubations, lignin analyses and field work . I really appreciated the help from the librarians as well, Tracy Hanning and Chri stine Alley. Team work is essential! I nearly forgot my brother, Marcelo and my sisters, Suzana and Adriana. Over the past four years I have missed you guys a lot. I love you. Pepito, my Spanish friend, your company and friendship here in NZ was really cool and appreciated. We travelled togeth er and got to know each other better day by day. I hope your dreams come true and good luck with writing your PhD thesis. Keep doing Capoeira. I will see you in Brazil at some stage. Tu serás para mi un amigo siempre. My friends from Brazil. When I came to New Zealand in 1998 I left my entire group of friends behind but the real friends always keep in touch. Thanks also to the Internet for allowing these friendships to keep running. I am not going to name anyone because 300 pages, of this thesis, are enough to read! iv Contents Ab stract ii Acknowledgments iii C ontents v List of illustrations vii Ab b reviations x Chapter 1 – General in troduction. 1 Chapter 2 – Literature review. 5 Chapter 3 – Chemical composition, in vitro and in sacco digestion kinetics 49 of perennial ryegrass as influenced by stage of maturity. Chapter 4 – Digestion kinetics of leaf, stem and inflorescence from five 116 species of mature grasses. Chapter 5 – Supplementing fresh pasture with maize, lotus, sulla and pasture 145 silages for dairy cows in summer. Chapter 6 – Supplementation of grazing dairy cows with sulla and maize 167 silages in summer. Chapter 7 – Simulation models for rati on balancing and an evaluation of the CNCPS model. 202 Chapter 8 – General discussion. 228 L ist of pub lications 2 3 3 Appendices APPENDIX 1: In vitro incubation. 235 APPENDIX 2: Ammonia determination method. 238 APPENDIX 3: VFA determination method. 240 APPENDIX 4: In vitro data from Chapter 3. 241 APPENDIX 5: Equations used to describe relationships between height, herbage mass and nutritive characteristics of ryegrass due to maturation (age; days of re-growth). 245 v APPENDIX 6: In sacco data from Chapter 3. 246 APPENDIX 7: In vitro data from Chapter 4. 252 APPENDIX 8: In sacco data from Chapter 6. 255 APPENDIX 9: A simplified method for lignin measurement in a range of forage species. 261 B ib liography 2 7 3 vi List of illustrations TABLES FIGURES Chapter 1 Page Page Figure 1.1 1 Chapter 2 Table 2.1 6 Figure 2.1 7 Table 2.2 10 Figure 2.2 8 Table 2.3 12 Figure 2.3 9 Table 2.4 13 Figure 2.4 11 Table 2.5 14 Figure 2.5 16 Table 2.6 15 Figure 2.6 22 Table 2.7 17 Figure 2.7 25 Table 2.8 19 Figure 2.8 32 Table 2.9 20 Figure 2.9 33 Table 2.10 21 Figure 2.10 44 Table 2.11 23 Table 2.12 28 Table 2.13 31 Table 2.14 34 Table 2.15 36 Chapter 3 Table 3.1 51 Figure 3.1 52 Table 3.2 69 Figure 3.2 53 Table 3.3 70 Figure 3.3 53 Table 3.4 71 Figure 3.4 54 Table 3.5 71 Figure 3.5 55 Table 3.6 72 Figure 3.6 56 Table 3.7 72 Figure 3.7 56 Table 3.8 73 Figure 3.8 58 Table 3.9 73 Figure 3.9 64 Table 3.10 81 Figure 3.10 65 Table 3.11 82 Figure 3.11 65 Table 3.12 83 Figure 3.12 74 Table 3.13 84 Figure 3.13 75 Table 3.14 85 Figure 3.14 76 Table 3.15 86 Figure 3.15 87 Table 3.16 92 Figure 3.16 88 Table 3.17 93 Figure 3.17 89 Table 3.18 95 Figure 3.18 90 Table 3.19 95 Figure 3.19 96 Table 3.20 108 Figure 3.20 97 Figure 3.21 98 Figure 3.22 99 Figure 3.23 100 vii TABLES FIGURES Page Page Chapter 3 Figure 3.24 101 Figure 3.25 110 Figure 3.26 112 Chapter 4 Table 4.1 123 Figure 4.1 126 Table 4.2 124 Figure 4.2 126 Table 4.3 127 Figure 4.3 129 Table 4.4 130 Figure 4.4 129 Table 4.5 132 Figure 4.5 137 Table 4.6 133 Figure 4.6 138 Table 4.7 135 Figure 4.7 139 Table 4.8 136 Figure 4.8 140 Table 4.9 136 Figure 4.9 141 Table 4.10 143 Chapter 5 Table 5.1 154 Figure 5.1 149 Table 5.2 155 Figure 5.2 150 Table 5.3 155 Figure 5.3 157 Table 5.4 156 Figure 5.4 160 Table 5.5 163 Figure 5.5 161 Table 5.6 164 Table 5.7 166 Chapter 6 Table 6.1 170 Figure 6.1 182 Table 6.2 176 Figure 6.2 185 Table 6.3 178 Figure 6.3 185 Table 6.4 179 Figure 6.4 190 Table 6.5 181 Figure 6.5 191 Table 6.6 183 Figure 6.6 193 Table 6.7 183 Figure 6.7 194 Table 6.8 186 Figure 6.8 195 Table 6.9 188 Figure 6.9 197 Table 6.10 190 Table 6.11 199 Table 6.12 200 Chapter 7 Table 7.1 216 Figure 7.1 224 Table 7.2 217 Figure 7.2 225 Table 7.3 218 Figure 7.3 226 Table 7.4 219 Figure 7.4 227 viii TABLES FIGURES Page Page Chapter 7 Table 7.5 220 Table 7.6 221 Table 7.7 222 Table 7.8 223 Appendix 1 Figure 1.1A 237 Appendix 4 Table 4.1A 241 Table 4.2A 242 Appendix 6 Table 6.1A 246 Figure 6.1A 248 Table 6.2A 246 Figure 6.2A 249 Figure 6.3A 250 Figure 6.4A 251 Appendix 7 Table 7.1A 252 Table 7.2A 253 Appendix 8 Table 8.1A 255 Figure 8.1A 260 Table 8.2A 256 Table 8.3A 257 Table 8.4A 258 Table 8.5A 259 Appendix 9 Table 9.1A 266 Figure 9.1A 271 Table 9.2A 268 Figure 9.2A 272 Table 9.3A 270 ix Abbreviations AA: Amino acids. AAC: Australian Agriculture Council. ADF: Acid detergent fibre. ADL: Acid detergent lignin. ADFIN: Acid detergent fibre insoluble nitrogen (the N present in ADF). BW: Body weight (kg). oC: Degree centigrade. CP: Crude protein, being total N x 6.25 (or x 6.38 for milk). CPI : Crude protein intake. CS: Condition score. CV: Coefficient of variation. DE: Digestible energy. GE of the feed minus the GE of the corresponding faeces. DIP : Degradable intake protein (%CP). DM: Dry matter. DMD: Dry matter digestibility. DMI: Dry matter intake. DOMI: Digestible organic matter intake. g: Grams. GE: Gross energy. Synonymous with heat of combustion. h: Hour(s). ha: Hectare. I N RA : Institut National de la Recherche Agronomique. in sacco : In bag. in vitro : In glass. i n vivo : In animal. x k i (subscript): The net efficiency of use by the animal of ME (i.e., NE/ME) for energy maintenance (k m), for NE gain in growth and fattening (k g), as milk produced (k L ) and pregnancy (k pr). kg : Kilograms. LIC : Livestock Improvement Corporation. LW: Live weight. LWC: Live weight change. LWG: Live weight gain. MCP: Microbial crude protein. ME: Metabolisable energy. M E I : Metabolisable energy intake. ME m: The ME required by the animal for maintenance, or the maintenance (support) metabolism. mg: Milligram (10 -3 g). mL : Millilitres. mm: Millimetres. MP : Metabolisable protein. MS: Milk solids (milk fat + milk protein). M S E : ean square error. NAN: Non-ammonia nitrogen (total N in digesta minus ammonia-N). NDF: Neutral detergent fibre. NDFIN: Neutral detergent fibre insoluble nitrogen (the N present in NDF). N E : Net energy. The NE value of a feed (MJ/kg DM) varies with the purpose for which its ME is used because of differences between the k m, k g and k L values for that feed. NFC: non-fibrous carbohydrates (calculated by difference: 100 – CP – lipid – NDF – ash). ng: Nanogram (10 -9 g). N I RS : Near infrared reflectance spectroscopy. NPN: Non-protein nitrogen. N SC : Non-structural carbohydrates. NV : Nutritive value. xi N Z : New Zealand. OM: Organic matter. OMD: Organic matter digestibility. O M I : Organic matter intake. pH : Whole number referring to the number of hydrogen ions present in a solution. Negative logarithm of the hydrogen ion concentration. RDP: Rumen degraded protein; that part of the CPI which is fermented and may be used by the microbial population in the rumen, yielding MCP (synonymous of DIP). R UD: Rumen undegraded protein. S E : Standard error. S R : Substitution rate (kg pasture/kg supplement) = (Pasture DMI in control (unsupplemented) group – pasture DMI in supplemented group)/supplement DMI. S TDEV : Standard deviation. TMR : Total mixed ration. UDP : Undegradable dietary protein (synonymous of RUD). µg: Microgram (10 -6 g). µL : Microliter (10 -6 L). VFA: Volatile fatty acids. xii Forage common and scientific names Birdsfoot trefoil, lotus Lotus corniculatus Browntop: Agrostia capillaris Chicory: Cichorium intybus Foxtail: Alopecurus arundinaceus Italian ryegrass: Lolium multiflorum Kentucky bluegrass: Poa pratensis L. Kikuyu: Pennisetum clandestinum Lotus major: Lotus penduculatus Lucerne, alfalfa: Medicago sativa Maize, corn: Zea maize Meadow bromegrass: Bromus biebersteinii Roem & Schreb. Orchardgrass or cocksfoot: Dactylis glomerata L . Paspalum: Paspalum dilatatum Red clover: Trifolium pratense Phalaris, reed canarygrass: P halaris arundinacea L. Ryegrass (perennial): Lolium perenne L . Smooth brome grass: Bromus inermis L . Sulla: Hedysarum coronarium Tall fescue: Festuca arundinacea Schreb. Timothy: P hleum pratense Yorkshire fog: Holcus lanatus White clover: Trifolium repens xiii Chapter 1 - General introduction The productivity and nutritive value of perennial ryegrass dominant pasture is a constraint to future increases in productivity in New Zealand dairy systems (Clark et al., 2001 ). Over the last 50 years, annual pasture production has not increased significantly on research or commercial farms. Cows have a rapid decline after peak milk production relative to cows fed roughage with concentrates and short lactation lengths (266 - 268 days between 1997 and 2002) which is dependent upon growing conditions and availability of silage and supplements. In 2001/0 2 New Zealand had over 3.5 million dairy cows; herd size averaged 271 milking cows on 103 hectares giving a stocking rate of 2.7 cows per hectare (Livestock Improvement Corporation (LIC), 2002). National cow production averaged 307 kilograms of milksolids per lactation (North and South Island, 301, 338 kg milk solids/year respectively). In 2001/02, three co-operatively owned dairy companies processed over 13 billion litres of milk in New Zealand. Over 1.1 billion kilograms of milk solids was processed from seasonal supply units into products predominantly for export (LIC, 2002). The trend of increased production per cow over the last 9 years (Figure 1.1) is due to genetic gain and improvements in farm management (LIC, 2002 ). However animal improvements are influenced by the effect of the weather on the quality and quantity of pasture available for the cows. Unfavourable weather conditions in 1998/9 9 resulted in the lowest production per cow since 1992. 200 250 300 350 92 94 96 98 00 Season A ve ra ge k g m ilk so lid s/ co w FIGURE 1.1 – Trend in milksolids production per cow since 1992. Data from LIC (2002). 1 The quality and quantity of nutrients av ailable to grazing animals is extremely variable because weather conditions affe ct growth, growth rate, onset of grass maturation, sward composition and feeding management practices. Ryegrass pastures can have low dry matter and high fibre concentrations which may restrict feed intake so cow nutrient requirements are often not met (Waghorn, 2002). Grasses commence spring growth with vigorous leaf production, which has a high feeding value (nutritive value x voluntary intake) for grazing cattle, but as daily temperatures rise, an increasing proportion of stem and inflorescence appears. Although grazing management can control th e extent to which grass is allowed to produce seed heads, nutritive value declines because of changes in chemical composition. The principal changes are in creased proportions of fibrous stem and decreased concentrations of leaf protein (Wilman and Agiegba, 1982), so that the maturing plant has higher proportions of fibre and lower proportions of protein in the dry matter. These changes reduce the amount of amino acids available to ruminants and may increase the proportions of acetate: pr opionate available for absorption (Russell and Strobel, 1993 ). However, the most significant effect of grass maturation is that the rate of digestion and clearance of residual forage fibre from the rumen is reduced, because mature forages are slower to digest and require more chewing to reduce the particle size of plant fragments to a size ab le to pass out of the rumen (Ulyatt et al., 1986). The slower rate of passage from the rumen reduces feed intake. Hence, maturation in a grass-dominant sward results in lowered intake as well as declining nutritive value (Kennedy and Mu rphy, 1988). Grass maturation has a major impact upon dairy cow productivity because grass-dominant pasture is unable to provid e sufficient nutrients to match the genetic merit of New Zealand dairy cows, as evidenced by low milk production compared to cows fed concentrate diets and anoestrus coinciding with pasture maturation in some situations (Verkerk et al., 2000). Improved nutrition for dairy cows requires an understanding of digestion kinetics, including the rate of degradation of plant constituents and the nutrients released from digestion. This is especially important wi th grass dominant pastures where composition changes with the time. Burke et al. (2000 ) defined the degradation kinetics of immature leafy material from a range of grasses, legumes and herbs to be used as a basis for formulating forage mixed rations. That work determined degradation rates for dry matter, protein, soluble carbohydrate, neutral detergent fibre (NDF; cellulose, 2 hemicellulose and lignin) and acid detergent fibre (ADF; cellulose and lignin). It also indicated the amount and proportions of volatile fatty acids (VFA) produced, and have provided a mathematical basis for comparing contrasting feed types, but only for immature forages. Accurate prediction of performance for animals fed ryegrass diets and options for supplementation with high-quality forage s requires a greater understanding of how digestion processes are influenced by the stage of maturity. The objective of Chapter 3 was to analyse and compare changes in composition and digestion of DM, CP NDF and ADF of ryegrass growing to maturity in spring. The work described in Chapter 4 examines grasses, which have not been grazed and are in an advanced stage of maturity, not senescent but with stems and nearly mature flowers. This study aimed to determine the digestive characteristics at the opposite end of the range to that of Burke et al. (2000), using five very mature grass species. This involved the separation of the five grass species into leaf, inflorescence and stem (including sheath) fractions for incubation in vitro and in sacco. In vitro incubations were conducted to determine the products of degradation (ammonia from proteolysis and VFA), whilst the in sacco technique was conducted to determine the rate at which dry matter (DM) and its constituent chemical fractions are degraded through microbial digestion. Since the early 1970’s, a large number of studies have quantified the response in milk production to supplementation with grazing dairy cows. The principal factors controlling the response to supplementation have been identified as the quantity and quality of the herbage allowed, milk yield potential and stage of lactation of cow plus the quantity and nature of the supplement (Bryant and Trigg, 1982; Holmes, 1987 ; Thomson et al., 1998; Woodward et al., 2002; Holmes et al., 2002; Penno, 2002). Chapter 5 and 6 summarises experiments designed to improve the nutrition of cows in summer when pasture allowances are often insufficient to meet nutrient needs. Supplementation will increase and sustain milk production through to the end of lactation but responses will depend on pasture quality and type of supplement offered. This work emphasises the im portance of meeting cow protein requirements especially as maize silage, commonly used as a forage supplement, has a very low protein concentration and is not suitable for feeding wi th low quality summer pasture. Sulla is of interest because it is a high yielding le gume containing condensed tannins (CT) and high concentration of readily fermented carbohydrates, which offer good potential for high quality silage production (Niezen et al., 1998). 3 Simulation through mathematical modelling has shown to be a powerful tool for effectively integrating and utilising current knowledge of digestion kinetics (e.g.: Baldwin et al., 1987; Dijkstra et al., 1992). Such models ca n be used theoretically to predict rumen fermentation and nutrient abso rption for improving animal performance. Chapter 7 describes the use of a dairy nutr ition model (the Cornell Net Carbohydrate and Protein System) to develop strategies fo r high milk production within a grazing system and predictions are evaluated against data from animal feeding trials at Dexcel, Hamilton, New Zealand de scribed in chapters 5 and 6. 4 Chapter 2- Literature review Introduction This review will define the nutritional problems experienced by dairy cows grazing pasture and indicate options for use of supplements for cows fed pasture. The focus will include effects of grass maturation and the need to complement changing pasture quality in order to maintain production. In New Zealand, cows that calve in spring can experience a very rapid decline in milk production post peak lactation. Do New Zealand dairy cows have the potential to achieve high levels of milk production from pasture based seasonal systems? Work by Kolver et al. (2002) compared performance of overseas (OS) and New Zealand (NZ) Holstein Friesian (HF) dairy cows fed either a pasture diet (predominantly perennial ryegrass) or a total mixed ration (TMR) in the New Zealand environment (Table 2.1). The aim of this study was to test the potential of both genetics (NZ and OS) and diet type when fed a balanced diet (TMR) or when fed only pasture, and the likely consequences of using overseas genetics in New Zealand systems. Pasture was fed generously to NZ and OS throughout lactation with a daily allowance of > 60 kg dry matter (DM)/cow. Feed availability was not restricted for any treatment. The NZ Friesian cows fed TMR produced 38% more milk, 29% more milk solids (MS) and weighed 12% more than NZ cows grazing pasture at the end of lactation. The findings of Kolver et al. (2002) support previous observations summarised by Ulyatt and Waghorn (1993) which clearly show a high genetic merit for milk production by New Zealand cows. The rapid decline of milk production after peak lactation of cows grazing pasture is a consequence of feeding and diet. 5 TABLE 2.1 – Average annual milk production, efficiency of milksolids production, liveweight and reproduction performance of New Zealand (NZ) and overseas (OS) Holstein Friesians grazing pasture (Grass) or fed total mixed ration (TMR) during the 2000/2001 season (Kolver et al., 2002). Genotype NZ OS Diet Grass TMR Grass TMR P diet Annual milk production (kg/cow) 5300 7304 5882 10097 <0.001 Milksolids (MS: kg/cow) 465 602 459 720 <0.001 4.42 5.26 3.97 5.72 <0.001 Efficiency (kg MS/kg LW ) 0.75 Mean LW (kg/cow) 495 556 565 634 <0.001 Gain LW during lactation (kg/cow) 44 92 -20 77 <0.001 Empty rate (%) 7 14 62 29 0.1 The principal causes for the rapid post peak decline in milk production appear to be the cow’s inability to consume sufficient nutrients in early lactation, compounded by lower feeding value of pasture in late spring. The decline in feeding value is a consequence of lower nutritive value of grass as it matures, compounded by a lower intake of mature forage. Low intakes may be due to increased concentrations of fibre in ryegrass as it initiates stem development. The fibre requires physical breakdown to pass out of the rumen, and thus limits intake. The increase in proportions of structural fibre is also associated with lower concentrations of protein, rapid digested soluble carbohydrates, and a low ME content of pasture DM. Hence the low feeding value of pasture is due to the dominance of grasses as well as effects of flowering, but under some conditions there may be insufficient feed in late spring/summer and supplementation will be essential in order to maintain milk production (Wilson and Moller, 1993; Moller, 1997). Intakes and pasture quality are inter-related and supplements must complement the pasture quality on offer. This review provides a brief summary of pasture and forage supply for dairy production, examines pasture and forage changes with maturation and consequences for dairy cow nutrition. Results of trials where supplements have been fed to dairy cows have been summarised. Opportunities for using in vitro and in sacco techniques are presented, to identify principal factors required for model inputs and also to design appropriate supplements for pasture based diets fed to cows. The review concludes 6 with a synopsis of the CNCPS model for predicting cow performance when pasture is supplemented with forage species. 2.1 - Impacts of pasture quality and forage supply on dairy production Pastures are managed swards, usually based on improved cultivars of grasses and legumes. Management will optimise feed quality and supply in a sustainable manner. New Zealand pastures are usually dominated by grasses but can include a range of forages including herbs, weeds as well as legumes. Forage is the edible part of plants, other than separated grain, that can provide feed for grazing animals, or that can be harvested for feeding. Pastures fed in New Zealand dairy systems are predominantly but not exclusively based on ryegrass. For example, a typical dairy botanical pasture composition in Northland and Waikato regions is illustrated in Figure 2.1. Waikato region- Summer W hite clover, 5% Weeds, 2% C 4 and other grasses, 11% Perennial ryegrass, 30%Dead matter, 51% Waikato region- Autumn Perennial ryegrass, 70% C 4 and other grasses, 9% White clover, 8% Weeds, 4% Dead matter, 9% North of Whangarei Clover 10% Ryegrass 40% Tall fescue 5% Kikuyu 45% South of Whangarei Tall fescue 5% Ryegrass 65% Clover 10% Kikuyu 20% FIGURE 2.1 – Typical botanical pasture composition in Waikato regions (summer and autumn) and in north and south of Whangarei – Northland – Sources Simons et al. (1998) and Freeman (unpublished). 7 New Zealand pasture is characterised by the change in composition over the grazing season and by variability in supply. The principal dairying regions have been located in areas of reliable rainfall and good pasture growth (Waikato, Taranaki) but improved farming practices, availability of supplements (maize and pasture silage; green crop and hay) has enabled dairying to expand from Northland to Southland. These regions have very different practices, climates and needs for supplementation. For example, farming in Northland includes C grasses (kikuyu; paspalum14 ) and can be influenced by toxins such as grass endophytes (Easton and Couchman, 1999) or facial eczema (Towers, 1986). In this region there is a requirement for substituting grass with supplements which are free of toxins. In contrast, dairying in Southland is sustained largely by ryegrass, with reliable rainfall but short growing seasons. There is a significant requirement for supplements in that region to prolong lactation. These examples demonstrate the diversity of pasture diets and suggest provision of supplements to achieve an optimal balance of nutrients for the lactating cow is not a simple matter. The situation is made more complex by the changes in grass composition over a growing season and the need to balance cow requirements with feed availability and pasture growth. Forage quality decreases as plants mature. Increased proportions of fibre result in lower quality forage with fewer and more slowly released nutrients. Figure 2.2 demonstrates seasonal changes in pasture nutritive value averaged over three years in the Manawatu region. 17 39 45 50 42 23 26 25 7.2 9.610 8.3 12 12 10 11 0 10 20 30 40 50 60 Winter Spring Summer Autumn Crude protein, g/100g DM Neutral detergent fibre, g/100g DM Non-structural carbohydrates, g/100g DM Metabolisable energy, MJ ME/kg DM FIGURE 2.2 – Changes in pasture composition thoughtout the season. Adapted from García (2000). Please see page xiii for all scientific names used in this thesis.1 8 Overall, the nutritive value of pasture is lowest during summer and reductions in quality are usually accompanied by reduced pasture growth under dry conditions. The relationship between pasture growth and feed demand will depend on location, but the general relationship is illustrated in Figure 2.3. It is apparent that supplementation is essential to maintain both diet quality and quantity, and the type and extent of supplementation will differ for all environments. The objective of the research present here is to formulate a system to accommodate changes in grass quality and to devise supplementation strategies most appropriate to the dairy cow. 0 20 40 60 80 J A S O N D J F M A M J Season kg D M /h e c ta re .d a y Pasture Growth Feed demand FIGURE 2.3 – Annual pattern of pasture growth and demand for a spring calving herd stocked at 2.7 cows/hectare (Brookes, 2003). 2.1.1 - Feed supply Factors associated with feed supply on farms include herd size, effective area, grazing management over the lactation and dry period, as well as bought in replacements and sources of supplementary feed. Sourcing supplementary feed outside the farm, use of off-farm winter grazing and purchase of supplements can give a falsely high value for stocking rate (cows per hectare). The management and effective utilisation of forages is more relevant to good dairy practice than comparisons of stocking rates. Producers try to match as closely as possible the seasonal pattern of feed production with feed required by the dairy herd (Holmes et al., 2002). Most seasonal dairy farms calve in winter (late July-August; called “spring calving”) to ensure peak lactation coincides with peak pasture growth. This system results in an inadequate feed supply in the first few weeks of lactation with a surplus in late spring which needs to be 9 controlled to maintain pasture quality. Summer pasture growth is often insufficient to meet feed requirements of the herd (Figure 2.3). Herbage mass will influence intake per animal together with pasture quality (pre-grazing) to affect herbage utilisation (pre-grazing minus post-grazing). Table 2.2 suggests targets for average pasture cover and pre-grazing mass throughout the season (Matthews et al., 1999). The post-grazing herbage mass indicates differences in pasture management needed to achieve optimum utilisation at different times of the year for a ryegrass based pasture. Low post-grazing herbage mass indicates an insufficient feed supply and hard grazing may slow pasture re-growth. TABLE 2.2 – Seasonal pasture targets for a seasonal supply dairy farm. Adapted from Matthews et al. (1999). Season Average pasture cover Pre-grazing herbage mass Post-grazing herbage mass ________________________kg DM/ha___________________ Winter 1900-2100 2500-2700 800-1000 Early spring 1900-2000 2500-2700 1300-1400 Late spring/summer 2000-2200 2500-2700 1500-1600 Late autumn 1750-2000 2400-2700 1200-1400 The relation between intake and forage allowance is generally curvilinear (Holmes, 1987; Poppi et al., 1987). Pasture intake increases with an increasing allowance until it reaches a plateau. Hodgson (1990) suggested that the herbage allowance should be two to three times the maximum daily herbage intake of the animals, but higher allowances result in pasture wastage associated with low utilisation. Figure 2.4 shows a relationship between pasture allowance and intake summarised from recent publications (Bargo et al., 2002; Dalley et al., 1999; Delaby et al., 2001; Kolver and Muller, 1998; Stockdale, 2000a; Wales et al., 2001; Wales et al., 1999b). Low allowances, when growth is slow (early spring), during drought (summer) or with overstocking reduces feed intakes and lowers the quality of feed consumed. The diet will contain a high proportion of sheath and stem (Hodgson, 1990; Minson, 1990) and ryegrass endophyte toxins from hard grazing in summer. Cows which are last to be milked, especially young animals, will be most disadvantaged because they may reach the paddock 1-2 hours after the first cows have commenced grazing. 10 A high allowance is synonymous with a low utilisation, but high residual DM does not always equate to wastage. In fact allowances below about 40 kg DM/cow.day in spring are likely to limit intakes even though only 40% of available feed may be utilised (Table 2.2). High allowances will benefit cow health and production but stocking rate may be increased to manage pasture quality and prevent wastage (Brookes, 2003). ___ Polynomial (overall) 0 5 10 15 20 25 15 25 35 45 55 65 75 Herbage allowance, kg DM/cow.day H e rb a g e in ta ke , kg D M /c o w .d a y FIGURE 2.4 – Pasture dry matter intake of lactating cows over a range of pasture allowance. Each symbol (∆, ○, ◊, +, x, -) represents one study. Pre-grazing pasture mass ranged between 2000 – 4900 kg DM/ha. Pasture types: mainly perennial ryegrass but included smooth bromegrass, cocksfoot, Kentucky bluegrass and weeds. When excess forage accumulates, this is normally conserved as silage or hay. Conservation can be used to manage pastures but mature (flowering) grass provides a low quality feedstuff which will not achieve high intakes or productivity by lactating cows. Forage conservation must be designed to achieve high quality silage/hay as well as good pasture management, and this is often difficult to achieve. With lower stocking rates (SR) MacDonald et al. (2001) showed that there was a need for increased topping and pasture conservation (Table 2.3). The lowest SR were topped 2.1 times and 1.54 t DM/ha was conserved as silage compared with no topping and little conservation at the highest SR (P < 0.02). Pasture utilisation increased as SR increased from 2.2 to 4.3 cows/ha (P < 0.001). Also, as SR increased, milksolids per cow declined but the efficiency of pasture utilisation increased. 11 TABLE 2.3 – The effect of stocking rate on lactation length, annual milksolids (MS) per cow and per hectare, the amount of pasture conserved and topped, and estimated economic farm surplus. Adapted from trials carried out in the Waikato region, New Zealand by MacDonald et al. (2001). Stocking rate (cows/hectare) 2.2 2.7 3.2 3.7 4.3 Comparative stocking rate (kg liveweight/t DM on offer) 62 76 90 103 120 Days in milk 296 278 260 238 222 Cow performance (kg MS/cow) 435 380 353 309 274 Productivity (kg MS/hectare) 967 1043 1105 1145 1168 Silage conserved (t DM/ha) 1.54 1.37 0.94 0.37 0.10 Percentage of farm area topped (%) 210 100 40 10 0 Pasture eaten: t DM per cow 5.06 4.65 4.24 4.01 3.71 t DM per hectare 11.1 12.5 13.6 14.9 16.0 Feed required for cow maintenance (t DM/ha) 4.7 5.7 6.7 7.7 8.7 Pasture utilisation (%) 63 70 72 81 81 Feed conversion (kg MS produced / t DM eaten) 86 82 83 77 74 2884 2960 3054 2940 2751 EFS - Economic farm surplus ($/ha)1 EFS calculated at $4.50/kg milksolids. 1 Stocking rate is an important variable determining pasture intake and responses of cows to supplementary feeds. Stocking rate should be high enough to ensure that the herd eats a high proportion of the pasture growth in spring, without the need for excessive pasture conservation or topping (Figure 2.2; Table 2.3) and without limiting intakes. Conserved feeds could be used to improve lactation in summer and extend lactation in autumn (MacDonald, 1999; Pinares and Holmes, 1996). The poor match between typical pasture growth and cow requirements are further complicated by climate. Droughts are an important problem and vary in severity, duration and timing. Other problems include high or persistent winds which can reduce pasture growth, whilst excessive rain will cause pasture damage by pugging. Menneer et al. (2001) reported a 40% decline in pasture re-growth due to excessive rain. Low temperatures limit growth rate, and farmers may use nitrogen fertilizers to stimulate growth, but forage nitrogen concentration may be excessive and clover growth is likely to be suppressed by rapid grass growth. The net effect of climate and seasonality is a requirement for supplementary feed. The choice of supplement will depend on pasture quality and quantity, cow requirements and cost. In a recent review, Bargo et al. (2003) has indicated that high producing dairy cows fed pasture-based diets need to be supplemented to achieve their genetic potential for pasture intake, and substitution rate (the reduction in pasture 12 intake per kilogram of supplement) was a major factor explaining the variation in milk response to supplementation. 2.1.2 - Pasture quality Pasture quality can be maintained by good management, but ryegrass will enter a reproductive phase in late spring (October-November) and have lower nutritive value (NV) than vegetative grass. The impact of flowering on NV of the sward will be influenced by grass cultivar, proportion of grass versus legumes and other species, climate and farmer management. Pastures which are dominated by grasses will have greater changes in NV over the cow’s lactation than pastures with a higher (and sustainable) proportion of legumes and herbs. Ulyatt (1981) used growing lambs to rank grasses in terms of feeding value (Table 2.4) and Burke et al. (2002b) ranked forage legumes using the same methodology (Table 2.5). Legumes were almost twice as effective for lamb growth as perennial ryegrass. Grasses showed a considerable range in feeding value. Increasing the legume content of the diet is the most effective way to improve the feeding value of New Zealand pastures (Harris et al., 1997a; Woodward et al., 1999; Marotti et al., 2002; Burke, 2003). TABLE 2.4 – Comparative feeding value in terms of sheep live weight gain of some pasture species grown in New Zealand. Adapted from (Ulyatt, 1981). Forage Feeding value Number of studies 1 Italian ryegrass cultivar Paroa 83 1 Timothy 67 5 Perennial ryegrass cultivar Ariki 58 2 Perennial ryegrass cultivar Ruanui 52 16 Browntop in spring 52 1 Browntop in summer 43 1 All values relative to white clover. 1 13 TABLE 2.5 – Comparative feeding value in terms of sheep live weight gain, forage dry matter (DM) content, composition (g/100g of DM), and metabolisable energy concentration for fresh species. Adapted from Burke et al. (2002b). Feeding valueForage DM(%) CP NSC NDF ADF Lignin CT MEP2 3 4 5 6 7 81 White clover 100 15 27 12 26 19 5.9 11.5 Sulla 100 12 23 18 22 18 8.6 5 12.7 Chicory 95 14 19 11 24 21 7.0 12.5 Lotus 87 16 22 13 28 20 7.1 4 11.0 Lotus major 84 16 22 12 33 22 17.1 5-9 12.0 Lucerne 82 24 30 9 30 21 6.1 10.9 Ryegrass 52 19 16 9 49 26 2.9 11.0 Maize silage 35 8 42 41 25 4.4 10.7 All values relative to white clover. 1 2 Crude protein. Non-structural carbohydrates. 3 Neutral detergent fibre (cellulose + hemicellulose + lignin). 4 Acid detergent fibre (cellulose + lignin). 5 6Values elevated due to condensed tannin and other phenolic compounds associated with lignin. Condensed tannins (phenolic compound that reduce rumen proteolysis). 7 Metabolisable energy (MJ ME/kg DM). 8 Moller (1997) monitored a range of NV parameters in pastures throughout the year. Pasture samples were collected at 2-4 week intervals from four sites (two each in the Manawatu and Waikato) and analysed to predict digestibility (OMD), CP, ADF, NDF and non-structural carbohydrate (NSC) concentrations. Pasture digestibilities, which are strongly associated with energy intakes, averaged 74% - 80% in spring and autumn but were lower in summer (70%). Pasture CP and NSC followed a similar pattern to digestibility and were inversely related to fibre concentration (ADF and NDF). The author suggested that the low NSC concentrations in pasture may be insufficient to capture ammonia arising from pasture breakdown in the rumen. Together with increasing structural fibre, these factors combine to lower intakes of energy and protein, and may contribute to an accelerated decline in lactation performance following peak production in October. 14 The proportion of plant components also change during growth. Studies with ryegrass under controlled conditions show an increasing proportion of stem at the expense of leaf as the plants enter the reproductive stage (Table 2.6). Hoogendoorn (1986) reported similar results for ryegrass pasture (Figure 2.5). When grazed in late winter, grass leaf accounted for about 52% of DM after 8 or 16 weeks of re-growth. Lax grazing decreased the legume content of the sward and increased the amount of dead matter. When grazed in late spring, the lax grazing again resulted in more dead matter in re-growth than intensive grazing, but it also reduced the amount of leaf to 34% of DM versus 46% of DM in more intensively grazed swards. Pasture management does have an influence on proportions of leaf and stem in ryegrass. TABLE 2.6 – Changing composition of Italian ryegrass components (% of plant DM) and in vitro digestibility of green and dead leaf, stem and inflorescence harvested at two week intervals after an initial cutting. Source Wilman and Agiegba (1982). Weeks after cutting 2 4 6 8 10 12 14 DM distribution (%) in: green leaf 82 63 37 25 15 8 4 dead leaf 0 0 3 6 8 11 14 Stem 18 37 57 63 67 65 64 Inflorescence 0 0 3 6 10 16 18 DM digestibility (%): green leaf 64 66 65 61 65 61 58 dead leaf 52 44 49 47 Stem 64 68 63 62 59 57 49 Inflorescence 65 62 65 63 Estimated metabolisable energy 12.0 11.5 11 10.5 10.1 9.6 9.2 (MJ ME/kg DM) 1 For the whole plant, based on expected in vivo digestibility. 1 15 A - Spring 0 10 20 30 40 50 6 grass leaf grass stem legume dead material % of total DM 0 Intensely laxly B - Summer 0 10 20 30 40 50 6 grass leaf grass stem legume dead material % of total DM 0 Intensel y laxl y FIGURE 2.5 – Composition of New Zealand pasture in spring and summer. Adapted from Hoogendoorn (1986). In contrast with grasses, the digestibility of white clover is less affected by maturity despite an increase in proportion of petiole. Legumes with stems, such as lucerne and red clover become less digestible as they mature, in association with increasing proportion of stem (Table 2.7). The botanical composition and proportions of grass versus legume and of leaf, stem and dead material in pasture (Figures 2.2 and 2.5) will determine the nutritive value over the lactation. Providing feed availability is not limiting, the impact of changes in plant components on nutritive value of pasture will depend on the extent and duration of elevated proportions of stem, inflorescence and dead matter. If climate, management and grass species enable a short flowering period then dairy nutrition could be maintained by substituting poor quality pasture for high quality legume or other supplement over that period. If flowering is prolonged, the effect on pasture feeding value will be more severe. The principal objective of pasture management could be to minimise the amount of stem and dead matter (Figure 2.5). Grass and legume leaf both contain high concentrations of CP, but the concentration of cell wall structural fibre (NDF) is much higher in grasses than legumes. These differences are even greater for C4 grasses. Legume leaves contain higher concentration of readily fermentable DM (pectin, organic acids, and soluble sugar) than grasses and are more rapidly degraded in the rumen (Moore and Hatfield, 1994). 16 TABLE 2.7 – Changing composition and in vitro digestibility of legume components harvested at two week intervals after cutting. Source Waghorn and Barry (1987). Weeks after cutting 3 5 7 9 11 13 Structural components (% of whole plant) White clover leaf (%) 75 68 61 55 petiole (%) 15 23 29 37 inflorescence (%) 8 11 11 12 Red clover leaf (%) 71 62 54 45 40 petiole (%) 25 26 28 30 34 stem (%) 4 11 16 19 17 flower (%) 2 6 9 Lucerne leaf (%) 70 50 44 39 33 stem (%) 30 50 56 61 67 Digestibilities white clover 82 81 80 76 74 70 red clover 77 77 71 67 64 54 lucerne plant 75 72 63 62 61 60 lucerne leaf 76 76 75 74 72 72 lucerne stem 73 65 58 54 54 52 2.1.3 - Interaction between availability and quality When pasture is offered as a sole diet, even under true ad libitum conditions, feeding value will be limited by both the balance of nutrients in pasture (compared to cow requirements) and by voluntary feed intake. The implications of high concentrations of rapidly degradable protein and low concentration of readily fermentable carbohydrate (Moller, 1997) are confounded by a high water content in spring growth (Ulyatt and Waghorn, 1993), excess concentration of slowly degradable structural fibre in early summer and in some situations inadequate forage supply in summer. High bulk, slow rates of fibre breakdown and insufficient forage are all associated with insufficient nutrient intake and a low feeding value of pasture for dairy cows. These limitations may be compounded by fungal toxins such as ryegrass endophytes (Thom et al., 1999) or facial eczema which become more important as cows are forced to eat a high proportion of feed on offer, typically during summer dry periods. The effect of low DM percentage spring pasture on voluntary feed intake was identified by Vérité and Journet (1970) who suggested water contents above 84% would lower feed intake. A relationship between intake and pasture DM content was identified by John and Ulyatt (1987) r = 0.79, confirming observations of Vérité and 2 17 Journet (1970). Moisture content is affected by rapid growth and forage species, but impacts upon feed intakes and lactation performance in New Zealand dairy systems have not been defined. In contrast, effects of maturation and summer drought on feeding supply are well documented. The drop in pasture quality in October/November coincides with the reproductive phase of ryegrass growth, rising soil temperature and increasing day length, and as long as ryegrass remains the base species of dairy pasture there would appear to be limited opportunity to manipulate this with grazing management. Rotational grazing (short rotation lengths), topping of seed heads, maintaining good fertility, taking paddocks out of the system for renewal or feed conservation (silage or hay) when the herbage mass is about 3000 kg DM/ha are the standard methods used for quality control but they can only reduce rather than prevent effects of flowering. Choosing later flowering ryegrass varieties may help slow the decline in quality. Easton et al. (2001) evaluated new perennial ryegrass cultivars from 17 trials undertaken throughout New Zealand since 1991 and concluded that the new cultivars released to the market yielded on average 6% more herbage annually, and 9% in summer, than previous cultivars. Use of legumes that are active in cooler conditions, can be grazed without damage or fed fresh, as well as silages (e.g. cultivars of red clover, lucerne and lotus silages) would alter opportunities for maintaining diet quality, providing they are persistent and productive. Kolver et al. (1999) showed that mowing pasture before or after grazing in summer improved pasture quality and milksolids production. The mowing gives a more uniform pasture for the next round of grazing, reducing the amount of stems and inflorescences. Although fungal endophyte is not normally considered a nutritive characteristic of pasture, presence of deleterious endophytes have major effects on dairy production, especially in areas affected by summer drought. In many cases, pastures have flowered, have reduced quality associated with high proportions of stem and low growth rates and are in short supply. Cows are forced to graze for longer periods and to low residuals but may not achieve adequate intakes. Feeding value is low and intakes of stem and sheath will often include endophyte fungus (Easton et al., 2001). The ergovaline present in both ryegrass and fescue infected with wild type endophyte causes heat stress in cattle and further depresses intake, production and profit (Table 2.8). Blackwell and Keogh (1999) found that in December, cows grazing endophyte- 18 free pasture produced 24% more milk than the cows grazing pasture with endophyte. This difference between performances increased progressively until mid-April. TABLE 2.8 – Effect of ryegrass endophyte on annual milk solids production in Northland. Adapted from Blackwell and Keogh (1999). Milk solids Milk valueTotal season 1 (kg/cow) 2 ($/cow) Cows grazing endophyte free pasture 335 251 828 Cows grazing pasture containing endophyte 295 211 696 Difference 40 40 132 % Difference + 19% + 19% Milk solids produced between 1 October and 29 April. 1 Calculated using $3.30/kg milk solids. 2 Inclusion of dead matter and litter in the diet of cows given insufficient feed is believed to further reduce feeding value. Quantitative data concerning NV of senescent forage is scarce, but dead matter is likely to contain other pathogenic fungi (Waghorn et al., 2002). 2.2 - Implications of forage maturation on feeding value for dairy cows Feeding value is a combination of intake (availability) and nutritive value. These factors are linked in the New Zealand situation, because climate and growth physiology of grasses alter nutritive value, which in turn limits the potential intake even when feed allowance is high. In reality, allowance is also reduced in some situations and this will further limit milk production. 2.2.1 - Nutritive value In a survey of pasture composition from the principal dairy regions in the North Island Prewer (unpublished) demonstrated a small increase in DM content, decline in concentration of NSC and a significant increase in NDF concentration in the DM in late August (corresponding with stem elongation) which continued through February-March (Figure 2.6). Predicted ME concentration and OMD declined from October through March and indicate a need for supplementation to maintain NV. 19 The data in Figure 2.6 also demonstrate wide ranges in values for pastures throughout the growing season, suggesting good opportunities for matching performance with feed quality. The changes in pasture composition (Figure 2.6) correspond to changes in proportion of leaf, stem, petiole and inflorescence of grasses and legumes (Tables 2.6 and 2.8), but effects are dominated by changes in grasses (Table 2.9). The increase in structural fibre (cellulose, hemicellulose and lignin) at the expense of other components reduces the rate and extent of physical breakdown and clearance from the rumen which in turn reduces supply of both energy (VFA) and protein to the animal. Lower protein content has important implications for nutrient supply for both young growing and lactating ruminants. The physical structure of grasses may have a greater influence on feeding value than suggested by chemical analyses. The stem of ryegrass comprises about 20% thick- walled vascular bundle and sclerenchyma cells, in contrast to about 14% in leaf (Table 2.10). These long narrow fibrous cells develop thick secondary walls which lignify with maturity and account for a high proportion of the force required to shear leaves and stems despite the relatively small cross section of these tissues (Table 2.10). Vascular tissues have important functions for plants, in transport of nutrients and maintaining plant vertical structure so leaves are able to intercept incoming energy for photosynthesis. However they also represent a significant inhibition to nutritive value for ruminants. TABLE 2.9 – Composition (% of dry matter) and digestibility of ryegrass at 4 stages of maturity. Adapted from Waghorn and Barry (1987). Young leaf Mature leaf Head emergence Seed setting Non-structural carbohydrates 14 12 11 10 Organic acids 4 5 5 3 Protein 15 12 11 6 Non-protein nitrogen 4 4 3 3 Pectin 2 2 2 2 40 45 47 60 NDF 24 26 28 34 ADF Lignin 3 4 4 7 Ash 8 8 7 6 Lipid 9 8 7 5 Digestibility of dry matter (%) 86 83 79 62 Abbreviations see Table 2.5. 20 TABLE 2.10 – Comparison of anatomy and nutritive characteristics of leaf and stem of C3 and C grasses and C legumes. Adapted from Wilson (1991). 4 3 Nutritive characteristics Tissue types (% cross-sectional area basis) (% of dry matter) Thin-walled Thick-walled NDF ADL OMD a b Leaf C tropical grass 68 32 57 2.4 61 4 C temperate grass 86 14 43 1.8 72 3 C temperate legume 96 4 11 1.4 81 3 Stem C tropical grass 80 20 65 6.8 50 4 C temperate grass 80 20 69 5.6 50 3 C temperate legume 75 25 42 9.9 58 3 Thin walled cells comprise epidermis, mesophyll and parenchyma. a Parenchyma bundle sheath, sclerenchyma and vascular tissue. b NDF=neutral detergent fibre; ADL=acid detergent lignin; OMD=organic matter digestibility. 21 0 7 14 21 28 35 May Jul Aug Oct Dec Jan Mar Apr Jun Pas ture d ry m atte r, % DM 0 10 20 30 40 May Jul Aug Oct Dec Jan Mar Apr Jun Crud e pro tei n, g/ 100 g DM 35 45 55 65 May Jul Aug Oct Dec Jan Mar Apr Jun Neutr al d ete rg en t fibr e, g/ 100 g DM Seed head emergence 0 4 8 12 16 20 May Jul Aug Oct Dec Jan Mar Apr Jun Non -str uctur al c ar bohyd ra tes , g/ 100 g DM Seed head emergence 10 11 12 13 14 May Jul Aug Oct Dec Jan Mar Apr Jun Meta bol is abl e en er gy, MJ ME/ kg DM 60 65 70 75 80 85 90 May Jul Aug Oct Dec Jan Mar Apr Jun Org an ic m atte r di ge sti bili ty, g/ 100 g DM Seed head emergence FIGURE 2.6 – Changes in pasture chemical composition thought the season in the North Island. Adapted from Prewer (unpublished). Tropical C4 grasses are increasing in pastures and have moved steadily south over the past decade (Bell and Keene, 1996; Davies and Hunt, 1983). They now comprise nearly half of the pastures north of Whangarei (Figure 2.1) and it is important to appreciate the very considerable differences in anatomy and feeding value of temperate and tropical grasses, as the latter are likely to become increasingly dominant in North Island pastures. They differ in basic biochemical pathways for photosynthesis, and carbon fixation as well as anatomical structures (Table 2.11). 22 TABLE 2.11 – Differences between C3 and C4 forages in habit and biochemistry. C Plant C Plant 3 4 Examples of forages Ryegrass, forage legumes; temperate grasses Maize, tropical grasses Temperature Optimum 15-25 Optimum 25-40 C o oC Photosynthesis decreases below 10 C Minimal 0 o C o Annual production DM 22 ± 3.3 t/ha year 39 ± 17 t/ha year Transpiration (g H2O/g DM synthesis) 450 to 900 g H O/g DM 250 to 350 g H O/g DM 2 2 Light Intensity Saturated at 1/3 of maximum radiation summer Rarely saturated Photo-respiration losses 1/3 total gross process Very low or absent 9~P 11~P Energetic requirements to fix CO2 (high energy phosphate bonds ~P) Source: Starr and Taggart (1989). Leaves of C3 grasses contain a high proportion of thin walled cells, which are loosely configured, so 10-35% of the tissue volume comprises intracellular air space (Hanna et al., 1973; Wilson et al., 1991). The thin walls and limited lignification enable rapid bacterial colonisation and rapid digestion (Chesson, 1988). In contrast, C4 leaves and stems of both C and C3 4 grasses contain higher proportions of thick walled cells (scherenchyma, vascular tissue and parenchyma bundle cells in C4 grasses) that are closely packed with only 3-12% intracellular air (Wilson, 1993) and have lignified cell walls (Table 2.10). This structure inhibits bacterial access to cell contents and to the secondary (interior) cell wall, resulting in slow rates of degradation. Bacteria are unable to digest lignin and cannot penetrate the primary (outer) cell wall unless it has been physically damaged allowing bacterial access to degrade internal secondary cell wall tissues. Both the leaf and stem of C4 grasses contain a high proportion of thick walled cells, similar to that of ryegrass stems. The implications of structural fibre, lignification and thick walled cells for feeding value is substantial, because of a lower nutrient release, slower digestion and especially because extensive chewing is required to enable clearance from the rumen. Maize silage is often used to supplement pasture, 23 but the stover will not provide a good source of nutrient for the cows fed poor quality grass. Maize stover is equivalent to any other C grass and has a low NV. 4 The proportion of soluble dry matter (DM) is highest in actively growing forage tissue and declines as plants become mature and dormant. Soluble nutrients move from leaves to roots, and may be leached by rain. The loss of soluble DM from leaves increases the percentage of cell wall. Cell wall thickness also increases as plants mature; organic matter digestibility is lowered and rate of fermentation decreases with both the proportion and thickness of the fibre. The net effect of maturation is a decreased nutrient availability for the ruminant. The effect of maturation is relatively minor for legumes (Table 2.7) and of little practical significance in dairy pasture because legumes usually account for less than 15% of the DM. The low proportions of legumes are usually a consequence of rapid grass growth and shading. Legumes have a high feeding value and because cows are able to achieve high intakes, they have a high NV. Hoffman et al. (1998) showed dairy cows fed lucerne silage produced more milk and had a high DM intake compared with cows fed ryegrass silage. In recent work Broderick et al. (2002) indicated that, relative to ryegrass silage, feeding alfalfa silage stimulated much greater feed intake, which supported greater milk production. 2.2.2 - Digestion of plant cells Study of histological changes in forage leaves after in vitro incubation showed that leaf mesophyll cells were the first component to be digested due to thinner walls, absence of lignin and less cutinised nature of these cells compared to other leaf cells (Hanna et al., 1973). The large intercellular spaces in mesophyll tissue of ryegrass enable easy access for rumen flora through damaged surfaces. Digestion begins adjacent to damaged areas and proceeded toward the more compact cells. C4 grass leaves contain lower proportions of mesophyll cells and the parenchyma bundle sheath cells in these grasses are lignified and slow to digest. Parenchyma bundle sheath cells may account for 20% of C4 leaf cross-sectional area and surround vascular tissues further limiting microbial access to these fibres. Unlike mesophyll cells, the parenchyma bundle sheath cells become more lignified with age so the nutritive value of tropical grass leaf declines rapidly with maturity, whereas ryegrass leaf digestibility does not change significantly until very mature (Wilman and Agiegba, 1982). 24 Akin (1979) ranked the digestibility of cell walls in the order mesophyll = phloem = undifferentiated parenchyma > epidermis > parenchyma bundle sheath > sclerenchyma > lignified vascular tissue. The initial three cell types are usually rapidly and completely digested, as are the epidermis cells, except for the wall adjacent to the cuticle. The structure of parenchyma bundle sheath cell walls in C4 grasses and their rate of digestion vary greatly between different taxonomic groups. It is mainly the cell walls of the sclerenchyma, vascular tissue, and sometimes the stem parenchyma and stem epidermis, which are digested very slowly and contain a substantial indigestible component. The degree of indigestibility depends largely on the chemical nature of their lignin polymers and linkage with fibre components (Waghorn and McNabb, 2003), but also on the extent of cell rupture. 2.2.3 - Maturation on rate and products of digestion Ruminants have evolved digesting forage plants. They consume fibrous feeds that are not suitable for consumption by humans and single stomached animals and convert them into nutritious feeds such as meat and milk. The rumen provides a site where the rumen microorganisms can digest structural and soluble carbohydrates, proteins, and fibre. Through this digestion process, energy (mainly volatile fatty acids) and microbial protein are produced that are utilised by the animal (Figure 2.7). However, very successful selection for improved milk production has exceeded the capacity of dairy cows to consume sufficient forage feeds (Kolver et al., 2002). This is especially true when forages mature. C rude p ro te in Suga r, s ta rch Fe rm en tab le fib re Fa t R U P R D P M icrob ia l g row th a nd fe rm en ta tion M icrob ia l p ro te in A m ino P rop iona te A ce ta te , bu tyra te Fa tty a c id s ac id s M ilk p ro te in M ilk la cto se M ilk Fa t Feed N u tr ien ts M ilk com ponen ts FIGURE 2.7 – Feed, nutrient flow from rumen, and milk components. Adapted from Ishler et al. (1996). 25 Principal factors that limit forage intake include availability, palatability, protein to energy ratio, distension of the reticulum and cranial sac of the rumen, and perhaps hormonal and chemical factors associated with nutrients (Forbes, 1986). There is general agreement that intake of poor quality forages containing large amounts of lignocellulose is limited by the capacity of the reticulum-rumen and the requirement to reduce particle size to pass out of the rumen (Mertens, 1994; Wilson and Mertens, 1995). Murphy (1990) concluded that when distension of the reticulum-rumen is limiting feed intake, either rate of digestion or passage from the organ must increase before further increases in consumption can occur. This means that chemical and physical properties of lignocellulose which affect digestibility and passage will determine intake and nutrient availability. The resistance of lignocellulose to particle size reduction is an important property, because digesta must be made small enough to pass from the rumen and alleviate the inhibition caused by distension on voluntary feed intake. Also reduction of digesta particle size by chewing and rumination increases the relative surface area exposed to microbial attack and digestion increases the functional specific gravity of the particles, which enhances their probability of passing from the rumen (Ulyatt et al., 1986; Waghorn and Barry, 1987). Additional chemical factors affecting clearance and particle breakdown include stiffness or brittleness, silica, and physical factors such as the orientation of cellulose fibrils in cell walls. Waghorn et al. (1986) indicated that the effectiveness of chewing was greater during eating than rumination for sheep fed chopped lucerne hay, perhaps because of the brittle nature of the feed and differences in feed and digesta composition. In addition, they reported that rumination up to 14 hours post feeding was more effective than rumination between 14 hours post-feeding and subsequent feeding. This result might suggest digesta particles become more resistant to reduction with time of fermentation but it remains to be seen whether this effect is due to lignocellulose composition and whether it can be correlated with other physical measures such as maximum bending stress. The combined effects of increasing fibre content and declining digestibility causes a dramatic decline in feed value of the whole plant and lower intake by grazing animals (Waghorn, 2002). Increased rumination chewing and gut motility to enhance particle passage may be equally important for affecting voluntary feed consumption (Beauchemin et al., 2003). However, these functions will only increase supply of digestible energy if the rate of digestion is sufficiently high relative to passage (or disappearance) rates, so that particles flowing to the intestines have a low content of potentially digestible fibre. 26 The only option for overcoming the effects of fibre which limits the feeding value of mature grasses is to substitute a portion of the diet with a rapidly digestible feedstuff. Hence supplements may be used as an additional source of nutrients, when pasture is in short supply and also as an essential alternative to mature pasture in order to maintain nutrient intakes. 2.2.4 - Carbohydrate digestion When carbohydrates, both structural (neutral detergent fibre) and non- structural (sugars and starches), undergo microbial fermentation, VFA are produced as the principal metabolites. The primary VFA in descending order of abundance are acetic, propionic, butyric, isobutyric, valeric and isovaleric acids. The VFA can provide up to 80 percent of the energy needs of the animal. Acetic acid (CH3COOH) can constitute 50 to 70 percent of the total VFA and predominates in a high forage diet. Acetate is an energy source and a major precursor for lipogenesis in adipose tissue and the mammary gland. Acetate production is essential to maintain adequate quantities of milk fat (France and Siddons, 1993). Propionic acid (CH CH3 2COOH) can make up 18 to 22 percent of the total VFA when forages are fed, but accounts for a higher proportion of VFA in cattle fed high grain diets. Propionic acid is converted to glucose in the liver and is essential for lactose and protein synthesis. Butyric acid (CH CH CH3 2 2COOH) accounts for 10 to 15 percent of the total VFA in forage fed animals. It is largely converted to B-hydroxybutyric acid (BHBA) during absorption through the rumen epithelium. BHBA is used for fatty acid synthesis in adipose and mammary gland tissues and as an energy source (Ishler et al., 1996). Fibre-degrading bacteria are the principal sources of acetate, propionate and butyrate, where as valerate and iso-acids are derived largely from amino acid digestion. The proportion of VFA is greatly influenced by diet. Despite differences in the microbial population and in feed intake, ruminal VFA proportions are fairly stable among forage diets, but differ for forage versus concentrate based diets. As the forage to concentrate ratio decreases, the acetate to propionate ratio also decreases (Agnew and Newbold, 2002), indicating a change in microbial populations. Although proportions of VFA produced from contrasting diets have been characterised, ruminal conditions (e.g.: ratio of fibre: readily fermentable substrates) can affect microbial populations and their production of VFA from dietary components. This is illustrated in Table 2.12 which shows that the acetate to propionate ratio resulting 27 from fermentation of hemicellulose in a high forage diet was 3.2, but only 2.2 when fermented in a high grain diet. The acetate to propionate ratio from cellulose fermentation also varied with diet (7.3 for a forage diet and 13.2 for a grain diet) and rates of production were much higher for hemicellulose digestion (Murphy et al., 1982). TABLE 2.12 – Proportions of VFA arising from digestion of contrasting carbohydrate fractions as affected by forage and concentrate diets. Adapted from Murphy et al. (1982). Proportion of carbohydrate converted to a Substrate Diet Acetate Propionate Butyrate A : P ratio b F 0.69 0.21 0.11 3.3 Soluble carbohydrate C 0.45 0.21 0.30 2.1 c F 0.59 0.14 0.21 4.2 Starch C 0.40 0.30 0.20 1.3 F 0.66 0.09 0.23 7.3 Cellulose C 0.79 0.06 0.07 13.2 F 0.57 0.18 0.21 3.2 Hemicellulose C 0.56 0.26 0.11 2.2 a Ratios do not add up to 100 because valerate and the iso-acids are not taken into account. b F=forage diets; C=diets containing more than 50 percent of a cereal-based concentrate diet. c Soluble carbohydrate fraction includes organic acids and pectin in this analysis. 2.2.5 - Protein digestion The utilisation of structural fibre by ruminants is dependent on microbial degradation, and therefore on microbial growth. The rumen fauna includes a diverse range of bacterial species, many which have proteolytic activity so there is extensive and rapid degradation of forage protein to peptides and amino acids. The high solubility of protein in fresh forages makes it vulnerable to degradation by a range of bacteria, including hyper-ammonia producing species (Attwood et al., 1998). About 70% of plant protein is degraded in the rumen when fresh forages are fed (Ørskov, 1999). Amino acids may be used for microbial growth but most bacteria utilise ammonia released from plant amino acids (AA) for microbial protein synthesis. Degradation of de-animated AA yield branched chain and other VFA. The rate of ammonia absorption through the rumen wall increases at higher ruminal pH values, but concentrations above 1000 mg/L or 20 mg/L blood are toxic (Annison et al., 2002). 28 Ammonia production in excess of bacterial utilisation is converted to urea at a net metabolic cost of about 12 kJ/g NH -N (Baldwin, 1995). 3 Although protein degradation provides substrates for bacterial growth, it can account for an excessive loss of forage protein and reduces AA available for absorption. Microbial protein has amino acid composition that is very close to casein and has higher biological value (BV) than plant protein. The BV of microbial protein has been reported to be from 66 to 87 BV. Protein supplements which are fed with concentrates may be heat treated to increase the proportion of undegradable protein (UDP). Inclusion of condensed tannins (CT) in forage diets have reduced the degradation of plant protein (McNabb et al., 1996) and increased the flow of plant protein to the abomasum from 0.30 to 0.44 of intake in sheep (Waghorn et al. , 1994). About 80% of microbial nitrogen is protein. When a good quality pasture is fed, with an OMD of about 75% and containing 25% CP in the DM the loss to degradation is about 175 g CP/kg DM or 233 g CP/kg digestible organic matter intake (OMI). Microbial growth in ruminants fed pastures is lower than for concentrate diets, where CP content is about 18% of the DM and comprises approximately 35% UDP. About 100 g microbial CP is synthesised per kg pasture digestible OMI (ARC, 1980). Equivalent values for concentrate diets are about 130 g microbial CP/kg digestible OMI. Table 2.13 summarises losses and availability of CP for cows fed pasture in spring and summer, compared to those given TMR, but it should be appreciated that predictions of microbial CP from cows fed pasture using National Research Council (NRC, 2001; pages 55-56) information suggest microbial CP will be lower than values from Agriculture Research Council (ARC, 1980). Data from NRC (2001) suggest only about 67 and 86 g microbial CP will be synthesised from each kilogram digestible organic matter from spring and summer pasture respectively. The excess ammonia will be excreted at an added cost to the cow. Data summarised in ARC (1980) and NRC (2001) suggest microbial growth is about 80 g microbial amino acid/kg pasture digestible OM, but the range could extend from 60 to 90 g. Values for concentrate based diets are about 100 g microbial AA/kg digestible OMI. The impact of high CP concentration in spring pasture and the high degradability of CP in both spring and summer forage is emphasised by comparison with the TMR diet (Table 2.13). Net absorption of AA required for metabolism and milk production is substantially less when grass is fed with similar quantities from plant and microbial origin, compared with TMR ration. Care must be exercised when supplements 29 are chosen at any time during lactation to complement the quality and CP content of pasture available to cows. Energy is often considered to be the first limiting nutrient for milk production in pasture based systems, but metabolisable protein supply can be limiting at some times. 2.2.6 - Lipids digestion Rumen microbes rapidly and extensively modify dietary lipids. Hydrolysis of galactolipids (from plant leaves) and triglycerides (from seeds) releases glycerol and galactose which are fermented to VFA. Liberated fatty acids may adhere to the surfaces of bacteria and feed particles but are not degraded by rumen microbes in significant amounts. Some fatty acids are incorporated into cells but most are hydrogenated prior to absorption from the intestine. 30 TABLE 2.13 – An illustration of nitrogenous fluxes and microbial growth in lactating cows fed spring and summer pasture typical of New Zealand farming and a total mixed ration (TMR). Spring pasture Summer pasture TMR a Milk (kg/day) 25 15 30 Milk solids (kg/day) 2.10 1.20 2.40 Dry matter intake (kg/day) 16 12 18 Organic matter digestibility (g/100g DM) 75 67 75 Digestible organic matter intake (DOMI; kg/day) 11.16 7.48 12.56 b Crude protein (g/100g DM) 25 15 18 Crude protein intake (kg/day) 4.00 1.80 3.24 Rumen degradable protein intake (kg/day) 2.80 (70%) 1.26 (70%) 1.13 (35%) Microbial crude protein 98 98 123 (g/kg DOMI) c Microbial crude protein 1094 733 1545 (g/day) Microbial AA absorbed 700 469 989 d Plant AA absorbed 840 378 1474 e Total AA absorbed 1558 847 2463 800 630 972 Faecal crude protein f Milk crude protein 900 540 1080 g 368 101 190 Urinary N h Cost of urea synthesis (MJ/day) 4.41 1.21 2.28 a Intakes and performance from Kolver et al. (2002). Begin of lactation for TMR cows. Assume organic matter = 0.93 of dry matter and 0.66 of digestion occurs in the rumen. b c Microbial crude protein yield assumes 186 g CP (29.7 g N)/kg OM fermented in the rumen, or about 123 g CP/kg DOMI for diets where N flow to the intestines equal N intake. Calculations here assume 0.20 of N is lost as ammonia absorption for cows fed pasture and there is no net loss for the TMR diet (NRC, 2001) but predictions from NRC (2001) page 56 suggest values for spring, summer and TMR would be about 67, 86 and 147 g/kg DOMI respectively. 0.80 of microbial CP is amino acids, of which 0.80 are absorbed. d e 0.70 of undegradable dietary protein is absorbed. Assume 0.80 digestibility for spring pasture, 0.65 for summer and 0.70 for TMR. f Milk protein is 3.6 g/100 g. g Assume no gain or loss of N to body weight and a cost of 12 kJ/g NHh 3-N for urea synthesis (Baldwin, 1995). 31 2.3 - Consequences of forage maturation for the lactating cow The potential for productivity by New Zealand cows has been illustrated by Kolver et al. (2002) (Table 2.1) who showed that a TMR supplied ad libitum was able to sustain lactation more efficiently than pasture given at a high allowance of 60 kg DM/cow.day (Figure 2.8). 1.2 1.8 2.4 0 100 200 300 Days in milk Milk s ol id s (kg /d ay) OS cows fed pasture OS cows fed TMR F I G U R E 2.8 – Milk solids (kg/day) production in early, mid and late lactation for overseas Holsteins (OS) cows fed pasture or total mixed ration (TMR) diet. Source Kolver et at. (2002). The low peak lactation of cows fed pasture may be a consequence of pre- calving nutrition, an inability to consume sufficient nutrients from lush spring grass, or the nutrient balance of the diet, but these effects are confounded by grass maturation in late spring. The impact of maturation is often made worse by insufficient supply (Figure 2.1), typical of many dairying regions which rely on rainfall, rather than irrigation. The impact of lowered NV (and possible feed shortage) typical in late spring/summer in many regions is made worse by substantial body weight losses after calving. Low feeding value can result in post-partum anoestrus (period of non-cycling) which has major implications for calving in the following season (Macmillan, 1997; Verkerk et al., 2000) as well as low milk production. The timing and choice of supplements will be crucial if cow performance is to be maintained and the opportunity to complement mature summer pasture is a focus of this thesis. 32 The rapid decline in milk production post peak lactation (Figure 2.8) is likely to result from limited ability to mobilise body tissue as well as lower NV of feed available to the cow. The mobilisation of body reserves during early lactation can result in daily live weight losses up to 1 kg, which are clearly unsustainable. Maximum energy deficit occurs within 2 - 3 weeks after calving and cows may achieve a positive energy balance approximately 60 days after calving (Holmes et al., 2002). The loss of body weight is indicated by condition score (CS) and in the post-calving period one body CS is about 25 kg of live weight for a mature Jersey cow and more likely 40 kg for a large Holstein-Friesian dairy cow (Grainger and McGowan, 1982). Although it is normal for cows grazing pasture to lose weight after calving, a loss of more than one CS during early lactation indicates the feeding programme or management are not adequate to allow cows to achieve high milk production and remain healthy (Figure 2.9). 0 7 14 21 1 5 9 13 17 21 25 29 33 37 Weeks of lactation kg /c ow .da y 360 400 440 480 Liv ew ei ght, kg /c ow Milk Liveweight Pasture intake FIGURE 2.9 – Phases of lactation cycle for New Zealand dairy cows fed ryegrass/white clover pasture ad libitum (> 60 kg pasture DM/cow.day). Adapted from Holmes et al. (2002). Pasture intake (kg DM/cow.day) was estimated by: (0.372 x FCM + 0.0968 x LW ) x (1 - e0.75 (-0.192x(WOL+3.67))) where FCM = four percent fat corrected milk (kg/day), LW = liveweight (kg), and WOL = week of lactation (NRC, 2001). If a lactating cow is neither gaining nor losing body weight (BW), its feed requirements will depend on its live weight and on its milk yield (Table 2.14). For example, to gain 1 body CS in 70 days, the feed requirement for a Friesian cow producing 1.4 kg MS/day is 15.8 kg leafy pasture DMI/day. 33 TABLE 2.14 – Feed requirements of lactating cows with various levels of milk production, losing, gaining or maintaining a constant level of body condition (kg leafy pasture DM eaten/day). Adapted from Holmes et al. (2002). Live weight of cow (kg) Milk solids 370 (Jersey) 450 (Friesian) 550 (Holstein) Condition score (CS) (kg/day) No change in body condition 0.7 8.6 9.6 10.5 1.4 12.6 13.9 14.9 2.1 16.6 18.2 19.2 a a a Losing 1 CS in 70 days 1.4 11.6 12.6 13.6 2.1 15.6 16.7 17.9 a a a Gaining 1 CS in 70 days 0.7 9.9 11.4 12.3 1.4 13.9 15.8 16.7 a Cows may be unable to eat these amounts of feed, especially in the first month of lactation. A primary goal is to manage the feeding program to properly manipulate body condition loss and minimize the duration and extent of negative energy balance. High milk yield does not cause excessive weight loss if the feeding program is well-tuned, but this can be difficult when grass is the only dietary component. Successful feeding management during spring will enable cows to get in calf and be gaining bodyweight at the beginning of summer. The genetic potential of modern cows exceeds the ability of all dairying systems in the world to meet nutrient requirements – the cow is designed to loose body condition in early lactation. This is due in part to the bulk of ryegrass (Ulyatt and Waghorn, 1993; Waghorn, 2002) as well as its rate of clearance from the rumen, microbial growth and composition of absorbed nutrients (Mertens, 1992b). Hodgson and Brookes (1999) described three main factors affecting pasture intake: 1. nutrient requirements of the cow; 2. factors associated with distension of the alimentary tract, digestibility and rate of digestion and passage of the feed; 3. limitation to the potential pasture intake resulting from a combination of pasture and animal factors affecting grazing behaviour. Several researchers have reviewed the factors controlling pasture intake by ruminants (Demment et al. 1995; Dulphy et al. 1989; Forbes, 2000; Mertens, 1994; Poppi, 34 1996) but none explain the constraints linked with intake of fresh ryegrass, although Waghorn (2002) identified the high moisture content and physical bulk of pasture as a likely constraint to intake by New Zealand dairy cows. He showed that rumen fill was up to 22% of bodyweight (BW) compared to 15-20% of BW in North American cows fed forages and suggested limited increases in milk production could be expected from current feeding practices. Low pasture DMI has been identified as a major factor limiting milk production of high genetic merit cows under grazing conditions (Clark et al. 2001; Clark et al. 1997b; Kolver and Muller, 1998; Waghorn, 2002). 2.2.7 - Options for maintaining cow performance It is essential to improve the intake and nutritive value of diets for dairy cows to reduce the post peak decline in milk production. Careful management of pastures is essential to minimise the intensity and extent of flowering that impacts on rumen fill and to provide sufficient forage during periods of poor growth. Use of supplements will be necessary to overcome detrimental aspects of grass growth on animal performance. The overriding considerations associated with dietary supplementation must be economic viability, and in New Zealand fresh or conserved forage crops will be the main source of supplements. Principal variables will be type of supplements, time of supplementation and extent to which the supplement complements or substitutes for pasture. Supplements vary in nutritive characteristics, including protein content and degradability, DM content, ME content and acceptability, so appropriate choices should be made to optimise provision of nutrients. Table 2.15 summarises cow responses to a range of forage-based supplements evaluated in New Zealand. Supplements may be used to provide more ME, more protein or both. Provision of additional ME will be possible if pasture supply is inadequate, but when a high pasture allowance is available, substitution is likely. When the supplement has a high water or fibre content, it may not increase intake, but use of silage (e.g.: maize silage) may increase total intake of lactation cows (Table 2.15). 35 TABLE 2.15 - Lactation responses of New Zealand cows in post peak lactation when fed supplements with pasture or fed high quality legumes. Adapted from Burke (2003). Diet DMI ME CP Response Milk yield (kg/ Milk solids (kg) (MJME/ kg diet DM) (g/100g DM) (g milk solids/ yield (kg/day) kg DM fed) day) Ryegrass + 20%WC 10.9 10.3 21.5 8.5 0.80 74 1 Ryegrass + 50% WC 12.1 10.5 22.6 10.0 0.93 77 Ryegrass + 80% WC 12.0 10.6 23.8 9.8 0.93 77 Ryegrass 12.1 9.5 14.3 10.2 0.96 79 2 Ryegrass + 25% WC 13.1 10.1 16.4 12.5 1.17 90 Ryegrass + 50% WC 14.8 10.5 18.4 13.6 1.24 84 Ryegrass + 75% WC 15.8 10.7 21.9 13.7 1.26 82 Pasture - 9.7 24.3 10.4 0.87 - 3 Pasture + turnips - 9.8 23.2 11.3 0.99 34 Pasture + chicory - 9.9 24.1 10.8 0.93 32 Pasture 9.6 10 19.7 8.6 0.73 76 4 Pasture + turnips 12.1 10.7 17.1 10.6 0.93 82 Pasture + sorghum 11.0 9.7 17.4 9.2 0.82 80 Pasture 12.4 10.4 11.6 12.4 0.93 75 5 Grass + 75% WC 15.0 11.3 19.1 16.6 1.26 84 Grass + 75% Lotus C 13.8 11.8 20.8 18.3 1.38 100 Ryegrass 14.2 10.6 18.2 10.0 0.83 59 6 Lotus C . – CT 16.7 11.4 25.6 13.8 1.13 68 Lotus C.+ CT 16.8 11.4 25.6 16.5 1.40 83 Pasture (restricted) 12.5 10.0 17.4 13.2 1.00 80 7 Pasture (full) 18.5 10.1 18 17.0 1.11 70 Pasture + PS 17.0 10.3 16.9 14.3 1.11 65 Pasture + MS 16.6 10.1 14.4 13.7 1.12 68 Pasture + Lotus CS 17.2 10.3 19.1 13.7 1.29 75 Pasture + SS 15.7 10.0 16.7 13.7 1.10 70 Pasture 16.9 9.2 19.7 11.0 0.87 52 8 Pasture + MS (66:34) 14.5 9.8 15.6 11.3 0.86 59 Pasture + SS (66:34) 15.1 9.6 19.3 12.4 0.97 64 Pasture + MS:SS (66:16:16) 15.5 9.7 18 13.2 1.02 66 Mean 73 Range 32-100 Abbreviations: DMI: DM intake; WC: white clover; Lotus C: Lotus corniculatus; MS: maize silage; SS: sulla silage; Lotus CS: Lotus corniculatus silage; CT: condensed tannin (– CT indicates inactivation by daily administration of polyethylene glycol). Harris et al. 1997b: ad libitum feeding indoors; Harris et al. 1997a: ad libitum grazing; 1 2 Waugh et al. 1998: 4-8 kg DM supplement + 25 kg DM pasture offered/cow; 3 4Clark et al. 1997a: 4 kg supplement + 25 kg DM pasture offered/cow; 5Harris et al. 1998a: ad libitum grazing ; Woodward et al. 1999: ad libitum feeding indoors; 6 7Woodward et al. 2002: 5kg DM supplement + 25 kg DM pasture allowance/cow; 8Woodward et al. unpublished: ad libitum feeding indoors. 36 Substitution rate has been reviewed by Penno (2002) and Bargo et al. (2003) for cows grazing pasture. In general terms, when cows are consuming high quality forages, SR increases as the energy intake of the cow increases from either forage or the supplement. Penno et al. (1998) found a greater MS response to supplements during winter and summer than spring and autumn with cows grazing pasture but stage of lactation did not affect MS response. They suggested supplementary feeding decisions were based on the level of under-feeding and in Waikato farming the response to supplement was directly proportional to the increase in ME supply by the supplement (Penno et al., 1998). They also concluded that the full lactation responses to supplements were two fold greater than those measured in short term feeding trials. The results of trials conducted by Penno et al. (1998) are due in part to an inadequate supply of pasture, so the supplements provide more ME, but efficient dairy production must provide sufficient feed and the choice of supplement should focus on dietary nutritive value as well as sufficiency. The high CP content of spring pasture and losses to ammonia become a substantial metabolic cost which can exceed 4.4 MJ/day (Table 2.13). Forage legumes that contain condensed tannins (CT) (Lotus corniculatus, Lotus pedunculatus, Hedysarum coronarium) may have potential in the pasture-based system because the CT can reduce protein degradation in the rumen and allow a greater passage of undegraded protein to the small intestine for absorption (McNabb et al., 1996). These forages could improve the UDP supply when used to supplement spring pastures and at others times of the year. Woodward et al. (1999) showed that cows fed Lotus corinculatus produced 51% more milk than pasture-fed cows, with the response due to the effect of CT (Table 2.15). The reasons why New Zealand farmers currently do not supplement with legumes (fresh or ensiled) are due to failure of legumes to produce competitive yields of dry matter, difficult agronomic or management requirements to maintain pure or mixed swards, and the high cost of ensiled these forages. It is clear from past experiments (Table 2.15) and the changing quality of pastures that the choice and supply of supplements is complex and must be based on the chemical composition, rates of digestion, and nutrient supply from existing pasture as well as characteristics of the supplement. These values can be measured in cow feeding trials, but these are very expensive and time consuming. An alternative procedure is to characterise both pasture (at different maturity and composition) and supplements using in vitro, in sacco and chemical analyses to predict cow responses. 37 This form of evaluation is indirect but cost effective. Results from indirect evaluations must be validated against cow production trials where supplements are given with pasture and a good relationship could form the basis for a predictive model to provide appropriate supplements for cows given a wide variety of pastures. 2.4 - Techniques for evaluating feeding value of pasture and supplements 2.4.1 - Chemical composition Measurements of chemical composition by wet chemistry or prediction with NIRS have a role to play in determining nutritive value, but they are not discussed here and on their own are unable to predict animal performance. 2.4.2 - In vitro incubations In vitro systems to study rumen fermentation have been used extensively for more than 40 years. They involve fermentation of different feeds, often with rumen fluid (microbial inoculum) but some times with enzymes in a buffered anaerobic environment. These systems differ in their degree of complexity and a brief indication of alternative methods with their principal attributes and weaknesses is given below, with several factors affecting the outcome of these systems. Many in vitro incubations are batch cultures where incubations proceed for set time periods without addition or removal of material from vessels. A common evaluation is that of protein degradation which can be calculated from the ammonia nitrogen released from protein degradation with losses to both ammonia nitrogen (NH3- N) and utilisation by ruminal microbes. Interpretation of proteolysis is affected by the nature of the substrate, incubation time and feed preparation. The problem of N incorporation into bacteria has been avoided through the use of chloramphenicol and hydrazine sulphate to inhibit microbial protein synthesis and prevent AA and ammonia nitrogen utilisation by ruminal microbes (Broderick, 1987). This method can give degradation rates similar to in vivo observations (Broderick and Albrecht, 1997) but values were about 30% lower than uninhibited systems (Hristov and Broderick, 1994). It is more logical to allow a normal microbial growth, without artificial inhibition and to use NH to quantify microbial nitrogen utilisation (Hristov and Broderick, 1994). 15 3 38 Protein degradation rates are computed as appearance of NH3 plus net microbial protein synthesis. A major limitation of the Tilley and Terry (1963) method is the need for rumen fistulated animals to provide rumen fluid. This can be avoided by using enzymatic method and good correlations can be obtained between enzymatic digestion and in vivo or in vitro digestibility for a variety of forages. Jones and Theodorou (2000) summarised enzymatic methods and indicated potential problems, especially with feeds having high levels of ammonia and free AA (as in forage silages). The breakdown of slowly degraded residual proteins (i.e. less than 0.01 h-1) must be computed from the appearance of additional ammonia and AA in the presence of high background nitrogen and is not accurate. Microbial activity tends to decrease over time, especially if pH is not maintained. The accumulation of ammonia and AA may also result in end-product inhibition of enzymes (Calsamiglia et al., 2000). It is important to realise that these problems apply equally to enzymatic or microbial systems. The value of either system will depend on research objectives so it is easier to obtain a ranking of plant material than to obtain absolute values which match in vivo degradation rates. 2.4.2.1 - Continuous culture The most physiologically appropriate approach to determining constituent degradation in vitro would be to design a system that carefully simulates ruminal fermentation. Various continuous-culture fermentation systems have been designed to simulate the ruminal environment, enabling the study of ruminal microbial ecology and digestion of nutrients, for example dual-flow continuous culture (Hoover et al., 1976) or single-flow Rusitec (Czerkawski and Breckenridge, 1977). Advantages of these systems compared with in vivo measurements include reduction in cost, time and variation among experimental units. Continuous cultures avoid problems of products inhibition or altered pH but the microbial population will not mimic in vivo over long periods. Compared to in vivo, there are no complications from endogenous N sources, and digesta flow markers are not required because passage rates are regulated and measured directly. However, similar problems exist as with in vivo measurements; reliable techniques are required for isolation of microbial cells and for differentiation of effluent digesta into microbial and dietary fractions. These systems are tools for research and modelling of ruminal fermentation, but are elaborate and expensive, require inoculation with ruminal digesta and may not be suitable for routine analysis of microbial digestion for individual feed ingredients (Calsamiglia et al., 2000). 39 2.4.2.2 - Products of fermentation In vitro incubation enables the products of digestion to be quantified. Measurement of VFA and ammonia production during microbial fermentation can indicate the nutritive value (NV) of forages in terms of protein losses to degradation and yields of VFA. Measuring ratios of VFA especially the proportion of glucogenic: lipogenic precursors (propionate: acetate and butyrate) also help understanding the NV of feeds. The measurements of ammonia yield indicate net proteolysis but omits the quantity of ammonia-N incorporated into bacteria, which Barrell et al. (2000) reported to be 10% of white clover N and almost 20% of degraded N after 24 hours. Hence a measure of microbial growth would add value to data sets because it will indicate the extent of N capture and also the nutritive value of the substrate for bacterial growth. The proportion of plant N converted to ammonia indicates susceptibility to degradation and can indicate effects of CT on proteolysis. Ammonia production will indicate relative ability to meet bacterial N requirements, the amount of NH3 needing to be detoxified post-absorption and the amount of protein which has not been degraded and is potentially available for AA absorption in vivo. Hence proportional loss of plant N to ammonia will indicate the likelihood of insufficient N for bacterial growth, for example when plant CP concentrations are less than 9% of the DM in vivo. Conversely a large ammonia production in vivo results in a large toxic load to be converted to urea for excretion. The measurements of net ammonia production during in vitro incubation provides a relative measure of protein breakdown in the rumen, nevertheless this does not account for the nitrogen taken up by the bacteria. 2.4.3 - I n situ or i n sacco incubations The in sacco technique is synonymous with in situ, dacron or nylon-bag technique. In sacco incubations measure the disappearance of feed components from a bag containing the test diet after incubation, for a variable period, in the rumen of an animal fitted with a rumen fistula. Degradability of DM, CP, fibre and energy can be measured against time. The technique was first used to provide a dynamic assessment of protein degradation by Mehrez and Ørskov (1977). Adesogan et al. (2000) suggest the in sacco rumen degradability technique to be theoretically superior to in vitro digestibility techniques because it provides information on the dynamics of forage digestion. Digestion kinetics are important, but 40 such information will be affected by the way material is prepared for incubation and no information is obtained on the products of digestion when the in sacco method is used. Several authors (Weiss, 1994; Nozière and Michalet-Doreau, 2000; Ørskov, 2000) have reviewed the methodology of the in sacco incubations and identified several important sources of variation in results from this technique. These are briefly described below, as they influence the interpretation of data obtained in this study. 2.4.3.1 - Animal and diet Mehrez and Ørskov (1977) concluded that three cannulated animals should be used and that replication within animals and between days was not worthwhile and made little difference to the total variance. However, if the objective is to rank the feed potential of forage selection for forage or quality, then only one animal needs to be used (Ørskov, 2000). 2.4.3.2 - Host dietary effects It is important to standardise diets of animals use for in sacco incubations (and provision of inoculum for in vitro incubations). Mould and Ørskov (1984) fed cattle a high-quality roughage diet with about 25 g nitrogen (N) kg-1 digestible DM. Reasons for different degradation rates observed with contrasting rations of roughage and grains are not well defined but probably relate to a combination of ruminal microbial (Weimer et al., 1999) and physical factors that are subject to dietary changes. Animals should be fed twice daily at a maintenance level or slightly above, with a minimum interval between meals of 8 hours. Weakley et al. (1983) showed clearly that diet affects the DM and N disappearance from dacron bags in the rumen for cows fed different hay: grain ratios. When a soybean meal was incubated, cows fed a high grain diet resulted in the slowest DM and N losses. The type of forage fed to the recipient animal can also affect in sacco disappearance of forages, but results have been inconsistent (Weiss, 1994). 2.4.3.3 - Bag type and sample size The type of material used for the in sacco bags is polyester, nylon or dacron, with the latter being readily available from the Ankon® Corporation. The principal requirement for bags, aside from indigestibility, is a consistent pore size, so a welded monofilament mesh is preferable to a woven multifilamentous mesh (Australian Agricultural Council (AAC), 1990). The pore size must permit free exchange of fluid and 41 microorganisms between the bag and the ruminal liquor and be small enough to prevent loss of indigestible particles or the entry of feed particles. The pore size recommended for in sacco incubations is between 30 and 50 µm (Hvelplund and Weisbjerg, 2000; Nocek, 1988). Sample behaviour and digestion kinetics will also be affected by the quantity in the bag. There should not be an excess of sample in the bag because this could affect the rate at which bacteria enter the bag and colonise the test material but, equally, sufficient sample needs to be incubated so sufficient residues are available for analyses. The latter point can be accommodated by using bags of different sizes (e.g.: 5 x 10 cm versus 10 x 20 cm) and this will be affected by sample preparation. Fresh minced forages may contain as little as 10% DM, so a substantial bulk must be placed in a bag, compared to the conventional use of freeze dried and ground material. Hvelplund and Weisbjerg (2000) recommend in sacco incubations contain 10 - 16 mg DM/cm2 of bag surface. 2.4.3.4 - Bag placement and incubation sequence The rumen environment is often compacted and layered with a raft or mat in the dorsal aspect with a liquid phase in the ventral rumen. This heterogeneous environment makes it difficult to obtain a consistent placement of bags, so typically they are contained within a large aperture lingerie bag and weighted to reduce mobility and prevent floatation. Without the weight the bags may float on the surface of the rumen and not become properly incorporated into the digesta. A cord length at least equal to the distance from the cannula to the bottom of the rumen is recommended for attachment to the cannula. For comparison of degradation rates, it is probably better to introduce all the bags at the same time in the rumen because there will be a similar microbial environment for all bags during initial stages of digestion, compared to the sequential placement method (Michalet-Doreau and Ould-Bah, 1992). 2.4.3.5 - Washing and drying procedure For washing, cold water should be used. Hvelplund and Weisbjerg (2000) recommend an automatic machine washing as preferable for standardisation, with a washing time of 10 - 15 minutes. Different methods of washing bags post incubation gave extremely variable results (Huntington and Givens, 1995). There appears to be a need for further research in order to standardise this procedure and more work is required to elucidate the optimum process. However, Cherney et al. (1990) compared 42 and evaluated effects of length of time of machine and hand rising on DM disappearance, DM remaining and their standard error (SE). Machine rinsing twice for two minutes or hand rinsing bags resulted in a similar SE. Machine rinsing twice for five minutes was too long and resulted in an under-estimate of DM remaining after incubation. 2.4.3.6 - Microbial contamination Microbial contamination arises from bacteria adhering to plant residues, which are not removed by washing. The extent of contamination is dependent upon incubation time and the extent to which forages have been degraded, so contamination is not constant. Correction for attachment is more important with high microbial contamination values and adds an additional variable to the assay. Attachment is measured using microbial 15N-labelled markers (Hvelplund and Weisbjerg, 2000) but such measurements are not practical during routine in sacco incubations. The importance of microbial contamination appears to be minor with concentrates (Nocek, 1985) but Mathers and Aitchison (1981) demonstrated about 20% of the residual N in lucerne samples arose from microbial contamination after 24 hours in the rumen. For lucerne hay, Blair and Cummins (1983) cited by Nocek (1985) reported 4.1, 11.7 and 8.7% of total N was microbial after 12, 18 and 30 hours of ruminal exposure with a subsequent water wash, respectively. These effects can influence calculation of degradation kinetics, with reductions of 0.5 – 5 hours in lag time of hay corrected for microbial N contamination compared with non-corrected ruminal N digestion (Nocek and Grant, 1987). The effects of correction on rates of forage digestion were variable and although microbial contamination of residues does underestimate N degradation, removal can be expensive, laborious or inaccurate (Olubobokun et al., 1990). 2.4.3.7 - Modelling in sacco degradation kinetics A number of methodological factors affecting the experimental measurements of in sacco disappearance have been considered (Huntington and Givens, 1995; Nocek, 1988; Nozière and Michalet-Doreau, 2000) and the choices of mathematical models used to fit curves and estimate rumen degradation parameters have been described by López et al. (1999). The characteristics of feed degradability have been defined as the soluble fraction, A; the insoluble but degradable fraction, B and the speed at which the B fraction is degraded (k). These rate constants have been used in an attempt to 43 develop a system that can predict not only feed nutritive value but consumption as well (Ørskov, 2000). McDonald (1981) was the first to define a lag phase indicating that the microbes take time to adhere to the substrate and for a time at the beginning of the in sacco incubation (up to 2 - 4 hours) there is no loss in DM. In fact for some feeds there may be a small increase in DM. The lag phase will be used in all digestion kinetics studies described in this thesis. The characteristics of the degradation curve are described for DM by equation (i): P = A + B (1-e-k (t-L)) Equation (i) where: P = potential degradability (%) t = incubation time (hour) A = soluble DM (% of DM washed out of bags at t = 0h) B = degradable insoluble DM (%) L = lag phase (hours) k = the fractional disappearance rate (k, %. hour-1). 1 – (A+B) = the undegradable portion of a sample (C; %). In the example given above and in Figure 2.10, the degradation referred to DM, but fibre (NDF and ADF) and protein are also evaluated in terms of digestion kinetics. Figure 2.10 illustrates the degradation of a typical forage. In c u b a tio n t im e (h ) 0 2 0 4 0 6 0 D M d eg ra da tio n (% ) 0 2 0 4 0 6 0 8 0 1 0 0 U n d e rg ra d a b le f ra c t io n (C ) 1 - (A + B ) S lo w ly d e g ra d a b le f ra c t io n (B ) S o lu b le f ra c t io n (A ) L a g t im e FIGURE 2.10 – Typical in sacco degradation curve of forage DM. 44 Both the lag time (L) and soluble (A) fraction are affected by the incubation and preparation procedure, so it is important that preparation is well defined and repeatable. The largest component of plant degradation is represented by the parameters B and k which result from microbial-plant interactions. 2.4.3.8 - Determination of effective degradability Effective degradability takes into account both extent and rate of digestion, to enable comparison between forages, and indicate relative nutritive value. To estimate the extent of ruminal digestion, mathematical methods for interpretation of the in sacco degradation profile may include a fixed turnover rate to indicate rate of outflow. Ørskov and McDonald (1970) proposed an integration of the in sacco degradation profile in relation to the ruminal particulate outflow rate. The effective degradability (E) can also be calculated from the kinetic parameters obtained from equation (ii): E = A + ((B * k) / (k + k )) Equation (ii) d Where A, B and k are defined for equation (i) and kd is fractional passage rate (or rumen small particle outflow rate; %.hour ). The value for k has been set up 6%.hour-1 -1d for analysis of forage digestion. The main methods used for fitting data to the kinetic model are: a) Logarithmic transformation followed by linear regression (lnLIN); b) Non-linear least square regression (NLIN). Both are procedures of SAS program (SAS, 2001). In most cases, NLIN regression methods are preferred because they result in the smallest residual sum of squared deviations from the model (Nozière and Michalet-Doreau, 2000). For the last 20 years, the in sacco technique has proved to be useful and robust technique for estimation of the feeding value of forages, and effects of the rumen environment and degradation. Description of degradation characteristics is inexpensive, repeatable but must relate to in vivo performance to be useful. 45 2.4.4. Sample preparation The type of feed preparation has very significant effects on degradation kinetics, in terms of constituent disappearance, proteolysis, VFA production and microbial growth. Although conventional feed preparation involves freeze drying and grinding, this is not appropriate for fresh forages which have been both minced and freeze dried for comparison in sacco (McNabb et al., 1996). Incubation of masticated feeds would be the optimal preparation for in sacco studies, but it is difficult to obtain masticated material. Barrell et al. (2000) and Cohen and Doyle (2001) demonstrated degradation kinetics of fresh chopped, minced and also freeze dried and ground material and suggested the minced preparation to be most appropriate for fresh grasses and legumes. In vitro and in sacco incubations should be carried out with feed prepared in a way that best mimics chewing by ruminants. Particle size has an influence on the accessibility of dietary components to the microflora enzymes. A few studies have shown that the lag time significantly increased with an increase in particle size (Barrell et al., 2000; Emanuele and Staples, 1988). This maybe overcome by incubating material of a similar size, rather than by passing through a single mesh to allow a wide range in particle sizes to be included in the incubator. However the technique must also relate to chewed material in vivo to reproduce the range of mean particle sizes in a swallowed bolus. Barrell et al. (2000) demonstrated that mincing fresh forage with a 12 mm screen aperture simulated particle distribution achieved by mastication in ruminants (Ulyatt et al., 1986; Waghorn et al., 1989). For wet materials, e.g. forage and silages, Ørskov (2000) also suggested that a mincing machine is the most appropriate form of preparation. 46 2.5 - Simulation modelling The purpose of a model is to describe mathematically the response of each compartment or several connected compartments to a variable or combination of variables. A model is considered mechanistic when it simulates behaviour of a function through elements at a lower level (Gill et al., 1989). Most biological responses are integrated and nonlinear and change over time (Sauvant, 1991 cited by Fox and Barry, 1994). Mechanistic models provide a useful means of integrating knowledge and formulating hypotheses. Thus mechanistic modelling is an integral part of a research programme, with experimental and modelling objectives highly inter-related (Dijkstra and France, 1994). Prediction of requirements and feed utilisation by ruminants is unique to each production setting, so models should integrate knowledge of feed, intake, digestion and passage rates in relation to feed composition, energy concentration, digestion and escape of dietary protein and microbial growth. (Fox and Barry, 1994). Several models used for ration balancing (Udder, GrazFed, Camdairy, feedTECH) are empirical rather than causal, and often they are not interactive or user friendly. Models used for ruminal studies, including Molly (Baldwin et al., 1987) and Dijkstra’s rumen model (Dijkstra et al., 1992), are complex, difficult to use and are not designed for interpreting cow performance in relation to feed characteristics. In most application systems (AFRC, 1993; AAC, 1990; Institut National de la Recherche Agronomique (INRA), 1989) the prediction of metabolism of nutrients is not as advanced as the prediction of ruminal fermentation, because the fate of nutrients in the animal is less well understood than their production in the rumen. Metabolic routes connect tissues and metabolic compartments, involve interactions among nutrients and include metabolic regulations which drive the partioning of absorbed nutrients between tissue maintenance and production (Sauvant, 1991 cited by Fox and Barry, 1994). Therefore a combination of mechanistic and empirical approaches must be used. Systems are generally steady state and static, and use statistical representations of data that represent the aggregated response of whole compartments. This approach was used in developing the Cornell Net Carbohydrate and Protein System (CNCPS) for evaluating cattle diets as described and validated by Russell et al. (1992); Sniffen et al. (1992); Fox et al. (1992); O'Connor et al. (1993); Pitt et al. (1996) and Tylutki and Fox (1997). The CNCPS model has been used in the interpretation of data obtained from experiments described in this thesis. 47 A problem with nutrition models is that they cannot predict intake well consequently giving a weak prediction of performance. They are designed to estimate nutrient supply not predict cow performance (St-Pierre and Thraen, 1999). The CNCPS (version 5.00.20) systems will be tested in Chapter 7 to determine the accuracy and utility of this model to predict and formulate diets that were based on pasture and forage supplements using data obtained in sacco and in vitro incubations and two dairy cow trials conducted in mid-lactation, where pasture was complemented with contrasting silages. 2.6 – Conclusion Dairy cow performance is affected by feed quality and availability. In New Zealand the ryegrass based pastures show a significant decline in quality in late spring in association with flowering. This change is bought about by increasing proportions of stem and reductions in leaf with a net lowering of both ME content and cow intake. The changes in ryegrass have a significant input upon cow performance, and this thesis quantifies these changes as ryegrass matures. The hypothesis is that both degradation rate and yield of nutrients diminish from ryegrass as it matures and these can be quantified to develop forage mixtures, which provide balanced diets for dairy cows. 48 Chapter 3 Chemical composition, i n vitro and i n sacco digestion kinetics of perennial ryegrass as influenced by stage of maturity 1 1 A small portion of these data has been previously published in the Proceedings of the New Zealand Society of Animal Production , 2002, 157-162. 3.1- Abstract Animal performance from grass-dominant pastures is affected to a large extent by pasture quality and especially maturation in relation to flowering. Maturation in late spring/summer lowers feed quality due to the changing proportions of leaf, stem, inflorescence and dead matter and the associated changes in chemical composition. Perennial ryegrass of increasing maturity has been used for in vitro and in sacco incubations to determine the net production of ammonia from protein degradation and yields of volatile fatty acids (VFA) and to determine rates of digestion of the dry matter (DM), crude protein (CP) and fibre fractions of the dry matter (DM). The principal finding was a rapid decline in crude protein content from about 24 g CP/100 g of DM harvested at 22 days to 10 g CP/100 g of DM for that harvested after 60 days. These changes were associated with increases in the neutral detergent fibre (NDF) fraction of the DM (43 to 54 g NDF/100 g of DM) and in lignin concentration (2.7 to 4.9 g lignin/100 g of DM). Changes were more rapid in late-cut than early cut forages. The principal consequences of increased maturity were slower degradation rates of DM (k = 0.11 to 0.03 h-1) and NDF (k = 0.14 to 0.03 h-1) and no effect on CP degradation rates. There was a significant change in the pattern of ammonia and VFA production when ryegrass matured. Keywords: ryegrass; forage maturity; digestion kinetics; in sacco; in vitro; dairy cows. Short title: Digestion kinetics of ryegrass 3.2 - Introduction The effect of maturity on digestion and animal performance arises mainly from changes in plant morphology and in cell wall components, which affect dry matter (DM) intake and digestibility (Van Soest, 1994). Maturity is the most important factor affecting pasture quality. From a nutrition perspective, forage quality relates to the feeding value or the ability to utilise feed for production (e.g. milk, meat, wool). Forage quality is never static; plants continually change in quality as they mature. As plant cell-wall content increases, indigestible lignin accumulates and in late spring, grass maturity changes so rapidly that it is possible to measure a significant decline in forage quality every two or three days (Cherney et al. , 1993). Changes occur in both the proportions of leaf, stem and inflorescence as well as the chemical composition of these structures. Both the extent and the rate of change in components are important and farmers face a considerable challenge 50 when they manage pastures to ensure a high nutrient intake for high producing animals. Accurate prediction of performance for animals fed ryegrass diets and options for supplementation with high-quality forages requires a greater understanding of how digestion processes are influenced by the stage of maturity. The objectives of this experiment were to analyse and compare changes in composition, digestion and its products when ryegrass (Lolium perenne L.) grows to maturity in spring. The rate of change in quality is also important and the effect of cutting date on these components was included in the evaluation. The hypothesis to be tested here was that ryegrass maturation will alter rates of degradation and products of digestion and these changes will be affected by initial cutting dates of the sward. Experiments were designed to create a dataset for use in dairy nutrition for grazing cows in New Zealand. 3.3 – Material and methods A one-year-old pure perennial ryegrass sward (Lolium perenne L. cv. Grasslands Samson) was grown at AgResearch Grasslands Aorangi Research Station. A 15 m by 25 m plot was divided into three equal areas separated by trimmed pathways and mown on either 21/08/2000 (area 1), 11/09/2000 (area 2) or 21/09/2000 (area 3) as illustrated in Figures 3.1 and 3.2. After the initial cut, the plots were allowed to grow for the duration of the trial until 11 December with about 2 kg of forage harvested from each area at 7 - 14 day intervals for analysis. Table 3.1 indicates the schedule of cutting dates and days of re-growth (age) of the forage harvested from each area. TABLE 3.1 – Cutting schedule days of re-growth and harvesting dates of ryegrass from three areas. The samples were used for in vitro and in sacco incubations. Area 1 2 3 Initial harvest date Sample dates 21 August 11 September 21 September 11 September 21 a 21 September 31 a 10 a 5 October 45 f 24 f 14 f 13 October 53 b 32 b 22 b 24 October 64 43 g 33 g 3 November 74 c 53 c 43 c 10 November 81 60 50 g 17 November 88 d 67 d 57 d 27 November 98 77 67 4 December 105 e 84 e 74 e 11 December 112 91 81 Superscripts indicate samples analysed in vitro (A to E) and in sacco (A to G). 51 Samples were cut with electric clippers approximately 5 cm above the soil level. Ryegrass re-growth was harvested on 11 occasions from area 1, ten occasions from area 2 and nine occasions from area 3 (Table 3.1). The areas harvested decreased as the ryegrass matured to maintain approximately the same amount of sample harvested for each period of ryegrass re-growth, and forage from each area was stored in plastic bags and frozen at -16oC for chemical analysis, in vitro and in sacco incubations. The area and weight of forage harvested were measured throughout the trial and the height of the sward determined using a ruler to estimate mean length of leaf or leaf and stem. This design enabled the nutritive value of ryegrass to be monitored in relation to age (length of re-growth period), harvest dates and in relation to herbage mass (t DM/ha). Although the effects of plant age (maturity) on nutritive characteristics are well known, this study was designed to determine whether the effects of initial cutting date and rate of growth affects changes in nutritive value. The chemical, in vitro and in sacco analyses provided information on the composition of the ryegrass, the ammonia and VFA production during in vitro digestion and the rate of DM, CP, NDF and ADF disappearance during in sacco digestion. In vitro and in sacco analyses were undertaken with 15 and 21 samples respectively (Table 3.1). All the samples were submitted for chemical analyses (wet and NIRS) and DM determination (AOAC, 1990). Once the grass was harvested and frozen it was maintained below 0oC throughout all preparations, including chopping, mincing, weighing into in vitro bottles and dacron in sacco bags. The material was thawed when in vitro incubations commenced or immediately prior to placement in the rumen for in sacco incubation. FIGURE 3.1 – View of the ryegrass plots midway through the maturation trial. 52 FIGURE 3.2 – The organization of plots showing areas 1, 3 and 2 (left to right) with newly mown paths separating each area. 3.3.1 - Preparation of fres h forage for incubations Ideally the sample preparation for incubations should mimic the particle distribution resulting from chewing during eating and rumination. Frozen forages were chopped into approximately 2 cm lengths (scissors) and minced in a Kreft Compact meat mincer R70 (Kreft, GmbH) fitted with a screen plate with 12 mm holes (Barrell et al., 2000; Burke et al., 2000; Chaves et al., 2001). The mincer components (screen plate, housing, screw, cutter and loading tray; Figure 3.3) were placed in a freezer prior to mincing to ensure the grass remained frozen and this enabled the forage to be macerated rather than squeezed and prevented excessive cell wall rupture during mincing. The process was designed to mimic effects of chewing by ruminants as far as possible (Ulyatt et al., 1986; Waghorn et al., 1989). This method is described by Waghorn and Caradus (1994) and Barrell et al. (2000), and is similar to the method used by Cohen and Doyle (2001). FIGURE 3.3 – Mincer used for fresh forage preparation. Approximately 500 g of wet material were used for each ryegrass sample for in sacco and in vitro incubations, chemical and particle size analysis. Mincing took place 53 within 1-3 days of incubation. The mincing procedure involved assembly of cold mincer parts and mincing the chopped frozen material. Care was taken to ensure the sample did not thaw, so aliquots of about 200 g were removed from the freezer, minced and the minced material returned to the freezer. With very mature forage, only 250 g could be minced before the mincer parts become warm. When this occurred the mincer parts were dismantled, washed with cold water, dried with paper tower and returned to the freezer. Once cold, the process was continued to provide sufficient material for analysis. Minced material was stored at -16oC in sealed plastic bags until the day prior to incubations. In vitro incubations required about 0.5 g DM (approximately 1.5 – 3.0 g wet weight (ww), depending on DM content) to be placed in incubation bottles and 5.0 g DM (15 - 30 g ww) into 100 x 100 mm dacron bags (35 μm pore size; Figure 3.4) for in sacco incubation. FIGURE 3.4 – Visual aspect of the ryegrass after mincing. 3.3.2 - Source of rumen fluid for i n vitro and i n sacco incubations One non-lactating Friesian cow fitted with a permanent rumen fistula was used for all the in sacco incubations and provided rumen liquor for in vitro incubations (Figure 3.5). Three forage samples were incubated as one batch and both in vitro and in sacco incubations commenced at the same time with rumen liquor removed immediately prior to placement of bags in the rumen. The liquor was used for inoculum of in vitro bottles within 20 minutes of collection. A single cow was used for all incubations to avoid effects of variation between animals in studies on digestion kinetics (Waghorn and Caradus, 1994; Nozière and Michalet-Doreau, 2000; Ørskov, 2000). Differences in microbial populations between individual cows can exceed differences attributable to contrasting diets (Weimer et al., 1999). The effects of diet on microbial population (Nocek, 1988; Weiss, 1994) were minimised by feeding a single diet of good quality lucerne hay to the cow 10 days prior 54 and during each incubation. The hay was fed at maintenance (Madsen and Hvelplund, 1994), at 07:00 and 19:00 h, with water available ad libitum . FIGURE 3.5 – The Friesian cow used in all incubation runs. Collecting rumen samples. 3.3.3 - Chemical analyses and particle distribution Samples of minced forage were retained for chemical analysis by Near Infrared Reflectance Spectroscopy NIRS (Corson et al., 1999), measurement of dry matter content by drying at 60oC for 24 hours and particle size distribution by wet sieving (Waghorn et al., 1986). 3.3.4 - In vitro incubations The in vitro incubations provided measures of net ammonia production, yield and proportions of VFA (acetate, propionate, butyrate, iso-butyrate, valerate and iso- valerate) and changes in pH over each incubation period. The net ammonia production and yield of VFA were expressed in terms of plant material incubated (e.g.: µMol NH3-N/mMol plant N) and plotted over time. In vitro incubations were undertaken in 50 mL Schott bottles with bicycle valves fitted into the lid (Figure 3.6) enabling fermentation gases to escape. Bottles were placed in an incubator (Gallenkamp orbital incubator. Cat. No. IOC400.XX1.C, Made in UK) with good temperature control which had been fitted with a rack to handle simultaneous incubations of up to 94 bottles (Figure 3.7). Incubator temperature was 55 maintained at 39 ± 0.5 oC and the rack was set at 90 oscillations per minute. Three forage samples were incubated simultaneously, with standards. Each minced forage was weighed into 24 Schott bottles and at each sample time three bottles of each forage were removed for sampling and analysis. Sampling times were 0, 2, 4, 6, 8, 10, 12 and 24 hours. On the day before the incubation, 1.5 - 3.0 g of sample, corresponding to 0.5 g DM frozen minced material was weighed into incubation bottles. On the day of incubation, the bottles containing forage substrate were warmed to 39oC in the incubator for 60 minutes, gassed with CO2 before adding 12 mL of artificial saliva (buffer, Appendix 1) saturated with carbon dioxide at 39oC and 0.5 mL cysteine sulphide reducing agent (Appendix 1). This process took about 40 seconds for each bottle, which was then capped and returned to the incubator at 39oC. Rumen liquor was obtained from the cow, strained through cheese cloth (Figure 1.1A, Appendix 1) into a two litre thermos® flask and 3 mL dispensed into each bottle. Addition of rumen liquor to 76 bottles took approximately 12 minutes. Appendix 1 shows details for incubation procedure. FIGURE 3.6 – Bottle used for in vitro incubation. FIGURE 3.7 – Gallenkamp incubator used for in vitro incubations. 56 The cow was fed 120 - 150 minutes before the rumen liquor was collected. The pH was measured at the time of collection and sub-samples of rumen liquor were taken for determination of ammonia and VFA concentrations. Three bottles of each forage preparation were removed from the incubator for determination of ammonia and pH concentrations at each sampling time. VFA concentrations were measured in pooled samples from three bottles at 0, 6, 12 and 24 hours. Incubation residues were retained for future measurements of microbial DNA to indicate microbial growth. Changes in pH were used to indicate the ability of the buffer to maintain the in vitro environment despite production of VFA. Low pH values (below 5.6) demonstrated an atypical situation for forage fermentation and data from these bottles were removed from the analysis. 3.3.4.1 - Standards for in vitro incubations Freeze dried and ground lucerne standards were included with all in vitro incubations to monitor variation between runs. Three bottles of the standards were removed at 2 and 8 hours of incubation. Freeze dried and ground lucerne was used as an internal standard because ryegrass occasionally demonstrated substantial microbial inhibition. It is most important that the internal standard demonstrates consistency of incubation characteristics. 3.3.4.2 - Determination of ammonia The procedure for sampling bottles was as follows: nine bottles were removed from the incubator and pH was measured. Sub-samples of 1 mL were added to 1.5 mL microcentrifuge tubes containing 15 µL concentrated HCl and centrifuged at 14000 rpm (14000 g) for 15 minutes (eppendorf – 5415 D Centrifuge ; Figure 3.8) to precipitate particulate matter. The supernatant was transferred to a 1.5 mL microcentrifuge tube and frozen for ammonia analysis (Chaney and Marbach, 1962; Appendix 2). 57 FIGURE 3.8 – Equipment for sampling in vitro bottles. 3.3.4.3 - Determination of VFA A further 1 mL of liquor was removed and centrifuged for determination of VFA concentrations. Two samples of approximately 0.4 mL were taken from each bottle and combined (from the triplicate bottles of each forage sample) to provide one sample for VFA analysis and a spare. Samples were centrifuged as above and the supernatant frozen for analysis by gas liquid chromatography described by Attwood et al. (1998) – Appendix 3. 3.3.4.4 – Fermentation kinetics Fermentation kinetics for each data set were used to evaluate in vitro incubations. Changes in pH were plotted against incubation time to remove data when pH concentrations dropped below 5.6. Excessively low values (below 5.6) probably mean that the data are no longer representative of in vivo digestion and were discarded. During the incubations, ammonia is produced by the microbes as a product of protein degradation during fermentation. Some of this ammonia is used by bacteria for growth. Thus ammonia measurements from in vitro incubations represent the net amount of ammonia produced rather than the gross amount. The measurements of net ammonia production during in vitro incubation provided a relative measure of protein breakdown in the rumen; nevertheless this does not account for the nitrogen taken up by the bacteria. Both ammonia concentration and net NH3-N production as a proportion of plant N were plotted against time to indicate rate of proteolysis for each incubation run (A, B, C, D and E; Table 3.1). 58 Measurement of VFA and ammonia production during microbial fermentation can indicate the nutritive value of a feed in terms of protein losses to degradation and yields of VFA from fermentation. Measuring ratios of VFA contribute to the understanding of the nutritive value of feeds, especially proportions of propionate: acetate. Concentrations of VFA expressed per gram of DM incubated were plotted against incubation time (0, 6, 12 and 24). VFA data have been combined across areas to give values for young (under 33 days), medium (43 - 57 days) and mature (over 67 days) ryegrass. 3.3.5 - In sacco incubations The in sacco incubations allowed the rates of disappearance of DM, CP, NDF and ADF to be calculated. Approximately 5 g DM was weighed into dacron bags, sealed and held at -16oC prior to incubation. Bags were incubated for 0, 2, 6, 12, 24 and 72 hours, with duplicate bags at each time for each forage sample. The procedure for incubation required duplicate bags for each forage sample to be placed in a weighted (350 g) lingerie bag, so that five lingerie bags were placed in the rumen at the commencement of incubation. The 0h bags were not placed in the rumen. Minced ryegrass standards were incubated and removed at 2 and 12 hours to monitor between-run variation. After removal from the rumen, dacron bags containing forage residues were washed under cold water until no further colour appeared (i.e.: all soluble material was removed) prior to drying at 60oC for 24 hours. Dried bags were weighed and analysed to determine disappearance of DM, residues were removed, ground and analysed by NIRS to determine CP, NDF and ADF concentration in the DM. Calibration curves had been developed for NIRS from samples of bags residues collected after 2 to 72 hours of incubation and analysed by wet chemistry during initial incubation runs (See Table 6.1A Appendix 6 for details). The quantity of DM digested, with composition of minced plant material less residues enabled loss of ryegrass constituents to be calculated and plotted to enable a visual appraisal and calculation of digestion kinetics. The use of reference samples as internal standards enabled inspection for variability between runs. Minced ryegrass collected from a different site was used as an internal standard for in sacco digestions. In most instances the three samples used for in sacco and in vitro incubations runs were from the three different areas (mowing dates 1, 2 and 3), and the 7 59 sequential incubation runs gave 21 data sets to define ryegrass maturation (7 ages for each area). 3.3.5.1 - Digestion kinetics Digestion curves for each data set were used to evaluate in sacco degradation kinetics of ryegrass in the rumen. The disappearance of DM, CP, NDF and ADF were analysed using a non-linear model (model no. 1) described by López et al. (1999) to determine fractional disappearance rate (k, %/hour) and potential degradation (P) according to: P = A + B (1-e-k (t-L)) Where A = soluble fraction (% of each constituent, washed out of bags at t = 0h), B = insoluble degradable fraction (%), t = time in hours, L = lag phase (hours). The effective degradability (E) was calculated from the kinetic parameters obtained from exponential adjustment assuming a fractional passage rate (kp) of 0.06 h-1: E = A + B * (k / (k + kp)) The turnover rate (kp) of 0.06 used here is based on measurements by Van Vuuren et al. (1993) who reported values ranging from 0.041 to 0.067/h for dairy cows fed ryegrass, and is commonly used to evaluate forages (Elizalde et al., 1999; Kolver et al., 1998). The metabolisable protein system (AFRC, 1992) for defining ruminal degradation was used to calculate protein degradability parameters: Quickly degradable protein (QDP, g/100g DM) = A * [CP] where [CP] is CP concentration (g CP/100g DM). Slowly degradable protein (SDP, g/100g DM) = [(B * k) / (k + k p)] * [CP], Effective rumen degradability of crude protein (ERDP, g/100g DM) = [(0.8 * QDP) + SDP], Rumen degradable protein (RDP, g/100 g DM) = QDP + SDP, Rumen undegradable protein (RUP, g/100g DM) = [CP] – RDP. Model parameters were estimated with the non-linear (NLIN) procedure of (SAS, 2001), which is appropriate for situations that do not have long digestion lag times. The model is able to reduce the residual deviations from the model equation for both 60 degradation rate and A and B estimates (Nocek and English, 1986). Separate curves were calculated for each stage of maturation for each area. 3.3.6 - Statistical analyses 3.3.6.1 - Fixed effects model analysis fo r grass growth and chemical composition The effect of initial mowing dates (areas 1, 2 and 3) were analysed in the context of a fixed effect model using PROC GLM (SAS, 2001). Linear, quadratic and interactions effects were tested using the following model: Y = bo + b1i*mowing date + b2i*X + b 3i*mowing date*X + b 4i*X 2 + b5i*mowing date*X 2 + e [equation 1], where: Mowing data is a categorical variable so bo is in fact the intercept for mowing date 3 (20 September). b1i is then the difference for mowing date 1 and 2. Similarly for interaction terms. X = days of re-growth (age) or herbage mass (HM). The complete outputs from the models using PROC PRINT option (SAS, 2001) are presented in the CD appendix. Statistical analyses for herbage mass and days of re-growth (age) Ryegrass production per hectare (herbage mass; HM) was regressed against age (days of re-growth) using the fixed effects model to check if the growth rates change with maturity between and within areas (mowing dates). It is important to express data in relation to both parameters (age and HM) because relationships between them may differ under contrasting climatic and environmental conditions, i.e., ryegrass cut at different dates may have different growth rates (or herbage mass) in different parts of New Zealand. Statistical analyses for chemical composition Using the same basis as HM and age, the NIRS chemical composition (CP, NSC, Lipid, NDF, ADF, Ash, OMD and ME) was regressed against both age and HM independently. In order to compare slopes and intercepts from the three areas (for each component) against age and against HM concentrations were analysed using the PROC GLM (SAS, 2001) described above. When there were no differences between slopes and intercepts for individual chemical components for the three areas, the fixed model analysis was used to 61 calculate overall slope and intercept for each component against age and against HM across all areas. Results from fitting a fixed model with plant age (10 days ≤ age ≤ 112 days) and herbage mass variables using the GLM procedure (SAS, 2001) are presented as equations (slopes, intercepts, r , error, significance level) and plotted to show relationships. See appendix CD (Chapter 3) for details of SAS procedures. 2 3.3.6.2 - Statistical analyses for i n vitro data In vitro incubations were usually carried out for samples taken from each area on the same harvest date (Table 1.1A - Appendix 1). Incubation runs are a continuous variable and we arbitrarily chose five time points. Ammonia concentrations in each incubation bottle were used to calculate net ammonia production (Appendix 2) and analyses were based on net conversion of plant N to ammonia N. The conversion of plant N to NH3-N was evaluated by comparing the effects of plant age between areas within the same incubation run and interactions of age x incubation time according to the model: Y = bo + b1i*age + b2i*time + b3i*age*time + e [equation 2], where: Y = response variable (e.g.: NH3-N, acetate, propionate), bo = intercept, age = age effect, time = incubation time effect, age*time = interact terms and e = error term. VFA analyses were based on single samples for each of the four periods for each forage age. VFA production was expressed as net production mMol/g grass DM incubated. Data are expressed as net production from 0 - 6, 6 - 12 and 12 - 24 hour, and as molar proportion including rates of acetate: propionate (A: P). Statistical analyses enabled a comparison of net VFA production between age at harvest and time of incubation using a similar model to that for testing ammonia concentration (equation 2, above). Data were analysed using general linear models procedures of SAS (SAS, 2001). Differences among means were detected using F tests. Significance was declared at P < 0.05 unless otherwise noted. See appendix CD for procedures details. Acidity (pH) was used primarily to monitor incubations and the rumen inoculum. When pH ≤ 5.6 data were not included in any analysis as this situation no longer represents in vivo rumen fermentation of animals fed forages. Mean pH values for each forage at each time point have been determined and plotted over incubation time for all 15 forage incubations. 62 3.3.6.3 - Statistical analyses for in sacco data Data from in sacco incubations were expressed as degradation curves using a non-linear least-square procedure (PROC NLIN; SAS, 2001) to provide estimates for A, B and k. Observations were weighted using the term 1/standard error2 to give a higher weighting to data having least variance about the curves, and therefore more precise estimates (Koong et al., 1975; Murphy et al., 1982). Evaluations were made for DM, CP, NDF and ADF for soluble (A) and degradable (B) pool, and rate of degradation (k) according to López et al. (1999). These analyses were made against area, age at harvest date and lignin concentration using a general linear model (Appendix CD). When treatment effects were significant for any parameter, those data were fitted using a fixed effects model analysis [equation 1] to determine slopes and inte rcepts and to determine if mowing date effect was statistically significant. For example k for DM was regressed against age at harvest for each area. Appendix CD (Chapter 3) shows details for statistical analysis SAS procedures and complete data set outputs. 63 64 3.4 - Results 3.4.1 – Growth of ryegrass sward Ryegrass grew to 70 - 90 cm in height over the experiment (Figure 3.9; Table 3.2). Herbage mass (above the 5 cm cutting height) increased from about 120 to over 4500 kg DM/ha (Table 3.2) and both height and mass were greatest for the first area mowed (Figure 3.9 and 3.10). _ _ Quadratic (overall) 0 25 50 75 100 0 20 40 60 80 100 120 Days of re-growth (age) Hei ght (cm ) F I G U R E 3.9 – Relationship between ryegrass height (cm) and days of re-growth (age) for area 1 (◊), area 2 (□) and area 3 (∆). Height (cm) = 5.6 + 0.5age + 0.002age2 (r2 = 0.92). The fixed effects model used to evaluate herbage mass (HM; t/DM) and chemical composition (g/100 g DM) of pasture samples from the three areas (Table 3.3) enabled both linear and quadratic relationships to be tested. Analysis of herbage mass with age (days of re-growth) showed that a significant amount of variation was accommodated by the model (P < 0.0001). The relationship was not linear (P = 0.14; Table 3.4) but a quadratic relationship accounted for 81% of variation in herbage mass with days of re-growth: HM = - 0.5 + 0.034age + 0.0002age2 (r2 = 0.81; Figure 3.10). The interaction (age2*mowing date) showed there was no effect of initial mowing date on herbage mass, in relation to days of re-growth (P = 0.25). The ryegrass in each area grew at similar rates irrespective of initial cutting date (Figure 2). Model outputs are calculated for the three areas (Appendix CD) and values for area 1 and 2 are tested in relation to area 3. The parameter estimates provided by the model have been presented for area 3 in Table 3.5 and show: HM = - 0.44 - 0.028age + 0.0011age2 (Table 3.5). 65 _ _ Quadratic (overall) 0.0 1.2 2.4 3.6 4.8 6.0 0 20 40 60 80 100 120 Days of re-growth (age) Her bag e m as s (t DM/ ha) FIGURE 3.10 – Relationship between ryegrass herbage mass (t DM/ha) and days of re- growth (age) for area 1 (◊), area 2 (□) and area 3 (∆). The height of the sward canopy can be used as an indicator for monitoring grazing management. The relationship between herbage mass and the height of the ryegrass sward showed linear relationship with maturity (Figure 3.11). Estimation of herbage mass based on the height of the pasture, with 92% of the variation between height and HM explained by the model. 0 1.2 2.4 3.6 4.8 6 0 20 40 60 80 100 Height (cm) Her bag e m as s (t DM/ ha) Linear (overall) FIGURE 3.11 – Relationship between herbage mass (HM: t DM/ha) and height (cm) for area 1 (◊), area 2 (□) and area 3 (∆). HM = - 0.95 + 0.065height (r2 = 0.92) 66 3.4.2 - Chemical composition Changes in chemical composition with maturity are tabulated (Table 3.3) and presented in Figure 3.12 with regression relationships to define changes over time. Single curves are presented when effects of initial mowing date were not significant; otherwise the curves relate to each data set. Crude protein (CP) concentration (g/100 g DM) declined rapidly with maturity (Table 3.3) from about 23.7 at 22 days to 6.8 after 112 days of re-growth in area one, 20.9 at 10 days to 6.5 after 91 days of re-growth in area two, and 18.5 at 14 days to 8.2 after 82 days of re-growth in area three. These changes were associated with increases in the neutral detergent fibre (NDF) fraction of the DM (42.7 to 64.8; 48.1 to 62.3 and 49.5 to 59.7 g/100 g DM for area one, two and three respectively) as well as the acid detergent fibre (ADF) fraction of the DM (Table 3.3). The concentration of acid detergent lignin (ADL) in the DM increased by a small amount as the ryegrass matured from about 2.74 up to 24 days to about 5.25 g/100 g DM at the final harvest (Table 3.3). Non-structural carbohydrate concentration increased with maturity until about 60 days of re-growth after which it declined. Both ash and lipid concentrations decreased with maturity. Predictions of organic matter digestibility (OMD; g/100 g DM) suggested decreases from about 84 at early stages of re-growth to about 62 in the last harvest date across the areas. The nutritive value of mature ryegrass was also interpreted in relation to the predicted metabolisable energy (ME content) of the DM. As ryegrass matured, the ME (MJ ME/kg grass DM) decreased from 12.3 MJ/kg at 22 days of age to 8.8 MJ/kg at 112 days in area one. Immature grass in area three had consistently lower nutritive value than immature grasses in areas one and two at similar days of re-growth (age). 3.4.2.1 – Chemical composition and days of re-growth (age) Model analysis of all nutritive value parameters (CP, NSC, Lipid, NDF, ADF, Ash, OMD and ME) have been fitted against days of re-growth (age) to demonstrate either linear or quadratic relationships and to identify significant differences between data sets (initial mowing dates) in Table 3.4. In all cases except NSC the model accounted for a significant amount of variation in the data (P < 0.0001; Table 3.4) but mowing date (age2*mowing date) only affected fibre concentration and predicted OMD (Figure 3.12). Significant fits to the data were made with a linear effects model for lipids and ash, and quadratic effects for CP, NSC, NDF, ADF, OMD and ME (Table 3.4; Figure 3.12; Appendix CD). Residual (predicted – actual values) plots from fitted data have been 67 presented in Figure 3.13 to indicate patterns associated with chemical composition and indicators of NV. These data relate to the choice of linear or quadratic fit to the data indicated in Table 3.4. Interpretation of data has been made on the basis of best (linear or quadratic) fit for chemical components (all g/100 g in the DM) in relation to days of re-growth (age) as follows: The decline in CP concentration with maturation was curvilinear (Figure 3.12) and the rate was not affected by initial mowing dates (P = 0.25). The overall model best able to predict CP concentration in ryegrass swards in relation to days of re-growth will be: CP = 25 - 0.29age + 0.001age2 (r2 = 0.82) Model results for non-structural carbohydrate (NSC) concentrations demonstrated curvilinear relationship with age but there were no significant difference between initial mowing dates. When non-significant terms were removed from the model only 28% of the total variance in NSC concentrations and days of re-growth (age) can be explained by the model (P = 0.012): NSC = 5.5 + 0.014age - 0.001age2 (r2 = 0.28) Changes in fibre concentrations due to maturity showed similar patterns for NDF and ADF. Mowing date 2 and 3 suggests a slightly more rapid increase in fibre concentration during maturation compared with the first mowing date (Table 3.4; Figure 3.12). The effect of initial mowing date on fibre concentration has little practical significance because the lower values for the first mowing date are due to the very early cutting in relation to flowering. The overall model to predict fibre from ryegrass pastures using days of re-growth will be: NDF = 46.6 - 0.016age + 0.0016age2 (r2 = 0.80) ADF = 25.5 + 0.013age + 0.0011age2 (r2 = 0.79) Both lipid and ash concentration in the DM did not demonstrate significant quadratic relationships with age (Table 3.4; Figure 3.12 and 3.13), and when the non- significant terms where removed from the models, the overall linear predictions for lipid and ash were: Lipid = 4.3 - 0.02age (r2 = 0.73) Ash = 13.5 - 0.07age (r2 = 0.76) 68 Concentration of metabolisable energy (MJ ME/kg DM) decreased when ryegrass matured but the difference between initial mowing dates were non-significant (P = 0.16). Data were explained using the quadratic component of the model (P < 0.0001) and ME concentration in ryegrass swards was predicted by: ME = 11.6 + 0.009age - 0.0003age2 (r2 = 0.73) Interpretation of chemical composition in relation to NV requires that analyses account for the majority of the DM. Mertens (1992b) reported that the concentrations of non-fibrous carbohydrates [NFC; calculated by difference: 100 – CP – lipid – NDF – ash] and non-structural carbohydrate [NSC ; as measured by enzymatic methods] are not equal for many feeds and the terms should not be used interchangeably (NRC, 2001). Pectin should be included in NFC but not in NSC. When the principal components (CP + NSC + lipid + NDF + ash) predictions by the model are added, the values total about 84 g/100 g DM. Data are given in Table 3.6 to illustrate the extent to which NIRS analyses account for DM in ryegrass at 31 and 60 days of re-growth compared with model predictions. As expected, model predictions were similar to measured values but constituents only accounted for about 78 - 85% of DM. Some of the discrepancy will be due to pectin and organic acids which were not used in calibration of the NIRS (i.e. are not part of the NSC calibration) and represent a weakness in the data set for NSC. Typically pectin accounts for about 1.5 ± 0.2 g/100 g of DM in ryegrass and organic acids 8 ± 1.5 g/100 g of DM (Ulyatt, 1984). When these components are added to model predicted values, total constituents are 94.2 g/100 g of DM for ryegrass at 31 days of re-growth but only 87.3 g/100 g of DM at 60 days of re- growth (Table 3.6). 3.4.2.2 – Chemical composition versus herbage mass (HM) Concentrations of CP, NSC, lipid, NDF, ADF, etc. have been evaluated in relation to HM (Table 3.7 and 3.8) and have been plotted in Figure 3.14. Overall, results were similar to relationships with age but all the components had a curvilinear response with HM as the quadratic term (HM2) and/or interaction with initial mowing date (HM2*mowing date) showing significant effects (Table 3.7). The equations enabling predictions of chemical composition from herbage mass are given in Table 3.8. This chapter has focused on relationships between herbage quality and age, but data in Table 3.8 demonstrate a good relationship with herbage mass as well. This has important implications for using the model, as either days of re-growth (age) or herbage mass could predict ryegrass quality. 69 Table 3.9 compares the prediction models for NV using HM and NIRS analyses. This comparison leads to a conclusion that HM would be a better predictor of NV than age (Table 3.6). Also the goodness of fit for all constituents and OMD were slighter higher when regressed against HM (Table 3.8) than age (Table 3.5), but the differences were small. TABLE 3.2 - Sward height (cm) and estimates values of total herbage dry matter production per hectare (herbage mass; t DM/ha). Area 1 2 3 1 2 3 Initial mowing dates 21/08 11/09 21/09 21/08 11/09 21/09 Sample dates Height (cm) Herbage mass (t DM/ha) 11 September 18 0.21 21 September 19 11 0.32 0.12 5 October 27 23 15 0.91 0.37 0.12 13 October 32 26 21 1.11 0.64 0.31 24 October 40 32 27 1.11 1.53 0.84 3 November 48 42 35 2.20 1.30 1.56 10 November 57 45 37 2.44 1.79 1.49 17 November 57 58 47 2.47 2.58 1.98 27 November 79 61 68 3.31 3.84 2.77 4 December 89 77 70 4.48 3.77 4.65 11 December 94 80 70 5.41 4.57 5.01 70 TABLE 3.3 - Chemical composition (g/100 g dry matter (DM)), and predicted nutritive value of ryegrass cut at different days of re-growth (age) and initial mowing dates (21/08 for area 1, 11/09 area 2 and 21/09 area 3). Area Date Age CP NSC Lipid NDF ADF ADL Ash OMD ME 1 11.9 22 23.7 10.9 4.2 42.7 21.8 2.66 11.6 87.7 12.3 1 21.9 31 19.1 9.0 4.0 45.9 24.8 2.38 10.9 83.8 11.9 1 5.10 45 16.8 9.4 3.8 48.7 26.4 2.50 10.8 81.6 11.6 1 13.10 53 18.0 6.1 4.0 51.5 29.2 2.50 12.3 80.4 11.3 1 24.10 64 12.5 11.4 3.1 48.0 27.4 3.91 9.3 80.5 11.7 1 3.11 74 12.6 9.9 3.2 51.8 30.2 2.62 9.2 76.0 11.0 1 10.11 81 9.5 10.3 2.8 53.7 32.1 5.04 8.5 72.6 10.6 1 17.11 88 8.9 9.8 2.5 55.2 33.8 3.02 8.3 71.1 10.4 1 27.11 98 9.4 9.7 2.7 55.7 34.2 5.04 7.4 68.5 10.1 1 4.12 105 8.0 8.6 1.9 60.7 38.8 4.35 6.7 61.2 9.1 1 11.12 112 6.8 7.3 2.4 64.8 39.4 5.31 7.1 59.9 8.8 2 21.9 10 20.9 7.5 4.1 48.1 25.8 2.68 11.7 82.7 11.7 2 5.10 24 18.6 3.3 3.9 44.8 30.1 2.94 15.3 76.8 10.4 2 13.10 32 17.4 8.1 3.6 49.3 27.9 2.66 11.7 82.6 11.7 2 24.10 43 16.7 10.4 3.5 47.1 26.3 2.70 10.5 83.2 11.8 2 3.11 53 13.0 9.7 3.3 49.8 28.9 2.62 10.4 79.8 11.4 2 10.11 60 10.1 11.3 2.8 50.8 29.7 4.62 9.0 77.0 11.2 2 17.11 67 9.7 10.7 2.8 51.1 30.9 3.07 9.3 76.0 11.0 2 27.11 77 7.9 9.5 2.5 56.5 34.4 5.12 7.3 68.0 10.0 2 4.12 84 7.3 8.9 2.1 60.2 37.4 3.74 6.6 63.2 9.4 2 11.12 91 6.5 8.3 2.3 62.3 37.9 5.40 6.5 61.5 9.1 3 5.10 14 18.5 7.6 3.7 49.5 28.5 2.95 12.7 82.2 11.5 3 13.10 22 19.5 8.4 3.9 48.9 26.1 2.49 11.4 82.4 11.7 3 24.10 33 13.5 8.9 3.0 47.4 28.4 2.44 11.8 79.3 11.1 3 3.11 43 13.1 10.9 3.0 48.8 28.2 2.49 9.6 79.4 11.5 3 10.11 50 10.2 11.4 2.7 51.8 29.6 2.65 9.1 77.8 11.3 3 17.11 57 10.5 11.1 2.9 50.4 30.0 2.81 9.2 76.7 11.1 3 27.11 67 8.9 10.8 2.8 53.7 31.7 4.94 8.0 71.4 10.5 3 4.12 74 8.9 9.1 2.4 56.7 35.6 4.56 6.9 65.2 9.6 3 11.12 81 8.2 8.5 2.3 59.7 37.2 5.04 7.6 63.6 9.4 Mean a 59 12.8 9.2 3.1 52.2 30.8 3.5 9.6 75.1 10.8 Abbreviations: CP, crude protein; NSC, non-structural carbohydrates; NDF, neutral detergent fibre; ADF, acid detergent fibre; ADL, acid detergent lignin; OMD, organic matter digestibility; ME, metabolisable energy (MJ ME/kg DM). a Mean across the three areas. 71 TABLE 3.4 – Significance levels of fixed model analysis for height, herbage mass and chemical components of nutritive value against days of re-growth (age). Model Pr > F Difference between intercepts Linear term age Interaction age*mowing date Quadratic term age2 Interaction age2*mowing date HM <.0001 0.66 0.14 0.46 <.0001 0.25 Height <.0001 0.49 0.10 0.16 0.0003 0.46 CP <.0001 0.08 <.0001 0.32 0.02 0.24 NSC 0.09 0.24 0.01 0.21 0.01 0.19 Lipid <.0001 0.75 0.01 0.63 0.87 0.54 NDF <.0001 0.09 0.01 0.12 <.0001 0.04 ADF <.0001 0.02 0.12 0.10 <.0001 0.06 Ash <.0001 0.64 0.20 0.52 0.44 0.56 OMD <.0001 0.12 0.01 0.17 <.0001 0.04 ME <.0001 0.28 0.01 0.36 <.0001 0.16 Abbreviations see Table 2 (and text). TABLE 3.5 – Parameter estimates of regression for height, herbage mass and nutritive value against days of re-growth (age) defined by the model for area (mowing date) 3. intercept coefficient age coefficient age2 Error r2 HM - 0.44±0.54 ns - 0.028±0.025 ns 0.00105±0.0003 ** 0.33 0.97 Height 7.60±6.05 ns 0.44±0.28 ns 0.005±0.003 ns 3.68 0.98 CP 24.32±1.93 ** - 0.36±0.09 ** 0.002±0.001 * 1.18 0.96 NSC 3.76±2.55 ns 0.28±0.12 * - 0.003±0.001 * 1.55 0.43 Lipid 4.27±0.36 ** - 0.04±0.02 * 0.22 0.92 NDF 52.45±2.63 ** - 0.27±0.12 * 0.004±0.001 ** 1.60 0.94 ADF 29.85±2.21 ** - 0.19±0.10 ns 0.003±0.001 ** 1.34 0.94 Ash 14.15±1.53 ** - 0.11±0.07 ns 0.93 0.86 OMD 79.78±3.22 ** 0.22±0.15 ns - 0.005±0.002 ** 1.96 0.96 ME 11.61±0.41 ** 0.009±0.015 ns -0.0003±0.0001 * 0.52 0.73 Estimates ± standard error (SE). ns = not significant; ** P < 0.01; * P < 0.05; t-test for intercept and regression coefficient differ from zero. Error = Error of the model (error of prediction). Root mean square error. Abbreviations see Table 2 (and text). The use of single or double asterisk in this and other tables denotes significance of estimates parameters from zero at 5 and 1% levels respectively. This terminology applies to all equations in this chapter. ns = non significant. 72 TABLE 3.6 – Example of model predictions of nutritive value against age compared with NIRS analyses at 31 and 60 days. Model predictions NIRS analyses Days of re-growth 31 60 31 60 CP 17.2 11.7 19.1 10.1 NSC 4.9 2.4 9.0 11.3 Lipid 3.7 3.1 4.0 2.8 NDF 47.6 51.3 45.9 50.8 Ash 11.4 9.3 10.9 9.0 Sub-total 84.7 77.8 88.9 84.0 Pectin and organic acids a 9.5 9.5 9.5 9.5 Total 94.2 87.3 98.4 93.5 Abbreviations see text. a Ulyatt (1984). TABLE 3.7 – Significance levels of fixed model analysis for chemical components of nutritive value against herbage mass (HM). Model Pr > F Difference between intercepts Linear term HM Interaction HM*mowing date Quadratic term HM2 Interaction HM2*mowing date CP <.0001 0.27 <.0001 0.88 <.0001 0.89 NSC <.01 0.03 0.0002 0.04 0.0002 0.07 Lipid <.0001 0.10 <.0001 0.77 0.003 0.99 NDF <.0001 0.05 0.04 0.05 0.05 0.09 ADF <.0001 0.003 0.005 0.01 0.28 0.02 Ash <.0001 0.35 <.0001 0.72 0.02 0.76 OMD <.0001 0.01 0.003 0.01 0.06 0.01 ME <.0001 0.02 0.31 0.01 0.01 0.02 Abbreviations see Table 2 (and text). 73 TABLE 3.8 – Parameter estimates of overall regression of chemical composition (g/100g in the DM) against herbage mass (HM; t/hectare) across all initial mowing dates (areas). intercept coefficient HM coefficient HM2 Error r2 CP 21.21±0.76 ** -6.48±0.75 ** 0.76±0.14 ** 1.75 0.88 NSC 7.37±0.64 ** 2.36±0.63 ** -0.45±0.12 ** 1.47 0.35 Lipid 4.10±0.12 ** -0.72±0.12 ** 0.07±0.02 ** 0.27 0.84 NDF 46.26±0.87 ** 2.45±0.86 * 0.17±0.16 ns 2.01 0.87 ADF 25.46±0.73 ** 2.37 ±0.73 ** 0.04±0.14 ns 1.70 0.87 Ash 12.90±0.40 ** -2.33±0.39 ** 0.23±0.07 ** 0.92 0.83 OMD 84.30±1.04 ** -3.89±1.03 ** -0.15±0.19 ns 2.42 0.91 ME 11.76±0.17 ** -0.27±0.17 ns -0.055±0.03 ns 0.39 0.85 Abbreviations see Table 2 and 4 (and text). TABLE 3.9 – Example of model predictions of nutritive value against herbage mass compared with NIRS analyses at 31 and 60 days (0.32 and 1.79 t DM/ha, respectively). Model predictions NIRS analyses Herbage mass (t DM/ha) 0.32 1.79 31 days 60 days CP 19.2 12.1 19.1 10.1 NSC 8.1 10.2 9.0 11.3 Lipid 3.9 3.0 4.0 2.8 NDF 47.1 50.7 45.9 50.8 Ash 12.2 9.5 10.9 9.0 Sub-total 90.4 85.4 88.9 84.0 Pectin and organic acidsa 9.5 9.5 9.5 9.5 Total 99.9 94.9 98.4 93.5 Abbreviations see text. a Ulyatt (1984). 74 0 5 10 15 20 25 0 20 40 60 80 100 120 Crud e pro tei n, g/ 100 g DM Poly. (Overall) 0 4 8 12 0 20 40 60 80 100 120 Non -str uctu ra l ca rbo hyd ra tes , g /10 0g DM Poly. (Overall) 0 1 2 3 4 5 0 20 40 60 80 100 120 Lipi d, g/ 10 0g DM Linear 40 50 60 70 0 20 40 60 80 100 120 Neu tra l d ete rg en t fi bre , g /10 0g DM Poly. (Mowing date 1) Poly. (Mowing date 2&3) 20 30 40 0 20 40 60 80 100 120 A ci d de ter ge nt fibr e, g/ 10 0g DM Poly. (Mowing date 1) Poly. (Mowing date 2&3) 4 8 12 16 0 20 40 60 80 100 120 A sh, g/ 100 g DM Linear (Overall) 60 75 90 0 20 40 60 80 100 120 Days of re-growth (age) Org an ic m atte r di ge sti bili ty, g/ 10 0g DM Poly. (Mowing date 1) Poly. (Mowing date 2) Poly. (Mowing date 3) 9 11 13 0 20 40 60 80 100 120 Days of re-growth (age) Met abo lis abl e en er gy, MJ ME/ kg DM Poly. (Overall) FIGURE 3.12 – Plots of chemical composition against days of re-growth (age) for area 1 (◊), area 2 (□) and area 3 (∆). Regression equation for area 3 is showed in Table 3. Poly. means quadratic. 75 -3 -2 -1 0 1 2 3 4 9 14 19 Predicted crude protein (CP) Res id ual s of CP -6 -3 0 3 6 8.5 9 9.5 10 Predicted non-structural carbohydrates (NSC) Res id ual s of NS C -1.0 -0.5 0.0 0.5 1.0 2 3 4 Predicted lipid Res id ual s of lipi d -6 -3 0 3 6 42 50 58 Predicted neutral detergent fibre (NDF) Res id ual s of NDF -6 -3 0 3 6 22 30 38 Predicted acid detergent fibre (ADF) Res id ual s of A DF -2 -1 0 1 2 6 8 10 12 14 Predicted ash Res id ual s of as h -6 -3 0 3 6 60 67 74 81 88 Predicted organic matter digestibility (OMD) Res id ual s of OMD -2.0 -1.2 -0.4 0.4 1.2 2.0 9 10 11 12 Predicted metabolisable energy (ME) Res id ual o f ME FIGURE 3.13 – Residual plots from the fixed model analysis to demonstrate tendency (linear or quadratic) in change in nutritive value with increase of days of re-growth (age) for area 1(◊), area 2(□) and area 3(∆). 76 0 5 10 15 20 25 0.0 1.2 2.4 3.6 4.8 6.0 Crud e pro tei n, g/ 100 g DM 0 5 10 15 0.0 1 .2 2 .4 3 .6 4 .8 6.0 Non -str uctur al c ar bohyd ra te, g/ 100 g DM 0 1 2 3 4 5 0 .0 1 .2 2 .4 3 .6 4 .8 6 .0 Lipid , g /10 0g DM 40 54 68 0.0 1.2 2.4 3.6 4.8 6.0 Neutr al d ete rg en t fibr e, g/ 100 g DM 20 30 40 0.0 1 .2 2.4 3.6 4.8 6.0 A ci d de ter ge nt fibr e, g/ 100 g DM 0 5 10 15 0.0 1 .2 2 .4 3 .6 4.8 6.0 A sh, g /10 0g DM 50 65 80 0.0 1.2 2.4 3.6 4.8 6.0 Herbage mass (t/ha) Org an ic m atte r di ge sti bili ty, g/ 100 g DM 8 9 11 12 0.0 1.2 2.4 3.6 4.8 6.0 Herbage mass (t/ha) Meta bol is abl e en er gy, MJ ME / k g DM FIGURE 3.14 – Plots of chemical composition against herbage mass (HM) for area 1 (◊), area 2 (□) and area 3 (∆). Estimates overall regression parameters across areas are presented in Table 6. Dot lines represent 95% confidence limits for individual predicted value. 77 3.4.3 - In sacco results The in sacco incubation measured rates of disappearance of DM, CP and fibre during digestion. The ryegrass had been minced to replicate effects of chewing, and the extent to which cells were ruptured and broken into small particles will influence the accessibility of plant components for bacteria. The particle size distribution of minced DM (Table 3.10) shows that about 31% was retained on sieves with 2 mm or greater aperture size and 35 - 38% was able to pass through sieves with 0.25 mm apertures. Particle size distribution was similar for the three areas (initial mowing date) and no effects were evident for changing maturity (age). The absence of changes with maturity for soluble DM (A fraction) indicated that the mincer performed in a consistent manner, although more mature grasses required a much greater mechanical effort to mince. The digestion of minced material reduced the quantity of residues in the in sacco bags, indicated by the rates of DM loss and differential rates of disappearance for soluble constituents, fibre and CP. Digestion will alter the composition of remaining (residual) DM and may affect microbial growth and products of digestion after 12 or 24 hours of incubation, relative to initial degradation (0 to 6 hours). The changing composition of residual DM is summarised in Table 3.11 and Appendix 6 (Table 6.2A). Most apparent was the loss of CP with washing minced forage (initial – 0h) and the decline in CP concentration in residues to 24 hours, after which it increased (at 72 hours). The fibre content increased as a consequence of CP and soluble DM loss during the incubation, reaching 70 - 80% of the DM with ryegrass of all ages. The decrease in CP concentration with digestion was greatest with young ryegrass (20.3 to 6.5% of DM at 24 hours) but the pattern was similar for all levels of maturity (Table 3.11). Fibre, indicated by NDF, showed a substantial (17 - 21 percentage units) increase in concentration with washing (initial – 0 hours) followed by small increases to 24 hours. Overall, the NDF concentration increased from about 49% of DM in the ryegrass, to about 78% at 24 hours, whilst ADF concentration increased by a lesser amount from about 27% to 45% of DM at the end of the incubation (Table 3.11). 3.4.3.1 – DM digestion kinetics The DM digestion kinetics have been summarised in Table 3.12 and Figure 3.15 for ryegrass harvested from the three areas (initial mowing date) and maturity (age). 78 The percentage DM in the soluble “A” fraction of grasses harvested from the three blocks averaged 41% of DM and was not affected by age (P = 0.39) or by initial mowing date (P = 0.63). Thirty four percent of the variation in the value for “A” (DM) was explained by age, mowing date and lignin concentration. The percentage of slowly degradable DM in the “B” fraction of ryegrass harvested from area one, two and three averaged 56, 51 and 52 respectively (Table 3.12) and was not affected by initial harvest date, days of re-growth or lignin concentration (P > 0.30). The model explained only 20% of the variation in the B fraction of ryegrass. Neither initial harvest date nor age had a major effect upon the proportion of DM in either A or B fraction, so the absence of a significant effect explained by the model may be expected. In contrast to the distribution of DM in the A and B pools, maturity had a major (P < 0.001) impact on the fractional disappearance rates (k) which declined from 0.11 h-1 to 0.03 h-1 with mature ryegrass. Rates of DM loss were not affected by either initial harvest date or lignin concentration and the model accounted for 66% of the variation in k. The DM effective degradability (E), which takes into account the effect of passage from the rumen, was calculated assuming an out flow rate of 0.06 h-1,and averaged 65, 67 and 67 for grasses harvested from areas one, two and three respectively. Values declined from about 77% from immature grass to 49% with mature grass harvested from area 3. In sacco dry matter disappearance data is illustrated in Figure 6.1A (Appendix 6). The impact of forage maturation on rates of disappearance is clearly evident with separate lines defining disappearance of each “age” over 72 hours. Also it shows similar DM degradation rates for mature ryegrass from each area. Differences in DM disappearance for grass harvested at different ages were most apparent at the 12 and 24 hour incubation times, which would affect rumen clearance. Total digestion (after 72h of incubation) was similar for all materials tested (averaging 92.6 %). 3.4.3.2 – CP digestion kinetics Table 3.13 presents kinetic data for in sacco digestion of ryegrass CP. Soluble CP ranged from 46 to 80% of total CP. In contrast to DM, the percentages of CP in the “A” fraction (average 61% of CP) was higher in older plants (P < 0.001) with 73 - 80% released by mincing on the most mature ryegrass (Table 3.13 and Figure 3.16). Initial mowing date did not affect the proportion of solubilised protein (P > 0.09) and the 79 model explained 78% of the variation in soluble CP. Parameters used to predict quantities of degradable protein (QDP, SDP, ERDP, and RDP) were higher in immature ryegrass because protein concentrations were highest in young material. Rumen degradable protein averaged 12.1 g/100 g DM (83% of total CP) across all ages with a range from 19.4 to 5.6 g/100 g DM and the percentage of protein in the RDP fraction was not affected by days of re-growth. The slowly degradable “B” fraction of CP had much smaller values than the soluble and rapidly degradable A fraction, averaging about 46% of total CP in immature forage and declining to 25% in mature ryegrass. Initial mowing date did affect the proportion of CP in this fraction (P < 0.01) and the model explained 88% of the variation in the proportion of CP in the “B” pool. Rates of CP degradation (k = 0.12 h-1) were about twice that for DM but rates were not affected by maturity (Table 3.13). Maturation, initial harvest date or lignin concentration did not affect degradation rates for CP (P > 0.16) and only 24% of variance was explained by the model. The effective degradability for CP averaged 83% across the three areas and the rapid and extensive degradation is illustrated in Figure 6.2 (Appendix 6). 3.4.3.3 – Fibre digestion kinetics The concentration of fibre increased with plant maturation from less than 50% to about 70% of the DM, and after mincing the distribution between soluble and insoluble but potential degradable pools was similar for all levels of maturity (age). The distribution of NDF and ADF between soluble and potential degradable pools, and degradations rates (k) are presented in Tables 3.14 and 3.15, and Figures 3.17 and 3.18. Both NDF and ADF were present mainly in the insoluble and potential degradable (“B”) pool. The distribution of NDF was similar for initial cutting date. About 21% of fibre was soluble and 10% undegradable. In contrast to the similar distribution of fibre fractions between pools for all samples, the rate of ADF degradation (k) declined as ryegrass matured (P < 0.001; Table 3.15). The degradation curves for both NDF (Figure 6.3A) and ADF (Figure 6.4A – Appendix 6) show clear effects of age on degradation rates and lag time. Principal effects were reduction in k values from about 0.11 h-1 with immature forage to about 0.04 h-1 for mature ryegrass. The model accounted only for 21% of the variation in degradation rates for NDF and 65% for ADF, but lignin concentration did not affect rates of fibre degradation or distribution of ADF between pools. Interpretation of rates 80 were complicated by lag times, which varied in accordance with the model fit to the data but averaged 4.2 and 4.9 hours for NDF and ADF, respectively. When equations incorporated a substantial lag period (e.g.: Figures 3.17 and 3.18) rates of degradation post lag were quite rapid, but when the fit did not demonstrate a substantial lag, rates of digestion were slower. Nevertheless, the fibre digestion was slow (k = 0.08 h-1) relative to CP (0.12 h-1) and shows that the microbial population were either slow to colonise or slow to degrade the fibre fractions of ryegrass, especially when mature (Figure 6.3A and 6.4A). The lags tended to be longer for ADF fraction (which has a higher proportion of lignin than NDF) but lignin did not affect the distribution of fibre between A and B pools (P > 0.77; Table 3.15). The effective degradability incorporates effects of both A and B pools and degradation rate, and calculations for NDF and ADF ranged from about 59% for ryegrass up to 35 days of re-growth to less than 40% for mature forage. These calculations assume a passage rate of 6%.h-1, and would probably be lower with very mature forage. 81 TABLE 3.10 – Particle size distribution of ryegrass dry matter for in sacco and in vitro incubations indicated by sieve aperture size either retaining or enabling material to pass. Area Date Age > 2mm 0.25 – 1mm < 0.25 mm 1 11/09/00 22 36.7 26.7 36.6 1 21/09/00 31 35.9 28.8 35.4 1 5/10/00 45 27.9 30.8 41.3 1 13/10/00 53 36.8 25.6 37.6 1 3/11/00 74 34.5 27.9 37.6 1 3/11/00 88 22.4 50.4 27.3 1 4/12/00 105 19.0 49.9 31.1 1 Average 30.4 34.3 35.3 1 STDEV 7.4 10.9 4.7 2 21/09/00 10 27.8 34.4 37.8 2 5/10/00 24 28.6 30.1 41.2 2 13/10/00 32 39.1 25.1 35.8 2 24/10/00 43 28.4 31.7 39.9 2 3/11/00 53 36.7 23.2 40.2 2 17/11/00 67 34.9 37.2 27.9 2 4/12/00 84 22.6 37.1 40.2 2 Average 31.1 31.3 37.6 2 STDEV 5.9 5.6 4.7 3 5/10/00 14 38.6 25.0 36.3 3 13/10/00 22 42.9 24.7 32.4 3 24/10/00 33 33.3 24.1 42.6 3 3/11/00 43 31.2 31.7 37.1 3 10/11/00 50 25.5 32.4 42.1 3 17/11/00 57 35.1 30.9 33.9 3 4/12/00 74 20.1 45.7 34.2 3 Average 32.4 30.7 36.9 3 STDEV 7.7 7.5 4.0 Average all areas 31.3 32.1 36.6 STDEV all areas 6.7 8.1 4.3 STDEV: standard deviation. 82 TABLE 3.11 – Composition of ryegrass (0 hour) and in sacco residues over the 72 hours digestion period, averaged for young, medium and mature forage. Data derived from appendix 6 (Table 6.2A) and averaged for area 1, 2 and 3. Units all in g/100 g in the DM. Duration of incubation (hours) Age Initial 0 2 6 12 24 72 CP 22-24 days 20.3 17.0 14.3 11.2 10.2 6.5 9.6 43-45 days 15.5 11.1 9.8 8.4 5.7 5.8 10.8 67-74days 10.4 5.0 5.2 5.5 4.5 4.2 7.3 NDF 22-24 days 45.7 62.3 65.4 66.9 73.2 78.0 73.0 43-45 days 48.2 69.3 70.7 66.3 74.1 76.2 69.2 67-74days 53.2 70.4 72.0 74.6 77.4 79.8 77.9 ADF 22-24 days 26.0 35.8 38.3 42.8 43.7 46.1 46.4 43-45 days 27.0 39.1 39.9 39.0 44.0 44.9 42.4 67-74days 32.2 42.3 43.3 46.2 48.4 49.5 46.6 Abbreviations see text. 83 TABLE 3.12 - Ryegrass dry matter (DM) degradation characteristics (% of total DM) as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h-1), lag time (Lag, hour), and effective degradability (E) which takes in account the effect of passage from the rumen1. Area Age A B C k Lag E 1 21 42 57 2 0.07 0.9 71 1 31 38 60 2 0.06 0.0 67 1 45 51 45 4 0.10 3.1 77 1 53 36 57 6 0.08 0.9 69 1 74 43 50 7 0.05 0.9 65 1 88 28 67 5 0.04 0.0 57 1 105 34 56 11 0.03 1.4 52 2 10 46 49 5 0.11 9.8 67 2 24 44 52 4 0.11 1.1 77 2 32 42 53 6 0.08 4.6 67 2 43 51 45 4 0.10 3.1 77 2 53 42 48 10 0.07 1.7 66 2 67 40 55 6 0.04 3.0 60 2 84 36 54 10 0.03 0.0 55 3 14 46 51 3 0.10 1.4 77 3 22 33 62 5 0.08 1.4 66 3 33 50 46 4 0.11 2.7 78 3 43 42 51 7 0.06 1.7 65 3 50 48 48 4 0.07 1.4 72 3 57 45 50 5 0.05 4.2 65 3 74 31 56 12 0.03 3.1 49 Average 49 41 53 6 0.07 2.2 67 Model P 0.1284 0.4298 0.0012 Age P 0.3927 0.8947 0.0017 Area P 0.6259 0.3005 0.6457 Lignin P 0.2698 0.5745 0.9033 r2 0.34 0.20 0.66 1 Passage rate set at 0.06 h-1. P: P values assessing goodness of fit for the overall model and tests (age, area and lignin concentration). Area relates to initial mowing date. 84 TABLE 3.13 - Ryegrass crude protein (CP) degradation characteristics (% of total CP) as defined by soluble (A), degradable insoluble (B), undegradable residual (C = 100 – A – B) pools as well as fractional disappearance rate (k, h-1), lag time (L, hour), effective degradability (E) which takes in account the effect of passage from the rumen. Crude protein is also expressed in terms of quickly degradable protein (QDP), slowly degradable protein (SDPa), effective rumen degradability of CP (ERDP), rumen degradable protein (RDP) and rumen undegraded protein (RUP), all in g/100 g DM. Area Age A B C k L Ea QDP SDP ERDP RDP RUP 1 22 53 47 0 0.10 0.5 81 12.5 6.9 16.9 19.4 4.4 1 31 51 48 1 0.08 0.0 78 9.8 5.2 13.0 14.9 4.2 1 45 55 43 2 0.18 0.5 87 9.2 5.5 12.8 14.6 2.2 1 53 55 41 4 0.13 0.0 83 9.9 5.0 12.9 14.9 3.1 1 74 61 33 7 0.16 3.3 83 7.6 3.0 9.0 10.6 2.0 1 88 69 30 1 0.03 0.0 78 6.1 0.8 5.7 7.0 1.9 1 105 78 12 10 0.06 12.0 80 6.2 0.5 5.4 6.6 1.3 2 10 56 44 0 0.06 0.0 79 11.8 4.7 14.1 16.5 4.4 2 24 53 45 2 0.19 0.1 87 9.8 6.4 14.2 16.2 2.4 2 32 50 46 4 0.12 2.3 79 8.8 5.3 12.3 14.1 3.3 2 43 70 28 2 0.13 0.1 89 11.7 3.2 12.6 14.9 1.8 2 53 57 35 8 0.34 4.5 85 7.4 3.8 9.8 11.2 1.8 2 67 73 22 5 0.06 4.9 82 7.2 1.1 6.8 8.2 1.5 2b 84 77 22 1 0.01 6.0 79 5.6 0.0 4.5 5.6 1.7 3 14 50 49 2 0.20 1.0 86 9.2 6.9 14.2 16.1 2.4 3 22 46 49 4 0.13 0.0 81 9.0 6.7 13.9 15.7 3.8 3 33 54 44 2 0.16 0.0 86 7.3 4.3 10.1 11.6 2.0 3 43 57 35 8 0.20 5.0 82 7.5 3.5 9.5 11.0 2.0 3 50 64 31 5 0.13 0.0 85 6.6 2.2 7.4 8.7 1.5 3 57 75 22 3 0.05 0.5 84 7.9 1.0 7.3 8.9 1.6 3 74 80 15 5 0.02 12.0 81 7.0 0.3 5.9 7.3 1.5 Mean 49 61 35 4 0.12 2.5 83 8.5 3.6 10.4 12.1 2.4 Model P <.0001 <.0001 0.3456 Age P 0.0006 <.0001 0.9780 Area P 0.0957 0.0079 0.6697 Lignin P 0.1036 0.0699 0.1614 r2 0.78 0.88 0.24 a Passage rate set at 0.06 h-1. b PROC NLIN failed to converge. P: P values assessing goodness of fit for the overall model and tests (age, area and lignin concentration). Area relates to initial mowing date. 85 TABLE 3.14 - Ryegrass neutral detergent fibre degradation characteristics (% of total NDF) as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h-1), lag time (Lag, hour), and effective degradability (E) which takes in account the effect of passage from the rumena. Area Age A B C k Lag E 1 22 29 64 7 0.14 9.4 61 1 31 20 77 2 0.05 0.8 55 1 45 19 75 6 0.10 1.5 65 1 53 23 66 11 0.09 3.6 58 1 74 30 61 10 0.05 2.2 53 1 88 19 70 12 0.07 10.1 41 1 105 b 16 71 13 0.03 2.0 37 2 10 30 63 7 0.09 10.7 53 2 24 17 76 7 0.11 1.9 65 2 32 27 64 9 0.08 7.1 54 2 43 21 73 6 0.09 1.3 63 2 53 24 65 11 0.04 1.3 50 2 67 19 68 13 0.07 10.1 40 2 84 b 20 38 42 0.08 1.5 40 3 14 26 69 5 0.10 1.5 67 3 22 13 78 9 0.08 3.3 53 3 33 22 73 5 0.10 1.2 65 3 43 25 67 9 0.04 2.6 50 3 50 24 69 6 0.07 1.6 60 3 57 24 67 9 0.07 9.0 47 3 74 7 77 16 0.03 3.9 29 Average 49 22 68 10 0.08 4.1 53 Model P 0.0962 0.2514 0.4448 Age P 0.9811 0.1633 0.1431 Area P 0.6126 0.2672 0.3597 Lignin P 0.0377 0.9703 0.3734 r2 0.39 0.29 0.21 a Passage rate set at 0.06 h-1. b PROC NLIN failed to converge. P: P values assessing goodness of fit for the overall model and tests (age, area and lignin concentration). Area relates to initial mowing date. 86 TABLE 3.15 - Ryegrass acid detergent fibre degradation characteristics (% of total ADF) as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h-1), lag time (Lag, hour), and effective degradability (E) which takes in account the effect of passage from the rumena. Area Age A B C k Lag E 1 22 17 75 8 0.15 10.2 54 1 31 20 73 7 0.08 8.8 48 1 45 20 74 6 0.09 1.6 63 1 53 19 69 12 0.10 4.1 58 1 74 29 62 10 0.04 2.7 52 1 88 19 70 11 0.08 10.4 41 1 105 19 74 6 0.02 3.0 37 2 10 21 71 8 0.10 11.6 45 2 24 30 64 6 0.10 1.7 69 2 32 21 69 10 0.08 7.4 51 2 43 19 75 6 0.08 1.1 61 2 53 24 66 11 0.04 1.7 49 2 67 18 69 13 0.07 10.5 39 2 84 22 41 38 0.05 1.3 40 3 14 31 63 6 0.10 2.5 67 3 22 4 86 10 0.08 4.0 48 3 33 28 67 5 0.09 1.1 66 3 43 23 68 9 0.04 3.3 48 3 50 25 69 6 0.07 1.6 59 3 57 22 69 9 0.07 9.2 46 3 74 b 10 77 13 0.03 6.0 28 Average 49 21 69 10 0.08 4.9 51 Model P 0.9863 0.2825 0.0024 Age P 0.9855 0.2814 0.0008 Area P 0.8651 0.2618 0.0567 Lignin P 0.9271 0.7704 0.6133 r2 0.02 0.27 0.65 a Passage rate set at 0.06 h-1. b PROC NLIN failed to converge. P: P values assessing goodness of fit for the overall model and tests (age, area and lignin concentration). Area relates to initial mowing date. 87 A = 47.62 - 0.13age r2 = 0.27 20 30 40 50 60 % o f s ol uble DM, A B = 50.11 + 0.06age r2 = 0.07 40 50 60 70 % o f in so luble DM, B C = 2.27 + 0.07age r2 = 0.46 0 4 8 12 %DM un de gr ad abl e re si dua l, C k = 0.11 - 0.0009age r2 = 0.63 0.00 0.05 0.10 0.15 DM d eg ra da tio n ra te, k (h -1 ) 0 4 8 12 0 20 40 60 80 100 120 Days of re-growth (age) DM la g tim e, hour E = 78.43 - 0.24age r2 = 0.55 50 60 70 80 0 20 40 60 80 100 120 Days of re-growth (age) %DM e ffec tiv e de gr ad abi lity, E FIGURE 3.15 - Dry matter (DM) degradation parameters during in sacco incubations of ryegrass in different ages (days of re-growth) for areas 1 (◊), 2 (□) and 3 (∆). 88 A = 44.97 + 0.33age r2 = 0.65 20 40 60 80 100 % o f s ol uble CP, A B = 54.10 - 0.39age r2 = 0.73 0 20 40 60 % o f in so lubl e CP, B C = 0.93 + 0.06age r2 = 0.25 0 5 10 %CP un de gr ad abl e re si due , C 0.0 0.1 0.2 0.3 0.4 CP d eg ra da tio n ra tes , k (h -1 ) 0 4 8 12 0 20 40 60 80 100 120 Days of re-growth (age) CP l ag tim e, hour 60 80 100 0 20 40 60 80 100 120 Days of re-growth (age) %CP e ffec tiv e de gr ad abi lity, E FIGURE 3.16 – Crude protein (CP) degradation parameters during in sacco incubations of ryegrass in different ages (days of re-growth) for areas 1 (◊), 2 (□) and 3 (∆). 89 0 10 20 30 % o f s ol uble NDF, A 30 50 70 90 % o f in so lubl e NDF, B 0 10 20 % un de gr ad abl e re si due NDF, C k = 0.11 - 0.001age r2 = 0.35 0.00 0.05 0.10 0.15 NDF d eg ra da tio n ra tes , k (h -1 ) 0 4 8 12 0 20 40 60 80 100 120 Days of re-growth (age) NDF l ag tim e, hour E = 67.32 - 0.30age r2 = 0.53 30 50 70 0 20 40 60 80 100 120 Days of re-growth (age) %NDF e ffe cti ve d eg ra da bili ty, E FIGURE 3.17 – Neutral detergent fibre (NDF) degradation parameters during in sacco incubations of ryegrass in different ages (days of re-growth) for areas 1 (◊), 2 (□) and 3 (∆). 90 0 10 20 30 % o f s ol uble A DF, A 30 45 60 75 90 % o f in so lubl e A DF, B 0 10 20 % un de gr ad abl e re si due A DF, C k = 0.11 - 0.0008age r2 = 0.50 0.00 0.05 0.10 0.15 A DF d eg ra da tio n ra te, k (h -1 ) 0 4 8 12 0 20 40 60 80 100 120 Days of re-growth (age) A DF l ag tim e, hour E = 63.30 - 0.25age r2 = 0.36 25 45 65 0 20 40 60 80 100 120 Days of re-growth (age) %A DF e ffec tiv e de gr ad abi lity, E FIGURE 3.18 – Acid detergent fibre (ADF) degradation parameters during in sacco incubations of ryegrass in different ages (days of re-growth) for areas 1 (◊), 2 (□) and 3 (∆). 91 3.4.4 - I n vitro results The buffered incubation media used for in vitro runs enabled the net yield of ammonia to be measured as well as both the production and proportions of VFA. All incubations included about 0.5 g minced ryegrass DM and yield of metabolites have been expressed in terms of plant N and plant DM. Calculations require the contribution of rumen inoculum to ammonia and VFA (Table 3.16) pools be subtracted from yields determined over the incubation period (Appendix 4). Net production of ammonia indicates both the extent to which protein degradation exceeds the capacity for microbial utilisation and the likelihood that plant N is insufficient for microbial growth. The in vitro incubations complement the rates of disappearance of feed components from in sacco bags. The in vitro incubations were buffered to minimise changes in pH and values over the first 12 hours of incubation always remained above 5.6, so the environment was similar to that of cows fed forage diets (normal range about 5.8 – 6.6). The decline in pH over the incubation was due to production of VFA in excess of the capacity for buffering. Most the incubations remained above 5.6 for all time points (Figure 3.19) but data from bottles where pH declined below 5.6 were discarded because the data are probably not representative of forage digestion. Few data were omitted in response to the pH criteria, all from 24 hour samples especially from run D where pH averaged 5.4 after 24 hour. 3.4.4.2 – Net ammonia production in vitro The concentration of ammonia exceeded 0 hour values for about 10 hours of incubation with young forages, harvested up to 32 days of re-growth (runs A and B), but older ryegrass resulted in an ammonia deficit after 3 - 4 hours of incubation (Figure 3.20). The mature ryegrass in runs C, D and E (averaging 57, 71 and 88 days of re-growth) had peak ammonia concentrations after 2 hours of incubation (Table 3.17) and the lowest concentrations after 6 hours of incubation (Figure 3.21). Immature ryegrass, with CP concentrations ranging from 14 to 24 g/100 g of the DM, appeared to have sufficient N to meet microbial requirements for up to 10 hours of incubation and CP concentrations below 10 g/100 g DM in mature ryegrass were inadequate for sustained microbial growth. Increasing ammonia concentration after 12 hours suggests bacterial lysis is exceeding utilisation for growth. Microbial yield of VFA may have been restricted by ammonia insufficiency when pasture over 32 days of age was incubated because microbial growth would be limited. 92 The extent of CP degradation was determined in sacco and these data can be used to provide an estimate of microbial growth in vitro when ammonia concentration in the buffered media declined to the 0 hour value. This is calculated (Table 3.17) by multiplying the proportion of plant protein degraded in sacco (at the time in vitro ammonia concentrations declined to the 0 hour value; Figure 6.2A – Appendix 6; 0.60 – 0.99) by the amount of plant CP incubated in vitro. The net release of plant N, assumed to be utilised for microbial growth, ranged from 0.28 to 0.85 mMol N. The calculation could not be made for highly degradable young (22 days) ryegrass, because the ammonia concentration remained elevated over the entire 24 hour of incubation. Microbial growth was not sufficient to utilise all of the N released from immature ryegrass. In summary, there was a changing pattern of NH3 production as ryegrass matured; young grass with CP concentration over 17 g/100g of DM (Table 3.3) had sufficient N to sustain microbial growth for at least 10 hours. However older material, even with similar CP concentrations (e.g.: 53 days of re-growth from Area 1), released insufficient CP to maintain ammonia concentrations after 2 hours of incubation (Figure 3.20). The relationship between N incorporated into microbial CP and days of re- growth is presented in Figure 3.24. TABLE 3.16 – pH, ammonia, volatile fatty acid concentrations and proportion of acetate to propionate in the rumen liquor of the cow used for in sacco and in vitro incubation runs. Runs: Rumen inoculum A B C D E pH 6.20 6.59 6.60 6.41 6.39 NH3 concentration (mMol/L) 21.60 21.40 15.70 20.20 29.30 Total VFA concentration (mMol) 109.00 90.90 72.17 109.10 99.93 Acetate: propionate ratio 4.01 3.90 4.29 3.34 2.94 93 TABLE 3.17 – Time for peak ammonia concentration and decline to initial (0 hour) values in vitro with an estimate of plant N incorporated to microbial N (in mMol). Time after start (h) Run Area Age peak NH3 Return to 0h NH3 conc. Plant N (mMol ) Proportion of N degraded 1 Plant N to microbial N (mMol) 2 A 1 22 6 NA 1.38 0.99 NA 1 31 6 12 1.08 0.76 0.82 2 10 10 12 1.17 0.73 0.85 B 1 53 2 9.0 0.92 0.82 0.75 2 32 6 10 0.98 0.79 0.77 3 22 6 10 0.97 0.8 0.78 C 1 74 2 3.5 0.65 0.64 0.42 2 53 2 3.5 0.69 0.62 0.43 3 43 2 3.0 0.74 0.6 0.44 D 1 88 2 4.5 0.63 0.72 0.45 2 67 2 3.5 0.47 0.74 0.35 3 57 2 4.5 0.53 0.78 0.41 E 1 105 2 3.0 0.32 0.79 0.25 2 84 2 3.0 0.36 0.77 0.28 3 74 2 3.5 0.36 0.77 0.28 NA – not applicable as in vitro ammonia remained elevated for the 24 hour incubation period. 1 Based on in sacco degradation after incubation times required for in vitro ammonia concentrations to return to initial (0 h) values. 2 Product of plant N and proportion degraded. 94 3.4.4.3 – VFA production in vitro Concentrations of VFA (Appendix 4) have been summarised for young, medium and mature ryegrass (Figure 3.22) and demonstrate higher total concentrations per gram DM incubated from mature versus young ryegrass. This was apparent for acetate and n-butyrate, but concentrations of propionate were similar for all three maturities at the three sampling times and the concentration of minor VFA were highest from young ryegrass. The rates of acetate: propionate average 3.1 for ryegrass having less than 51 days of re-growth and 3.4 for more mature forage. When expressed as rates (Table 3.18; Figure 3.23), VFA production declined from an average of 2.17 mMol total VFA/g DM from 0 - 6 hours to 1.14 between 6 - 12 and 0.84 between 12 - 24 hours. During the first 6 hours of incubation VFA production was higher for mature (over 57 days) forage, at 2.69 mMol/g DM than medium or young ryegrass (2.03 and 1.79 mMol//g DM respectively), although effects of maturity on VFA production had disappeared by 12 hours of incubation. Effects of maturity on rates of production were greatest for acetate and minor VFA, and least for propionate. Ratios of A: P produced were similar throughout the 24 hours incubation for young ryegrass (3.19), highest for medium maturity (3.74) and mature ryegrass showed a decline in A: P produced from 3.5 at 0 - 6 hours to 1.9 at 12 - 24 hours. The yield of VFA over 24 hours was most consistent for young ryegrass (1.79 declining to 1.19 mMol/g DM) and least consistent for mature ryegrass (Table 3.18). When the yield of each VFA was expressed as mg, the proportion of DM degraded and released as VFA was calculated (Table 3.19). These data show on average 14.6, 7.7 and 5.6% of DM converted to VFA after 0 - 6, 6 - 12 and 12 - 24 hours time period respectively and after 24 hours the percentage of young, medium and mature ryegrass DM released as VFA were 26.7, 26.3 and 30.4 respectively. 95 TABLE 3.18 – Rates of volatile fatty acid (mMol/g DM) production per time period. Data have been combined across areas to give values for young (under 33 days), medium (43 - 57 days) and mature (over 67 days) ryegrass. Period Acetate Propionate n-butyrate Minor Total Ratio A:P Young 0-6 h 1.12 0.39 0.22 0.06 1.79 3.01 6-12 h 0.66 0.21 0.11 0.03 1.01 3.10 12-24 h 0.82 0.25 0.07 0.04 1.19 3.47 Medium 0-6 h 1.31 0.46 0.24 0.02 2.03 3.01 6-12 h 0.73 0.23 0.17 0.03 1.16 4.33 12-24 h 0.50 0.14 0.05 0.02 0.72 3.89 Mature 0-6 h 1.81 0.53 0.32 0.04 2.69 3.48 6-12 h 0.79 0.25 0.16 0.04 1.24 3.15 12-24 h 0.43 0.09 0.08 0.02 0.62 4.57 TABLE 3.19 – Amounts (mg) of volatile fatty acids produced per in vitro incubation time period. Approximately 0.5 g dry matter was used for all incubations. Period Acetate Propionate n-butyrate Minor Young 0-6 h 67.14 29.13 19.59 5.87 6-12 h 39.87 15.50 9.49 2.51 12-24 h 49.53 18.56 6.20 4.07 Medium 0-6 h 78.59 33.74 21.48 2.24 6-12 h 43.96 16.68 15.26 3.22 12-24 h 30.31 10.06 4.73 2.25 Mature 0-6 h 108.58 38.95 27.86 4.02 6-12 h 47.71 18.40 13.91 3.44 12-24 h 25.86 6.92 6.66 2.19 Abbreviations see Table 3.18. 96 Run A 5.0 5.5 6.0 6.5 7.0 7.5 0 4 8 12 16 20 24 pH 22d mowing date 1 31d mowing date 1 10d mowing date 2 Standard Run B 5.0 5.5 6.0 6.5 7.0 7.5 0 4 8 12 16 20 24 pH 53d mowing date 1 32d mowing date 2 22d mowing date 3 Standard R un C 5.0 5.5 6.0 6.5 7.0 7.5 0 4 8 12 16 20 24 pH 74d mowing date 1 53d mowing date 2 43d mowing date 3 Standard Run D 5.0 5.5 6.0 6.5 7.0 7.5 0 4 8 12 16 20 24 pH 88d mowing date 1 67d mowing date 2 57d mowing date 3 Standard R un E 5.0 5.5 6.0 6.5 7.0 7.5 0 4 8 12 16 20 24 Incubation time (hour) pH 105d mowing date 1 84d mowing date 2 74d mowing date 3 Standard FIGURE 3.19 – pH during in vitro incubations for ryegrass at different ages (days – d) in five incubations runs (A to E). 97 Run A -25 15 55 95 0 6 12 18 24 µMo l N H 3/ m Mol pla nt N 22d mowing date 1 31d mowing date 1 10d mowing date 2 Run B -60 -30 0 30 60 0 6 12 18 24 µMo l N H 3/ m Mol pla nt N 53d mowing date 1 32d mowing date 2 22d mowing date 3 Run C -70 -35 0 35 70 0 6 12 18 24 µMo l N H 3/ m Mol pla nt N 74d mowing date 1 53d mowing date 2 43d mowing date 3 Run D -120 -60 0 60 120 0 6 12 18 24 µMo l N H 3/ m Mol pla nt N 88d mowing date 1 67d mowing date 2 57d mowing date 3 Run E -400 -250 -100 50 200 0 6 12 18 24 Incubation time (hour) µMo l N H 3/ m Mol pla nt N 105d mowing date 1 84d mowing date 2 74d mowing date 3 FIGURE 3.20 – Net ammonia production expressed in terms of plant N during five incubations runs (A to E). 98 Run A 0 3 6 9 0 6 12 18 24 m Mol NH 3/ L 22d mowing date 1 31d mowing date 1 10d mowing date 2 Run B 0 2 4 6 0 6 12 18 24 m Mol NH 3/ L 53d mowing date 1 32d mowing date 2 22d mowing date 3 Run C 0.0 1.5 3.0 4.5 0 6 12 18 24 m Mol NH 3/ L 74d mowing date 1 53d mowing date 2 43d mowing date 3 R un D 0.0 1.5 3.0 4.5 6.0 0 6 12 18 24 m Mol NH 3/ L 88d mowing date 1 67d mowing date 2 57d mowing date 3 R un E 0 2 4 6 8 10 0 6 12 18 24 Incubation time (hour) m Mol NH 3/ L 105d mowing date 1 84d mowing date 2 74d mowing date 3 FIGURE 3.21 – Pattern of ammonia concentration mMol/L for in vitro incubations. 99 0.0 2.0 4.0 0 6 12 18 24 m Mol to tal VFA /g DM Young Medium Mature 0.0 1.0 2.0 3.0 4.0 0 6 12 18 24 m Mol a ce tate /g DM Young Medium Mature 0.0 0.4 0.8 1.2 0 6 12 18 24 Incubation time (hour) m Mol pr opi on ate /g DM Young Medium Mature 0.0 0.4 0.8 0 6 12 18 24 Incubation time (hour) m Mol n -butyr ate /g DM Young Medium Mature 0.00 0.08 0.16 0 6 12 18 24 Incubation time (hour) m Mol m in or VFA / g DM Young Medium Mature FIGURE 3.22 – Concentration of total VFA, acetate, propionate, n-butyrate and minor VFA (iso-butyrate, n-valerate and iso-valerate) expressed in terms of substrate DM after 6, 12 and 24 hours of incubation in vitro. Data have been combined across areas to give values for young (under 33 days), medium (43 - 57 days) and mature (over 67 days) ryegrass. 100 0 - 6 h 6 - 12 h 12 - 24 h 0.0 2.0 4.0 m Mol to tal VFA /g DM Young Medium Mature 0 - 6 h 6 - 12 h 12 - 24 h 0.0 1.0 2.0 m Mol a ce tate /g DM Young Medium Mature 0 - 6 h 6 - 12 h 12 - 24 h 0.0 0.4 0.8 m Mol pr opi on ate /g DM Young Medium Mature 0 - 6 h 6 - 12 h 12 - 24 h 0.0 0.3 0.5 m Mol n -butyr ate /g DM Young Medium Mature 0 - 6 h 6 - 12 h 12 - 24 h 0.00 0.04 0.08 Incubation time (hour) m Mol m in or VFA /g DM Young Medium Mature 0 - 6 h 6 - 12 h 12 - 24 h 3.47 3.10 4.33 3.01 3.89 3.153.48 1.89 0.0 2.5 5.0 Incubation time (hour) Ra tio n A :P Young Medium Mature FIGURE 3.23 – Production of total VFA, acetate, propionate, n-butyrate, minor VFA (iso- butyrate, n-valerate and iso-valerate) and the ratio of acetate: propionate expressed in terms of substrate DM available for fermentation from 0 - 6, 6 - 12 and 12 - 24 hours of incubation. Data have been combined across areas to give values for young (under 33 days), medium (43 - 57 days) and mature (over 67 days) ryegrass. 101 0.0 0.2 0.4 0.6 0.8 1.0 0 20 40 60 80 100 120 Days of re-growth (age) Esti m ate d m ic ro bia l N (m Mol ) FIGURE 3.24 – The relationship between N incorporated into microbial CP and days of re-growth for areas 1 (◊), 2 (□) and 3 (∆). Microbial N (mMol) = 0.92 – 0.007*age (r2 = 0.73; P < 0.001). Data from in vitro incubations and assume that all N released from forages, where ammonia concentrations declined to 0 hour values, was incorporated into microbial N. 102 3.5 - Discussion The primary purpose of this study was to define rates and products of digestion from ryegrass as it matures. The ryegrass used in this study was allowed to become very mature and normal grazing would be within 50 days of re-growth but a broad range of composition will provide a sound base for diet formulation. Furthermore, mature forage resembles summer pasture in many situations, so the entire data set will enable diet formulation using appropriate feedstuffs to complement ryegrass pastures. Our immediate plans are to apply this information to dairy cow nutrition. Grass maturation can be defined as the proportion of leaf, stem, inflorescence and dead matter on the basis of chemical composition and in terms of digestibility of components. Data presented here complement work of Armstrong (1964; 1982), Elizalde et al. (1999), INRA (1989), Jacobs et al. (1998), Mambrini and Peyraud (1994), Waite et al. (1964) and Wilman and Agiegba (1982). Wilman and Agiegba (1982) described the changes in digestibility and proportion of leaf, stem, inflorescence and dead material from 2 – 14 weeks of re-growth. Elizalde et al. (1999) compared ruminal DM and CP degradation kinetics of fresh bromegrass, tall fescue and lucerne at different stages of maturity and showed an analogous relationship between forage quality and degradation parameters. Mambrini and Peyraud (1994) studying effects of two stages of ryegrass maturity (28 and 49 days of re-growth) on cow performance showed that Increasing maturity reduced grass N content, organic matter digestibility and non-ammonia N flow to the duodenum. Mambrini and Peyraud (1994) also found that total retention time in the gastrointestinal tract increased (43 - 48.8 hours) with a more mature grass and a longer chewing time was associated with coarse particles. The increase in proportion of stem, with associated increase in fibre content of the DM, and reduction in digestibility are well known (Cowan and Lowe, 1998; Hodgson, 1990; Minson, 1990; Sheaffer et al. , 1998; Wilson et al. , 1995), but less information is available concerning the rates and products of digestion, composition of in sacco residues and the effects of initial cutting date (relative to flowering) on these changes. The data presented here are based on perennial ryegrass cultivar Grasslands Samson and are unique to New Zealand. Although herbage mass and days of re-growth (age) were evaluated against chemical composition and rates of digestion, these may not be an appropriate base for ration balancing. Herbage mass will depend on soil fertility and growing conditions, so digestion characteristics of ryegrass grown at sites other than that used for this experiment must be based on composition rather than growth. This trial also showed 103 that the relationship between days of re-growth (age) and digestion was affected by initial mowing date (in relation to flowering), so effects of age should be interpreted in relation to flowering date. Even though chemical composition has not been a good predictor of feeding value on its own (Cherney and Mertens, 1998; Poppi, 1996; Ulyatt, 1973) it provides the underlying basis for ration balancing. This is shown in formulation of total mixed rations, where chemical composition is supported by requirements for effective fibre, buffers and processing to create a ration able to match cow nutrient requirements and sustain stable rumen fermentation (Mertens, 1997). In a similar way, the chemical composition of ryegrass can form the basis of ration formulation but it needs to be supported by digestion kinetics and this will be affected by time from flowering. Mathematical modelling, together with laboratory analysis of forage, forms the basis of ruminant feeding systems (Dynes et al. , 2003) but many models have been developed to formulate diets from chopped forages fed with grains. These models are essential to formulate balanced diets (Johnston and Shivas, 1999), but data are needed to improve model predictions when forages are grazed and form a major component of the diet (Alderman et al. , 2001). Kinetics of degradation and microbial growth will form a significant part of any system designed to predict cow performance from forage diets. Future systems will need to consider constraints to voluntary intake of ryegrass and especially implications of substitution when other forages are offered (Dixon and Stockdale, 1999; Stockdale, 1996; Wales et al. , 1999b). Prediction of chemical composition is also vital for model inputs and the use of NIRS (Adesogan et al. , 2000; Corson et al. , 1999; Givens and Deaville, 1999; Reeves III, 2000) has enabled economical determination of forage composition and products of digestion. Chemical composition and nutritive value Fibre was the principal component of ryegrass, with high concentrations of NDF and ADF in the DM even with young ryegrass (NDF and ADF ≥ 42.7 and 21.8 g/100 g DM, respectively). This is typical of pure perennial ryegrass species (Ulyatt, 1984), which differ from pastures that contain clovers and weed species. Cherney et al. (1993) measured changes in forage quality with increased maturation for five perennial grasses (tall fescue, foxtail, timothy, reed canarygrass and meadow bromegrass) and showed average NDF concentrations of fibre of young grasses was 40% and increased to 63% of DM after 42 days of re-growth. Chilibroste et al. (2000) reported a linear increase in NDF concentration of ryegrass over 30 days of growth from 44.6 to 53.2% at 6 to 30 days of re-growth respectively. The NDF forms a significant portion of ryegrass, 104 even when vegetative and the increase with maturation occurs before the appearance of the stem or inflorescence (Ulyatt et al. , 1988). Once flowering commences the proportion of leaf declines and fibre dominates ryegrass structure (Wilman and Agiegba, 1982). Conversely, CP concentrations decreased from 23.7 g/100 g DM in young material to only 5 g/100 g DM in mature grass. High values are associated with a high proportion of leaf, and low values with a high proportion of stem. Typical concentrations in ryegrass leaf are 19 g/100 g DM, and in stem 8 g/100 g DM. The increase in stem contributed to reduced plant CP in addition to the changing concentration in plant structure. Concentrations of NSC (excluding pectins and organic acids) measured in this study did not demonstrate consistent trends with age. The NIRS was calibrated to measure soluble sugars and starches based on chemical analysis (Englyst et al., 1987) but calibration did not include pectin which accounts for about 1 – 2 % of DM or organic acids, fructans and sucrose. These are major components of NSC, particularly in pasture grasses with typical values of 9 – 14 % in DM for water soluble sugars and 5 – 9 % in DM for organic acids (Ulyatt, 1984). Hall (1998) showed that NSC should include all carbohydrates not found in neutral detergent fibre (NDF) so an alternative estimate of NSC could be derived from 100 – CP – NDF – lipid – ash, however this prediction sums errors associated with each analysis. Digestible NDF and NSC comprise the main energy sources available in the rumen. Digestible nutrient pools The distribution of DM and its constituents between soluble (A), insoluble but degradable (B) and insoluble undegradable (C) pools will depend on both the composition of the forage and the method of preparation. The absence of age or initial mowing date effects on DM distribution between pools suggests a similar cell rupture and breakage using fresh minced technique, but there were differences in the DM and fibre (NDF and ADF) degradation rate with maturity (k; P < 0.001). Rates of DM degradation for young ryegrass were similar to those reported for immature ryegrass (k = 0.114) by Burke et al. (2000). The fibre content of grass increased as it matured, but both fibre and DM degradation rates declined as grass matured (Tables 3.12 – 3.14), probably in association with lignification (Jung et al. , 1997). There were no meaningful effects of initial mowing dates on rates of fibre degradation. Although lignin concentration increased from 2.4 to 5.4 g/100 g of DM with ryegrass maturation, concentration was not associated with changes in degradation 105 rate (k) of any constituent tested (DM, CP, NDF and ADF). Nevertheless, lignification increases the requirement for chewing to achieve the desired degree of cell rupture and increased bonding between lignin and structural carbohydrate appears to account for slower digestion of mature forage (Waghorn and McNabb, 2003). The increased bonding and presence of schlerenchyma and vascular tissues that are tough and resistant to microbial enzymes appear to reduce the effectiveness of chewing, rumen clearance and yield of nutrients during digestion. The degradation rates for CP were not reduced with age (P > 0.97; Table 3.13), and the distribution in minced DM did not match initial expectations. The proportion of CP in the readily fermentable pool increased from about 50% with young ryegrass to about 75% with very mature material. This distribution shows that a higher proportion of protein was released from mature plant cells which were resistant to mincing, but once released (solubilised), plant age had no affects on rate of disappearance in sacco. However the amount of CP was affected by plant maturity and this has implications for microbial growth and feed quality. Predicted OMD declined with maturity, from 83 to about 60 g OMD/100 g of DM. Similar values were summarized by Waghorn and Barry (1987) showing young leafy ryegrass digestibility decreased from 86% to 62% digestibility when mature. Low digestibility is associated with slow degradability and this is indicated by the in sacco data. Digestibility is also positively correlated with voluntary feed intake (VFI) (Minson and Wilson, 1980; Minson and Wilson, 1994). However this relationship is empirical and OMD accounts for only about 60% of the variation in VFI of forages for ruminants (Dynes et al. , 2003). These relationships are altered by changing particle size (Minson, 1990) and demonstrate the difficulty of predicting the intakes of ruminants fed roughages. 106 Forage preparation for in sacco and in vitro incubations The implications of sample preparation for in sacco procedures have been recently reviewed by Cohen and Doyle (2001) who concluded that concentrates should be dried and ground, but fresh mincing was more suitable for fresh forages. This supports the conclusion of McNabb et al. (1996) that mincing was more suitable for evaluating protein solubilisation and degradation than drying and grinding fresh forages. Barrell (2000) compared in vitro and in sacco digestibility of fresh minced, chopped and freeze dried and ground forages and showed mincing minimized lag times relative to chopping or drying and grinding. Preparation affected the ranking of forages and the author concluded mincing was the preferred preparation. Our results from in vitro and in sacco incubations also indicate minimal lag times, but the extensive release of CP from mature grasses required a comparison between particle size distribution (Table 3.10) and chewed boli or rumen contents from ruminants fed immature versus mature forage to demonstrate the suitability of mincing. While the general objective for sample preparation is to imitate chewing, this is obviously not possible because once forage is placed in the bags it can not be chewed further. The initial mincing is a compromise between initial masticating during eating and further chewing during rumination. The initial particle size of material can not be a perfect representation of in vivo particle size because it could be “over chewed” (i.e. smaller particle size than chewed during eating) which will result in degradation that may be too fast (compared to in vivo). However, the alternative of “under chewed” would result in an unrealistically slow degradation in the latter part of the in sacco period (i.e. after 12 hours of incubation). Ruminants chew plant material several times before it exits the rumen and in sacco preparations can only involve a single chewing. Data in Table 3.20 summarise particle size data for either young or mature grass, leaf and stem or contrasting diets (e.g.: ryegrass versus legumes). This information, from sheep and cattle clearly demonstrates a similar degree of particle size reduction for immature grass versus mature and for legumes, stems and hay. Although the distribution of particle size between large (> 2.0 mm), medium and small (< 0.25 mm) pools differed for individual studies, it is clear that particle size of mature ryegrass was not larger than for immature ryegrass or legumes. These observations were supported by trials with 6 or 12 week re-growth of leaf or stem from tropical grasses which showed no difference in particle size distribution in rumen contents of sheep or cattle (Poppi et al. , 1985). Modulus of fineness measurements for rumen contents of sheep were 2.78 107 mm for 6 week re-growth, and 2.67 mm for 12 week re-growth. Comparative values for cattle were 2.98 and 2.90 mm, respectively. The percentage of DM able to pass a 0.25 mm sieve after chewing during eating (38%) was similar to that for rumen content (35%; Table 3.20). These values are similar to mean particle size of minced ryegrass (Table 3.10) where 37% of DM passed a 0.25 mm sieve and show the mincer provided a good representation of rumen of diets and maturities. Particle size distribution in rumen contents was similar to swallowed boli, but proportions on each sieve will depend on outflow rates, intake and time since eating (Ulyatt et al. , 1986; Waghorn, 1986). The similar particle size in swallowed boli and rumen contents of sheep and cattle fed diverse diets is a function of time spent chewing, with more chewing required for fibrous than succulent material. Dado and Allen (1994) showed that chewing time was a function of NDF content in diet fed to cows. The speed and ease with which the mincer ground young versus mature ryegrass in this study, showed maturity demanded a much greater energy input for similar quantities of DM but the particle size distribution of young and mature forages were similar (Table 3.10). The mincing preparation used for incubations presented here showed a high degree of uniformity and the proportional distribution of DM across particle sizes resemble these for chewed forage and rumen contents (Table 3.20). 108 TABLE 3.20 – Particle size distribution in swallowed boli and rumen content of sheep and cattle fed contrasting diets to indicate effects of grass maturation or differences between forages containing high and medium concentrations of fibre. All perennial ryegrass (Lolium perenne L.) unless indicated. Particle size (sieve aperture; mm) Source Feed type > 2 2 - 0.25 < 0.25 Sheep chewing during eating 1 Early vegetative ryegrass a 48 14 38 1 Early bloom ryegrass a 33 32 35 2 Vegetative ryegrass b 43 16 41 2 Poor quality (mature) meadow hay b 49 29 23 3 Young ryegrass (51 g NDF/100 g DM) 47 14 39 3 Mature ryegrass (61 g NDF/100 g DM) 31 32 36 Sheep rumen contents 3 Young ryegrass (51 g NDF/100 g DM) 15 18 67 3 Mature ryegrass (61 g NDF/100 g DM) 13 46 41 4 Tropical grass leaf c 12 70 18 4 Tropical grass stem c 12 57 31 Cattle rumen contents 4 Tropical grass leaf c 15 40 45 4 Tropical grass stem c 20 46 34 5 Fresh lucerne after a 2 hours meal 29 31 40 5 Lucerne hay after a 2 hours meal 27 38 35 6 Fresh ryegrass after a 2 hours meal 51 12 37 6 Fresh lucerne after a 2 hours meal 39 24 37 Abbreviations: NDF, neutral detergent fibre a Assume 3 and 6% of DM passed a 0.5 mm and was retained in a 0.25 mm sieve for vegetative and pre bloom ryegrass, respectively b Data from publication and also the original data set. c Pangola (Digitaria decumbens ) and Rhodes (C hloris gayana ) grasses fed in separated trials, with no significant species effects and averaged for presentation. 1 - Ulyatt et al. (1986). 2 - Ulyatt (1982). 3 - Dellow et al. (unpublished). 4 - Poppi et al. (1985). 5 - Waghorn (1986). 6 - Waghorn et al. (1989). 109 Similar particle size distribution across grass maturities in vivo (Table 3.20) and in samples incubated here (Table 3.10) suggest the slower DM degradation rates with increasing maturity is primarily a function of chemical composition rather than particle size. This is also evident with fibre fractions but CP degradation rate was independent of maturity. The higher proportion of small losses (including soluble) DM arising from chewing or mincing does limit the use of in sacco procedures because up to 50% of DM (and up to 80% of CP) is able to leave the bags immediately and does not contribute to measurement of degradation kinetics. Soluble DM will degrade very rapidly with only a small fraction leaving the rumen unchanged (Ørskov, 2000). Use of in vitro fermentation in conjunction with in sacco digestion provides an indicator of products arising from digestion of soluble (A) and insoluble but degradable (B) DM and constituents. The preparations used here support the idea of Michalet-Doreau and Ould-Bah (1992) who suggested inclusion of large particles were important in forage incubations and acknowledge the requirement to compromise between the mincing preparation and the need to mimic mastication by ruminants. Mincing resulted in about 37% of DM able to pass a 0.25 mm sieve and 31% unable to pass a 2 mm sieve. Table 3.20 shows this distribution is similar to that of rumen contents in sheep and cattle fed forages but the soluble (A) fraction from ryegrass (Table 3.12) accounted for a slightly higher proportion of DM than identified by sieving (Table 3.10). The higher value for A fraction does support studies showing chewing ruptured about 60% of fresh forage cell (Reid, 1984; Waghorn et al., 1986). The technique for measuring particle size and release of cell contents may contribute to these anomalies. DM loss from in sacco bags require washing through pores of about 35 µm, compared to the 0.25 mm sieve used to determine particles size; nevertheless the overall distribution of DM of minced ryegrass does resembles the in vivo situation. The soluble fractions of forages are important for ruminant digestion as they represent readily fermentable energy, and proteins are very rapidly degraded (McNabb et al. , 1996). Data from this experiment suggest extensive protein degradation from mature ryegrass and this may be essential for maintaining the microflora when poor quality diets are fed. Differences in proportions of protein release from immature and mature ryegrass is supported by more extensive chewing when mature or fibrous forages are fed (Dado and Allen, 1994). Cows consuming early cut, high quality lucerne spend less time chewing than those fed full bloom lucerne and the higher intakes and milk production support the impact of energy cost associated with 110 fibre degradation (Brouk and Belyea, 1993; Martz and Belyea, 1986; Robinson and McQueen, 1997). Particles retained on larger sieves are not soluble and further physical degradation size is important for increasing tissue surface area and exposure to adherent fibrolytic bacteria (Weimer, 1996). Continued particle size reduction, to less than 2 mm (Waghorn, 1986) is essential to enable passage from the reticulum-rumen and avoid constraints of rumen fill and to maximise forage intake. However simple reduction in particle size does not mean the residues will have rapid rates of digestion because many small particles deeply stain with acid phloroglucinol, indicating a high proportion of lignin in the DM and have limited digestibility despite a high surface area exposure (Pond et al. , 1984). The Figure 3.25 illustrates fractions affecting passage of forage particles through the ruminant digestive tract. Grazing animals consume forage plants which are broken down by mastication, whereas intensive feeding involves mechanical chopping before ingestion (e.g. maize silage) and these diets are characterised by high voluntary intakes. FIGURE 3.25 – Diagram of passage of forage particles in the ruminant. Adapted from Martz and Belyea (1986). Whole plant Mastication Swallow particles 1 – 2 mm or less Rumination of particles > 1.2 mm Mechanical reduction of particle size Rumen fermentation and digestion Wet sifting and mixing, particles < 1.2 mm escape Fecal output usually particle < 1.2 mm Regulators of passage: 1. Size of particles 2. Density of particles 3. Fibre content 4. Rate of particle size reduction 5. pH and osmotic pressure 6. Distention of rumen and abomasum 7. Strength and frequency of abomasal and rumen contraction 111 Crude protein degradation Chewing during eating or mincing released about 40% of DM into soluble or very small particulate matter pools which included about 50% (or more) of CP and non- structural carbohydrates and less than 30% of fibre. These components would enable rapid microbial growth but the release of insoluble (B pool) protein was more rapid (12%.h-1) than fibre (8%.h-1), which is likely to result in a low efficiency of CP utilisation and excessive formation of ammonia when immature ryegrass is offered. The protein consumed by the animal should be partly degradable in the rumen; peptides and amino acids derived from proteolysis stimulate microbial growth and rumen fermentation (Beever and Cottrill, 1994). It is, therefore, important to define the CP degradability (Table 3.13) and ammonia production (Figure 3.20) for ryegrass at different maturities and provide appropriate readily fermentable carbohydrate to limit wastage. It is difficult to quantify the metabolic cost of ammonia conversion to urea (Lapierre and Lobley, 2001) but there is no doubt that a rapid release of soluble protein from immature ryegrass is wasteful and energetically expensive. Excessive CP degradation may be the most limiting nutritional factor in high-quality temperate forages (Broderick, 1995). Analyses of protein kinetics across all levels of maturity (Table 3.13) showed the proportion of CP as RDP averaged 0.83 and maturation did not affect the RDP: RUP ratio. Regression analyses from 38 studies with dairy cows (206 treatments) were used to evaluate lactation responses to proportions of RDP and RUP (Figure 3.26; NRC, 2001). Those analyses, when applied to cows producing up to 32 kg milk/day suggested a dietary requirement of 12.2% RDP with 6.2% RUP in the DM. The relationship between milk production and dietary protein was: Milk = - 55.61 + 1.15DMI + 8.79RDP – 0.36RDP2 + 1.85RUP (r2 = 0.52) where DMI and milk are kg.day-1, and RDP and RUP are percent of diet DM. 112 FIGURE 3.26 – Milk production response to rumen undegradable and degradable protein (RUP and RDP, respectively). Dry matter intake held constant at 20.6 kg/day. Source NRC (2001). Although some success has been demonstrated in complementing rapidly degradable nitrogenous forages with readily fermentable energy (Carruthers et al. , 1997; Nocek and Russell, 1988) in most of cases there has been little benefit to cow performance (Huntington and Archibeque, 1999; Kolver et al. , 1998; Robinson and McQueen, 1994; Shabi et al. , 1998). An alternative strategy may be the reduction in protein solubility through complementing pasture with forage containing condensed tannins (CT). The CT from one forage have been shown to reduce protein degradation in other forages (Waghorn and Jones, 1989) and when fed as mixed diet (Waghorn and Shelton, 1995). Also CT have increased milk production from dairy cows with reductions in fat and increases in protein content in some instances (Harris et al. , 1998a; Woodward et al. , 1999). As ryegrass matured the rate of CP degradation was not affected, but the proportion of protein released by mincing increased, so ration balancing would require different criteria to that for immature ryegrass. The lower concentration of CP in mature ryegrass suggests a need for supplementation with RUD to ensure adequate supply for lactation. The in vitro incubations of ryegrass older than about 50 days of re-growth demonstrated a very rapid disappearance of CP released by mincing so the microbial growth will not be sustained. The in vitro incubations complemented in sacco data and indicated the extent of protein degradation, as well as the rates and proportions of VFA produced. In vitro ammonia concentrations represented a net production, so the short peak of NH3 from mature forages (Figure 3.21) indicated the extent of release by mincing and the rate of 113 ammonia utilisation by bacteria. However the in vitro ammonia concentration was above 3.5 mMol/L, considered minimal concentration for growth (Satter and Slyter, 1974) for the initial 3 – 4 hours of incubation, even with mature forages. Calculations of microbial growth from Table 3.17 and OMD suggests digestion of medium – mature grasses could yield 80 – 120 g microbial CP/kg DOM, which is similar to estimates for cows grazing pasture (98 g/kg DOM; Table 2.13). These calculations are supported by the similar amount and rate of in vitro VFA production from grasses at all stages of maturity (Figure 3.22), so the predominant limitation to performance of cows grazing mature ryegrass may be their capacity for particle size reduction, rumen fill and intake. Recycling of endogenous N would help maintain good levels of microbial growth in vivo (Lapierre and Lobley, 2001) but extensive and rapid protein degradation of both immature and mature ryegrass will limit amino acid availability, despite a liquid turnover of 9 – 18%.h-1 (Berzaghi et al. , 1996; Reis and Combs, 2000). There is a need for either supplementation with readily fermentable carbohydrates or protected protein when cows are given immature pastures, or provision of RDP with mature pastures for milk production greater than 30 kg milk/cow.day. These options should be evaluated with feeding trials to determine to extent of response as pastures mature. Fibre degradation and VFA production NDF comprises cellulose, hemicellulose and lignin and represents the cell wall of plants. About 20% of the NDF appeared in the soluble “A” fraction after mincing, but the majority is digested slowly to provide a main portion of the energy derived from rumen fermentation. Rates of digestion of NDF are slow (3 to 14%/h; Table 3.14) compared to CP. The model demonstrated significant reductions in the rate of NDF digestion as ryegrass matured which will limit both the rate and supply of VFA for absorption and possibly feed intake. The filling effect of mature grass (or higher fibre content diets) is due to resistance to particle size reduction during rumination or chewing (Dado and Allen, 1995; Dulphy et al. , 1989). However lactating cows often have high intakes that result in passage of particles at a similar rate as digestion rates of fibre (Mertens, 1992b). Some fibre is not digested and will pass out of the rumen before microbial fermentation is completed. The impact of ryegrass fibre on rumen fill, function and passage requires careful evaluation. Although NDF usually accounts for 45 - 50% of ryegrass DM, which is well above the 35% upper limit recommended for lactating dairy cows, the effective fibre 114 (eNDF) content in pastures has been estimated as 43% to maintain rumen pH between 6.0 – 6.2 even though pasture samples ranged from 17 to 78% across studies (Kolver and de Veth, 2002). The fibre in ryegrass is flexible and resilient (Pond et al. , 1984) but does not have the “scratch factor” typical of chopped and stalky lucerne or maize, so the assessment of effective fibre given for North Hemisphere TMR do not seem relevant for ryegrass. The analyses reported here have shown the slow degradation of NDF in mature grass is likely to limit both feed intake and the rate at which VFA are supplied to the cow. Dietary NDF from ryegrass may pose a substantial limitation to production from cows grazing pasture (Kolver et al. , 2002; Waghorn, 2002). The proportions and amounts of VFA produced from fermentation are a result of substrate (additional to NDF) and the rate at which degradation occurs. Cellulytic bacteria adhere to particles (Weimer et al. , 1999) so the extent to which forage is damaged during chewing (or mincing) will affect rates of VFA production. The in vitro incubations showed similar proportions of VFA over the 24 hours duration and this was not affected by grass maturation (Table 3.18). Typically, acetate accounted for about 63 – 67 % and propionate 19 – 22 % of total VFA, which is in accordance with in vivo concentrations and production (Berzaghi et al. , 1996; Kolver et al. , 1998; Reis and Combs, 2000). These data show that maturity does not affect proportions of VFA and the proportions are similar at the commencement of fermentation and after 12 hours. Most surprising was the rapid rate of VFA production from mature ryegrass during the first six hours of incubation. Production was about 30% greater than immature ryegrass (Figure 3.22). It is possible that the mature grass had a higher proportion of ruptured cells relative to younger material, as indicated by the high proportion of CP released into the soluble (A) fraction or that high N content of immature grass limited microbial activity. These data suggest the changes in chemical composition associated with maturity (Table 3.3) affect the ease and extent to which cell wall are damaged rather than the potential degradability of the cell walls. When ground sufficiently, virtually all plant cell wall can be digested, including lignified structures (Wattiaux et al. , 1991). Yield of VFA after 6 or 12 hours declined with all grass maturities. This was unlikely to be affected by pH of the media for the initial 12 hours of incubation, but low CP concentrations in mature forages must have limited bacterial growth. The soluble DM (especially CP and NSC) would enable a rapid initial fermentation; however incubation of mature forage after 6 hours would not represent in vivo conditions where endogenous urea would sustain bacterial growth. The cow is faced with a substantial task in chewing mature forages during eating and rumination This is time consuming, 115 energy demanding and likely to provide a major constraint to intake as well as release of nutrients. 3.6 – Conclusion This research illustrated that DM and fibre digestion rates, along with other established forage quality parameters, clearly declined with increased maturity. Plant maturity had no effect on CP degradation rates or VFA yields but affected ammonia production in vitro and probably limited microbial growth when very mature. Particle size distribution of ryegrass provided by the mincer indicated a high degree of uniformity and the proportional distribution of DM across particle sizes resembled chewed forage and rumen contents in vivo. Changing maturity of ryegrass pastures demands a flexible system of supplementation to accommodate the excessive ammonia production in new growth and inadequate ammonia release from mature grass. The data provided here are able to set the foundation for a model able to predict optimal types and levels of supplementation for dairy cows grazing pasture in New Zealand. The hypothesis was proven in part; maturation did alter rates of degradation but products of digestion were less affected by maturation and the proportion of N degraded was similar for all maturities. Maturation lowered microbial yield and lessened ammonia production but yield and proportion of VFA in vitro was not affected by maturation. Initial cutting dates did not affect rates or products of digestion after a similar number of days re-growth. Chapter 4 Digestion kinetics of leaf, stem and inflorescence from five species of mature grasses. 1 1 A small portion of these data were previously published in the Proceedings of the New Zealand Society of Animal Production , 2001, 8-12. 4.1 - Abstract Digestion kinetics were measured for mature (green and non-senescent) components of five grass species using in sacco and in vitro incubations to define rates of degradation and nutrient release. Mature perennial ryegrass, tall fescue, Yorkshire fog, phalaris and paspalum were hand separated into leaf, stem and inflorescence for incubations. Concentration of fibre (NDF) in dry matter (DM) fractions ranged from 49 - 68 (leaf), 63 - 72 (stem) and 50 - 68 g/100 g of DM (inflorescence). Crude protein concentrations in the DM of the respective fractions were 7.0 - 23.6, 3.3 - 7.7 and 7.5 - 12.0 g/100 g of DM. Soluble DM (% of the total) determined after mincing accounted for 31 - 54% of leaf, 26 - 56% of stem and 20 - 49% of inflorescence, and fractional (h-1) degradation of the insoluble DM was very slow, ranging from 0.04 - 0.11 (leaf), 0.03 - 0.05 (stem) and 0.03 - 0.08 (inflorescence). After 24 hours of in vitro incubation plant nitrogen content become limiting for fermentation in most instances, especially with tall fescue and paspalum. Volatile fatty acid (VFA) production appeared to be similar for leaf, stem and flower fractions, but the proportion of plant DM released as VFA after 48 hours was only 7 – 12%, with a higher value (19%) for tall fescue. Nitrogen concentration in forage DM was not directly related to VFA yield in vitro. Keywords: forages; digestion kinetics; in sacco; in vitro; plant maturity; dairy cows. Short title: Digestion kinetics of mature grasses 4.2 - Introduction A major issue facing the New Zealand dairy industry is the rapid decline in milk production after peak lactation and anoestrous corresponding with grass maturation in October-December. Grasses commence spring growth with vigorous leaf production, which has a high feeding value (nutritive value x voluntary intake) for grazing sheep and cattle, but as daily temperatures rise an increasing proportion of stem and inflorescence appears (Fulkerson et al. , 1998). Although the extent to which grass is allowed to produce seed heads can be controlled by grazing management, nutritive value declines because of changes in chemical composition. The principal changes are increased proportions of fibrous stem (Wilman and Agiegba, 1982) and decreased concentrations of leaf protein, so that the maturing plant has higher proportions of fibre and lower proportions of non-structural (readily fermentable) carbohydrate and protein in the dry matter. These changes reduce the amount of amino acids available to 117 ruminants and may increase the proportions of acetate: propionate available for absorption (Russell and Strobel, 1993). Nevertheless, a significant effect of grass maturation is that the rate of digestion and clearance of residual forage fibre from the rumen is reduced, because mature forages are slower to digest and require more chewing to reduce the particle size of plant fragments to a size able to pass out of the rumen (Chapter 3). The slower dry matter passage rate from the rumen may reduce feed intake. Consequently, maturation in pasture will result in lowered intake as well as declining nutritive value. Supplementation of dairy cows on pasture is often undertaken without knowledge of the nutrients available from the pasture base. Burke et al. (2000) defined the degradation kinetics of immature leafy material from a range of forages to be used as a basis for formulating forage mixed rations (FMR). That work determined degradation rates for dry matter, protein and fibre. They also indicated the amount and proportions of volatile fatty acids (VFA) produced, and have provided a mathematical basis for comparing contrasting feed types, but only for immature leaves. The work described here examines grasses which have not been grazed and are in an advanced stage of maturity, not senescent but with stems and nearly mature flowers. This study aims to determine the digestive characteristics at the opposite end of the range to that of Burke et al. (2000), using five very mature grass species. This experiment involved the separation of the five grass species into leaf, inflorescence and stem (including sheath) fractions for incubation in vitro and in sacco. In vitro incubations were conducted to determine the products of degradation (ammonia from proteolysis and VFA), whilst the in sacco technique was conducted to determine the rate at which dry matter (DM) and its constituent chemical fractions are degraded through microbial digestion. The primary purpose of this work was to define the degradation kinetics and products of digestion from components of mature grasses. The grasses form a cross section of species found in New Zealand, and are also used in grazing systems overseas. Perennial ryegrass forms the dietary basis of New Zealand dairy systems. Tall fescue is used in New Zealand but covers over 14 million hectares in the USA. Tall fescue has many desirable agronomic and forage attributes, forms a dense persistent sward and is tolerant of a wide range of management regimes. Paspalum is a subtropical dairy pasture that occurs in Northland and dominates most irrigated swards during summer and autumn in northern Victoria (Australia). It is a poor quality forage (Stockdale, 1997) but with a wide spread distribution. Phalaris and Yorkshire fog are considered weed grasses in New Zealand, but phalaris (Reed canary grass) is used in 118 North America and Australia where it is able to produce high biomass and dominate native grasses (Carlson et al., 1996). Yorkshire fog is common in wet, acid and infertile areas and frequently occurs in New Zealand pastures. Measurements were made of rates and products of degradation from leaf, stem and flowers from five mature grass species to define nutritional responses to flowering. 4.3 – Material and methods The mature grass species used in this study were perennial ryegrass (Lolium perenne L. cv. Grasslands Samson), tall fescue (Festuca arundinacea), Yorkshire fog (Holcus lanatus), phalaris (P halaris arundinacea) and paspalum ( Paspalum dilatatum ). About 2 kg (fresh) of each species was harvested in the summer of 1999/2000, refrigerated and immediately separated (by hand) into leaf, inflorescence and stem (with sheath) fractions, and stored at -16oC until incubation. Dead matter was discarded. Approximately 500 g wet material was obtained for each plant fraction for in sacco and in vitro incubations of minced material as well as measurements of dry matter content, particle size distribution of minced fractions and chemical composition by near infra red reflectance spectroscopy (NIRS). These procedures have been described in detail in section 3.3. Frozen forages were chopped into approximately 2 cm lengths (using scissors) and minced (whilst frozen) in a Kreft Compact meat mincer with 12 mm holes in the sieve plate. Minced material was stored at -16oC until the day prior to incubations when about 2.5 g wet weight (ww; 0.5 g DM) was placed in incubation bottles and 25 g ww (5.0 g DM) into 100 x 100 mm dacron bags for placement in the rumen of a fistulated cow. In sacco and in vitro incubations were carried out simultaneously for leaf, stem and inflorescence fractions of one species during each incubation. One ruminally cannulated non-lactating Friesian cow was fed lucerne hay for all incubations in order to maintain a similar rumen environment over the period of evaluation and inclusion of ryegrass standards enabled variations between runs to be monitored. Fourteen bags of each forage constituent were placed in the rumen and duplicates removed after 0, 2, 6, 12, 24, 48 and 72 hours for washing, drying (60 °C), weighing and analysis of residues by NIRS. Disappearance of DM, CP, NDF and ADF fractions were analysed using the non-linear model (No. 1) described by López et al. (1999) to determine fractional disappearance rate (k, %.hour-1), lag time (L, hour) and potential degradation (P) according to: 119 P = A + B (1-e-k(t-L)) where A = soluble fraction (% of DM, CP, NDF or ADF at t = 0 hour), B = degradable insoluble fraction and t is time in hours. Incubations showed a significant initial lag period for the components so the effective degradability was calculated by incorporating a lag phase with a DM fractional passage rate (kp) of 0.06 h-1 (Hoffman et al. , 1993) into the model. Analyses were also made for fibre fractions using a slower outflow rate which may be more typical of poor quality diet (kp = 0.02 h-1). E = A + B*k/(k + 1/((1/kp) - L)) Effective rumen degradability of CP (ERDP, g/kg DM) was calculated as: ERDP (g/kg DM) = CP [(0.8 x A) + (B x k)/(k + kp)] (AFRC, 1992). Relationships between degradability, nutritive characteristics of the forage components in in sacco samples and ERDP were analysed by regression (PROC GLM; SAS, 2001) using the model: ERDP (g/kg DM) = a + bX where X is the effective degradability of DM or NDF content of forage components for in sacco residues. Twenty-one bottles were prepared with each forage constituent for in vitro incubations by adding 12 mL buffer, 0.5 mL reducing agent and 3 mL rumen fluid to the plant material to 50 mL bottles (Burke et al. , 2000; Chapter 3). Bottles were made anaerobic by flushing with carbon dioxide and held at 39°C in an oscillating incubator for the duration of each incubation. Triplicate samples were removed after 0, 2, 6, 8, 12, 24, 36 and 48 hours of incubation for determination of ammonia concentrations and VFA at 0, 6, 12, 24 and 48 hours. See appendices 1, 2 and 3 for incubation details, ammonia and VFA analyses. Statistical analysis Chemical composition and in sacco data from leaf, stem and flowers were analysed by the general linear model (PROC GLM; SAS, 2001). Fixed model analysis was designed to test and detect differences among means for each data set (i.e. concentrations and digestion parameters). The terms tested were: “Forage” (sum of leaf, stem and flower for each forage) and “Components” (means of leaves, stem and flower among forages). As all components of each forage were incubated together there may be run effects in addition to forage effects, so the term “Forage” is in reality a forage-run effect. For in vitro data, the following terms were added to the in sacco model: “Time” (effect of incubation time on concentration of products of 120 fermentation); and interactions “Forage*components” (differences among components for each forage specie), “Component*time” and “Forage*time”. Data from in sacco incubations were expressed as degradation curves using a non-linear least-square procedure (PROC NLIN; SAS, 2001) to provide estimates for A, B, C, k and lag time. Plots of DM, CP and fibre disappearance against incubation time (in hours) were averaged by component (leaf, stem and flower). No attempt was made to weight degradation estimate parameters, as was done for in sacco data in Chapter 3. Ammonia concentrations in each incubation bottle were used to calculate net ammonia production (Appendix 2) and plotted against incubation time. Analyses were based on net conversion of plant N to ammonia N and also concentration of ammonia in media at each incubation time (mMol NH3/L). The conversion of plant N to NH3-N was evaluated by comparing plant components, forages and interactions described above. VFA analyses were based on single samples for each of the five periods for each forage component. VFA production was expressed as net production mMol/g DM incubated and as molar proportion of acetate: propionate (A: P). Mean pH values for each forage component at each time point were determined and plotted over incubation time for the 15 forage components incubated. Appendix CD (Chapter 4) has details of SAS procedures used for statistical analyses and complete data outputs. 4.4 – Results 4.4.1 - Chemical composition Although dead matter was discarded during the hand separation of leaf, stem and flower, the high DM percentage of most components, together with NDF concentrations in excess of 50 g/100g of the DM, demonstrate the maturity of the forages collected (Table 4.1). Paspalum had the highest NDF concentrations (over 65 g/100 g of the DM for all constituents), with a low crude protein (CP) concentration (below 8 g/100 g of the DM). Leaf of temperate forages (ryegrass, tall fescue, Yorkshire fog and phalaris) had at least 15 g CP/100 g of DM, with lower concentrations in flowers and low (< 10 g/100 g of DM) and variable CP concentrations in stem. Non-structural carbohydrate concentration was low (< 10 g/100 g of DM) in all plant components, except tall fescue flower (Table 4.1). Stem accounted for 40 - 60% of plant DM across all grasses but leaf ranged from 8 - 40% and flower 12 - 30% of the DM. Mature ryegrass 121 and Yorkshire fog had over 35% leaf when mature, whereas fescue and phalaris had less than 15% leaf and over 30% flower dry matter. 4 .4 .2 - Particle size The particle size distribution of minced material (% of DM) showed that 20 - 26% of leaf, stem and flower fractions of perennial ryegrass, tall fescue and Yorkshire fog were retained on sieves with 2 mm or large aperture sizes, although a higher proportion of tall fescue flower (38% of DM) was retained on these sieves (Table 4.2). In contrast only 7 - 13% of paspalum fractions were 2 mm or larger in size. Dry matter passing sieves with a 0.25 mm aperture size (fine particulate and soluble DM) accounted for 29 - 57% of plant DM, with a narrow range across the five grass species for flowers (35 - 47%) relative to leaves (tall fescue 30% - paspalum 57%) and stems (tall fescue 30% - Yorkshire fog 46%). Intermediate sized DM (0.25 – 1.0 mm) accounted for 37 - 43% of DM across leaf, stem and flower fractions of the five grasses (Table 4.2). 122 TABLE 4.1 – Dry matter concentration (g DM/100 g material) at harvest, chemical composition (g/100 g of the DM), predicted organic matter digestibility (g/100 g) and metabolisable energy (ME; MJ/kg DM) content of leaf, stem and flower fractions of five mature grasses used for measurement of digestion kinetics. Forage DM CP NSC Lipid NDF ADF ADL Ash OMD ME Perennial ryegrass leaf 21.9 18.9 8.7 4.0 49.2 23.7 4.7 11.5 79 8.5 stem 30.2 7.7 6.5 1.6 62.6 38.0 5.7 8.9 56 8.2 flower 45.5 12.0 8.3 3.4 52.8 31.5 5.2 8.1 61 8.8 Tall fescue leaf 30.5 15.0 4.5 3.1 56.0 33.9 3.6 12.1 64 9.0 stem 36.1 6.0 7.1 1.4 65.7 39.0 6.3 6.5 53 7.9 flower 46.7 8.7 18.5 2.2 50.6 32.5 3.4 6.6 62 9.1 Yorkshire fog leaf 20.3 23.6 6.2 4.2 50.8 24.7 4.3 8.9 78 9.9 stem 34.1 5.0 4.8 1.4 68.3 42.1 6.5 6.8 52 7.7 flower 54.3 8.6 5.6 5.3 59.0 34.0 4.6 6.2 57 8.2 Phalaris leaf 32.2 21.6 5.5 4.3 49.7 30.8 1.6 12.0 71 9.7 stem 34.6 6.1 0.0 1.3 71.8 48.3 6.5 8.9 46 6.7 flower 37.6 10.1 8.3 4.3 50.1 32.3 4.6 8.8 59 8.4 Paspalum leaf 39.3 7.0 0.1 3.1 67.7 42.4 6.2 10.0 54 7.7 stem 28.6 3.5 6.6 1.3 65.4 41.5 5.6 8.5 56 8.2 flower 50.7 7.5 0.0 3.3 67.7 43.1 7.7 5.5 39 5.8 Mean across species leaf 28.8 17.2 5.0 3.7 54.6 31.1 4.1 10.9 69.1 8.9 stem 32.7 5.7 5.0 1.4 66.7 41.8 6.1 7.8 52.7 7.7 flower 47.0 9.4 8.2 3.7 56.0 34.7 5.1 7.1 55.7 8.0 Abbreviations: CP, crude protein; NSC, non-structural carbohydrates; NDF, neutral detergent fibre; ADF, acid detergent fibre; ADL, acid detergent lignin; OMD, organic matter digestibility; ME, metabolisable energy. 123 TABLE 4.2 – Dry matter particle size distribution of leaf, stem and flower fractions of five mature grasses for in sacco and in vitro incubations indicated by sieve aperture size either retaining or enabling material to pass. Sieve size ≥ 2 mm 0.25 - 1 mm < 0.25 mm Perennial ryegrass leaf 22.2 37.6 40.2 stem 25.8 34.0 40.2 flower 24.1 28.8 47.1 Tall fescue leaf 19.9 50.6 29.6 stem 23.8 47.2 29.0 flower 38.0 22.1 39.9 Yorkshire fog leaf 26.0 21.4 52.6 stem 20.2 34.3 45.5 flower 20.6 32.3 47.1 Phalaris leaf 22.7 39.8 37.5 stem 13.1 47.7 39.2 flower 12.1 50.5 37.4 Paspalum leaf 9.6 33.8 56.6 stem 6.6 53.0 40.4 flower 12.5 52.4 35.1 Average ≥ 2 mm 0.25 - 1 mm < 0.25 mm leaf 20.1 36.6 43.3 stem 17.9 43.3 38.9 flower 21.5 37.2 41.3 STDEV ≥ 2 mm 0.25 - 1 mm < 0.25 mm leaf 6.3 10.6 11.1 stem 8.0 8.6 6.0 flower 10.6 13.5 5.5 STDEV = standard deviation. 124 4 .4 .3 - DM digestion kinetics In sacco dry matter distribution across fractions and disappearance rates are summarised in Table 4.3 and illustrated in Figure 4.1. DM in the soluble (A) fraction averaged across forage species and components was 38% with flowers having lower values than leaves and stems (P < 0.0001). The model explained 44% of variation in soluble DM content for forage species and components. The insoluble and slowly degradable DM fraction (B) differed between components with higher values for leaves followed by flowers and stems (P < 0.0001; Table 4.3). The model explained 57% of the variation in the B fraction for all forages. The undegradable (C) fraction was higher for stems and flowers (23%) compared to leaves (11%). Kinetic data derived from curves fitted to in sacco data (Table 4.3) show rapid degradation rates of ryegrass, paspalum and fog leaf and very slow degradation of phalaris and fescue leaf. However the rapid degradation of paspalum leaf was preceded by a 10 hour lag period, whereas the model did not fit a lag period to phalaris leaf. Leaf, stem and flowers had similar, slow degradation rates for phalaris but stem degradation was preceded by an 8 hour lag. Differences in effective degradability (a prediction of degradation in vivo) incorporate effects of lag phase with degradation rate and outflow (0.06 h-1) and show similar low values for both stem and flower relative to leaf. Effective degradability will be determined by particle size and release of soluble DM during mincing and/or chewing as well as the physical and chemical structure of leaf, stem and flower. Table 4.3 and Figure 4.1 show similar rates of degradation for both flower and stem fraction DM averaged for the five forages which were about half that for leaf (P < 0.0001). The error bars indicate substantial differences between grass species in degradation rate of individual constituents, and this is further demonstrated by the DM disappearance from the leaf fraction of each species in Figure 4.2. The model accounted for sixty seven percent of the variation in rate of DM digestion (k). Paspalum flower was predicted to be especially indigestible (E = 35%) compared to flower from temperate grasses (E = 51 - 59%). Most constituents of mature grasses had substantial lag periods prior to DM loss, suggesting a relatively slow colonisation of particles by rumen bacteria and fungi. 125 020 40 60 80 100 0 12 24 36 48 60 7 FIGURE 4.1 - In sacco dry matter (DM), crude protein (CP), neutral and acid detergent fibre (NDF and ADF) disappearance from leaf, stem and flowers from five mature grasses (mean ± standard error bar). FIGURE 4.2 - Disappearance of leaf dry matter (DM) during in sacco digestion of five mature grasses. FIGURE 4. 2 40 60 80 100 0 12 24 36 48 60 7 % CP d is appe ar an ce % DM d is appe ar an ce Leaf 2 Leaf Stem Flower Stem Flower 0 20 40 60 80 100 0 12 24 36 48 60 7 2 0 20 40 60 80 100 0 12 24 36 48 60 7 Incubation time (hours) % A DF d is appe ar an ce Incubation time (hours) % NDF d is appe ar an ce Leaf 2 Leaf Stem Flower Stem Flower 20 40 60 80 100 0 12 24 36 48 60 7 Incubation time (hours) % o f le af DM lo ss 2 Ryegrass Tall fescue Yorkshire fog Phalaris Paspalum 126 TABLE 4.3 - Mature grass dry matter (DM) degradation characteristics (% of DM) defined as soluble (A), degradable insoluble (B) and undegradable residue (C = 100 – A – B) as well as fractional degradation rate (k, h-1), lag time (Lag, hours) and effective degradability (E) which takes into account the effect of passage rate from the rumen. Forage A B C k Lag E1 Perennial ryegrass leaf 38 53 9 0.09 3.7 67 stem 47 30 23 0.04 9.5 54 flower 49 34 17 0.04 4.6 59 Tall fescue leaf 31 51 17 0.06 4.1 54 stem 26 33 41 0.05 8.7 36 flower 26 47 27 0.08 0.0 53 Yorkshire fog leaf 38 51 11 0.11 4.0 68 stem 35 45 20 0.03 6.0 46 flower 44 29 26 0.05 3.2 56 Phalaris leaf 43 51 5 0.04 0.0 63 stem 36 46 18 0.03 8.3 46 flower 31 55 14 0.03 0.0 51 Paspalum leaf 54 35 11 0.09 10.4 66 stem 56 31 13 0.05 8.1 66 flower 20 43 37 0.04 4.6 35 Leaf 41±1.00 48±0.76 11 0.079±0.002 4.5±0.22 64 Stem 40±0.96 37±0.72 23 0.042±0.002 8.1±0.21 50 Flower 34±0.96 42±0.72 24 0.048±0.002 2.5±0.21 51 1 Calculated using a fractional passage rate of 0.06h-1. 4 .4 .4 - CP digestion kinetics The soluble CP (A) fraction varied from 35 – 61% for mature leaves, 41 – 74% in stems and 41 – 67% in flowers and averaged 55% across grass species and components. There were significant differences between components and forages (P < 0.0001; Table 4.4) in the % of CP released into the soluble pool. Appendix CD (Chapter 4) presents the complete data set outputs (e.g.: Least Squares Means for effect of Forages and Components). The insoluble degradable (B) pool contained on average 46, 18 and 31% of CP for leaf, stem and flower across forages and also varied between forages (P < 0.001). 127 The distribution of CP between A and B fractions was similar for leaf of all grasses but when stem was minced most of the protein was released into the A fraction, except for paspalum. As with DM distribution, stems had higher values for undegradable CP (fraction “C” averaged 21% compared to flowers (13%) and leaves (7%). The model accounted for 41 and 55% of the variation in distribution of CP within the A and B fractions, respectively. CP degradation rates differed across forages species and components (Table 4.4; P < 0.001). Tall fescue had lower degradation rates (k) for leaf CP (0.05 h-1) and higher values for stem and flower compared to other forages. Lag phases resembled patterns for DM, with longer lags for stems than leaves or flowers. Effective degradability (E; Table 4.4) calculated using a fractional passage rate of 0.06 h-1 shows similar values of about 71% for leaves, stems and flowers in all forages, despite differences in distribution between pools, degradation rates and lag times. Protein degradability has been expressed in terms of DM (ERDP) and plotted against DM degradability and NDF content. Figure 4.3 demonstrates similar ERDP for stem of all species but diverse values for leaf, resulting in a poor positive relationship (r2 = 0.27) between effective rumen degradability of CP and DM across forages and components (P = 0.049). A stronger relationship is evident when ERDP is plotted against NDF concentration for the five forages and components incubated in sacco (P < 0.001; Figure 4.4). Leaf, stem and flower varied in NDF content from 492 to 718 g/ kg DM and the concentration of NDF accounted for 56% of the variation in effective rumen degradability for protein across forage components (Figure 4.4). 128 050 100 150 200 30 50 70 Effective DM degradability (E DM ) ERDP, k p=0.0 6 h -1 Leaf Stem Flower FIGURE 4.3 – Relationship between effective degradability of dry matter (EDM) and crude protein (ERDP, g/kg DM) of forage components. When rumen DM outflow is 0.06 h-1, the relationship is described as: ERDP (g/kg DM) = - 50.1 + 2.18 (±1.00) x EDM (r 2 = 0.27; Root MSE = 40.1; CV = 58.2%; P=0.0488). 0 50 100 150 200 450 550 650 750 Neutral detergent fibre, g/kg DM ERDP, k p=0.0 6 h -1 Leaf Stem Flower FIGURE 4.4 – Relationship between neutral detergent fibre content (NDF, g/kg DM) and effective rumen degradability of crude protein (ERDP, g/kg DM) of forage components. When rumen DM outflow is 0.06 h-1, the relationship is described as: ERDP (g/kg DM) = 313 - 0.41(±0.10) x NDF (r 2 = 0.56; Root MSE = 30.9; CV = 44.8%; P = 0.0012). 129 TABLE 4.4 - Mature grass crude protein (CP) degradation characteristics (% of CP) defined as soluble (A), degradable insoluble (B) and undegradable residue (C = 100 – A – B) as well as fractional degradation rate (k, h-1) and lag time (Lag, hours). These data are used predict effective degradability (E), which takes into account the effect of passage rate from the rumen and the effective rumen degradability protein (expressed on a dry matter basis (ERDP; g/kg DM). Forage A B C k Lag E1 ERDP1 Perennial ryegrass Leaf 42 52 6 0.10 1.0 73 124 Stem 70 12 18 0.09 10.2 74 49 Flower 67 20 13 0.06 4.7 75 76 Tall fescue Leaf 53 46 1 0.05 0.0 75 96 Stem 74 6 20 0.13 0.0 78 38 flower 41 46 13 0.23 0.0 78 60 Yorkshire fog leaf 46 48 7 0.15 3.8 77 167 stem 59 19 22 0.04 11.9 76 27 flower 50 33 18 0.08 3.9 67 51 Phalaris leaf 61 37 2 0.07 0.0 81 149 stem 60 28 12 0.04 2.0 70 36 flower 56 41 3 0.03 0.0 68 58 Paspalum leaf 35 45 20 0.09 12.0 48 39 stem 41 29 31 0.12 12.0 51 18 flower 65 18 17 0.11 5.2 75 48 Leaf 47±1.18 46±1.12 7 0.09±0.03 3.4±0.27 71 115 Stem 61±1.13 18±1.07 21 0.08±0.03 7.2±0.25 70 34 Flower 56±1.13 31±1.07 13 0.10±0.03 2.8±0.25 73 59 1 Calculated using a fractional passage rate of 0.06h-1. 130 4 .4 .5 - Fibre digestion kinetics Fibre accounted for 50 – 70% of DM for the five mature grasses and their components (Table 4.5). ADF accounted for a slightly lower percentage of leaf NDF (57%) compared to stem and flower (62%), and values were lowest for ryegrass and Yorkshire fog leaf (48%). Fibre digestion will contribute a major portion of energy for the animal but rate and extent of digestion will be affected by particle size and microbial colonisation. In contrast to protein, an average of 26% of NDF and 22% of ADF were released into the “A” fraction by mincing. There were differences between forages and components in the distribution of both NDF and ADF (Table 4.5 and 4.6) with 65% and 53% of variation in “A” pool of the respective components explained by the model. The insoluble degradable (B) fraction accounted for about half of the NDF and a similar proportion of ADF, but with higher values for leaf than stem and flower (Table 4.6). Both NDF and ADF degradation rates were twice as fast for leaves compared to stem and flower (P < 0.0001; Figure 4.2). The lag time was highest (8 – 9 hours) for stems compared to leaves and flowers and virtually all components of all plants exhibited a substantial lag period prior to degradation. When fractional outflow rate was 0.06 h-1 the net effect of lag and degradation rate resulted in an effective degradability for leaf NDF and ADF averaging 56 and 48% respectively, and lower values for stem and flower fibre (E; Table 4.5 and 4.6). Fibre was poorly degraded, but calculation using a slow outflow from the rumen (0.02 h-1) increased degradation of NDF and ADF by 15 – 20 percentage units. One consequence of a slow outflow rate would be very low feed intake. The model explained 53% and 59% of variation in the “B” pool for NDF and ADF respectively. 131 TABLE 4.5 - Mature grass neutral detergent fibre (NDF) degradation characteristics (% of NDF) defined as soluble (A), degradable insoluble (B) and undegradable residue (C = 100 – A – B) as well as fractional degradation rate (k, h-1), lag time (Lag, hours) and effective degradability (E) which takes into account the effect of passage rate from the rumen. Forage A B C k Lag E2% E6% Perennial ryegrass leaf 28 59 13 0.08 4.0 75 58 stem 36 38 27 0.04 10.9 58 42 flower 31 46 23 0.03 5.0 59 44 Tall fescue leaf 18 51 30 0.11 8.2 61 43 stem 5 45 50 0.04 10.0 34 15 flower 16 55 29 0.03 0.0 49 34 Yorkshire fog leaf 31 54 14 0.09 3.5 76 62 stem 22 55 23 0.03 6.0 52 35 flower 35 32 33 0.04 3.9 56 46 Phalaris leaf 30 62 8 0.03 1.8 68 50 stem 24 58 17 0.03 5.5 55 37 flower 9 66 25 0.04 3.9 51 31 Paspalum leaf 53 40 7 0.06 9.9 81 65 stem 27 51 22 0.05 8.3 62 42 flower 27 39 34 0.03 4.8 51 39 Leaf 32±0.87 53±0.86 14 0.08±0.02 5.5±0.25 72 56 Stem 23±0.83 49±0.81 28 0.04±0.02 8.1±0.23 52 34 Flower 24±0.83 48±0.81 29 0.04±0.02 3.5±0.23 53 39 E2% and E6% were calculated using a fractional passage rate of 0.02 and 0.06h-1 respectively. 132 TABLE 4.6 - Mature grass acid detergent fibre (ADF) degradation characteristics (% of ADF) defined as soluble (A), degradable insoluble (B) and undegradable residue (C = 100 – A – B) as well as fractional degradation rate (k, h-1), lag time (Lag, hours) and effective degradability (E) which takes into account the effect of passage rate from the rumen. Forage A B C k Lag E2% E6% Perennial ryegrass leaf 6 77 17 0.09 4.2 69 48 stem 31 42 27 0.04 10.7 56 39 flower 30 49 21 0.03 4.6 60 44 Tall fescue leaf 5 59 35 0.12 9.1 55 34 stem 0 51 49 0.04 11.0 32 5 flower 26 57 17 0.02 0.0 53 39 Yorkshire fog leaf 9 72 19 0.11 3.7 69 52 stem 18 60 22 0.03 6.0 51 32 flower 31 39 30 0.04 1.8 56 45 Phalaris leaf 21 70 9 0.03 1.4 65 45 stem 26 57 17 0.03 6.0 57 39 flower 19 61 21 0.04 4.0 58 39 Paspalum leaf 53 37 11 0.08 11.2 80 64 stem 29 50 21 0.05 8.8 62 43 flower 33 37 30 0.04 7.0 56 43 Leaf 19±1.21 63±1.04 18 0.09±0.002 5.9±0.26 68 48 Stem 21±1.15 52±0.99 27 0.04±0.002 8.5±0.25 52 31 Flower 28±1.15 48±0.99 24 0.03±0.002 3.5±0.25 57 42 E2% and E6% were calculated using a fractional passage rate of 0.02 and 0.06h-1 respectively. 133 4 .4 .5 - In vitro incubations The buffered media used for in vitro incubations showed moderate decreases in pH, suggesting a limited production of VFA. The only value below 5.6 was for tall fescue flowers after 12 hours of incubation (Figure 4.5). Tall fescue flower contained a high concentration of NSC (18.5 g NSC/100 g DM) relative to other grass components. Protein degradation is indicated by ammonia concentration in the incubation media (mMol NH3/L; Figure 4.6) and net ammonia production from plant components for each forage species (Figure 4.7 and 4.8). Concentrations generally peaked at or before 12 hours of incubation (except for phalaris leaf) and values above 0 hour suggest protein degradation is exceeding microbial utilisation. Inadequate ammonia yield from all components of tall fescue and paspalum are apparent (Figure 4.6 and 4.7; Table 4.7) after 12 hours whereas concentrations remained elevated for incubations of leaf from ryegrass, phalaris and Yorkshire fog. When NH3 concentrations are expressed as a proportion plant N, the values are affected by N in plant DM as well as proteolysis and incorporation into bacterial protein (Barrell et al, 2000). Some leaves contained sufficient N for microbial growth, but stem and flower had insufficient N, especially after 12 hours of incubation (Figure 4.8). Negative values indicate ammonia uptake from the rumen inoculum and this was apparent for all stem material except ryegrass. These data show insufficient N was released from all components of tall fescue and paspalum after 6 hours to sustain microbial growth. Nitrogen insufficiency may limit production of VFA in vitro and limit degradation rates in situ. The rates of VFA production over 48 hours have been summarised for leaf, stem and flower in Figure 4.9. A rapid initial VFA yield was apparent for all constituents (to 12 hours) although propionate production appeared to be more rapid from flower than stem or leaf. Net production of all VFA were sustained over 48 hours when stem was incubated, whereas the yield did plateau or became negative after 24 hours for the leaf and flower fractions. The average rates of VFA production were similar for leaf, stem and flower fractions from 0 – 12 hours (Table 4.8) with yields of 112, 105 and 119 µMol/g DM per hour for the respective fractions. Calculations of yield over the entire 48 hours incubation across the five species (Appendix 7 – Table 7.2A) averaged 1.61, 2.12 and 1.46 mMol/g DM for leaf, stem and flower respectively. 134 There was a much greater rate of VFA production from tall fescue and phalaris compared to ryegrass, Yorkshire fog and paspalum (Table 4.9; Appendix 7 Table 7.2A). The differences are illustrated in Table 4.9 at 12 hour, but the absence of replicates prevented a robust statistical validation of forage effects (not significant). By 48 hour, the net yield of VFA (% of DM assuming a mean molecular weight of 67) for the five grass species (assuming leaf, stem and flower each contribution are third of the DM) was perennial ryegrass, 6.8%; tall fescue, 18.9%; Yorkshire fog, 12.1%; phalaris, 10.5% and paspalum, 9.7%. TABLE 4.7 – Mean values for net ammonia production and concentration above 0 hour values in the incubation buffer and from leaf, stem and flower and for the five grass species at intervals during the 48 hour incubation. mMol NH3/mM plant N mMol NH3/L 2 - 6 8 - 12 24 - 48 2 - 6 8 - 12 24 - 48 Leaf 0.053 0.065 0.003 6.7 7.1 4.7 Stem 0.054 -0.030 -0.122 4.5 2.9 0.7 Flower 0.082 0.090 -0.045 6.8 6.8 2.3 Perennial ryegrass 0.071 0.107 0.022 4.6 6.0 2.4 Tall Fescue -0.009 -0.107 -0.113 5.7 1.5 0.8 Yorkshire fog 0.074 0.081 -0.078 6.6 6.8 2.2 Phalaris 0.084 0.079 0.043 6.9 7.4 6.8 Paspalum 0.087 0.027 -0.154 5.9 5.8 0.6 135 TABLE 4.8 – Mean values for pH and rates of volatile fatty acid production from leaf, stem and flower and for the five grass species incubated for 48 hours. Data for VFA are mean hourly rates averaged for 0 – 6 and 6 - 12 hour incubation period. Main effects and interactions have been tested for significance. µMol/g DM.hour pH Acet Prop Buty Minor Total Leaf 6.90 79.4 18.7 11.3 2.8 ac 112.1 Stem 7.08 74.5 25.1 13.1 1.2 bc 105.4 Flower 6.90 68.6 16.7 19.9 5.9 a 119.4 Perennial ryegrass 7.22 63.6 16.7 b 13.3 3.4 97.0 Tall Fescue 6.35 108.0 40.3 a 23.5 3.2 175.0 Yorkshire fog 7.22 57.5 12.8 b 9.0 2.0 81.2 Phalaris 6.95 86.5 14.2 b 14.7 5.3 120.7 Paspalum 7.05 55.2 16.7 b 13.1 2.7 87.6 Model P 0.000 0.976 0.223 0.651 0.518 0.947 Forage P <.0001 0.678 0.058 0.311 0.593 0.476 Components P 0.167 0.946 0.449 0.242 0.046 0.947 Time P 0.000 0.777 0.080 0.080 0.973 0.466 Forage*components P 0.002 0.900 0.629 0.932 0.707 0.907 Components*time P 0.949 0.988 0.405 0.777 0.962 0.928 Forage*time P 0.150 0.815 0.225 0.850 0.473 0.884 r2 0.82 0.47 0.82 0.69 0.73 0.52 Acet, acetate; Prop, propionate; Buty, n-butyrate; Total, total VFA. Other abbreviations see text. TABLE 4.9 – Total yield of volatile fatty acids from five grasses, assuming equal proportion of leaf, stem and flower in the DM after 6, 12, 24 and 48 hours of incubation, and rates of VFA production. Cumulative VFA (mMol/g DM) Rates (µMol/g DM per hour) 6 h 12 h 24 h 48 h 0-6 h 6-12 h 12-24 h 24-48 h Perennial ryegrass 0.83 1.16 1.57 1.01 138 55 34 -23 Tall Fescue 1.01 2.1 2.62 2.82 168 181 43 8 Yorkshire fog 0.46 0.82 1.34 1.8 77 60 43 19 Phalaris 0.77 1.45 1.92 1.57 128 68 39 15 Paspalum 0.67 1.05 1.03 1.45 112 63 -2 18 136 Perennial ryegrass 6.0 6.5 7.0 7.5 8.0 0 1 2 24 36 FIGURE 4.5 – pH during in vitro incubations of mature grass leaf, stem and flower. 48 Tall fescue 0 3 6 9 12 0 12 24 36 4 Incubation time (hours) m Mol NH 3/ L 8 Leaf Stem Flower Incubation time (hours) pH Leaf Stem Flower Yorkshire fog 6.0 6.5 7.0 7.5 8.0 0 1 2 24 36 48 Phalaris 5.5 6.0 6.5 7.0 7.5 8.0 0 1 2 24 36 4 pHpH 8 Leaf Stem FlowerLeaf Stem Flower Paspalum 6.0 6.5 7.0 7.5 0 1 2 24 36 4 Incubation time (hours) pH 8 Leaf Stem Flower 137 Perennial ryegrass 0 3 6 9 0 12 24 36 4 FIGURE 4.6 – Pattern of ammonia concen 48 hours expressed in mMol/L for in vitro incubations. tration over 8 Tall fescue 0 3 6 9 12 0 12 24 36 4 Incubation time (hours) m Mol NH 3/ L Leaf Stem Flower 8 Leaf Stem Flower Incubation time (hours) m Mol NH 3/ L Yorkshire fog 0 3 6 9 12 0 1 2 24 36 48 Phalaris 0 4 8 12 0 12 24 36 4 m Mol NH 3/ L m Mol NH 3/ L 8 Leaf Stem Flower Leaf Stem Flower Paspalum 0 4 8 12 0 12 24 36 4 Incubation time (hours) µMol NH 3/ L 8 Leaf Stem Flower 138 Perennial ryegrass -100 0 100 0 12 2 4 36 139 FIGURE 4.7 – Net ammonia production expressed in terms of plant nitrogen during in vitro incubations for mature grasses in five incubations runs. 48 Tall fescue -200 -100 0 100 0 1 2 24 36 Incubation time (hours) µMol NH 3/ m M pla nt N Incubation time (hours) µMol NH 3/ m M pla nt N Leaf Stem Flower 48 Leaf Stem Flower Yorkshire fog -200 -100 0 100 200 300 0 12 24 36 4 8 Phalaris -200 -100 0 100 200 0 1 2 24 36 µMol NH 3/ m M pla nt N µMol NH3 /m M pla nt N 48 Leaf Stem FlowerLeaf Stem Flower Paspalum -300 -150 0 150 300 0 12 24 36 Incubation time (hours) µMol NH 3/ m M pla nt N 48 Leaf Stem Flower -150 -75 0 75 150 0 12 24 36 4 Incubation time (hours) µMol NH 3/ m M pla nt N 8 Leaf Stem Flower FIGURE 4.8 – Net ammonia production expressed in terms of plant nitrogen averaged by components (leaf, stem and flower) for five mature grasses during in vitro incubations. 140 01 2 0 1 2 24 36 FIGURE 4.9 - Volatile fatty acids production expresses in terms of dry matter (mMol/g DM) and ratio acetate: propionate averaged by components (leaf, stem and flower) for five mature grasses during in vitro incubations. 48 0.0 0.3 0.5 0.8 0 1 2 24 36 m Mol pr opi on ate /g DM m Mol a ce tate /g DM Leaf Stem Flower 48 Leaf Stem Flower 0.0 0.3 0.5 0 1 2 24 36 48 0.0 0.1 0.3 0 1 2 24 36 m Mol m in or s VFA /g DM m Mol n -butyra te/ g DM Leaf Stem Flower 48 Leaf Stem Flower 0 1 2 3 0 1 2 24 36 48 0 4 7 0 1 2 24 36 Incubation time (hours) Rati o ac eta te: pro pio na te Incubation time (hours) m Mol to tal VFA /g DM Leaf Stem Flower 48 Leaf Stem Flower 141 4.5 - Discussion The chemical composition and digestion data from the contrasting species, and their components are intended to complement the digestion kinetic data for ryegrass presented in Chapter 3. Separation into leaf, stem and flower constituents provides information on individual fractions of mature grass, excluding dead material. The decline in nutritive value of mature grass is well known (Wilson, 1993; Wilson et al., 1995), and will be a consequence of both changing proportions of leaf, stem and inflorescence and can alter composition of these components. Data presented here confirm the slow degradation of stem, but also suggest a relatively low nutritive value of the flower, despite moderate concentrations of non-structural carbohydrates in perennial ryegrass and tall fescue. In contrast, leaf from ryegrass and Yorkshire fog had high DM degradation rates (k = 0.09 and 0.11) which were similar to values reported by Burke et al. (2000) for respective species (k = 0.114 and 0.092). The chemical composition of young and old ryegrass and fog leaves was also similar, whereas tall fescue and paspalum leaves contained substantially more NDF and less soluble carbohydrate when mature (Table 4.1 and Burke et al., 2000). Reductions in nutritive value appear to be a consequence of decreasing proportion of leaf and increasing proportion of stem and flower with less effect due to composition of each component. Stockdale (1999b) reported chemical composition of leaf, stem and flower from ryegrass and paspalum grown under irrigation in Northern Victoria (Australia; Table 4.10), which were similar to values in this study. Stem and flower had a very different chemical composition to leaf. Under irrigation, the proportion of ryegrass and paspalum components differed throughout the year, but the nutritive characteristics of leaf, stem and flower of both species remained relatively constant. The proportion of stem in paspalum pasture reached 35% of the DM in summer and flower accounted for a further 10% (Stockdale, 1999b). 142 TABLE 4.10 - Chemical composition (g/100 g of the DM; mean ± standard error) and estimated metabolisable energy (ME; MJ/kg DM) content of leaf, stem and flower fractions of ryegrass and paspalum irrigated pastures in summer-autumn in northern Victoria (Australia). CP NDF ADF ME Ryegrass leaf 14.3±1.09 57.2±0.71 30.9±0.64 10.6±0.24 stem 7.4±0.39 64.7±0.55 30.8±1.11 9.9±0.30 flower 11.1±0.65 59.9 34.2 8.8±0.20 dead matter 12.0±0.65 66.4±0.88 43.5±0.72 7.1±0.17 Paspalum leaf 13.8±0.72 64.0±0.56 36.8±0.38 8.6±0.12 stem 5.8±0.57 68.2±0.95 37.1±0.49 9.7±0.18 flower 8.4±0.38 73.0±1.29 38.2±1.36 7.6±0.23 dead matter 10.6±0.43 69.0±0.73 45.8±0.44 5.9±0.13 From Stockdale (1999b). The changing proportions of ryegrass leaf, stem and flower, and their nutritive value was reported by Wilman and Agiegba (1982; Table 2.6). Stem DM increased to over 60% of the DM, whilst green leaf declined to very low levels. In vitro digestibility of leaf declined by a small amount as grass matured (64 to 58%) whereas stem was 49% digestible when mature. As grasses flower and mature, there is a decline in forage quality caused by the translocation of soluble carbohydrates from stem and leaves to the flowers, and an increased lignification of cell walls (Hacker and Minson, 1981). Nitrogen concentration in leaves decline with maturity, but leaf always contains a higher protein concentration and less NDF (Tables 4.1 and 4.10) than stems and flowers. In general, leaf was more rapidly digested than stem and flower fractions, but with phalaris the difference among all three constituents was relatively minor. In vitro incubations revealed insufficient nitrogen in paspalum and fescue leaves (and stem and inflorescence of all species) for sustained microbial growth, as evidenced by the very low ammonia concentration following 12 hours of incubations. Volatile fatty acid production was very low for all grasses except for tall fescue, despite the apparently inadequate supply of NH3-N for microbial growth (Satter and Slyter, 1974). The slow rate of VFA production with ryegrass was particularly surprising given that adequate N was present in leaf and flower. Incubations of young ryegrass leaf have yielded 36% (Barrell, 2000) and 27% (Chapter 3) of DM as VFA after 24 hours. There was a poor relationship between feed quality characteristics and in vitro production of VFA in this study and reasons for this are not understood. 143 The principal factor affecting nutritive value of mature forages is the slow rate of physical degradation and clearance from the rumen of animals unable to select leafy material from the sward (Cherney et al. , 1993; Waghorn, 2002). A requirement to eat stem and flower is likely to restrict feed intake, and intake will be further affected by the extent to which animals chew and reduce particle size of the forage during eating and rumination. These factors were overcome by the mincing procedure to achieve a particle size distribution (Table 4.2) similar to that of rumen content (Table 3.20) but substantially more effort is required to chew mature versus immature forages, especially stem fractions. This was the main reason for the use of a slow (0.02 h-1) outflow rate typical of a poor quality diet to calculate effective degradability (E) of fibre (Table 4.5 and 4.6), for comparison with higher outflow rates (0.06 h-1) used in previous analyses (Chapter 3). When the outflow rate is 0.06 h-1 the effective degradability for leaf fibre reached about 52% but a longer residence in the rumen (i.e. slower passage rate: 0.02 h-1) improves the degradability to about 70%. A rapid particle size reduction will facilitate both digestion and clearance of residues from the rumen. 4.6 - Conclusion This assessment of grasses and their components showed stem and flower to have slower rates of digestion than leaf, with higher proportions of indigestible fibre. In sacco kinetics suggested slow colonisation of all components, but especially stem and this will limit the rate of nutrient production as well as voluntary intakes. Effective degradability was highest for leaf but rates in vivo will depend on the speed and extent of particle size reduction by chewing during eating and ruminating. High quality pastures are achieved by good pasture management that prevents flowering by removing the apical meristem from reproductive tillers using a rapid stock rotation or topping (Korte, 1982). 144 Chapter 5 Supplementing fresh pasture with maize, lotus, sulla and pasture silages for dairy cows in summer. 1 1 A small portion of these data were previously published in the Proceedings of New Zealand Grassland Association , 2002, 85-8 9. 5.1 - Abstract Forages suitable for supplementing pasture- fed dairy cows over summer should provide adequate nutrients and increase milk yield and liveweight above that produced by cows grazing ryegrass/white clover pasture. A trial was conducted in January – February 2001 to compare benefits obtained from feeding four types of silage. There were two silages that contained condensed tannins (CT): lotus ( Lotus corniculatus ) and sulla ( Hedysarum coronarium ), maize silage or traditional ryegrass dominant pasture silage, all fed at 5 kg dry matter (DM)/cow.day with restricted pasture (RP). Cows on the RP (control) treatment and those fed the silage treatments were offered an allowance of 25 kg pasture DM/cow.day, while the full pasture (FP) cows were offered 50 kg pasture DM/cow.day. Silage supple mentation, regardless of silage type, increased both DM intake and milk yield compared with cows given RP only. Cows on the lotus silage supplement and the FP treatment had significantly higher milk production than the other silage supplemented cows, all of which had similar milk yields. For cows given lotus silage, the high milk yield was probably due to a combination of the higher nutritive value of the silage and to the action of CT, because the total DM intake of cows fed the lotus silage was the same as that of cows given the pasture and maize silages. The high milk yield of the FP treatment was mainly a result of the cows having a higher intake of pasture than cows on all the other treatments. This study demonstrated the potential benefit of silage supplementation, particularly with lotus silage, for increased milksolids yield in summer when low pasture growth rates and quality may otherwise limit production. Keywords: condensed tannins; dairy cows; grazing; pasture; silages. Short title: Silages for grazing cows in summer. 5.2 - Introduction In New Zealand most dairy farming is in regions with sufficient rainfall to maintain pasture growth but summers often have pe riods without adequate precipitation and pasture growth is reduced (McGrath et al. , 1998). Pasture quality can also be reduced by dry conditions after flowering, with re duced digestibility, lower protein content (Figures 2.5 and 2.6) and an increased proportion of dead matter (Gray and Lockhart, 1996). Dead matter can result in fungi growth with toxicity problems in some situations. Summer feeding often results in harder grazing, with lower residual pasture dry matter and cows may be forced to consume stems and leaf sheaths. Ryegrass endophyte 1 46 ( Neotyphodium lollii ) grows in the reproductive stem and leaf sheath, with peak toxicity during seed head emergence and during dry summers in response to water stress and high temperatures (Easton, 1999). Supplements are usually provided to make up a shortfall in pasture supply, but they may also improve the nutritive value of pasture on offer. Supplements should be chosen to complement the pasture and to meet nutritional demands of lactating cows, but frequently maize and grass silage are chosen because it is available. In some instances supplements are given to reduce intakes of ryegrass infected with Neotyphodium lollii to avoid toxicity which can lead to an early end of lactation (Hamilton-Manns and Crothers, 1999). The ty pe and extent of supplementation should be determined by the specific or dominant constraint to feeding and dairy production. Trials undertaken here have required pasture supply to be constrained to mimic normal situations in many dairy regions. The experimental plan anticipated a dry summer (typical of the Waikato region) so legume silages were chosen for supplementation, as well as maize and pasture silages which are frequently used by farmers. These treatments were intended to alter diet composition and to provide sufficient feed to meet cow requirements. Issues associated with endophytes and/or other toxins were not addressed. The objective of this work was to evaluate contrasting silages as supplements for cows fed restricted amounts of summer pasture. The use of lotus ( Lotus corniculatus) and sulla ( Hedysarum coronarium) silage was based on a need to provide adequate protein to lactating dairy cows, especially as summer pastures usually contain less than 16% CP in the DM (Moller et al. , 1996; Wilson et al. , 1995). The CT in these forages protects protein from degradation during ensiling and may reduce proteolysis in the rumen (Niezen et al. , 1998 ). The hypothesis from this chapter was th at provision of about 35% of ME from silages with contrasting chemical and botanical composition to cows grazing a restricted pasture allowance would result in a range of milk solids responses in a short- term trial. Data obtained from this trial, with kineti c information, were used as inputs to the Cornell Net Carbohydrate and Protein System (CNCPS) model to determine the first limiting nutrient and provide information concerning rumen digestion parameters. 1 47 5.3 – Material and methods This experiment was carried out at Dexcel’s No. 5 dairy farm at Hamilton (Waikato region) and involved 60 cows in mid lactation. The trial comprised a seven day uniformity period when all cows were grazing a ryegrass/white clover pasture, followed by a treatment period when either pasture or pasture with supplements were offered for four weeks commencing 15 th January 2001. The 60 Friesian cows were allocated to six treatments and balanced for milksolids yield and liveweight measured during the uniformity period. 5 .3.1 - Treatments During the uniformity period cows we re grazed together with a pasture allowance of about 40 kg DM/day. Milk yi eld was measured on two consecutive days and liveweight on three mornings of the uniformity week. These data enabled allocation to treatments on the basis of performance and bodyweight and also a covariance analysis of treatment effects. There were six dietary treatments, enabling the effects of either lotus or sulla silage to be compared with the more typical pasture or maize silages. One group of 10 cows was given a high ( ad libitum ) pasture allowance (full pasture) and another was given a pasture allowance similar to cows receiving silage supplements (restricted pasture): 1. Pasture only – full allowance (50 kg DM/cow.day) 2. Pasture only – restricted allowance (25 kg DM/cow.day) 3. Pasture (restricted) + pasture silage (5 kg DM/cow.day) 4. Pasture (restricted) + maize silage (5 kg DM/cow.day) 5. Pasture (restricted) + lotus silage (5 kg DM/cow.day) 6. Pasture (restricted) + sulla silage (5 kg DM/cow.day) Pasture restricted and full allowance treatments enable comparisons between this trial and previous work (Harris et al. , 1998a), and calculation of substitution rates when silages were offered. Cows were grazed throughout the treatment periods in treatment groups of 10 animals but each group was divided into two groups for measurement on Tuesday, Wednesday and Thursday of each week. On these three days each herd of 10 cows were split into two herds of five (same cows in each group 1 48 each week) when milk and pasture measur ements and samples were collected in order to replicate the treatments. On the remaining four days each week (Friday, Saturday, Sunday, Monday) the replicate groups were combined into treatment groups (six groups of 10 cows) for ease of management. Cows in each treatment group were grazing similar pastures in adjacent plots (Figure 5.1). FIGURE 5.1 – Overall view of the random six treatments in separated paddocks. Cow management Each treatment group was given a new break of pasture on a daily basis with a back-fence using electric tape. Daily pasture allowances for each treatment group were estimated using visual assessment of pre-grazing herbage mass and allocation of break size accordingly. Silages were fed to cows on a group basis from portable feed troughs (one trough per five cows; Figure 5.2) once cows returned to the paddock after morning milking. Silage DM was determined by quick drying (microwave) confirmed by drying for 24 hours at 100 oC, and sufficient placed in troughs to provide five kg supplement DM/cow.day. Troughs were colle cted from paddocks when cows were at afternoon milking and any re fusal weighed and sub-sampled for DM determination. Water was available ad libitum . 1 49 FIGURE 5.2 –Trough used to feed silage supplements (in detail: lotus silage). 5 .3.2 - Measurements Pasture intakes : pasture intakes of each treatment group were estimated by using a rising plate meter to measure pre- and post-grazing herbage mass (50 measures per 24 hour break for each group). This was done three times per week during the measurement period for each group of 10 cows to coincide with milk sampling days. Quadrants of pasture were cut one day of each week pre- and post-grazing (on representative pasture) to calibr ate the rising plate meter (Hodgson et al., 1999 ). Silage intakes : silage intakes of each group were measured by weighing silage pre- and post-feeding. Pre-feeding weighing was done every day, but post-feeding weighing of refusals was done only during the measurement period to coincide with days on which pasture intakes were estimated. For all silages except sulla, refusals were very small, but cows did reject sulla stem. Pasture quality : samples were collected for analyses of pasture quality using an electric clipper. Pre-grazing pasture samples were cut at 5 cm above ground level to replicate grazing height from each treatment break on each day during the measurement period (15 samples/break). Samples were bulked over the three days to provide one sample for each group of five cows per week. Pasture was sub-sampled and dried to determine dry matter content (10 0 oC; 24 hours) and for NIRS analyses (60 oC; Corson et al., 1999). Silage quality : sub-samples of the silages offered were taken on measurement days and bulked to provide one sample per we ek. Silage refusals were sub-sampled to 1 50 determined DM refused. An additional sub- sample of lotus and sulla silages was freeze dried and the CT concentration measured using the butanol-HCl colorimetric procedure (Terrill et al. , 1992 ). Milk yield and composition : Milk yield and composition (fat and protein concentration) were determined on the three measurement days (Tuesday PM – Friday AM) using an automated milk sampler and infr ared milk analyser (Milkoscan 133B, Foss electric, HillerØd, Denmark). Milk yield and composition were also measured on two days during the uniformity period and milk soli ds yields were used in allocation of cows to treatments and for covariate analyses to determine treatment effects. Liveweight : Liveweight (LW) was measured immediately before each AM milking for three consecutive days during each measurement period and during the uniformity period. Liveweight was also used to allocate cows to treatments and for covariate analyses. 5 .3.4 - Statistical analysis Two replicate groups of five cows on each of the six treatments provided 12 groups in total and enabled a statistical evaluation of treatments effects. Data were analysed using PROC GLM (SAS, 2001) with group to group variation used as the error term. The uniformity data collected before th e feeding trial started was also used as a covariate for analysis of the milk paramete rs. See Appendix CD (Chapter 5) for a complete data set of results and SAS procedures used in the analyses. 5.4 - Results The conditions which are typical of the Waikato region in summer did occur in January of the year when this trial was carried out so the quality of pasture was very representative of a large proportion of New Zealand dairy farms that manage pastures well during summer (Table 5.1). However, pasture quality declined over the experimental period, especially during the last week of the measurements (Figure 5.3). The quality of the pasture is indicated by the concentration of CP in the pasture offered to cows in each treatment group (17.8 – 19.0 g CP/100 g DM; Table 5.1). Concentrations of fibre were moderate for ryegrass and similar to ensiled pasture and the average predicted OMD for pasture exceeded 70%, which was adequate for dairy cows. 1 51 5 .4 .1 – Pasture and silage composition The pasture composition was constant for the first three weeks of the treatment period, but dry conditions resulted in a decline in quality during week 4 (Figure 5.3). The pasture DM increased from 21.5 to 27.0 g DM/100 g material over the trial period, with a decline in CP (19.0 to 16.0 g CP/100 g DM) and increasing NDF concentration to 49.2 g NDF/100 g DM (Figure 5.3). The pasture quality in week 4 was reasonable but more typical of summer dairy pastures in New Zealand. The chemical composition of silages (Table 5.2) demonstrates a wide variation in quality. Maize silage had a very low CP and high NSC concentration, in contrast with sulla silage that contained low-medium CP and CT concentrations (14.5 and 1.6 g/100 g of DM) compared to the lotus silage (23.4 and 3.4 g/100 g of DM). In this trial, the lotus silage was of excellent quality (odour, colour, nutritive value (NV)) where as the sulla silage was stemmy (50.8g NDF/100 g of DM) and had lower ME content and estimated DM digestibility. The sulla was of poor nutritional quality compared to the other silages and about 25% was rejected (s tem) by the cows (Table 5.3). This happened because sulla was past optimum maturity when harvested for ensiling and the contractor did not have machinery for chopping the sulla prior to wrapping. The low proportion of sulla stems consum ed by the cattle resulted in a lower daily intake of sulla than other silages but the leaf fraction that was eaten had a higher quality than the silage on offer. Data from a separate trial showed the composition of ensiled sulla stem (g/100 g DM) was about 75 NDF, 52 ADF, 8 CP with a ME content of 8.6 MJ/kg DM. When components of the stem (rejected by cows) were subtracted from the material on offer, the sulla DM eaten comprised (g/100 g DM) 17 CP, 41 NDF and 36 ADF. Estimated ME content was 10.8 MJ/kg DM. These data were used to calculate the composition of the diet eaten by cows given restricted pasture with sulla silage (Table 5.4). Pre-grazing pasture mass for all treatments was about 3600 kg DM/ha and the larger area offered to cows given the full pasture treatment resulted in higher post- grazing herbage mass (2150 kg DM/ha) than these given the other treatments. Pasture residual DM was similar for cows fed silages (1920 ± 560 kg DM/ha) and averaged 1790 kg DM/ha for cows given restricted pasture as a sole diet. The lower residual DM for cows given restricted pasture compared to full pasture resulted in a higher proportion of stem being eaten and a lower diet quality. 1 52 5 .4 .2 – Cow performance Silage supplementation, regardless of type, increased both DMI and milk production when compared with the restricted pasture treatment (Table 5.3). Cows fed the lotus silage and full pasture treatments had significantly higher milk and milksolids (MS) yields than the other silage supplemented cows (P < 0.001). Cows fed pasture, maize and sulla silage supplements all had similar milk yields (Table 5.3) although intake of sulla silage was significantly less than other three. The higher milk yield of the full pasture treatment compared with the restricted pasture treatment was associated with a higher DMI (Table 5.3) due to the high pasture allowance and the quality of the diet eaten by cows given a full pasture allowance would be substantially better than for cows fed restricted pasture which were forced to eat a higher proportion of feed on offer. However the difference in MS yield (0.29 kg/cow.day) was small compared to the 6 kg difference in DMI of the two groups fed pasture. Cows given the lotus silage produced more milk than those receiving the other silages (Table 5.3) probably in response to an improved diet quality. DM intakes were similar for lotus, pasture and maize si lages (16.6 – 17.2 kg/cow; Table 5.3). Feeding value (intake x NV) of silages did not affect milk fat concentrations in any of the treatments over the four weeks of the trial. However, cows fed the lotus silage had a slightly higher milk protein co ncentration (P < 0.05) than those in other treatments over the entire trial period (Table 5.3). The lotus silage contained 23 g CP/100 g DM and the 3.4 g CT/100 g DM wo uld limit protein degradation and improve amino acid supply for absorption. Table 5.4 summarises the feeding value of the diets eaten by cows given the six treatments. The lotus silage showed the lowest dietary NDF among all treatments. The maize silage supplementation in this experi ment was sufficient to increase dietary NSC levels compared to pasture, but concentrations were well below optimum level of 30 to 40 percent of the ration DM (Nocek, 1997). Substitution rate (SR) defines the extent to which a supplement replaces pasture in the diet and is expressed as the decrease in pasture intake/kg of supplement fed. Estimation of SR requires a control or un-supplemented treatment (restricted pasture). SR (kg pasture/kg silage) is calculated as: (p asture DMI in restricted group – pasture DMI in supplemented group)/silage DMI. A zero value means that pasture intake remains the same, while a value of 1.0 means that the supplement completely replaces the 1 53 pasture. Low pasture SR must be achieved if feeding silage is to have an additive effect with pasture to increase milk production. The pasture DMI of cows receiving silage supplements was similar to the pasture DMI of cows fed the restricted pasture (control) treatment, and all SR values were below 0.15. As a result, the silage supplements reduced pasture DMI only slightly, particularly for cows in the lotus silage treatment which had the lowest SR (Table 5.3). This result suggests pasture availability was limiting cow intake and although a pasture allowance of 40 kg DM/cow.day would avoid this limitation (Table 2.2; Figure 2.4) pasture shortages do occur in many summer situations. The cows weighed 531 kg with a range of group means from 512 to 545 kg. Although statistical analysis did not show significant live weight changes between treatment (P = 0.06; Table 5.3), higher LW gains were obtained for cows receiving a high pasture allowance and those fed with pasture, lotus and sulla silages compared with cows fed low pasture allowance or maize silage. TABLE 5.1 – Chemical composition of pasture dry matter (DM; g/100 g) cut to 5 cm for each dietary treatment averaged over the four weeks of grazing. Cows feeding treatments FP RP PS M LS SS P LSD a r2 SE DM 23.6 23.4 24.9 25.0 23.2 23.0 0.2604 2.3 0.91 0.45 CP 18.8 18.5 17.8 18.0 18.6 19.0 0.0011 0.4 0.88 0.30 NSC 10.1 9.7 10.5 10.4 9.9 10.4 0.1935 0.8 0.95 0.12 Lipid 3.9 3.8 3.8 3.7 3.8 3.9 0.1100 0.1 0.75 0.04 NDF 46.4 47.3 46.8 47.4 45.8 45.5 0.1223 1.6 0.89 0.46 ADF 25.3 25.9 25.7 25.8 25.5 24.8 0.1154 0.8 0.84 0.27 Ash 8.9 9.1 8.7 8.6 8.9 8.9 0.1186 0.3 0.77 0.10 OMD 71.0 70.3 70.5 70.2 70.5 72.2 0.2593 1.9 0.92 0.40 ME b 10.1 10.0 10.1 10.1 10.1 10.3 0.4786 0.3 0.94 0.05 Abbreviations: FP, full pasture; RP, restricted pasture; PS, pasture silage; M, maize silage; LS, lotus silage; SS, sulla silage. Others see text and tables in previous chapter. P: P values assessing diffe rence between treatments. a Least significant difference alpha 0.05. SE = standard error. b ME, MJ ME/kg DM. 1 54 TABLE 5.2 –Average dry matter (DM) content and nutritional composition of the feeds offered to cows in the six treatment groups. Units are all g/100g DM unless stated otherwise. Full Restricted Pasture Maize Lotus Sulla pasture pasture silage silage silage silage DM 23.6 23.4 32.6 33.7 33.1 35.4 CP 18.8 18.5 15.6 6.9 23.4 14.5 NSC 10.1 9.7 5.9 31.6 2.6 3.6 NDF 46.4 47.3 46.8 44.5 35.5 50.8 ADF 25.3 25.9 31.1 27.0 27.5 40.6 OMD 71 70.3 71 NA 69.2 63.4 ME (MJ/kg DM) 10.1 10.0 11.4 10.4 10.9 10.2 CT 3.4 1.6 Abbreviations see text and tabl es in previous chapter. NA = not applicable. TABLE 5.3 – Daily pasture and silage allowances, dry matter intakes (DMI), substitution rate (SR), milk production and live weight (L W, in kg) change over 4 weeks trial for cows given six dietary treatments. Units are all kg/cow/day unless stated otherwise. Cow feeding treatments FP RP Pasture Maize Lotus Sulla LSD P< Silage Silage silage silage Pasture allowance 50 25 25 25 25 25 Silage allowance 0 0 5 5 5 5 Total DMI 18.5 12.5 17.0 16.6 17.2 15.7 0.89 0.001 Total DMI (%L W) 3.4 2.4 3.2 3.2 3.2 3.0 Pasture DMI 18.5 12.5 12.0 11.8 12.2 12.1 0.93 0.001 Silage DMI 5.0 4.8 5.0 3.6 0.34 0.001 Milk fat, % 4.29 4.48 4.20 4.29 4.12 4.19 ns Milk protein, % 3.29 3.23 3.24 3.21 3.37 3.23 0.05 Milk 17.0 13.1 15.0 15.0 17.2 15.1 0.54 0.001 Milk solids 1.29 1.00 1.11 1.12 1.29 1.10 0.07 0.001 L W change 9.5 1.6 4.8 0.7 5.8 8.2 7.50 0.06 LW 542 512 534 524 545 529 7.83 0.001 SR 0.10 0.15 0.06 0.11 0.09 0.001 Abbreviations see text and previous tables. 1 55 TABLE 5.4 – Composition of the diets eaten by cows. Units are all g/100g DM unless stated otherwise. Cow feeding treatments Full Restricted Pasture Maize Lotus Sulla pasture pasture silage silage silage silage Dietary CP 18.0 17.4 16.9 14.4 19.2 18.9 Dietary NSC 10.2 9.7 8.6 16.0 7.6 8.3 Dietary NDF 45.3 46.5 46.6 45.9 43.3 45.9 Dietary ME (MJ/kg DM) 10.1 10.0 10.4 10.1 10.3 10.2 Abbreviations see text and previous tables. 1 56 cb c b a 20 24 28 g DM/10 0g m ate ri al b a b c 16 18 20 g CP / 100 g DM a b bc c 8 10 12 g NS C / 100 g DM b b a a 3.0 3.5 4.0 g Lipid / 100 g DM a b b c 40 43 45 48 50 g NDF/10 0g DM a b b b 24 26 28 g A DF/10 0g DM b ab a b 66 70 74 15-Jan 22-Jan 29-Jan 5-Feb Summer effects g OMD/10 0g DM b a a a 9.4 9.8 10.2 10.6 15-Jan 22-Jan 29-Jan 5-Feb Summer effects MJ ME/ kg DM FIGURE 5.3 - Changes in chemical composition of pasture cut to 5 cm over four weeks of the grazing trial. Data are averaged across all treatment groups. a, b, c Least square means with different superscripts differ (P < 0.05) within plots. 1 57 5.5 - Discussion The principal result from silage suppl ementation of restricted pasture was increased DM intakes with minimal substitution. The lotus silage provided a much greater increase in milk production compared to maize, pasture or sulla silages. Silages accounted for only 23 – 30% of dietary DM, so effects on composition of the whole diet were minor. The principal re sponse to silage supplementation supports the suggestion of Penno et al. (19 98 ) that energy was limiting milksolids production. These authors used concentrate based supplements offered to cows consuming a similar intake of pasture as cows in this experiment and reported an increase of about 220 g milksolids from 50 MJ supplemental ME given to cows in mid lactation and grazing summer pasture. The response to pasture, maize and sulla supplements fed in this experiment was about 110 g MS from about 57, 50 and 40 MJ ME supplied by the respective silages. Supplementation with lotus silage resulted in an additional 290 g milksolids from 54 MJ ME. Lotus appeared to complement pasture to improve both the nutritive value of the diet and cow intake. An outstanding feature of the lotus silage was its acceptability. The cows actually ran down th e race to access lotus silage, ate all that was on offer and ate as much pasture as other treatments. The high acceptability of lotus silage has been reported previously by Woodward et al. (2001 ) when cows ate 38% more than moderate quality pasture silage, both given as a sole diet. Fresh lotus is also highly acceptable to dairy cows with intakes 8 - 15 % higher than pasture (Harris et al. , 1998a; Woodward et al. , 1999 ). Pastures intakes and residual DM Although summer feed shortage is common in several dairy regions (Clark, 1995; Clark et al. , 1997a; Penno et al. , 1998 ; Shaw et al. , 1997 ; Thom et al. , 1998) we anticipated a higher intake of pasture by cows on the restricted pasture treatment compared with those given silage supplements. Comparison with published reports support an allowance of 25 kg DM/cow.day to achieve a moderate, but not excessive restriction on pasture availability (Auldist et al. , 1998; Harris et al., 1998b ; Shaw et al., 1997 ; Suksombat et al., 1994) and the DM residuals after grazing of 1787 ± 368 kg/ha for restricted pasture and 1916 ± 350 kg/ha fo r silage treatments support a moderate restriction. These values exceed target post-grazing herbage mass for summer ryegrass based pasture of 1500 – 1600 kg/DM suggested by Matthews et al. (1999). Comparable residual herbage masses after grazing a moderate herbage allowance (18 - 30 kg 1 58 DM/cow) have ranged between 1120 and 2200 kg DM/ha (Suksombat et al. , 1994 ; Thomson, 1996). Pasture quality and composition of residual DM will influence acceptability, but the high proportion of dead matter typical of dry summer pastures (Figure 2.1) was not evident in this study. Hence cows given restricted pasture could have consumed more DM but the balance between demand for nutrients and the effort required to harvest, or quality of pasture available has resulted in a low intake, low milk production and no change in live weight. The residues will have a higher proportion of stem than the pre- grazed sward and future studies need to measure the composition of residual DM, especially when cows respond to supplements and leave a substantial pasture residue. Butler and Hoogendoorn (1987) showed herbage intake of dairy cows was better related to leaf allowance than to green or total herbag e and suggested that differences in performance may be affected to a greater extent by the level of leaf mass and dead matter in pasture than by green grass stem. More recently Clark et al. (1999) and Thom et al. (1999 ) have demonstrated detrimental effects of toxins produced by fungal endophyte ( Neothphodium lolli ) on intake and milk production. Endophytic and other fungi are concentrated in the base of the sward and in dead forage (Easton, 1999) and will reduce intakes and lower productivity. These factors may have contributed to a reluctance by cows in all treatment groups given restricted pasture to consume more than 12.5 kg DM/day. Irrespective of the reason for pasture intakes of only 11.8 – 12.5 kg DM/day, provision of silages did enable increased intakes, to levels approaching that of cows given unrestricted pasture. Pasture intake was not limited by rumen fill or metabolic factors and either medium quality leafy pasture or chopped silages was able to increase DM intakes by about 35% even though NDF accounted for 43 – 47% of the dietary DM. The chopping of pasture, maize and lotus silages, together with high DM percentage may have enabled high intakes despite the concentration of fibre because a substantial reduction in particle size and cell damage had occurred during processing and ensiling. The lo w substitution of silages contrasts with values of 0.3 – 0.4 when similar quantities of turnips or sorghum were fed to cows under a similar feeding regime (Harris et al. , 1998b) and with 0.14 – 0.40 kg DM reduction in white clover pasture intake per kg of maize silage eaten at two levels of pasture allowance (Stockdale, 1996). 1 59 The total intakes of cows given unrestricted pasture or silage supplements ranged from 30 – 34 g DM/ kg liveweight, which exceeds predicted maxima of about 14.0 kg/day on the basis of NDF concentrations (Figure 5.4; Cherney and Mertens, 1998). FIGURE 5.4 – Illustration of neutral detergent fibre NDF-Energy Intake System for predicting intake for optimal and low-fibre rations. BW = body weight of animal. Source Cherney and Mertens (1998). Fibre based predictions of DM intake derived from Northern hemisphere models for cows fed TMR (Kolver et al., 2002) do not apply to New Zealand pasture based diets but the NRC (2001 ) dairy model did predict intakes on the basis of milk production and body weight with reasonable accuracy (Figure 5.5). DMI (kg/d) = (0.372 x FCM + 0.0968 x BW 0.75 ) x (1 - e (-0.192x ( WO L + 3.67 )) ) Where FCM = four percent fat corrected milk (kg/day), BW = body weight (kg), and WOL = week of lactation. Mean predictions of cow DMI for all six treatments based on the equation above were compared with actual values (T able 5.3) and are shown in Figure 5.5. The mean bias, that represents the average inaccuracy of predictions across treatments, was only 0.4 kg/day DMI. So using NRC (2001 ) equation could lead the reader to conclude that predictions are precise for all treatments with silage 1 60 supplements. However restricted pasture showed the highest residual (predicted – actual) for DMI of 3.3 kg/day, because this group had insufficient feed to meet requirements. Lotus silage Pasture silage Maize silage Full pasture Sulla silage Restricted pasture -1.0 0.5 2.0 3.5 15.5 16.0 16.5 17.0 17.5 18.0 Predicted DMI (kg/day) by NRC (2001) Res id ual s DMI (k g/ da y) F I G U R E 5.5 - Residuals (predicted – actual) versus predicted values of dry matter intake (DMI) from NRC (2001 ). Line ( _ _ _ ) indicates mean bias. Diet quality and cow performance Pasture offered to cows in this study was better quality (19 g CP/100 g DM, 71 g OMD/100 g DM and 10.1 MJ ME/kg DM) than typical summer pasture (Figure 2.1) which would contain 14 – 16 g CP/100 g DM with a DM digestibility of about 65 g OMD/100 g DM (Clark, 1995; Clark et al. , 1997b ; Simons et al. , 1998). The use of forage supplements limited opportunities for major changes in chemical composition of diets, but the maize silage resulted in a higher NSC and lower CP and a higher substitution rate than other diets. The lotus silage did increase CP and lowered NDF concentrations in the whole diet (Tables 5.3 and 5.4). These diets did affect cow performance, with maize silage resulting in the lowest milk protein concentr ation and no change in body weight whilst the lotus silage achieved the highest milk protein concentration (P < 0.05), the highest milk production and smallest increa se in liveweight over the trial. The lotus silage also contained condensed tannins which are able to protect dietary proteins from degradation in the rumen and increase amino acid flow to the intestine for absorption (Waghorn et al., 1997; Wang et al. , 1996 b ). Even though lotus silage accounted for 29% of DM intake, the CT is able to affect both lotus and pasture protein (Min et al. , 2003; Waghorn and Jones, 1989) and this would complement the effect of this high protein silage supplement. 1 61 Wang et al. (1996a) demonstrated a 21% increase in milk production and a 14% increase in milk protein production in the second half of lactation due to the effects of CT in Lotus corniculatus grazed by sheep. Harris et al. (1998a) fed a high allowance (60 kg DM/cow.day) of either medium-poor pasture (11 g CP; 53 – 56 g NDF/100 g DM) or Lotus corniculatus (22 g CP; 25 g NDF/100 g of DM) to cows and reported a 8 – 15% increase in DM intakes with lotus and a 42 – 54% increase in milksolids production. Later studies identified the contribution of CT in fresh lotus to milk production. Woodward et al. (2001 ) showed CT in lotus did not affect voluntary intakes but accounted for 42% of the increase in milk yield and 57% of the incr ease in milk protein percentage, relative to ryegrass pasture. Cows fed lotus ate 15% more DM than cow fed medium quality (53 g NDF/100 g DM) ryegrass. The impact of forage quality on intake and performance has been demonstrated for the lotus silage, with an additional 290 g milksolids/day compared to about 110 g MS/day for other silages. Howe ver, sulla silage quality was poor so the cows only ate 3.6 kg DM/day, and the re sponse to silage (g MS/kg silage DM consumed) was 28 g compared to 58 g for lotus, 22 for pasture and 28 g for maize silages. We believe the sulla silage has been under-valued in this experiment because it was of poor quality. Sulla is able to yield in excess of 18 t DM/ha.year (Rys et al. , 1988 ) compared to a maximum of about 13 t DM/ha.year for Lotus corniculatus and this biannual crop requires fu rther evaluation. Sulla frequently contains higher concentrations of CT than the 1.6 g CT/100 g DM reported here (Waghorn et al. , 1998), with good concentrations of CP and NSC. It ensiles very well (Niezen et al. , 1998) but techniques need to be developed to ha rvest and chop sulla on a farm scale. CNCPS diet evaluation The CNCPS model has equations for predicting nutrient requirements (first limiting nutrient: ME or MP), feed intake, and feed utilisation over wide variations in cattle (frame size, body condition and stage of growth), feed carbohydrate and protein fractions and their digestion and passage rates (Sniffen et al., 1992). However, little research has been conducted to evaluate the CNCPS predictions in dairy cattle consuming fresh forages (Kolver et al., 1996). Data from this experiment were used in the CNCPS model to provide an explanation for cow responses to the six dietary treatments. Inputs to the model included feed composition collected in this experiment; degradation rates for carbohydrate and protein from CNCPS feed li brary and Burke (unpublished) (Table 5.5) and cow parameters (Table 5.3). CNCPS li brary parameters enabled diet composition 1 62 to be generated (Table 5.6) and to predict milk production and live weight change. The model also generated estimates of microb ial growth, nitrogen kinetics and passage rate (Table 5.6). TABLE 5.5 – Individual feed composition and degradation rates used in the CNCPS evaluation of pasture and silage diets fed to cows in this experiment. General characteristics Individual feeds DM (%) NDF (%DM) peNDF (%NDF) Lignin (%NDF) Fat (%DM) Ash (%DM) Starch (%NFC) Full pasture 17.4 55.9 40.0 7.23 3.9 11.7 48.0 Restricted pasture 23.4 47.5 40.0 9.31 4.1 11.0 48.0 Pasture silage 32.6 46.8 95.0 5.50 2.6 7.2 63.0 Maize silage 33.7 44.5 85.0 10.59 3.0 4.0 80.0 Lotus silage 33.1 35.5 80.0 20.30 3.2 10.0 64.0 Sulla silage 35.4 50.8 92.0 25.60 5.2 10.1 64.0 Energy and protein values Individual feeds CP (%DM) UIP (%DM) Sol-P (%CP) NPN (%Sol-P) NDFIP (%CP) ADFIP (%CP) Full pasture 19.0 22.4 55.0 4.76 9.1 3.04 Restricted pasture 17.4 22.4 55.0 4.76 9.1 3.04 Pasture silage 15.6 22.4 50.0 100.0 31.0 10.0 Maize silage 6.9 20.9 58.0 100.0 16.0 7.0 Lotus silage 23.4 17.6 50.2 28.0 13.0 9.0 Sulla silage 14.5 38.4 54.0 28.0 15.0 10.0 Degradation rates Individual feeds Degradation rates (%.h -1 ) Carbohydrate Protein A B1 B2 B1 B2 B3 Full pasture 85.3 19.2 14.0 200 12.0 2.00 Restricted pasture 85.3 19.2 14.0 200 12.0 2.00 Pasture silage 10.0 25.0 4.0 200 9.0 1.75 Maize silage 10.0 30.0 5.0 300 15.0 0.25 Lotus silage 10.0 25.0 9.0 150 9. 0 1.25 Sulla silage 10.0 25.0 9.0 150 9.0 1.25 peNDF, physical effective NDF; UIP, undegradable intake protein; Sol-P, soluble protein; NPN, non-protein nitrogen; NDFIP, neutral detergent fibre insoluble protein; ADFIP, acid detergent fibre insoluble protein. Carbohydra te A fraction relates to degradation of sugars and organic acids, B1: starch and solu ble fibre and B2: available NDF; Protein B1 fraction: rapid degradable protein; B2: intermediately degraded protein and B3: slowly degraded protein. Other abbreviations see text and previous tables. 1 63 TABLE 5.6 – CNCPS predictions of nutrient comp osition, cow performance and rumen parameters of treatment diets. Feeding treatment FP RP Pasture Maize Lotus Sulla silage silage silage silage Diet nutrient composition ME, MJ/kg DM 9.67 10.24 9.95 10.03 9.84 9.42 CP, g/100g DM 19.0 17.4 16.9 14.4 19.1 16.7 Soluble CP, %CP 55.0 55.0 53.6 55.4 53.3 54.8 NDF, g/100 g DM 55.9 47.5 47.3 46.6 44.0 48.3 peNDF, g/100 g DM 22 19 26 24 22 25 Total NFC, g/100 g DM 11.2 21.6 24.8 27.7 24.3 21.6 Total fat, g/100 g DM 3.9 4.1 3.7 3.8 3.8 4.4 Performance predictions ME allowable milk, kg/day 20.5 11.0 19.3 19.2 19.2 15.1 MP allowable milk, kg/day 24.7 14.0 18.5 17.5 19.5 15.3 Daily weight change due to reserves, kg/day 0.5 -0.4 0.7 0.8 0.3 0.1 Rumen digestion, metabolism and passage MP from bacteria, g/day 1013 764 978 1038 953 873 MP from undeg. feed, g/day 980 447 615 485 752 567 MP from undeg. feed, %MP total 49 37 39 32 44 39 Total DIP, %CP 78.0 79.7 75.3 78.1 75.9 77.7 Ruminal N balance, % of req. 160 146 146 125 167 151 Total bacterial nitrogen, g/day 270 204 261 277 254 233 Urea cost, MJ/day 5.0 2.0 3.0 1.0 4.2 2.3 Urea cost, %ME intake 2.8% 1.6% 1.8% 0.6% 2.5% 1.6% Excess N excreted, g/day 221 101 150 94 190 124 Liquid passage rate, %.h -1 11.2 9.3 10.8 10.7 10.7 10.3 Pasture passage rate, %.h -1 6.51 5.54 6.46 6.42 6.45 6.20 Silage passage rate, %.h -1 NA NA 5.02 5.30 5.83 4.73 Predicted ruminal pH 6.37 6.23 6.46 6.46 6.34 6.46 NFC, non-fibrous carbohydrates. DIP, degradable intake protein. Undeg, undegradble. Other abbreviations see text and previous tables. Cows fed pasture and pasture with lotus or sulla silages had ME as the first limiting nutrient but MP was the first limiting nutrient with the pasture silage and maize silage treatments. The low CP in past ure and maize silages lowered dietary CP compared to the pasture alone, lotus silage and sulla silage treatments. Although the maize silage diet had the lowest CP%, it resulted in excellent microbial growth (277 g/day) due to high NFC content of the diet. 1 64 The CNCPS provides a system in which microbial protein production and undegraded feed protein values are predicted mechanistically, based on the integration of feed carbohydrate and protein fraction pool sizes, microbial growth on fibre and non fibre fractions, digestion and pa ssage rates. A goal in formulation with the CNCPS is to have at least 50% of th e total MP be of microbial origin (Fox et al., 2003 ). All diets fed in this trial achieved that goal. Maize silage supplementation resulted in the highest proportion of MP from bacteria (68%). All diets met requirements for ruminal microbial nitrogen with ruminal N balance ranging between 125 – 167% of requirements. Diets with higher CP concentrations (full pasture and lotus silage treatments) resulted in higher cost for disposal urea and higher excess N excreted than other treatments. These diets also resulted in the highest milk production (actual and predicted; Tables 5.3 and 5.6). Complementary effects (FP versus silages supplements) The value of forage supplements to complement pasture can be determined by comparing FP with silage supplement di ets in the CNCPS diets evaluations. FP provided more nutrients for greater milk production because co w grazing FP had a greater DMI compared to restricted pasture with silages. As a result ME and MP supplies were highest when cows were fed FP. The effect of tannins in lotus and su lla was able to be modelled by using degradation rates determined in sacco, however the total DIP content of the diet was not changed markedly by the tannin-containing supplements. The lotus silage diet had a similar CP and DIP content to FR treatment an d not surprising resulted in a similar loss of N as urea. Maize silage, which had lower CP content, had a lower level of urea excretion. The model predicted that pasture would have a lower passage rate when fed with pasture, maize, lotus and sulla silages. This would have contributed to an improved MP yield from undegraded feed with the supplemented diets, but this effect was small compared to the generally lower CP and ME contents, and DMI of diets containing silage supplements. If cows can be well fed on good quality pasture, similar to that used in this experiment, the complementary benefits of forage silages are minimal as nutrient supply and milk production would be higher on the pasture only diet. 1 65 Supplementary effects The value of forage supplements to su pplement pasture can be determined by comparing RP with the forage supplement diet s. Increasing DMI by supplementing with forages increased ME and MP supply, and mi lk production. This was a function of higher DMI, as the pasture generally contained more ME and CP than the silage diets. Lotus silage gave the greatest milk solids response (5.3 g MS/MJ ME) which would be expected for a cow restricted to 12.5 kg pasture DM intake and fed a high quality supplement (Penno, 2002). The responses of the pasture, maize and sulla silages were much lower than fro lotus (Table 5.7). TABLE 5.7 – Cow responses to supplements expressed in grams of milk solids per silage dry matter intake (g MS/kg silage DMI) and in grams of milk solids per metabolisable energy consumed (g MS/MJ ME). Response to supplement: Pasture Maize Lotus Sulla silage silage silage silage Grams of milk solids per kg silage DM intake 22 25 58 28 Grams of milk solids per MJ ME 1.9 2.4 5.3 2.7 Cow responses showed the silages gave p oor responses except for lotus silage. The greater response of lotus may have been due to an effect of tannin on pasture protein. However CNCPS predictions suggested MP was not first limiting nutrient for milk production on the lotus diet, and the model was unable to explain the response to lotus versus other silages. 5.6 - Conclusion The benefits of lotus silage for milk production demonstrated in this experiment support previous observations wi th fresh lotus fed to sheep or cattle as a sole diet. The principal constraint to use of lotus in dairying systems is its inability to compete with other forages under high fertility situations and moderate DM yield. Future work should focus on ruminal digestion of lotus and supply of nutrients for absorption. The sulla silage used in this trial was of poor quality, but the cr op has good potential to complement pasture because it is able to produce very high DM yields, with a low fibre and high readily fermentable carbohydrate concentration. The hypothesis was proven, with lotus silage providing a significant greater response in milk and milk solids yield co mpared to pasture, maize and sulla silage supplements. The responses appear to been due to the high CP content of the silage, and possibly the effects of CT on protein degradability. 1 66 Chapter 6 Supplementation of grazing dairy cows with sulla and maize silages in summer. 1 1 A small portion of these data were previously published in the Proceedings of New Zealand Grassland Association , 2002, 125-128. 6.1 - Abstract A trial was carried out at Dexcel in Hamilton to investigate the effects of silage supplementation of grazing dairy cows in summer. Forage mixtures used in the four week trial were based on previous trial results (Chapter 5) but inclusion of rumen fistulated cows in five treatments enabled rumen sampling and use of in sacco incubations to determine effects of diet on digestion kinetics. Sulla and maize silages were used to supplement pasture and to meet minimum requirements for dietary protein concentration. Five groups of ten cows were grazed on a restricted daily allowance of 18 kg dry matter (DM) pasture/cow to simulate a summer pasture deficit, and four groups received sulla silage (S) or maize silage (M) alone or in mixtures of 25M:15S or 15M:25S to make up 40% of total DM intake. A sixth group was given a relatively unrestricted (38 kg DM/cow.day) pasture allowance. The silage mixtures and pasture were incubated in sacco during the final week of the trial. The pasture was of high nutritive value and not typical of usual summer conditions, which favoured a response to quantity rather than quality of silage supplements. There were no differences in cow performance with the four silage supplements and the low milksolids (MS) production (about 1.0 kg/day) relative to full pasture (1.3 kg MS/day) showed the principal limitation to performance was feed quantity. Milk composition was not affected by silage type and the low level of pasture substitution (0.29) suggested ME was the principal limitation to performance. Samples of rumen liquor and in sacco data demonstrated significant effects of supplement. In sacco data showed highest DM degradation rates (k, h-1) when cows were fed 6 kg sulla silage (0.08) whereas diets with a high proportion of maize silage were degraded slowly (P < 0.01). Supplementation with sulla may increase digestion rate and rumen clearance and reduce the effect of fibre in ryegrass diets. Keywords: dairy cows; maize; pasture; silage; sulla. Shot title: Sulla and maize silages for grazing dairy cows. 6.2 - Introduction The previous cow trial (Chapter 5) which examined responses to maize, pasture, lotus and sulla silage supplementation showed clear benefits of lotus silage for milk production when pasture supply was restricted. Although maize silage did maintain milk production over the four week experimental period the cows did not gain weight (in contrast to other silages) and the dietary CP was less than cow requirements (NRC, 2001). The potential of sulla silage was not expressed because the long stalks and sub- 168 optimal quality resulted in substantial refusals. These silages provided a foundation for further evaluation, with more focus on achieving adequate dietary CP using fewer silages types. The poor cow response to maize silage (M), in contrast with lotus silage (Chapter 5; Woodward et al., 2002) emphasised the importance of meeting cow protein requirements especially as M has a very low protein concentration and is not suitable for feeding with low quality summer pasture. However when maize silage is fed in a balanced diet, it provides a low cost source of non-structural carbohydrates (NSC; mainly starch) which complements pasture for much of the year (Kolver et al., 2001). Sulla is of interest to us because it is a high yielding legume (Waghorn et al. , 1998) containing condensed tannins (CT) and high concentration of NSC which offers good potential for high quality silage production (Niezen et al., 1998). Balancing dietary protein deficiency by feeding sulla and improving readily fermentable carbohydrate intake with maize silage may optimise milksolids production from cows grazing poor- medium pasture in summer. This trial also anticipated a dry summer with restricted pasture availability, typical of the Waikato dairy environment. Maize and sulla silages were fed alone or as mixtures to account for about 40% of DM intake. The use of maize silage as a sole supplement enabled data from this trial to be compared with results in Chapter 5 and literature reports (Kolver et al. , 2001; Phillips, 1988; Stockdale, 1995). The hypothesis to be tested here was that mixtures of maize and sulla silages fed to cows grazing restricted pasture would provide a more balanced diet and improve milk solids production relative to either silage fed as a sole diet. Data obtained from this trial, with kinetic information, were used as inputs to the CNCPS model to determine the first limiting nutrient and provide information concerning rumen digestion parameters. 6.3 - Material and methods Sixty Friesian cows [15 primiparous and 45 multiparous including 10 with rumen fistulae; 483 kg liveweight; 14.3 kg milk /day; 156 days in milk] were allocated to six treatments and balanced for milksolids yield and liveweight. The overall design comprised a uniformity (covariance) period of one week, when all cows were grazed on pasture enabling their subsequent allocation to six groups to be fed the experimental diets for three weeks. Table 6.1 presents schedule of events during this experiment. 169 TABLE 6.1 - Schedule of events for uniformity and feeding trial from 24 December 2001 to 26 January 2002. Day Event Uniformity period 24th December 2001 to 6th January 2002 (Days 1 - 14) 8 - 10 Milk yield and liveweight measured on all 60 cows Diet adaptation 7th to 14th January 2002 (Days 15 - 22) 15 - 22 All cows allocated to treatments and grazed with appropriate silage supplements. Diet evaluation Week 3 – 14th to 21st January 2002 (Days 22 - 29) 22 - 29 All cows allocated to treatments and grazed with appropriate silages. 23 - 26 Liveweight, milk yield and composition measured over a 4- day period. 22 - 26 Pasture mass measured by rising plate meter, with silage intakes and sampling for dry matter and NIRS analyses. Rumen samples taken from fistulated cows to measure digesta rumen pH, ammonia, volatile fatty acid concentrations. 25 - 28 Sampling and preparing forages for in sacco incubation Week 4 – 21st to 26th January 2002 (Days 29 - 33) 30 - 33 Liveweight, milk yield and composition measured over a 4- day period. 29 - 33 Pasture mass measured by rising plate meter, with silage intakes and sampling for dry matter and NIRS analyses. Rumen samples taken from fistulated cows to measure digesta rumen pH, ammonia, volatile fatty acid concentrations. In sacco (72 hours) incubation of pasture and pasture – silage mixtures corresponding to diets fed to each of the 10 fistulated cows. 170 6 .3.1 - Treatments The six treatments enabled the effects of supplementing a pasture diet with either maize and/or sulla silages to be compared with un-supplemented pasture. Use of full and restricted pasture allowance allowed substitution rates to be calculated as well as impacts of pasture availability. The pasture allowance was intended to be 25 kg DM/cow.day, but calculations after sub-division of plots using a visual assessment of pasture mass in the last week of the trial resulted in less herbage DM mass on offer than expected. 1. FP: full pasture allowance (38 kg DM/cow.day) 2. RP: restricted pasture allowance (18 kg DM/cow.day) 3. PMS: pasture (restricted) + maize silage (4 kg) + sulla silage (2 kg) 4. PSM: pasture (restricted) + sulla silage (2 kg) + maize silage (4 kg) 5. PS: pasture (restricted) + sulla silage (6 kg) 6. PM: pasture (restricted) + maize silage (6 kg) 6 .3.1.1 - Feeding The dietary regime was similar to that used in the previous trial (Chapter 5) where pasture and milk yield measurements were made on Tuesday, Wednesday and Thursday of each week. On these three days each group of 10 cows were split into two groups of 5 cows (with the same cows in each group each week) in order to replicate the treatments. On the remaining four days of each week the replicate groups were combined into treatment groups (six groups of 10 cows). The full allowance (38 kg DM/cow.day) of pasture was intended to provide unrestricted feed, while the restricted pasture allowance of 18 kg DM/cow.day was intended to mimic summer conditions with feed shortages. Cows in each treatment group were given a new break of pasture once daily using electric fences, and water was always available. Daily pasture allowances for each treatment group were estimated by visual assessment of pre-grazing herbage mass and allocation of the appropriate area to be grazed. Silage was fed to four groups of cows in portable feed troughs (one trough per 5 cows), after the cows returned to the paddock following the morning milking. Silage offered to cows was based on a rapid DM determination of DM using a microwave oven, to provide 6 kg silage DM/cow.day. Troughs were removed 171 from paddocks when cows were at afternoon milking and refusals weighed. The dry matter contents of the silages offered and refused were determined by drying at 100oC for 24 hours (See procedure details in Chapter 5). 6 .3.2 - Measurements Pasture intakes were determined by measurement with a rising plate meter before and after grazing as described in section 5.3.2. Silage DM intakes were also determined so total daily DM intakes for each sub-group of cows were calculated over the three day measurement period in weeks three and four of the trial. Pasture quality : two types of samples were collected using an electric clipper, for analyses of pasture quality: 1. Pre-grazing pasture samples were cut to ground level from each of 12 breaks on each day during the measurement period. Samples from each break were bulked within weeks to provide one sample for each group of five cows per week to indicate the quality of feed on offer. 2. A second pre-grazing pasture sample was cut to estimated grazing height at about five cm above ground level from each break on each day during the measurement period. Samples were bulked to provide one sample for each group of five cows per week to indicate quality of pasture consumed. Both pasture sample types were sub-sampled to determine DM content (100oC; 24 hours) and for NIRS analyses (60oC; Corson et al., 1999). Silage quality : sub-samples of the silages offered were taken on measurement days and bulked to provide one sample per week. Sub-samples were taken of silage refusals over the same period and bulked for analyses. Silages and refusals were sub- sampled for dry matter measurement (100oC; 24 hours) and NIRS analysis. Composition of silage offered and refused enabled calculations of the dietary composition (nutrient x intake) for each constituent (DM, CP, lipid, fibre and ME). Dietary composition = ((kg offered * concentration constituent offered) – (kg refused * concentration constituent refused)) / kg intake. Liveweight (LW) was measured before milking on three mornings per week during weeks one, three and four of the trial. Milk yield was measured for each cow on three days per week, and sub-samples taken to measure fat, protein and lactose concentration as described in Chapter 5. 172 6 .3.3 – In sacco incubation and digestion kinetics Two rumen fistulated cows were included in each treatment except those given restricted pasture and enabled in sacco incubations of both pasture and of the dietary mixtures fed to each cow to be conducted. Pasture used in incubations was obtained by cutting pasture five centimetres above ground level before grazing, and frozen prior to chopping and mincing for in sacco incubations. The pasture was incubated in all cows as a single constituent and also combined with sulla and maize silages in similar proportions to the diet eaten by cows. Samples of maize and sulla silages were collected and frozen in proportion for mixing with chopped pasture prior to mincing. Mixtures used for in sacco incubations comprised about 60% pasture DM and 40% silage DM. The four silage mixtures (PM, PS, PMS and PSM) were only incubated in cows which were fed the same dietary mixture; i.e. the two cows fed pasture with maize silage were used to incubate pasture and maize silage; those fed PMS incubated PMS in sacco. Duplicate bags of dietary mixtures were removed at each time from all cows as well as duplicated bags of ryegrass pasture. The only exception were cows fed full pasture, where duplicate bags of pasture where removed at each sampling time from the two cows In sacco incubations for each diet were prepared by chopping pasture and sulla silage to 2 cm length, mixing frozen chopped pasture, sulla and maize silages as required, and mincing to resemble chewed forage as described in section 3.3.1. Samples were weighed into in sacco bags and also used to determine DM, chemical composition and distribution of particle size. A total of 28 bags (four per lingerie bag) were placed in the mid-ventral rumen of cows, and were removed at 2, 6, 9, 12, 24, 48 and 72 hours. Bags, including 0 hour samples which were not put in the cow, were washed, dried (60oC) and analysed to determine DM, CP, NDF and ADF content to calculate rates of disappearance during digestion. Procedures for in sacco calculations and data analyses are given in section 3.3.5. The disappearance of DM, NDF and ADF were analysed using a non-linear model described in Chapter 3. The effective degradability (E) was calculated from soluble and degradable pools and kinetic parameters, by fitting equations to in sacco data assuming a fractional passage rate (kp) of both 0.06 and 0.08 h-1. A kp value of 0.08 h-1 was used (AFRC, 1992) for comparison with high producing dairy cows (Hoffman et al., 1998; Kolver et al., 1998) even though intakes in this study were relatively low. 173 Disappearance of CP was calculated using a similar procedure but additional definition was applied in relation to the degradability of protein and CP content of the DM. The RDP and RUP values for the diets (percent of CP) were estimated from the model describing degradation and ruminal escape of feed proteins (NRC, 2001; Ørskov and McDonald, 1970) using the equations: RDP = A + B [k / (k + k DM)] RUP = B [k DM /(k + kDM)] + C Where k is fractional rate of protein degradation and DM passage (kDM), and C is the undegradable fraction. Equation for estimating KDM (rate of DM passage from the rumen, %/h) of wet forages = 3.054 + 0.614X 1 where X 1 = DMI, percentage of body weight (NRC, 2001). The metabolisable protein system (AFRC, 1992) for defining ruminal degradation was used to calculate protein degradability parameters as described in section 3.3.5.1. Effective rumen degradability of CP (ERDP, g/kg DM) was calculated as: ERDP (g/kg DM) = CP [(0.8 x A) + (B x k)/(k + k p)]. In this chapter, values for effective degradability of CP and ERDP were calculated using outflow rates of digesta DM from the rumen of 0.02, 0.05 or 0.08/h, which approximate to the outflow rates in dry cows at maintenance, cows producing less than 15 L milk/day, or cows fed a good quality diet that enabled a rapid passage through the rumen (Wales et al. , 1999a). Relationships between degradability, nutritive characteristics of the diet in sacco samples and ERDP were analysed by regression (SAS, 2001) using the model: ERDP (g/kg DM) = a + bX where X is the effective degradability of DM or NDF content of diet in sacco residues. 6 .3.4 – Rumen samples Rumen fluid was also collected 5 times during the first day of the in sacco incubation to check the pattern of rumen metabolite concentration over 12 hours from 07:00 h (pre-feeding) to 19:00 h. Rumen fluid samples were collected twice daily on the measurements days before morning and afternoon milking (Table 6.1). On each occasion, about 1 kg of rumen contents was taken from the mid-ventral rumen and strained though a cheese cloth to collect 100 mL of rumen fluid. Rumen pH was determined at collection 174 (PHM210, Radiometer Pacific Limited, Copenhagen) before samples were centrifuged and prepared to determine ammonia and VFA concentrations (sections 3.3.4.2 and 3.3.4.3). 6 .3.5 - Statistical analysis Procedures for statistical analyses of in sacco data are given in section 3.3.6.3. For each degradation parameter (A, B, k and E) fixed model effect tested differences between diets incubated, cow/diet effects upon pasture digestion kinetics and cow effects. A general linear model procedure of SAS (PROC GLM; SAS, 2001) was used for analysis of the milk parameters and liveweight change with uniformity period data as a covariate (section 5.3.4). Rumen ammonia and VFA concentrations were analysed using mixed model procedure of SAS (PROC MIXED; SAS, 2001) to calculate treatment means. Rumen pH analysis included day, treatment and treatment x day interaction. Pearson correlation coefficients were estimated using the CORR procedure of SAS. Effects were declared significant at P < 0.05 unless otherwise noted. See Appendix CD (Chapter 6) for a complete data set of results and SAS procedures used in these analyses. 6.4 - Results This trial had similar responses to those in the previous year (Chapter 5) because pasture quality in both studies were very representative of well-managed summer pasture in the Waikato, Taranaki and South Island in most years. The quality of pasture on offer was better (Table 6.2) than the previous year (Table 5.1) and prevented any benefits associated with silage quality to be expressed even though feed availability was very low. The CP concentration in pasture exceeded that in sulla silage, which had a higher NDF concentration than pasture. The main effect of silage supplementation was through provision of additional ME although the NSC content of maize silage did alter diet composition. 6 .4 .1 – Pasture and silage composition The chemical composition of pasture and silages are summarised in Table 6.2. Dry matter of pasture cut to 5 cm above ground level ranged from an average of 15.9 to 20.5 g DM/100 g across treatments. There were no significant changes in pasture quality over the four week duration of the trial, with average DM of 18.6 in week three and 17.9 g/100 g in week four. Crude protein was 21.1 and 21.4 g/100 g DM in weeks three and four with NDF contents of 47.3 and 43.3 g/100 g DM in the respective periods. 175 The mean values for the pastures offered to each treatment group were consistent, averaging 21.2 g CP/100 g DM and 45.3 g NDF/100 g of the DM (Table 6.2). Predicted organic matter digestibility exceeded 73 g/100 g for all pastures cut at 5 cm above ground level. Maize and sulla silages contrasted in chemical composition, with concentrations (g/100 g DM) of 6.6 and 15.7 (CP) and 39.1 and 50.0 (NDF) for maize and sulla silage respectively. The low NSC concentration in sulla silage (3.6 g/100 g DM) was due in part to the high lactic acid content. Sulla silage contained 3.5 g CT/ 100 g DM. Pasture cut to 5 cm left about 1600 kg residual DM per hectare. Residual DM was poor quality with about 71 g NDF/ 100 g DM. Fibre (NDF) and CP concentrations of the sward cut to ground level averaged 51.0 and 19.6 g/100 g DM respectively. TABLE 6.2 – Chemical composition of pasture dry matter (DM; g/100 g) cut to 5 cm above ground for each dietary treatment, and of maize and sulla silages averaged over the three week experimental period. Cows feeding treatments Maize Sulla FP RP PMS PSM PS PM silage silage DM 19.4 20.2 20.5 18.2 19.9 18.4 35.9 34.1 CP 21.7 21.5 21.0 21.7 21.0 20.6 6.6 15.7 NSC 9.2 9.1 8.5 8.9 9.4 9.0 41.4 3.6 Lipid 4.0 3.9 4.0 4.0 3.9 3.8 2.9 5.2 NDF 43.9 44.1 46.3 45.8 45.4 46.5 39.1 50.0 ADF 24.4 24.9 25.4 24.8 25.2 25.9 23.8 40.8 Ash 10.1 10.2 9.9 9.9 9.6 9.9 4.0 10.1 Lignin 4.0 4.2 3.7 4.0 4.0 4.3 3.4 8.0 OMD 75.6 74.9 74.1 74.8 73.6 74.6 NA 65.9 ME (MJ/kg DM) 10.7 10.6 10.5 10.6 10.5 10.6 10.8 10.5 pH 3.9 3.9 Lactic acid 2.3 11.4 Ammonia-N (% total N) 0.9 4.7 Total CT 3.5 Treatments: FP: full pasture; RP: restricted pasture; PMS (RP + 4 kg maize + 2 kg sulla silages/cow.day); PSM (RP + 4 kg sulla + 2 kg maize silages/cow.day); PS (RP + 6 kg sulla silage/cow.day); PM (RP + 6 kg maize silage/cow.day). Other abbreviations see text. 176 6 .4 .2 – Cow performance Maize, sulla and silage mixtures increased DMI but milk and milksolids (MS) production were similar to the cows given restricted pasture as a sole diet (Table 6.3). The main impact of silage was to maintain liveweight. There was insufficient feed available to maintain both milk production and liveweight (LW) of cows given the RP treatment. Cows fed the full pasture treatments had significantly higher milk and MS yields than the silage supplemented cows (P < 0.001) and LW of all cows except those given RP were similar at the beginning and end of the trial. Cows given the high pasture allowance ate 5.3 kg more pasture DM and produced additional 4.2 kg milk compared to RP cows, but the difference in milk production would have been greater if LW changes were similar for both groups. The RP treatment group lost 12 kg liveweight during the three week feeding period. Milk composition was not affected by treatment. Pasture intake by cows given the four supplements were about 1.6 kg DM/day lower than these given RP. On average 5.5 kg silage DM was eaten so the substitution rate was about 1.6/5.5 = 0.29 (Table 6.3). All supplements had a similar substitution rate (0.26 - 0.33) which showed provision of silages resulted in a substantial increase in feed intake when pasture allowance was 18 kg DM/cow.day. The pasture 5 cm above ground level was of unusually high quality (Table 6.2) but stubble below 5 cm had a higher fibre content. This level of substitution is similar compared to grazing cows were offered grains or hay (Stockdale, 1999a) with pasture allowance of 32 kg DM/cow.day. Also the SR in this trial was comparable with a level of substitution of maize silage for white clover pasture of 0.14 and 0.40 kg DM reduction in pasture intake per kg DM of silage eaten at pasture allowance of 19 and 39 kg DM/cow.day (Stockdale, 1996). 177 TABLE 6.3 – Daily pasture and silage allowances, dry matter intakes (DMI), substitution rate (SR), milk production and live weight (LW, in kg) change over 4 weeks trial for cows given six dietary treatments. Units are all kg/cow/day unless stated otherwise. LSD and significance are for treatments within rows. FP RP PMS PSM PS PM LSDa P Pasture allowance 38 18 18 18 18 18 Silage allowance 0 0 6 6 6 6 Total DMI 15.7 10.4 14.6 14.4 13.9 14.3 2.3 0.001 Pasture DMI 15.7 10.4 8.8 9.0 8.7 8.8 1.6 0.001 Silage DMI 5.8 5.4 5.2 5.5 0.3 0.001 Pasture utilisationb (%) 41 58 49 50 48 49 3.8 0.003 Milk lactose 0.83 0.64 0.71 0.66 0.66 0.67 0.05 0.002 Milk fat 0.73 0.57 0.56 0.56 0.54 0.55 0.06 0.001 Milk protein 0.57 0.42 0.46 0.44 0.44 0.46 0.02 0.001 Milk 17.2 13.2 14.3 13.7 13.7 13.7 1.00 0.001 Milk solids 1.30 0.99 1.02 1.00 0.98 1.01 0.07 0.001 LW change -3.8 -12.4 1.0 -4.5 -0.8 -1.4 3.4 0.06 LW, kg 497 471 480 485 475 489 3.5 0.001 Substitution rate 0.28 0.26 0.33 0.29 a Least significant difference (P < 0.05) b (kg DM pasture eaten/kg DM pasture offered) x 100. Abbreviations see text and Table 6.2. Pasture DM on offer averaged 3141 kg/ha for all treatments, with post grazing residuals of 1525 kg/ha for cows given a restricted allowance with silages and 1814 kg DM/ha for cows given a full pasture allowance. Cows given restricted pasture as a sole diet left 1298 kg residual DM. Pasture DM utilisation was 58% for RP and 49% across the four silage supplemented treatments (Table 6.3). Cows given full pasture allowance ate 41% of that on offer. Although all silages were acceptable, with refusals of 0.2 – 0.8 kg DM/cow.day, the preferred supplement comprised 4 kg maize silage and 2 kg sulla silage. The refusals from this treatment group comprised mainly sulla stem and proportions refused increased with the amount of sulla silage offered (about 0.6 kg with 4 kg sulla DM and 0.8 with 6 kg sulla DM; Table 6.3). PS treatment had the lowest proportion of sulla silage eaten compared to other supplemented treatments (P < 0.001; Table 6.3). The sulla was harvested about 6 weeks later than optimal because of difficulties obtaining services of a contractor. It was quite stalky as indicated by the high fibre content, but 178 was well ensiled and had a high concentration of lactic acid. Crude protein concentration (Table 6.2) may have been higher with an earlier harvest. There were significant differences between diets in the concentration of most constituents (Table 6.4). Provision of silages increased the dietary DM concentration (P = 0.005). The supplementary silages were intended to meet cow requirements for crude protein (CP), and the values in Table 6.4 show this was achieved with the exception of the PM diet which was only slightly less than a desirable 16 - 17% CP in DM for cows in mid lactation (NRC, 2001). The non-structural carbohydrates were highest when maize silages were used reaching 22% of dietary DM but were lowest with sulla. Dietary lipid averaged 4.0 g/100 g DM for all treatments and ash was lowest for cows fed elevated proportions of maize silage. There were no differences (P > 0.64) between diets in concentration of NDF or estimated ME (Table 6.4), although ADF concentrations were higher (P < 0.001) in treatments where sulla silage was used. TABLE 6.4 – Average dry matter content and nutritional composition of diets eaten by cows. Data are based on composition of pasture cut to 5 cm and proportion of silages in the diet. Data are all in g/100 g DM unless stated otherwise. Dietary composition FP RP PMS PSM PS PM LSDa P DM, % 17.4 15.9 26.5 24.6 25.2 25.4 4.85 0.005 CP 21.7 21.5 16.4 18.4 19.0 15.4 3.91 0.035 NSC 9.2 9.1 16.8 11.8 6.9 22.0 0.78 <.001 Lipid 4.0 3.9 3.8 4.2 4.3 3.6 0.29 0.006 NDF 43.9 44.1 44.8 45.8 47.5 42.7 6.94 0.647 ADF 24.4 24.9 26.9 28.6 31.3 24.5 3.00 0.007 Ash 10.1 10.2 8.4 9.3 10.0 7.4 0.69 <.001 ME (MJ/kg DM) 10.7 10.6 10.6 10.6 10.6 10.6 0.49 0.994 Abbreviations see Table 6.2 and text. Given that all silage treatments provided a similar level of nutrition, indicated by LW change and milk yield, and that 86 – 96 % of silage offered was eaten, the responses to performance of cows given full pasture were probably due to higher feed intakes. In contrast the cows offered 18 kg pasture DM/day (RP treatment) only ate 10.4 kg DM which was 58% of DM on offer and was probably too restricted. However pasture DMI is closely related to pasture allowance (Holmes et al., 2002) and the 179 restrictions did mimic pasture deficits faced by farmers when pasture is in short supply and silages are given to increase feed availability. Although there were no differences in responses between the individual silage supplement treatments, the PMS resulted in the highest milk yield without liveweight loss (Table 6.3). This diet also met cow requirement for CP and provided 16.9 g non- structural carbohydrates per 100 g of DM, suggesting a relatively good nutritive value despite its relatively high NDF concentration (44.8 g NDF/100 g DM). The PM diet also achieved a similar level of production, probably because the high pasture CP content enabled good complementation with low protein in maize silage. 6 .4 .3 – In sacco incubations The complete data set for degradation characteristics for DM, CP and fibre (NDF and ADF) are presented in Appendix 8 (Tables 8.1A to 8.4A). Data presented here are averaged for incubation of pasture in each of the 10 fistulated cows and for pasture/silage mixtures incubated in two cows fed each treatment: PMS, PSM, PS and PM. The key tests of significance are indicated on each table as follows: fit of the model (“Model”), comparison between pasture/silage mixtures regardless of the cow diet (“Forage”) and differences between cows based on in sacco digestion of pasture only (“Cow/diet”). Effects of individual cows on degradation rates were also tested in the model (“Cow”). 6 .4 .3.1 – DM digestion The amount of DM in the soluble “A” fraction was similar for all diets (Table 6.5; P > 0.67) and the slowly degradable B fraction was slightly higher for the PMS than other diets (43; P = 0.06). The PS was most rapidly degraded and the PM and PMS were slowly degraded (Table 6.5; Figure 6.1). The rate of pasture DM degradation rate (k = 0.067) was intermediate and there were no differences due to cow/diet (P = 0.17). There were significant differences in k values between cows (P < 0.019). 6 .4 .3.1 – CP digestion The effective degradability of CP (Table 6.6) was higher than that of DM (Table 6.5) for all diets, probably because about 61% of CP was solubilised and the undegradable CP fraction was smaller that that for DM. Diets with a high proportion of sulla silage had slightly higher effective degradability and a high proportion of maize silage lowered effective degradability. Cow/diet had no effects on protein degradability, measured with pasture (P = 0.5183). 180 TABLE 6.5 – In sacco degradation characteristics of dry matter (DM), in cows fed full pasture and four silage supplements. Kinetics are defined by soluble (A), degradable insoluble (B) and undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h-1), and effective degradability (E) which takes in account the effect of passage from the rumen1. DM A B k C E6% E8% Pasture 44 39 0.067 17 64 61 PMS 42 43 0.049 15 61 58 PSM 43 38 0.068 19 63 60 PS 46 36 0.084 17 68 65 PM 45 36 0.054 19 62 59 Model P 0.67 0.18 0.010 Forages P 0.67 0.06 0.004 Cow/diet P 0.31 0.170 Cow P 0.51 0.019 r2 0.17 0.89 0.98 Abbreviations are given in Table 6.2. 1 Passage rate set at 0.06 h-1 and 0.08 h-1. P: assessing goodness of fit for the overall model and tests of forage, cow/diet and cow effects. 181 40 60 80 0 12 24 36 48 60 72 % DM di sa ppe ar an ce Pasture PMS PSM PS PM 40 60 80 0 12 24 36 48 60 72 % CP di sa ppe ar an ce Pasture PMS PSM PS PM 0 20 40 60 80 0 12 24 36 48 60 72 Incubation time (h) % NDF d is app ea ra nc e Pasture PMS PSM PS PM 0 20 40 60 80 0 12 24 36 48 60 72 Incubation time (h) % A DF d is app ea ra nc e Pasture PMS PSM PS PM FIGURE 6.1 – Dietary dry matter (DM), crude protein (CP) and fibre (NDF and ADF) disappearance during in sacco incubations from fistulated cows fed full pasture and four silages supplements. Cow diets and forages in sacco bags are indicated on each figure. 182 TABLE 6.6 – Crude protein degradability coefficients (A, B and k) and effective degradability using fractional passage rates of 0.02, 0.05 or 0.08 h-1 (E2%, E5% or E8%) for pasture and mixtures of pasture and silages incubated in sacco. CP A B k C E2% E5% E8% Pasture 56 36 0.121 8 86 81 78 PMS 61 32 0.070 7 86 80 77 PSM 65 26 0.084 8 87 82 79 PS 66 28 0.090 6 89 84 81 PM 56 33 0.128 11 85 80 77 Model P 0.0009 0.1467 0.0675 Forages P 0.0009 0.0464 0.0204 Cows/diet P 0.1695 0.5183 Cow P 0.8067 0.1272 r2 0.74 0.91 0.94 Abbreviations are given in Tables 6.2 and text. TABLE 6.7 – Rumen degradable protein (RDP) and rumen undegradable protein (RUP) as a percentage of crude protein concentration, and rate of dry matter passage from the rumen (kDM) (NRC, 2001) for pasture and mixtures of pasture and silages incubated in sacco. % of crude protein RDP RUP KDM Pasture 81.4 18.6 0.049 PMS 79.9 20.1 0.050 PSM 82.1 17.9 0.049 PS 83.8 16.2 0.049 PM 79.7 20.3 0.049 Model P 0.0336 0.0336 Forages P 0.0099 0.0099 Cow/diet P 0.1492 0.1492 Cow P 0.1057 0.1057 r2 0.96 0.96 Abbreviations are given in Tables 6.2 and text. 183 In contrast to DM disappearance, the rate of protein degradation in sacco was reduced when sulla was mixed with pasture (Table 6.6; P = 0.02), possibly in response to the protection conferred by condensed tannins in sulla. Reduced protein degradation rate is likely to increase protein availability for absorption and increase nutritive value for cows. Calculations from NRC (2001) showed an average of 81.4% RDP and 18.6% RUP across all diets including pasture. Diets containing sulla had the highest proportion of RDP (Table 6.7). Crude protein degradability has also been evaluated by estimating dietary QDP, SDP, ERDP, RDP, RUP expressed as g/kg DM from pasture and mixtures of pasture, maize and sulla silage at three contrasting rumen outflow rates (Table 6.8). The slow, medium and fast (0.02, 0.05 and 0.08 h-1) rates correspond to cows fed at maintenance and at medium and high levels of intake. Outflow rate of about 0.05 h-1 correspond to cows at mid lactation with limited fed availability, as with silage supplementation in this study. Expression in terms of dietary DM provides a realistic assessment of mixtures with varying proportions of supplement containing a range of CP concentrations. Forage mixtures varied in NDF content from 430 to 479 g/ kg DM and the concentration of NDF accounted for 68% of the variation in effective rumen degradability for protein across diets. These data demonstrate a linear positive (P < 0.0001) relationship between dietary NDF concentration and ERDP across the five diets incubated in sacco (Figure 6.2). There was also a significant positive relationship between the effective DM degradability of pasture/silage mixtures and ERDP (Figure 6.3) but the variation about the regression (r2 = 0.29) confirmed forage differences between DM and CP degradation rates. Cow/diet may also affect the extent of pasture DM digestion in sacco (Appendix 8) and predicted degradability of pasture DM (Figure 6.3). 184 Pasture PMS PSM PS PM 0 40 80 120 160 200 425 445 465 485 NDF content (g/kg DM) ERDP (g /k g DM), k p=0 .05 FIGURE 6.2 – Relationship between neutral detergent fibre content (NDF, g/kg DM) and effective rumen degradability of crude protein (ERDP, g/kg DM) of ryegrass pasture, and pasture plus sulla and/or maize silages. When rumen DM outflow is 0.05/h, the relationship is described as: ERDP (g/kg DM) = - 352 + 1.1(± 0.18) x NDF (r 2 = 0.68; Root MSE = 11.1; CV = 8.05%; P < 0.0001). Abbreviations are given in Table 6.2 and text. Pasture PMS PSM PS PM 100 120 140 160 60 65 70 Effective degradability of DM (E DM ) ERDP (g /k g DM), k p=0.0 5 FIGURE 6.3 – Relationship between effective degradability of DM (EDM) and crude protein (ERDP, g/kg DM) of ryegrass pasture and pasture plus sulla and/or maize silages. When rumen DM outflow is 0.05/h, the relationship is described as: ERDP (g/kg DM) = - 176.7 + 4.80 (± 1.88) x E DM (r2 = 0.29; Root MSE = 16.4; CV = 11.9%; P = 0.0212). Abbreviations are given in Table 6.2 and text. 185 TABLE 6.8 –Crude protein degradability parameters QDP, SDP, ERDP, RDP, RUP in g/kg DM (AFRC, 1992) for pasture and mixtures of pasture and silages incubated in sacco. kp=0.02 QDP SDP ERDP RDP RUP Pasture 122 66 164 188 29 PMS 100 42 122 142 23 PSM 117 39 133 156 24 PS 124 43 142 167 21 PM 91 40 112 130 24 Model P 0.1018 0.0616 <.0001 <.0001 0.0063 Forages P 0.0186 0.0109 <.0001 <.0001 0.0012 Cow/diet P 0.5726 0.3766 0.0309 0.0263 0.0263 Cow P 0.9222 0.9306 0.7267 0.4764 0.4764 r2 0.93 0.94 0.99 0.99 0.98 kp=0.05 QDP SDP ERDP RDP RUP Pasture 122 55 153 177 40 PMS 100 32 112 132 33 PSM 117 31 125 148 32 PS 124 33 133 158 30 PM 91 32 105 123 31 Model P 0.1018 0.0746 <.0001 <.0001 0.0038 Forages P 0.0186 0.0132 <.0001 <.0001 0.0006 Cow/diet P 0.5726 0.5420 0.0711 0.0942 0.0942 Cow P 0.9222 0.8344 0.0570 0.0949 0.0949 r2 0.93 0.94 0.99 0.99 0.99 kp=0.08 QDP SDP ERDP RDP RUP Pasture 122 47 145 169 48 PMS 100 26 106 126 39 PSM 117 26 119 143 37 PS 124 27 127 152 36 PM 91 28 100 118 36 Model P 0.1018 0.0813 <.0001 <.0001 0.0055 Forages P 0.0186 0.0146 <.0001 <.0001 0.0009 Cow/diet P 0.5726 0.6376 0.0431 0.1327 0.1327 Cow P 0.9222 0.7518 0.0142 0.1361 0.1361 r2 0.93 0.93 0.99 0.99 0.98 Abbreviations are given in Tables 6.2 and text. K = fractional passage rate. p The parameters QDP (quick degradable protein), SDP (slow degradable protein), ERDP (effective rumen degradable protein), RDP (rumen degradable protein), RUP (rumen undegradable protein) all in g/kg DM were calculated as describe in the Material and methods of Chapter 3 (Item 3.3.5.1). 186 6 .4 .3.1 – Fibre digestion Although there were no differences in NDF concentration between the pasture/silage mixtures (Table 6.4) when they were incubated in sacco there were differences in rates of NDF and ADF digestion (P < 0.05; Table 6.9). Both NDF and ADF where present mainly in the slowly degradable fraction (B) and undegradable residues. Rates of fibre degradation were slowest for mixtures containing a high proportion of maize silage and tended to be most rapid for PS (Table 6.9). The degradation curves for DM, NDF and ADF (Figure 6.1) show clear effects of maize on rates and extent of degradation of each constituent. The fibre effective degradability for pasture and pasture supplemented with 6 kg DM sulla silage/cow.day were about 30% higher than other supplemented cows treatments (Table 6.9). Incubation of pasture showed that cow/diet did not affect rates of fibre degradation when pasture was incubated in sacco (P = 0.3389). 187 TABLE 6.9 – Neutral and acid detergent fibre (NDF and ADF) degradation characteristics (% of total DM) in cows fed full pasture and four silage supplements. Kinetics are defined by soluble (A), degradable insoluble (B) and undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h-1), and effective degradability (E) which takes in account the effect of passage from the rumen1. NDF A B k C E 6% E 8% Pasture 22 50 0.053 28 45 42 PMS 13 64 0.033 23 35 31 PSM 11 56 0.054 33 37 33 PS 27 43 0.058 30 49 46 PM 14 53 0.037 33 35 31 Model P 0.0001 0.0696 0.0803 Forages P <.0001 0.0152 0.0248 Cow/diet P 0.0038 0.3919 0.3389 Cow P 0.3947 0.1884 r2 0.94 0.94 0.94 ADF A B k C E 6% E 8% Pasture 23 49 0.050 28 44 40 PMS 14 62 0.031 25 34 30 PSM 14 52 0.053 33 38 34 PS 26 44 0.065 31 49 45 PM 13 53 0.030 34 32 29 Model P 0.0001 0.1367 0.0290 Forages P 0.0002 0.0556 0.0085 Cow/diet P 0.0009 0.2384 0.1666 Cow P 0.3019 0.0825 r2 0.94 0.91 0.96 Abbreviations are given in Tables 6.2 and text. 188 6.4.4 – Rumen pH, ammonia and VFA concentrations Rumen liquor pH averaged across all cows (Figure 6.4) shows higher values (P < 0.001) prior to AM feeding (mean 6.7 ± 0.1) compared to values after the PM milking (16:00 h) which average 5.6 ± 0.1. There was no effect of diets on either morning or afternoon pH, or diurnal pattern (Figure 6.5). Mean concentrations of rumen VFA were similar across treatments, averaged 101 mMol/L, with about 0.69 acetate, 0.17 propionate and 0.11 butyrate (Table 6.10). There were no treatment effects on concentration or molar proportion of VFA. The ratio of acetate: propionate averaged 4.1 and was similar for all diets. In contrast, the pasture diet resulted in highest concentrations of rumen ammonia (Table 6.10) and lowest values were measured when maize silage was included in the diet. Cows fed either pasture or pasture plus sulla silage had higher rumen NH3 concentrations than other supplemented treatments (P = 0.03). The diurnal variation in VFA concentrations (Appendix 8 – Figure 8.1A) showed peak concentrations about 6 hours after morning feeding. The diurnal range in total VFA concentrations was greatest with the PM diet and least with PMS. Dietary effects on the extent of diurnal variation was similar for acetate and n-butyrate but diets containing sulla appeared to have least diurnal variation in concentrations of minor VFA. Ammonia concentrations followed a similar diurnal pattern as VFA but the variation was much smaller with PMS and PSM, than other diets. This can be explained by grazing behaviour because cows fed PMS and PSM chose to eat supplements first, followed by pasture while cows given PS or PM grazed pasture first and ate supplements after grazing. 189 TABLE 6.10 – Rumen pH and concentrations (mMol/L) of metabolites for individual cows grazing pasture with and without maize and/or sulla silage supplements. Volatile fatty acids Diet Cow pH Acetate Propionate n-buty Minor Total NH3 Pasture 5272 6.1 69.4 17.5 11.7 4.0 102.6 12.0 Pasture 7915 6.2 71.9 17.8 11.7 4.3 105.6 10.8 PMS 3343 6.2 66.6 16.2 9.8 3.2 95.7 4.8 PMS 3792 6.1 64.7 15.6 9.4 3.4 93.2 6.2 PSM 5774 6.2 72.4 17.8 10.6 4.1 105.0 8.1 PSM 7920 6.2 70.8 16.1 10.3 3.2 100.5 5.9 PS 3788 6.3 66.5 15.8 9.8 3.7 95.7 9.2 PS 7926 6.2 66.5 18.7 11.1 4.2 100.5 8.6 PM 5756 5.9 76.4 17.3 11.7 4.2 109.7 7.0 PM 7912 6.1 71.6 16.0 10.3 3.2 101.1 5.1 Mean 6.1 69.7 16.9 10.6 3.8 100.9 7.8 Diet P 0.4079 0.5551 0.9373 0.6278 0.9303 0.6063 0.0156 Week P 0.0008 0.0062 0.1297 0.7804 0.8746 0.0171 0.0351 Diet*week P 0.8832 0.9707 0.8701 0.7218 0.0482 0.9935 0.4806 NH3 P 0.7739 0.0766 0.0163 0.0158 0.0003 0.0368 Abbreviations are given in Tables 6.2 and text. n-buty, n-butyrate; Total, total VFA. 6.7 6.8 6.8 6.8 6.6 5.6 5.5 5.6 5.7 5.4 5.0 5.5 6.0 6.5 7.0 Pasture PMS PSM PS PM Treatments pH Morning Afternoon FIGURE 6.4 – Average morning (AM) and afternoon (PM) rumen fluid pH from fistulated cows fed full pasture and four silages supplements. Abbreviations are given in Table 6.2. 190 5.0 5.4 5.8 6.2 6.6 7.0 pH Pasture PMS PSM PS PM * ** 0 10 20 30 40 m Mol NH 3/ L Pasture PMS PSM PS PM n.s. * † † 0 50 100 150 200 0 3 6 9 Hours after feeding m Mol to tal VFA /L 1 2 Pasture PMS PSM PS PM FIGURE 6.5 – Pattern of pH and rumen fluid ammonia and total VFA concentration between 07:00 h and 19:00 h, averaged from fistulated cows fed full pasture and four silages supplements on the first day of the in sacco incubation. † P < 0.10, *P < 0.05, and **P < 0.01. Abbreviations are given in Table 6.2. 191 6.5 - Discussion The impact of silage supplementation has been examined in Chapters 5 and 6 in terms of production response and extent of pasture substitution by silage. Interpretation has been based on DM and nutrient intake and in sacco digestion kinetics. Both trials were constrained by the high quality pasture on offer, which exceeded the quality of silage supplements in some instances. The extent of production response will depend upon the amount of pasture available, which was probably insufficient in the second study, and the nutritional value of the diet. Substitution of pasture by supplements can be an important consideration when optimising production against feeding costs and it is important to ask whether both the extent of substitution and the response to supplementation for several silages and mixtures can be addressed in a single trial. An initial requirement will be an accurate measurement of cow intakes. Dry matter intake prediction Estimation of cow intakes is difficult, whether on a group (Berchielli et al. , 2000) or on an individual basis using indigestible markers and faecal sampling. A brief evaluation of intakes measured in both experiments (Chapter 5 and 6) has been made by comparison with predictions of DMI using the AAC equations for dairy cows (AAC, 1990). Inputs for the model are cow breed, days of pregnancy, liveweight (LW) and LW change, age, condition score, milk production and composition, energy content and digestibility of the feeds, and initial estimate of DMI. DMI predictions were highly correlated with actual values for DMI (P < 0.0001; r2 = 0.61) and showed that DMI was underestimated by only 0.12 kg across all diet treatments (Figure 6.6). Cows fed restricted pasture as a sole diet in these short term trials tended to maintain milk production but lost liveweight, so removal of these data from the analysis resulted in a more precise model prediction of DMI (r2 = 0.90) for cows either grazing a high pasture allowance or given restricted pasture plus supplemental silages. 192 Actual = Predicted Restricted pasture only 10 12 14 16 18 20 10 12 14 16 18 20 Predicted DMI, kg/day A ctua l DMI, kg /d ay (A) Restricted pasture only -3 -2 0 2 3 10 12 14 16 18 20 Predicted DMI, kg/day Pre di cte d - a ctua l DMI, kg /d ay (B ) FIGURE 6.6 – Plot of actual and predicted values for dry matter intake (DMI) using the AAC (1990) model (A) and the mean bias calculated from the model prediction minus actual DMI (B). The relationship between actual and predicted values (without the restricted pasture treatment) is given by: Actual DMI = 2.72 + 0.86 (± 0.10) x Predicted DMI (r2 = 0.90). Data are kg/day. The over prediction of intakes for cows given restricted pasture as a sole diet (Figure 6.6) suggest an inability of the model to account for minimal residual DM when pasture allowance is low. When pasture on offer was only 18 kg DM/cow.day it would be impossible for cows to eat 15 kg pasture DM/day. Model prediction when pasture was offered as a moderate - high allowance or when grazing cows were offered silage supplements, are close to DMI measured in both experiments. This is reassuring and suggests the pre and post-grazing pasture cuts enabled an accurate estimate of group intakes. The allocation of very low quantities of pasture does not enable good model prediction and excessive feed restriction is inappropriate for maintenance of milk production. 193 Pasture allowance and substitution Similar cow responses to all silage supplements in this experiment suggests ME was more limiting than other nutrients and differences in silage composition were small to effect milk production. The high quality pasture on offer probably contributed to the lack of response to specific silages. The impact of pasture allowance has been summarised by Stockdale (2000b) using data from 17 Australian and 3 Irish studies. These data, with those from experiments conducted here have been plotted (Figures 6.7 and 6.8) to show the relationship between pasture allowance and both intake and substitution of pasture by supplement. Figure 6.7 shows a typical curvilinear relationship between pasture allowance and pasture DM intake without supplements but intakes of supplements with pasture were higher than of pasture alone, especially at very low allowances (Peyraud et al., 1998). When 20 – 30 kg pasture DM was allowed per cow, with supplements, the data did not demonstrate any relationship between pasture intake and allowance, so intakes would have been determined by quality and quantity of feedstuffs and the cow’s metabolic requirements. 0 5 10 15 20 25 0 20 40 60 8 Pasture allowance, kg DM/cow.day Inta ke , k g DM/c ow .da y 0 Australia New Zealand Ireland Pasture intake only Pasture intake plus supplements FIGURE 6.7 – Relationship between dry matter intake and pasture allowance (PA) of cows offered only pasture (PDMI; PDMI = 4.0 + 0.47 x PA - 0.0046 x PA2; r2 = 0.45) and total intake (TDMI) of cows offered pasture with supplements (symbol: ◊; TDMI = 14.2 + 0.08 x PA; r2 = 0.13; P < 0.001). Data from Stockdale (2000b) and Chapters 5 and 6. 194 0.0 0.2 0.4 0.6 0.8 1.0 0 10 20 30 40 50 6 Pasture allowance, kg DM/cow.day S ubsti tutio n ra te 0 FIGURE 6.8 – Relationship between substitution rate (SR) and pasture allowance (PA) by cows grazing pastures in New Zealand, Ireland and Australia (SR = 0.1 + 0.009 (± 0.002) x PA; r2 = 0.20, Root MSE = 0.16, CV = 46.4%, SR mean = 0.35, P < 0.0001). Intake of supplements will also vary with quality and quantity of both the pasture and the supplements and the metabolic drive of cows to eat. Although substitution rate (SR) increased with pasture allowance (Figure 6.8) the coefficient of variation was large (46.4%). When pasture allowance was above 35 kg DM/cow.day, SR increased significantly. These relationships cannot define likely degrees of substitution for specific supplements and use of both low (20 – 30 kg DM/cow.day) and high (over 35 kg DM/cow.day) allowances may be required to define SR for contrasting supplements. Absolute levels of substitution vary with stage of lactation, quality and quantity of both pasture and supplements. There was a lower substitution in the first experiment (0.06 – 0.15) when pasture was restricted to 25 kg DM/cow, compared to 0.26 – 0.33 when only 18 kg pasture DM/cow was available, but 5 and 6 kg silage DM were available in the respective trials. Effects of diet and supplement have been shown by Stockdale and Dellow (1995) where lactating cows fed white clover pasture with 3.5 – 5.0 kg maize silage had an average SR of 0.49. More recently, Walker et al. (2001) reported substitution of paspalum pasture for grain increased from 0.02 – 0.28 as concentrate increased from 3 – 10.4 kg DM/day (experiment 1) and when un- supplemented pasture intake was 15 kg DM/cow and 3 – 5.9 kg of concentrates were fed, the SR ranged from 0.23 – 047 kg/kg (experiment 3). They concluded that high rates of substitution were responsible for diminishing financial returns to farmers. The acceptability of both supplements and pasture may provide an insight into the nutritional balance required by the cows, and suitability of silages for supplementing 195 pasture. In this study the cows ate the silage mixtures (PMS and PSM) in preference to pasture and this resulted in lower 12 hour fluctuations in rumen ammonia and VFA concentrations (Figure 6.5). In the previous experiment (Chapter 5) the cows actually ran to access the lotus silage. Lotus silage, maize and sulla silages mixtures were eaten before pasture whereas pasture was eaten before either maize or pasture or sulla silages in both experiments. Acceptability is likely to affect both DMI and the extent of substitution. Other factors include quality and quantity of pasture and supplements, cow demand for nutrients and consequences of grazing close to the ground or near faeces and urine patches when pasture availability is restricted. Rumen degradation in sacco When maize silage was added to pasture the extent and rate of DM degradation declined, whereas the reverse was true for sulla. With highly productive cows the rate and extent of DM degradation will affect performance and sulla silage may be more appropriate than maize silage, especially as CP degradation rate was reduced when sulla silage was incubated with pasture. Slower CP degradation is consistent with the presence of CT (Broderick and Albrecht, 1997; Burke et al. , 2002a; Messman et al. , 1996) Pasture incubation in cows given five diets allowed cow/diet effects to be measured, and different diets affected degradability of pasture DM and fibre (NDF and ADF; Figure 6.9). There appeared to be less DM and fibre degradation when pasture was incubated in cows fed diets including maize silage but addition of sulla to cow diets increased effective degradability. Effects of diet on CP degradation were less apparent (Table 6.8). Impact of diet on degradation rates has been observed by Mertens and Waghorn (unpublished) where diets based on maize silage reduced rates of both maize and lucerne degradation relative to diets based on lucerne. Results here support this observation and show the effects of diet on effective rumen degradation were greater with higher intakes and/or high outflow rates (Tables 6.6 and 6.8). The effect of diet on in sacco degradation requires more investigation. 196 Diets fed to cows: 35 55 75 95 0 12 24 36 48 60 72 % Pa stur e DM l os s Pasture PMS PSM PS PM Diets fed to cows: 50 70 90 0 12 24 36 48 60 72 % Pa stur e CP l os s Pasture PMS PSM PS PM Diets fed to cows: 0 20 40 60 0 12 24 36 48 60 72 Incubation time (h) % Pa stur e NDF l os s Pasture PMS PSM PS PM Diets fed to cows: 0 20 40 60 0 12 24 36 48 60 72 Incubation time (h) % Pa stur e A DF l os s Pasture PMS PSM PS PM FIGURE 6.9 – Pasture dry matter (DM), crude protein (CP) and fibre (NDF and ADF) disappearance during in sacco incubations from fistulated cows fed full pasture and four silages supplements. Why did this study use different ways to assess dietary protein? Metabolisable protein is an important component of diet quality and is the sum of digestible microbial true protein and digestible undegraded feed protein (AFRC, 1992). Digestible microbial true protein is calculated from the microbial CP supply which is, in turn, calculated from the total fermentable energy in the diet, providing there is sufficient effective rumen degradable protein (ERDP) for unrestricted microbial growth. The degradation coefficients (A, B and k) of feed proteins incubated in dacron bags in the rumen are required for the determination of ERDP (section 3.3.5.1). Clark et al. (1997b) suggest there is limited understanding of the nutritional constraints to dairy production from summer pastures, but effects of grass maturation and pasture supply can limit MP for milk production. There is little information on rumen degradability of forage CP (Chapters 3 and 4; Barrell et al. , 2000 and Burke et al. , 2000) so several approaches (AFRC, 1992; NRC, 2001) were used to evaluate this problem. Proportion of RDP and RUP differed across diet treatments (Table 6.7 and 6.8) and 197 estimates of effective rumen degradability of crude protein (ERDP, g/kg DM) showed strong positive relationship with dietary CP content. CP degradation characteristics reported here are in agreement with an Australian study of perennial pastures through the year (Wales et al., 1999a). They showed that summer pasture in Victoria, Australia, provided a surplus of MP of 0.14 to 0.23 kg/cow.day compared to 0.48 to 1.21 kg/cow.day in spring and suggested MP was unlikely to limit milk production of cows eating 17 kg pasture DM/day and producing up to 30 kg milk/day. These data support the argument that the good quality pasture available for both trials presented here resulted in insufficient energy rather than protein for cows producing 1.0 – 1.2 kg milksolids/day. Rumen digesta and cow feeding behaviour Diets did not affect concentration or molar proportions of VFA (Table 6.10). The 84% increase in dietary NSC concentration when maize silage was added to pasture did not affect either propionate concentration or acetate: propionate ratio. In contrast cows fed pasture as a sole diet had a higher concentration (P = 0.0156; Table 6.10) of ammonia compared to other diets. Although there were no effects of diet on rumen VFA, diets with pasture, maize and sulla silages (PMS and PSM treatments) provided a stable rumen environment possibly due to feeding behaviour. The lower variation over the 12 hours period from 07:00 – 19:00 h for VFA was also evident for NH3 (Figure 6.5), probably because the cows given these treatments ate supplements first (PMS and PSM) and then grazed pasture. Cows grazed pasture first with PM and PS diets. The PMS treatment was the most acceptable diet for the cows, with least refusals and provided the highest NSC dietary concentration. Visual observations suggest the cows preferred mixtures of sulla and maize silages compared either silage fed alone, but the short duration of the trial and restricted pasture availability prevented further evaluation of silage effects on total DMI or performance. Future progress to maximise milk production from cows grazing pasture and supplemented with silages should investigate the ability of diets to alter fermentation rate and be nutritionally balanced. Studies should also test both low and high pasture allowances to obtain information on choice of nutrient from pasture versus the supplement. This information will provide a better understanding of substitution than can be obtained with a single pasture allowance and a focus on choice and cow behaviour may provide a better understanding of rumen stability and cow performance. 198 CNCPS diet evaluation The CNCPS model was used to evaluate dietary mixtures of maize and sulla silages with pasture to identify limitations to performance and predict rumen degradation parameters. The model provided useful information concerning microbial growth, rumen passage rates, nutrient limitations and predictions of performance of cattle supplemented with single silages in Chapter 5. Silages mixtures were fed to provide a more balanced nutrient supply to dairy cows in this experiment and ruminal measurements indicated some effects of diet on fermentation. The CNCPS model has been used to further evaluate these diets and to identify sources of variation in cow performance. This information will enable improved dietary formulations for grazing cows given supplements. Table 6.11 presents feed composition collected in this experiment and degradation rates (CNCP feed library, section 6.4.3 and Burke, unpublished) used in diets evaluated with CNCPS. TABLE 6.11 – Feed composition and degradation rates used into CNCPS evaluation. General characteristics Individual feeds DM (%) NDF (%DM) peNDF (%NDF) Lignin (%NDF) Fat (%DM) Ash (%DM) Starch (%NFC) Full pasture 23.60 45.30 40.0 7.23 3.90 11.70 48.0 Restricted pasture 15.90 45.30 40.0 9.31 4.10 11.00 48.0 Pasture - PMS 20.47 46.33 40.0 7.50 3.98 9.90 48.0 Pasture - PSM 18.19 45.80 40.0 6.22 4.03 9.91 48.0 Pasture - PS 19.92 46.53 40.0 7.40 3.83 9.87 48.0 Pasture - PM 18.42 45.43 40.0 8.02 3.90 9.55 48.0 Maize silage 33.00 49.00 90.0 10.59 3.00 4.00 80.0 Sulla silage 34.00 50.00 85.0 22.72 5.20 10.10 70.0 Energy and protein values Individual feeds CP (%DM) UIP (%DM) Sol-P (%CP) NPN (%Sol-P) NDFIP (%CP) ADFIP (%CP) Full pasture 18.00 22.4 55.0 4.76 9.1 3.04 Restricted pasture 21.90 22.4 55.0 4.76 9.1 3.04 Pasture - PMS 20.98 22.4 61.0 4.76 9.1 3.04 Pasture - PSM 21.70 22.4 65.0 4.76 9.1 3.04 Pasture - PS 20.63 22.4 66.0 4.76 9.1 3.04 Pasture - PM 21.00 22.4 63.0 4.76 9.1 3.04 Maize silage 9.50 20.9 58.0 100.0 16.0 7.0 Sulla silage 15.70 29.1 66.0 28.0 24.0 16.0 Degradation rates Individual feeds Degradation rates (%.h-1) Carbohydrate Protein A B1 B2 B1 B2 B3 Full pasture 85.3 19.2 14.0 200 12.0 2.0 Restricted pasture 85.3 19.2 14.0 200 12.0 2.0 Pasture (PMS, PSM, PS, PM) 85.3 19.2 14.0 200 12.0 2.0 Maize silage 10.0 30.0 5.0 300 15.0 0.25 Sulla silage 10.0 25.0 9.0 150 11.0 1.75 Abbreviations see text and Table 5.5. 199 TABLE 6.12 – CNCPS predictions of nutrient composition, cow performance and rumen characteristics of treatment diets. FP RP PMS PSM PS PM Nutrient composition ME, MJ/kg DM 10.41 10.53 10.03 9.99 9.52 10.34 CP, g/100g DM 18.0 21.9 17.2 18.6 18.8 16.6 Soluble CP, %CP 55.0 55.0 61.1 64.7 66.0 61.9 NDF, g/100 g DM 45.3 45.3 47.5 47.2 47.8 46.8 peNDF, g/100 g DM 18 18 28 28 28 28 Total NFC, g/100 g DM 22.7 19.7 25.2 23.2 21.7 27.4 Total fat, g/100 g DM 3.9 4.1 3.9 4.2 4.3 3.6 Performance predictions ME allowable milk, kg/day 18.2 8.1 16.7 15.4 13.4 16.7 MP allowable milk, kg/day 20.4 13 14.6 14.1 12.1 14.1 Daily weight change due to reserves, kg/day 1.3 -0.9 -1.9 0.3 0.0 0.5 Rumen digestion, metabolism and passage MP from bacteria, g/day 942 585 855 814 745 886 MP from undeg. feed, g/day 652 473 454 470 450 413 MP from undeg. feed, %MP total 41 45 35 37 38 32 Total DIP, %CP 78.2 80.4 79.3 79.9 79.5 80.5 Ruminal N balance, % of req. 123 207 151 168 177 142 Total bacterial nitrogen, g/day 251 156 228 217 199 236 Urea cost, MJ/day 4.8 3.6 2.3 3.3 3.4 1.8 Urea cost, %ME intake 2.9% 3.3% 1.6% 2.3% 2.6% 1.2% Excess N excreted, g/day 204 167 118 151 154 103 Liquid passage rate, %.h-1 10.6 8.7 10.5 10.3 10.2 10.2 Pasture passage rate, %.h-1 6.40 5.22 6.23 6.14 6.07 6.12 Maize silage passage rate, %.h-1 NA NA 4.83 4.76 NA 4.73 Sulla silage passage rate, %.h-1 NA NA 4.90 4.82 4.78 NA Predicted ruminal pH 6.19 6.19 6.46 6.46 6.46 6.46 NFC, non-fibrous carbohydrates. Other abbreviations see text and previous tables. ME was the first limiting nutrient for cows fed pasture alone and MP was the first limiting nutrient when cows were fed pasture plus silages. At least 50% of the total MP was of microbial origin for all diets, with highest percentage for the maize silage treatment (PM: 68% of MP of microbial origin; 886 g/day). The recommendation for ruminal N balance is 100 to 110% of requirements (Fox et al., 2003) and this was easily achieved from all diets (123 to 207% of requirement). Development of rations that will improve efficiency of nutrient use and reduce nutrient wastage on farm are requirements for economic and environment sustainability. Provision of silage mixtures attempted to balance nutrient supply with demand and the CNCPS simulations predicted nutrient requirements, balances, 200 excretion of N and urea cost for each treatment group. Higher concentration in CP in pasture resulted in high urea costs (2.9% and 3.3% of ME intake for FP and RP respectively). This represents an increased cow maintenance cost, associated with removal of excess nitrogen, and the high predicted MP from undegraded feed (652 g/day) with the FP diet was a association with the highest N excretion (204 g/day). This will have a negative impact on the environment in terms of nitrogen leaching and generation of nitrous oxides (Carran, 2002) relative to diets containing maize and sulla silages. However, the low CP% of the maize and sulla has caused MP to be the first limiting nutrient and milk production drooped by feeding these silage supplements compared to FP treatment. Although CNCPS provided useful information concerning nitrogen fluxes with similar predictions of total DIP to effective CP degradability measured in sacco (Table 6.6), the predictions of mean rumen pH differed from observed values (Figures 6.4 and 6.5). 6.6 - Conclusion Dairy farmers face the task of maintaining a desired level of production often when the quality of pasture is less than optimal. As pasture is likely to provide the cheapest source of nutrients, it is important to maximise feed intake at minimal cost. Silage supplements can be used to fill summer feed deficits when pasture quality declines due to maturation of ryegrass. To achieve positive responses from supplements, supplements should be of higher quality (nutritive value) than pasture and the supplements must be chosen to complement the pasture on offer. The differences between digestion kinetics of maize and sulla silage supplements demonstrate the importance of selecting an appropriate supplement to complement the pasture on offer. The hypothesis was not proven, but the low pasture allowance may have prevented a response due to insufficient ME intake. Dietary mixtures did reduce diurnal variation in rumen fermentation parameters, without affecting changes in milk or milk solids production. 201 Chapter 7 Simulation models for ration balancing and an evaluation of the CNCPS model. 1 1 Part of these data was previously published in the Proceedings of the New Zealand Society of Animal Production , 2003, 91-95. 7.1 - Abstract The importance of mechanistic models for ration balancing with forages is indicated and physical limitations to intake emphasised, because these limit nutrient supply to cows grazing forages, especially grass. Ration balancing models using fresh or ensiled forages to complement pasture will need to accommodate intake limitations, due to rumen fill, clearance, chewing or other criteria. The potential of the Cornell Net Carbohydrate and Protein System (CNCPS) model to predict milk production from diets based on pasture and forage supplements was tested using information presented in this thesis. Data were obtained from studies in which pasture was complemented with contrasting silages including maize, pasture, sulla, lotus and forage mixtures (Chapters 5 and 6), comprising 30 - 40% of dry matter intake (DMI). Twelve diets were used in this evaluation. DMI, live weight (LW), days in milk, and diet composition were determined during the trials and used as inputs in the model. Across all diets, a significant (P < 0.01) relationship existed between predicted and actual values for DMI (r2 = 0.63), milk yield (r2 = 0.64) and LW change (r2 = 0.57) but there were still large unexplained sources of variation. No significant mean bias was observed for any of the variables, but the slope of residual differences against predicted values was significantly different from zero for milk yield, LW change (P < 0.01) and for DMI (P < 0.06). The results indicate a satisfactory prediction of milk production when cows are neither gaining nor losing weight, but that a systematic bias exists probably because of the CNCPS model’s failure to account for nutrient partitioning. Keywords: CNCPS; dairy cows; diet formulation; modelling; nutrients requirement. Short title: Evaluation of the CNCPS for dairy cows. 7.2 - Introduction Improvements to the nutrition of dairy cows fed pasture will require a balance between the amount of pasture and supplements eaten. The biggest problems for achieving a nutritionally balanced ration are the continuous change in pasture composition and the effects of fibre in pasture on voluntary feed intake. Changing composition will require monitoring (available through NIRS analytical services) and ration balancing will need models with a strong mechanistic component which are suitable for New Zealand pastoral feeding. Ration balancing, to optimise performance, cow welfare and profitability should be able to: 1) indicate optimal mixtures of diet components to achieve intended levels of production and 2) incorporate an intake 203 regulatory component, because feed type will affect potential intake as well as nutrient supply from feed degradation. Intakes of forages are likely to be regulated by physical constraints (chewing, rumen fill and rumen clearance) rather than metabolic feedback. Ration models used in the dairy industry serve a number of functions. In the Northern Hemisphere they are used for ration balancing to formulate least cost diets. These models are designed to meet cow nutrient requirements for specific stages of lactation, and take into account pregnancy, age, liveweight, milk production and composition. These systems assume feed is not limiting and intake is regulated by metabolic rather than rumen fill parameters. These models focus on cow requirements and the National Research Council (NRC, 2001) dairy model and the Cornell Net Carbohydrate and Protein System (CNCPS) indicate whether metabolisable protein or energy is the first limiting nutrient for milk production. The models are based on cow needs for specific nutrients as well as overall demands for protein and energy. The rumen component of the model uses feed composition, degradability and empirical formulae to calculate outputs such as ME, rumen microbial growth, nitrogen and peptide balance and predicted ruminal pH but does not predict VFA yield or methane production. The model is evolving to meet changing needs of the North America dairy industry, for example limiting excessive waste of nitrogen and phosphorus because of environmental concerns. These mechanistic models have potential benefits for dairy nutrition, compared to empirical systems which tend to be specific to the data set used for model development, because they offer potential for accommodating diverse dairy feeding and production regimes. Modelling digestion parameters, especially supply of specific nutrients and consideration of rumen function enables diverse feed types to be accommodated, at least in theory. However no model can balance feed inputs with production indices when feed intake is limiting, as is often the case under grazing systems. The North America models do have constraints concerning intakes, but these do not apply easily to New Zealand pastures. Intake of cows fed pasture will be constrained by rumen outflow (clearance; Waghorn, 2002) and this has not been included as a regulatory component of the CNCPS or NRC models. The long, high fibre, leafy (some times stemmy) grass-based pastures grazed by New Zealand cows are very different from low fibre (below 35% NDF in the DM) chopped, grain and silage based diets for which North American models were developed. 204 Pasture based models have usually been empirical in nature, but they do accommodate feed supply, which has a major effect on performance of grazing animals. Examples of these systems include those developed in Australia (Camdairy; Udder), New Zealand (feedTECH), United States (Dafosym) and complex models which attempt to include metabolism and performance (Molly). These all have specific applications for the dairy industry, but many also have weaknesses. Some factors which are crucial for successful modelling to understand fresh forage utilisation include: o Ready availability for evaluation, for example downloading from the Internet for a trial period (normally 30 days) or even available with purchase of feeding information (NRC, 2001). Models will only be useful if they easily accessed for evaluation. o Model should be interactive and user-friendly. There is no point in developing a computer program (model) that is hard to use and does not interact with the user. The success of Microsoft Windows® is due to its interactive nature; people are discouraged by systems requiring specialist skills (e.g.: DOS system); difficult models are less likely to be tested and utilised than Windows based systems. o The model should be based on scientific evidence. Although most dairy models are empirical, based on a summary of equations from different studies, mechanistic models need to be tested and information provided to explain the basis of their function prior to release. o A mechanistic nutrition model needs to incorporate: feed composition and degradability parameters; animal characteristics (type, age, breed, liveweight, stage of lactation and pregnancy, milk production and composition) and environment factors. These should be interactive because each factor impacts on the others. o The inputs required by the model should be readily available and be unambiguous. Inputs should not have to be derived from a unique data basis or involve expensive measurements. If inputs are too complicated, models frequently use default values which may not be appropriate for the feeding systems being tested, and the predictive capability will be compromised. The NRC and CNCPS models offer good potential for predicting responses (e.g. milk yield) to feeding regimes (e.g.: use of silages supplements), and there is an increasing demand for this type of information in the New Zealand dairy industry. The industry needs systems for calculating production responses to supplementation, for 205 example enabling lactation to be extended into autumn or substitution of pasture in early lactation to achieve high energy intakes and persistent milk production. It is essential that appropriate supplements be given to complement ryegrass pastures and that responses to supplementation be predictable. The research summarised in this thesis has provided information concerning the composition and degradation kinetics of pasture, especially in relation to maturation and this will ultimately provide a basis for model predictions. The rumen component in the CNCPS model makes this a logical choice for using the in sacco and in vitro data from these and other trials (Burke et al., 2000). In sacco digestion kinetics comprise two principal fractions: 1) a soluble rapid degradable fraction released by eating and chewing (A) typically comprising about 40% of the DM, and 2) a slowly degradable insoluble fraction (B) which disappears in response to microbial activity (k). The proportion of insoluble degradable (B) fraction has a significant impact on predictions for performance. This is logical, but this fraction and its degradation rate will be influenced by particle size reduction by chewing to affect microbial colonisation and also rumen clearance. These parameters apply to all components of the diet and suggest application to the CNCPS model should be straight forward, but the model requires inputs which are not readily obtainable for fresh forages. Values for these fractions are more available for grains and silages than fresh forages. The model is valuable in that it can contribute to our understanding of nutrition, ration evaluation, research planning and understanding ruminal kinetics, rather than simply predicting cattle requirements (Fox et al., 1992; Kolver et al., 1996; Russell et al., 1992; Sniffen et al., 1992), but users must be aware of model limitations. The CNCPS model had been validated against results from grazing trials (Chapters 5 and 6) to indicate strength and weakness in the current model format. There are numerous statistical tools available for evaluating the accuracy and precision of models that predict animal performance. These include plots of predicted and observed values (Kolver et al., 1996), mean squares prediction error analysis (Bateman et al., 2001; Kohn et al. 1998; Smoler et al., 1998), and analysis of residuals (predictions from the models minus actual data) against predictions (Kohn et al., 1998; St-Pierre, 2003). The aim of this study was to determine the utility and accuracy of the CNCPS model to predict milk production based on pasture and silage supplements, using data obtained from the two dairy cow trials conducted in mid-lactation (Chapter 5 and 6) 206 when pasture was complemented with contrasting silages. Further evaluations were made by predicting performance of cows fed low, medium and high quality pasture with or without silage supplements to identify first limiting nutrients and indicate rumen digestion parameters. These evaluations utilized composition and kinetic data obtained in this thesis and from Burke (unpublished). The hypothesis was that the Cornell model (CNCPS) would predict cow performance and indicate nutritional limitations from intake and digestion parameters when cows grazing pasture were supplemented with contrasting silages. 7.3 - Material and methods 7.3.1 - Cow trials used for model evaluation The data against which the model predictions were tested were derived from twelve rations (treatments means) in two trials carried out in Hamilton (Chapters 5 and 6). Each trial comprised 60 Friesian cows (10/treatment) averaging 528 ± 17 kg live weight (LW); 17 ± 2.4 kg milk/day; 156 ± 15 days in milk. Cows grazed ryegrass ( Lolium perenne) and white clover (Trifolium repens ) pasture complemented by contrasting silage supplements contributing 30 - 40% of DMI. Each trial was four weeks in duration and silage supplements included maize (M; Zea mays ), sulla (S; Hedysarum coronarium), pasture, lotus (LC; Lotus corniculatus) and mixtures of M and S. Pasture intakes by each treatment group were estimated by using a rising-plate meter to estimate pre- and post-grazing herbage mass. This was done three times per week for each treatment group. Weekly pasture cuts (pre- and post-grazing of representative pasture) were made to ground level for calibrating the rising-plate meter and determining chemical composition (e.g.: protein, fibre, ME, minerals) of material on offer by NIRS analyses (Corson et al., 1999). Digestion kinetic data (rates and extent of digestion) were obtained from in sacco incubations (Burke et al., 2000; Chapter 6). Both kinetic data and chemical composition of feeds were entered into the feed library of the model. Neutral and acid detergent fibre insoluble nitrogen required by the model to estimate the amount of slowly degraded and unavailable protein in each feed, ruminal rates of soluble carbohydrate and protein fermentation, and amino acid composition (Tedeschi et al., 2000) for pasture and silages for all treatments were obtained from the CNCPS library files for similar feeds. 207 7 .3.2 - CNCPS evaluation In this study, principal outputs assessed from the model were: DMI, milk production predicted from the first limiting of either metabolisable energy (ME) or metabolisable protein (MP), LW change and dietary ME concentration. Animal characteristics (animal type, age, breed type, days pregnant and since calving, number of lactation), milk production, LW, management practices, environmental aspects and feed composition (e.g.: protein, fibre, lignin, starch, mineral concentrations) from the twelve rations were used as inputs in the CNCPS model. The data were used to examine the model predictions for trial means (four weeks each trial) over all treatments. 7 .3.3 - Model evaluation and statistical analysis The model used in this evaluation was CNCPS Version 5.00.20 (updated August 2002). Model evaluation should include a rigorous statistical component and in this study three different methods have been used to evaluate the CNCPS predictions. Method 1: Linear regression. Most often, predictions are evaluated by regressing actual values versus predicted responses. Method 2: Measures of deviation. Alternatively, Kohn et al. (1998) showed that a measure of how well model predictions fit observed data can be calculated as the root mean square prediction error (RMSPE): RMSPE = √ [ ∑(predicted – actual)2/number of observations] This term is the square root of the estimate of variance of actual values about the predicted values. The RMSPE is comprised of two terms that identify systematic problems with models: the mean bias and the residual error. The mean bias represents the average inaccuracy of model predictions across all data and the residual error is the remaining error in model prediction after accounting for the mean bias. The residual error is also referred to as prediction error excluding mean bias. Mean bias = ∑ (predicted – actual)/number of observations Residual error = ∑ [RMSPE 2 – (mean bias)2] As a summary measure of the relative degree of deviation, either mean bias or RMSPE can be used (Mayer and Butler, 1993). 208 Regressions of the residuals (predicted values minus actual values) against the predicted values were used to identify whether or not the magnitude of the bias increases or decreases with the magnitude of the predicted values (Draper and Smith, 1981). Method 3: Systematic bias. This method of evaluating model prediction is based on milk yield predicted from the first limiting of either ME or MP available. The difference between milk predicted from allowable ME or MP and actual milk (residual) was regressed against dietary variables affecting milk production. These included dry matter intake (DMI), LW change and dietary composition (crude protein (CP); fibre (NDF); ME; and fermentable carbohydrates). 7.4 - Results Mean predictions (Pr) of cow performance, based on model simulations from dietary composition, DMI, and LW, compared with actual (A) values are shown in Tables 7.1 and 7.2. The model under-predicted mean DMI (13.69 vs. actual 15.07 kg DMI/cow.day) and mean dietary ME (9.98 vs. actual 10.5 MJ ME/kg DM) and over- predicted milk production based on ME or MP content of the diet (15.08 vs. actual 14.85 kg milk). ME was first limiting for cows fed restricted and unrestricted pasture allowance and pasture with lotus silage, whereas MP limited milk production by cows given pasture with maize silage. Predicted milk yields by the CNCPS model from diets with sulla or pasture silages was limited to a similar extent by ME and MP. Method 1 – Linear regression of actual against predicted values Predicted values were significantly (P < 0.01) correlated with actual values for DMI (r2 = 0.63;), milk production (r2 = 0.64;), and LW change (r2 = 0.57; Figure 7.1), but predicted ME concentrations were not correlated with values measured by NIRS (r2 = 0.03). However, there were still large unexplained sources of variation (residual variance or mean-square error (MSE)), and the slopes of the regression lines were significantly different than the theoretical value of 1.0 (Table 7.1). Information provided by simple regression analysis can be ambiguous and lack sensitivity (Mitchell, 1997; St- Pierre, 2001), and, thus, was not able to provide a proper interpretation of these relationships. Method 2 – Deviation of predicted from actual values When model predictions were tested using measures of deviation, mean bias was not statistically significant from zero for DMI, milk production, LW change or dietary 209 ME. The residual error terms represent the error in prediction after accounting for the mean bias (Table 7.1; Figure 7.2). The slope of the regression line was significantly (P < 0.01) greater than zero for milk production, LW change and ME concentration, and this difference was close to significance (P < 0.06) for DMI. This indicates a systematic bias, in which the residual differences increase at higher predicted values. For instance, the model under- predicted milk production of cows fed restricted pasture (PR) and overestimated performance when fed unrestricted pasture (FP) and pasture silage. Predictions of milk production for high pasture allowance and maize silage supplements were inconsistent. The relationships are shown graphically in Figure 7.2. Method 3 - Systematic bias Examination of systematic bias provides an insight into factors responsible for the deviation between predicted and actual values. The regressions of the residuals of milk production (limited by energy or absorbed protein) against actual DMI and LW change are plotted in Figure 7.3. A significant slope (different from 0) for the regression indicates a systematic bias in the model prediction, and the r2 represents the fraction of the error (excluding mean bias) that can be explained by the slope bias (Draper and Smith, 1981). Significant biases for CNCPS predicted milk production were observed for DMI and LW change (Figure 7.3; r2 = 0.85 and 0.67, respectively; P < 0.01). The differences between predicted and actual milk production increased by 1.19 kg milk/kg DMI and 11.91 kg milk/kg LW change (slope bias; Figure 7.3). 7.5 - Discussion The CNCPS model is designed to predict nutrient supply, in terms of ME and MP, from rumen parameters, and recent data on digestion kinetics of fresh and conserved forages have been used as model inputs in this study. Because nutrient supply is difficult to measure in grazing animals, validation relies on a comparison of animal performance predicted from these estimates against that observed in practice. The lack of a significant mean bias for any of the parameters examined would suggest very good model prediction. However, the analyses carried out showed that accurate prediction of mean values does not necessarily demonstrate good predictability for individual diets (residual error is large; Table 7.1), and may limit the 210 utility of the CNCPS model for fresh forages. These concerns are illustrated by predictions of milk production. The mean actual and predicted milk yields were similar (14.85 and 15.08 kg/day) and the regression explained 64% of the variance across the diets. However, a regression equation with a slope of + 0.31 (theoretical value = 1.0) and an intercept of 10.14, (theoretical value = 0.0) has little biological meaning. When residuals were regressed against predicted milk production (Method 2), there was no significant mean bias, but residual differences increased for values above and below mean predicted milk production. Further analysis (Method 3) demonstrated a systematic bias in predicted milk production with changes in DMI and LW change. Good predictions were obtained for a small number of diets, whilst a substantial under-prediction was evident for cows fed restricted quantities of pasture, and a substantial over-prediction occurred with high DMI. This approach assumes that actual DMI and LW change are measured without errors and attention should be taken to avoid flawed conclusions because it is difficult to obtain accurate measurements of both parameters, especially with outdoor grazing. The inability of the model to predict milk production either side of the mean is a cause for concern (Table 7.1). The CNCPS model uses inputted milk production as a driving variable to calculate the ME, MP and other nutrients required to achieve that level of production. In this evaluation, the inputted milk production values were those observed for cows fed the experimental diets (Chapters 5 and 6). The nutrients available to meet the requirements for these milk yields are estimated from DMI and predicted dietary ME or MP concentrations, less the amounts required for maintenance and pregnancy. The predicted milk production is determined by the first limiting nutrient (ME or MP). If ME supply is insufficient to meet the inputted milk yield, then the extent of liveweight loss required to fill the ME deficit is calculated. However, the extent of liveweight loss that will actually occur is not predicted. When the supply of available nutrients is insufficient to meet the specified milk yield inputted (for example by feeding a restricted pasture allowance), the model does not allow extra nutrients to be partitioned between body reserves and milk production. This inability of CNCPS to account for partitioning, accounts for the systematic bias in performance identified by the analysis. Model predictions of milk production are much closer when cows do not gain or lose LW (Figure 7.3B). St-Pierre and Thraen (1999) highlighted that the CNCPS is a requirement system, not a response system. CNCPS will calculate the nutrients required to support a given 211 level of milk production and composition (Figure 7.4A). Milk production is an input and is used to estimate DMI but constraints of digesta clearance from the rumen and the ability of cows to convert body reserves into milk may account for poor DMI prediction. In addition, the production responses (e.g.: milk production, live weight change; Figure 7.4B) is a function of feeding value (nutritive value of feeds x intake), animal response (genetic merit), environment and management factors and interactions between those. Nutrient requirement systems (e.g.: CNCPS) are unable to predict responses because they cannot account for partitioning of nutrients between the various productive processes (e.g.: milk production; LW change). The CNCPS model showed poor cow performance prediction when cows are gaining or losing weight, however when the energy balance is close to zero the model presented satisfactory prediction. The model is useful to evaluate diet composition (DM, ME, amino acids, calcium, phosphorus, potassium) in relation to requirements and it identifies the first limiting nutrient for milk production (ME, MP, methionine or lysine). An improved understanding of how effective (eNDF) and physical effective fibre (peNDF) varies with pasture quality and how it is correlated with pH in the rumen is necessary (Kolver and de Veth, 2002). Prediction of rumen parameters (e.g.: peptide balance) are not easily validated because of the difficulty in measuring actual production in vivo. Dry matter intake based on liveweight may not apply to a forage-based system. For example, intakes of low quality hay would be different to intakes of high quality ryegrass/white clover pasture for cows with the same liveweight. CNCPS scenarios evaluation Additional tests of model outputs are based on low, medium and high quality ryegrass pasture fed with and without silages. These pastures will differ in fibre degradation rates as well as fibre and crude protein content. Simulations will predict dairy cows response to pasture composition, typical of changes from spring to summer. Table 7.3 presents feed composition and degradation rates used as inputs to CNCPS. Low, medium and high quality ryegrass pasture was fed alone at restricted (12 kg DM intake cow/day) and un-restricted allowances (18 kg DM intake cow/day). Restricted allowances of the three pasture qualities were also fed with 6 kg DM of pasture, maize, lotus and sulla silages as well as mixtures of maize and sulla silages (PMS, PSM). 212 Scenarios with pasture as a sole diet Table 7.4 presents CNCPS simulations for cows fed low, medium and high quality pasture and at restricted and un-restricted allowances. ME was the first limiting nutrient for cows grazing ryegrass based-pasture, at both allowances irrespective of quality (Table 7.4). With incremental increases in CP content (low to high quality pasture) the model showed a significant decrease of the percentage of metabolisable protein (MP) coming from bacterial flow and an increased proportion of MP from undegraded feed. Higher concentrations of CP in the pasture decreased total bacterial N production and resulted in large increases in cost of urea excretion, reaching 12 MJ/day (Tables 2.13; 7.4). Because ME was limiting, cow milk production only increased by a small amount in response to pasture protein concentration, so the cows fed high quality pasture had very high N losses to urea. The milk production response with increasing pasture quality was associated with a lower dietary NDF, but this caused a large increase in environmental pollution with urinary-N. The total degradable intake protein (DIP, % of CP) increased with increased pasture quality in association with decreased forage passage rates. DIP estimation is similar to the CP effective degradability calculated in previous chapter (Table 6.6). Predicted pH values agree well with observed values in other pasture studies (Kolver and Muller, 1998; Kolver and de Veth, 2002) but were higher than values reported in Chapter 6 (Table 6.10). Scenario of low pasture quality at restricted allowance and single silage supplementation ME was the first limiting nutrient for cows fed pasture with lotus or sulla silages but MP was the first limiting nutrient with the pasture silage and maize silage treatments (Table 7.5). Pasture and maize silages had a low CP content relative to other silages and the maize silage diet provided only 91% of rumen N requirements (Table 7.5). These simulations show no benefits in supplementing low quality pasture with any of the silages used in simulations. Provided the cow could eat 18 kg pasture (48% NDF) per day, predicted performance from pasture alone (Table 7.4) exceeded predictions with restricted pasture and silages (Table 7.5). 213 Scenario of medium pasture quality at restricted allowance and single silage supplementation In this scenario, MP was the first limiting nutrient for cows fed medium quality pasture (MQP) with pasture silage, but ME was the first limiting nutrient for milk production with other silages (Table 7.6). Again, maize silage resulted in the lowest CP and soluble-CP in the diet but highest NFC, which contributed to a higher daily total bacterial N (312 g) compared to other diets (275 – 293 g). The cows fed MQP supplemented with maize silage would excrete less urea than others treatments (150 versus 226 – 322 g/day) and could potentially produce most milk (Table 7.6). Silages did not complement restricted pasture for milk production, compared to un-restricted MQP, but if pasture supply were insufficient, CNCPS suggested maize silage to be the optimal supplement (Table 7.6). Scenario of high pasture quality at restricted allowance and single silage supplementation When high quality pasture (HQP) was fed with single silages, MP was again the first limiting nutrient for all treatments except lotus silages (Table 7.7). HQP plus maize silage again had highest NFC and lowest soluble-CP contents compared to other silage treatments (Table 7.7). It is likely that this maintained good bacterial growth (278 g/day), compared to-un-restricted HQP with 263 g total bacterial N per day (Table 7.4). When HQP was fed with lotus silage, predicted cost for urea synthesis were 5.6% ME intake. This is similar to values for HQP fed alone (5.7% of ME intake) and probably overestimates true costs of urea disposal because the CT in lotus will divert feed N to faeces rather than urine (Waghorn et al., 1994). None of the silages fed with HQP would result in better cow performance than un-restricted HQP fed alone (Table 7.4) according to CNCPS predictions. 214 Scenario of low, medium and high quality pasture (LQP, MQP and HQP) with mixture of maize and sulla silages. In this scenario, MP was the first limiting nutrient for cows fed high quality pasture (HQP) with silage mixtures, but ME was the first limiting nutrient for milk production with LQP and MQP with silage mixtures (Table 7.8). Again, diets with higher CP concentrations (HQP with PMS or PSM) resulted in higher cost for urea disposal and higher urinary N excretion than other treatments. All diets resulted in similar predicted milk production, passage rates from rumen and pH. Future studies should include long-term trials with continuous assessment of actual and predicted DM intakes, milk production and live weight change in association with dietary composition to enable a rigorous evaluation of model predictions. This information may provide credibility for the model predictions and add confidence to predictions of rumen function in cows fed fresh forage diets. 7.6 - Conclusion The need for mechanistic models to predict responses of lactating cows fed pasture and forage supplements have been indicated, and the CNCPS model was used to evaluate dairy cow performance. This evaluation included feed composition, soluble fraction (A) and degradation rates from in sacco data. The results of model predictions were analysed with a more rigorous statistical analysis than simple linear regression of actual versus predicted values and demonstrated systematic biases in the predictions. Milk production was either over- or under-estimated, depending on the level of feeding. This probably results from model inability to account for partitioning of nutrients between milk production and liveweight change. Evaluation of low, medium and high quality ryegrass pasture fed alone and with supplements demonstrated contrasting outputs, especially for nitrogen. The poor suitability of feeding maize silage with low quality pasture was not demonstrated by model predictions of cow performance despite insufficient N for microbial growth. The hypothesis was proven in part. The model was able to predict cow performance when intakes were restricted and it did indicate first limiting nutrients for all diets but and milk production predicted from pasture/silage diets appeared unrealistically high, compared to measurements presented in Chapter 5 and 6. The model predicted rumen parameters, especially nitrogen fluxes but the validity may be reduced if model behaviour differed substantially from cow performance. 215 TABLE 7.1 – Actual (A) and CNCPS predictions (Pr), regressions, correlations, bias and errors for dry matter intake (DMI), milk production, live weight (LW) change (all kg/cow.day) and dietary metabolisable energy (ME) concentration (MJ ME/kg feed DM). Predictions for milk production are based on the first limiting factor: allowable metabolisable energy (ME) or allowable metabolisable protein (MP) for all diets. Method 1 (Linear regression) Method 2 (Measures of deviation) Mean value SD a A versus Pr r2 MSE b P1 Mean bias c Residual error d RMSPE e r2 P2 DMI A 15.07 2.2 Pr 13.69 y = –12.79 + 2.04x 0.63 1.98 < 0.01 -1.40ns 1.6 2.1 0.31 0.06 Milk A 14.85 1.6 production Pr 15.08 y = 10.14 + 0.31x 0.64 0.96 < 0.01 0.23ns 2.75 2.76 0.89 < 0.01 LW change A 0.01 0.2 Pr 0.15 y = – 0.03 + 0.28x 0.57 0.02 < 0.01 0.14ns 0.39 0.42 0.9 < 0.01 Diet ME A 10.5 0.3 concentration Pr 9.98 y = 12.15 – 0.16x 0.03 0.62 ns -0.54ns 10.52 10.53 0.58 < 0.01 a Standard deviation. b Mean square error (estimate of variance). c Mean predicted minus mean actual. t-test (5%, n-2) for mean bias different from zero. d Model prediction error excluding that due to the mean bias. e Root mean square prediction error. P1: P value of F-statistic for slope = 1. P2: P value of F-statistic for slope ≠ 0. ns = not significant. 216 TABLE 7.2 – Actual and CNCPS predictions for dry matter intake (DMI), milk yield (MY) (both kg/cow.day), dietary metabolisable energy (ME) concentration (MJ ME/kg feed DM) and first limiting nutrient (ME or MP) for individual diets (Chapters 5 and 6). DMI MY ME Limiting Actual CNCPS Actual CNCPS Actual CNCPS ME or MP Chapter 5 treatments: FP 18.5 15.1 17.0 20.5 10.1 9.7 ME RP 12.5 13.4 13.1 11.0 10.0 10.2 ME PP 17.0 14.1 15.0 18.5 11.1 10.0 MP PM 16.6 13.9 15.0 17.5 10.4 10.0 MP PL 17.2 15.0 17.2 19.2 10.9 9.8 ME PS 15.7 14.0 15.1 15.1 10.2 9.4 ME Chapter 6 treatments: FP 15.7 14.4 17.2 16.0 10.7 10.4 ME RP 10.4 12.8 13.2 7.5 10.6 10.5 ME PMS 14.6 12.9 14.3 14.7 10.5 10.0 MP PSM 14.4 13.0 13.7 14.2 10.6 10.0 MP PS 13.9 12.6 13.7 12.0 10.6 9.5 MP PM 14.3 12.9 13.7 14.2 10.6 10.3 MP Average 15.1 13.7 14.9 15.0 10.5 10.0 Abbreviations: Chapter 5 treatments, FP, full pasture (50 kg pasture DM/cow.day); RP, restricted pasture (25 kg pasture DM/cow.day); PP, RP + 5 kg DM of pasture silage; PM, RP + 5 kg DM of maize silage; PLS, RP + 5 kg DM of lotus silage; PS, RP + 5 kg DM of sulla silage. Chapter 6 treatments, FP: full pasture (38 kg pasture DM/cow.day); RP, restricted pasture (18 kg pasture DM/cow.day); PMS, RP + 4 kg maize + 2 kg sulla silages/cow.day; PSM, RP + 4 kg sulla + 2 kg maize silages/cow.day; PS, RP + 6 kg sulla silage/cow.day; PM, RP + 6 kg maize silage/cow.day. MP, metabolisable protein. 217 TABLE 7.3 – Feed composition and degradation rates used into CNCPS scenarios evaluation. General characteristics Feed name DM (%) NDF (%DM) peNDF (%NDF) Lignin (%NDF) Fat (%DM) Ash (%DM) Starch (%NFC) Pasture quality: Low 20.0 48.0 40.0 5.14 3.0 9.0 45.0 Medium 17.0 46.5 40.0 5.80 4.0 9.4 48.0 High 15.0 40.0 60.0 6.00 6.9 10.7 48.0 Silage: Pasture 32.6 46.8 95.0 5.50 2.6 7.2 63.0 Maize 33.7 44.5 85.0 10.59 3.0 4.0 80.0 Lotus 33.1 35.5 80.0 20.30 3.2 10.0 64.0 Sulla 35.4 36.2 92.0 20.00 5.2 10.1 64.0 Energy and protein values Feed name CP (%DM) UIP (%DM) Sol-P (%CP) NPN (%Sol-P) NDFIP (%CP) ADFIP (%CP) Pasture quality: Low 15.0 29.3 54.0 4.76 24.0 2.2 Medium 20.0 21.9 54.0 3.41 12.5 2.6 High 25.0 0.0 54.0 2.44 4.54 1.65 Silage: Pasture 15.6 22.4 55.0 100.0 31.0 10.0 Maize 6.9 20.9 10.0 100.0 16.0 7.0 Lotus 23.4 17.6 51.0 28.0 13.0 9.0 Sulla 21.2 38.4 55.4 28.0 15.0 10.0 Degradation rates Feed name Degradation rates (%.h-1) Carbohydrate Protein Pasture quality: A B1 B2 B1 B2 B3 Low 350 40.0 9.5 200 14.0 2.00 Medium 350 45.0 11.00 200 16.0 2.00 High 350 21.5 12.0 200 18.0 2.00 Silage: Pasture 10 25.0 4.7 200 10.4 1.75 Maize 10 30.0 4.1 300 3.4 0.25 Lotus 10 25.0 12.3 150 15.0 1.25 Sulla 10 25.0 6.3 150 7.4 1.25 Abbreviations see text and Table 5.5. 218 TABLE 7.4 – Pasture diets; CNCPS predictions of nutrient composition, cow performance and rumen parameters. Restricted pasture allowance Un-restricted pasture allowance Simulated pasture DMI, kg/cow.day 12 12 12 18 18 18 Pasture quality Low Medium High Low Medium High Nutrient composition ME, MJ/kg DM 10.7 11.0 12.0 10.3 10.7 11.8 CP, g/100g DM 15 20 25 15 20 25 Soluble CP, %CP 54 54 54 54 54 54 NDF, g/100 g DM 48.0 46.5 40.0 48.0 46.5 40.0 peNDF, g/100 g DM 19 19 24 19 19 24 Total NFC, g/100 g DM 28.6 22.6 18.5 28.6 22.6 18.5 Total fat, g/100 g DM 3.0 4.0 6.9 3.0 4.0 6.9 Performance predictions ME allowable milk, kg/day 12.3 12.8 14.6 23.6 24.4 27.3 MP allowable milk, kg/day 16.8 16.3 16.6 26.4 26.7 28.9 Daily weight change due to reserves, kg/day 0.0 0.0 0.3 1.8 1.9 2.3 Rumen digestion, metabolism and passage MP from bacteria, g/day 902 796 688 1296 1143 985 MP from undegraded feed, g/day 399 468 564 681 842 1060 MP from undeg. feed, %MP supplied 31 37 45 34 42 52 Total degradable intake protein, %CP 74.4 80.3 85.8 72.2 78.1 83.4 Ruminal N balance, % of requirement 103 157 266 104 156 264 Total bacterial nitrogen, g/day 240 212 183 346 305 263 Urea cost, MJ/day 1 3.3 6.9 3.3 6.8 12 Urea cost, %ME intake 0.8% 2.5% 4.8% 1.8% 3.5% 5.7% Excess N excreted, g/day 42 151 339 123 285 563 Liquid passage rate, %.h-1 9 9 9 11.3 11.3 11.3 Forage passage rate, %.h-1 5.38 5.42 5.11 6.74 6.79 6.40 Predicted ruminal pH 6.24 6.21 6.44 6.24 6.21 6.44 Abbreviations see text and previous tables. 219 TABLE 7.5 - Low quality pasture (LQP) with single silages; CNCPS predictions of nutrient composition, cow performance and rumen parameters. LQP + pasture silage LQP + maize silage LQP + lotus silage LQP + sulla silage Simulated pasture DMI, kg/cow.day 12 12 12 12 Simulated silage DMI, kg/cow.day 6 6 6 6 Simulated total DMI, kg/cow.day 18 18 18 18 Nutrient composition ME, MJ/kg DM 10.13 9.80 10.00 9.99 CP, g/100g DM 15.2 12.3 17.8 17.1 Soluble CP, %CP 54.3 45.8 52.7 54.6 NDF, g/100 g DM 47.6 46.8 43.8 44.1 peNDF, g/100 g DM 28 25 22 24 Total NFC, g/100 g DM 29.9 33.3 29.4 29.2 Total fat, g/100 g DM 2.9 3.0 3.1 3.7 Performance predictions ME allowable milk, kg/day 23.3 22.4 22.4 22.6 MP allowable milk, kg/day 22.4 21.9 23.4 22.9 Daily weight change due to reserves, kg/day 1.7 1.5 1.6 1.6 Rumen digestion, metabolism and passage MP from bacteria, g/day 1192 1153 1147 1124 MP from undegraded feed, g/day 610 661 713 722 MP from undeg. feed, %MP supplied 34 36 38 39 Total degradable intake protein, %CP 71.3 65.9 73.8 71.8 Ruminal N balance, % of requirements 114 91 134 130 Total bacterial nitrogen, g/day 318 308 306 300 Urea cost, MJ/day 2.4 2.3 4.3 3.6 Urea cost, %ME intake 1.3% 1.3% 2.4% 2.0% Excess N excreted, g/day 125 44 193 173 Liquid passage rate, %.h-1 11.3 11.3 11.3 11.3 Pasture passage rate, %.h-1 6.74 6.74 6.74 6.74 Silage passage rate, %.h-1 5.25 5.58 6.11 5.82 Predicted ruminal pH 6.46 6.46 6.37 6.44 Abbreviations see text and previous tables. 220 TABLE 7.6 - Medium quality pasture (MQP) with single silages; CNCPS predictions of nutrient composition, cow performance and rumen parameters. MQP + pasture silage MQP + maize silage MQP + lotus silage MQP + sulla silage Simulated pasture DMI, kg/cow.day 12 12 12 12 Simulated silage DMI, kg/cow.day 6 6 6 6 Simulated total DMI, kg/cow.day 18 18 18 18 Nutrient composition ME, MJ/kg DM 10.39 10.32 10.25 10.26 CP, g/100g DM 18.5 15.6 21.1 20.4 Soluble CP, %CP 54.3 47.5 52.9 54.5 NDF, g/100 g DM 46.6 45.8 42.8 43.1 peNDF, g/100 g DM 27 25 22 24 Total NFC, g/100 g DM 25.9 29.3 25.4 25.2 Total fat, g/100 g DM 3.5 3.7 3.7 4.4 Performance predictions ME allowable milk, kg/day 23.7 23.9 22.7 22.8 MP allowable milk, kg/day 22.5 24.5 24.1 23.4 Daily weight change due to reserves, kg/day 1.8 1.8 1.6 1.6 Rumen digestion, metabolism and passage MP from bacteria, g/day 1099 1171 1054 1030 MP from undegraded feed, g/day 696 739 827 830 MP from undeg. feed, %MP supplied 39 39 44 45 Total degradable intake protein, %CP 75.7 72.2 77.2 75.7 Ruminal N balance, % of requirements 150 117 181 174 Total bacterial nitrogen, g/day 293 312 281 275 Urea cost, MJ/day 5.2 2.9 7.5 6.9 Urea cost, %ME intake 2.8% 1.6% 4.1% 3.7% Excess N excreted, g/day 226 150 322 292 Liquid passage rate, %.h-1 11.3 11.3 11.3 11.3 Pasture passage rate, %.h-1 6.79 6.79 6.79 6.79 Silage passage rate, %.h-1 5.25 5.58 6.11 5.82 Predicted ruminal pH 6.46 6.46 6.35 6.42 Abbreviations see text and previous tables. 221 TABLE 7.7 - High quality pasture (HQP) with single silages; CNCPS predictions of nutrient composition, cow performance and rumen parameters. HQP + pasture silage HQP + maize silage HQP + lotus silage HQP + sulla silage Simulated pasture DMI, kg/cow.day 12 12 12 12 Simulated silage DMI, kg/cow.day 6 6 6 6 Simulated total DMI, kg/cow.day 18 18 18 18 Nutrient composition ME, MJ/kg DM 11.12 11.06 10.98 10.99 CP, g/100g DM 21.9 19.0 24.5 23.7 Soluble CP, %CP 54.2 48.7 53.0 54.4 NDF, g/100 g DM 42.3 41.5 38.5 38.7 peNDF, g/100 g DM 31 29 25 27 Total NFC, g/100 g DM 23.2 26.6 22.7 22.5 Total fat, g/100 g DM 5.5 5.6 5.7 6.3 Performance predictions ME allowable milk, kg/day 25.6 25.9 24.7 24.8 MP allowable milk, kg/day 22.7 24.3 24.9 24.2 Daily weight change due to reserves, kg/day 2.1 2.1 1.9 2.0 Rumen digestion, metabolism and passage MP from bacteria, g/day 969 1041 924 900 MP from undegraded feed, g/day 800 829 967 963 MP from undeg. feed, %MP supplied 45 44 51 52 Total degradable intake protein, %CP 80.3 78.2 81.2 80 Ruminal N balance, % of requirements 214 162 268 257 Total bacterial nitrogen, g/day 258 278 246 240 Urea cost, MJ/day 8.9 6.5 11.1 10.5 Urea cost, %ME intake 4.4% 3.3% 5.6% 5.3% Excess N excreted, g/day 377 268 515 473 Liquid passage rate, %.h-1 11.3 11.3 11.3 11.3 Pasture passage rate, %.h-1 6.4 6.4 6.4 6.4 Silage passage rate, %.h-1 5.25 6.58 6.11 5.82 Predicted ruminal pH 6.46 6.46 6.46 6.46 Abbreviations see text and previous tables. 222 TABLE 7.8 – Low, medium and high quality pasture (LQP, MQP and HQP) with mixture of maize and sulla silages. CNCPS predictions of nutrient composition, cow performance, and rumen parameters. LQP + PMS (4M2S) MQP + PMS (4M2S) HQP + PMS (4M2S) LQP + PSM (4S2M) MQP + PSM (4S2M) HQP + PSM (4S2M) Simulated pasture DMI, kg/cow.day 12 12 12 12 12 12 Simulated maize silage DMI, kg/cow.day 4 4 4 2 2 2 Simulated sulla silage DMI, kg/cow.day 2 2 2 4 4 4 Simulated total DMI, kg/cow.day 18 18 18 18 18 18 Nutrient composition ME, MJ/kg DM 10.04 10.30 11.04 10.02 10.28 11.01 CP, g/100g DM 13.9 17.2 20.6 15.5 18.8 22.1 Soluble CP, %CP 49.4 50.3 50.9 52.2 52.6 52.8 NDF, g/100 g DM 45.9 44.9 40.6 45.0 44.0 39.7 peNDF, g/100 g DM 25 25 28 24 24 28 Total NFC, g/100 g DM 31.9 27.9 25.2 30.6 26.6 23.9 Total fat, g/100 g DM 3.2 3.9 5.8 3.5 4.2 6.1 Performance predictions ME allowable milk, kg/day 22.9 23.6 25.5 22.8 23.2 25.2 MP allowable milk, kg/day 23.9 24.0 24.2 23.4 23.7 24.1 Daily weight change due to reserves, kg/day 1.6 1.8 2.1 1.6 1.7 2.0 Rumen digestion, metabolism and passage MP from bacteria, g/day 1218 1124 994 1171 1077 947 MP from undegraded feed, g/day 679 766 868 699 796 913 MP from undeg. feed, %MP supplied 36 41 47 37 42 49 Total degradable intake protein, %CP 68.3 73.6 78.9 70.2 74.8 79.5 Ruminal N balance, % of requirements 102 132 187 115 151 218 Total bacterial nitrogen, g/day 325 300 265 312 287 253 Urea cost, MJ/day 2.8 4.2 7.8 2.7 5.5 9.2 Urea cost, %ME intake 1.5% 2.3% 3.9% 1.5% 3.0% 4.6% Excess N excreted, g/day 98 190 324 133 237 392 Liquid passage rate, %.h-1 11.3 11.3 11.3 11.3 11.3 11.3 Pasture passage rate, %.h-1 6.74 6.79 6.40 6.74 6.79 6.40 Maize silage passage rate, %.h-1 5.58 5.58 5.58 5.58 5.58 5.58 Sulla silage passage rate, %.h-1 5.82 5.82 5.82 5.82 5.82 5.82 Predicted ruminal pH 6.46 6.46 6.46 6.46 6.44 6.46 Abbreviations see text and previous tables. 223 FIGURE 7.1 - Actual versus predicted values for A: dry matter intake (DMI), B: milk production and C: live weight (LW) change using CNCPS. (◊) = individual treatments. PR = restricted pasture, FP = unrestricted pasture. F P R P10 12 14 16 18 20 10 12 14 16 18 20 Predicted DMI (kg DM/day) A ctua l DMI (kg DM/ da y) A Actual = predicted FP RP under over estimated 4 12 20 4 1 2 20 Predicted milk (kg/day) A ctua l m ilk (k g/ da y) B Actual = p redicted FP RP -1.0 -0.5 0.0 0.5 1.0 -1 -0.5 0 0.5 1 Predicted LWchange (kg) A ctua l LW cha ng e (kg ) Actual = predicted C 224 FIGURE 7.2 - Residual (predicted – actual) versus predicted values for A: dry matter intake (DMI), B: milk production and C: live weight (LW) change using CNCPS. (◊) = individual treatments. Line ( _ _ _ ) indicates mean bias. PR = restricted pasture, FP = unrestricted pasture. P value of F-statistic for slope ≠ 0. P R F P A Slope = -1.04 r2 = 0.31 P = 0.06 -3.5 -2 -0.5 1 2.5 12.5 13 13.5 14 14.5 15 Predicted DMI (kg/day) Res id ual s DMI, k g/ da y B FP PR Slope = 0.69 r2 = 0.89 P < 0.01 -6 -2 2 6 6 1 2 1 8 Predicted milk (kg/day) Res id ual s m ilk , k g/ da y 24 PR C FP Slope = 0.72 r2 = 0.90 P < 0.01 -1.0 -0.5 0.0 0.5 1.0 -1 -0.5 0 0.5 1 Predicted LW change (kg/day) Res id ual s LW c han ge , k g/ da y 225 FIGURE 7.3 - Milk production (kg/cow.day) predicted by the CNCPS model minus actual milk production (Y axis) versus A: actual dry matter intake (DMI) and B: liveweight (LW) change. PR = restricted pasture, FP = unrestricted pasture. A FP PR Slope = 1.19 r2 = 0.85 -6 -3 0 3 6 10 12 14 16 18 20 Actual DMI (kg/cow.day) Pr ed ic ted - ac tual , kg m ilk /c ow .da y B FP PR Slope = 11.91 r2 = 0.67 -6 -3 0 3 6 -0.5 -0.3 -0.1 0.1 0.3 Actual LW change (kg/cow.day) Pre di cte d - a ctua l, kg m ilk /c ow .da y 226 FIGURE 7.4 - Schematic representation of the CNCPS requirement-based system (A) and B: production response system. Adapted from St-Pierre and Thraen (1999). (A) Milk production Liveweight Growth Pregnancy Energy requirements Protein requirements Energy supply Protein supply Pasture Silages Factors Nutrient requirements Nutrient supplies Feeds (B) Pasture Energy supply Protein supplySilages Response Milk components Liveweight change Manure Feeds Nutrients Animal Products 227 Chapter 8 - General discussion The research undertaken for this thesis represents a comprehensive analysis of ryegrass maturation on nutritive value, covering chemical composition, rates of degradation and products of digestion. Th is has formed a data base for New Zealand pastoral systems and is intended to provide information to improve supplementation of dairy cows. The well known and rapid chan ges in ryegrass composition and feeding value with maturity require constant re-evaluation of supplementation to meet feeding objectives. It is not possible to provide a balanced diet without knowing the digestion kinetics of ryegrass. Part of this research involved separation of leaf, stem and flower to provide basic data concerning the contribution of each component of ryegrass. One objective of this work was to provide data enabling appropriate supplementation of summer ryegrass dominant pastures for dairy cows. Poor nutritive value (NV) of pasture in summer contributes to the decline in milk production and can also lead to liveweight losses and shortened lactation especially with restricted pasture availability. Two dairy trials were carried out to examine effects of contrasting silage supplements for lactating cows fed pasture in mid summer. The silages included conventional (maize, pasture) and legume (lotus and sulla) silages as single supplements and mixtures to provide data showing cow responses and substitution. These trials provided background informatio n against which future supplements can be designed to complement pasture on offer. There are several models to aid interp retation of cow grazing data, and the CNCPS system was chosen to compare measured and predicted data because it predicts the nutrient supply from digestion kinetics data. The evaluation showed energy to be a prime limitation when pasture was fed, but the model did not take into account effects of liveweight change. The model provides a good basis for diet evaluation using forages with contrasting nutritive characteristics but ruminal parameters in cows fed fresh forages have required detailed inputs of digestion kinetics which are not always available. Examples include a poor relationship between rumen effective fibre and pH, inadequate estimation of DM intake (based on liveweight) and complex feed degradability component inputs. 228 An important aspect of the in vitro and in sacco experiments was the use of the fresh mincing procedure to provide material for in vitro and in sacco incubations which mimicked cow digesta. The data from in vitro and in sacco incubations were complementary and provided a simple and logical data set to show the way maturation influenced microbial digestion. Chemical composition, in sacco digestion kinetics and in vitro products of digestion helped expl ain effects of maturation, and this was made more obvious when grass leaf, stem and flower were incubated separately. The slow degradation of mature grass components highlighted the importance of clearance of undigested residues from the rumen, and provided good evidence for diminished feeding value (FV) from flowering ryegrass. The energy required for mincing ryegrass of increasing maturity was not measured, but increasing effort would be required for particle breakdown of maturing ryegrass. Slow particle size reduction of mature forage will reduce microbial growth and outflow from the rumen, with lower intakes and nutrient yields per unit of intake. Cow productivity could only be maintained by substituting mature grass with more rapidly digested feeds. The information presented here will affect both the level and quality of supplements needed to maintain production from cows fed summer pasture. Comparison of initial cutting dates showed minor effects on the rate of change in chemical composition, with late cut grass maturing slightly more rapidly than early cut grass. This affected fibre but not CP or NSC content. Protein degradation was not affected by maturation, but a higher proportion was released into the soluble (A) fraction with mature grass. This appeared to indicate more extensive cell rupture of mature grass and is likely to reflect the in vivo situation where more extensive chewing is needed to swallow mature, compared to succulent forages. In contrast to protein, fibre degradation rate was halved as ryegra ss matured, probably as a result of cross linkages between fibre components and lignific ation. Estimated ME values for ryegrass were reduced from about 12.8 to 8.8 MJ/kg DM as ryegrass became mature, but effects of maturation on FV would be substantially greater because of effects on intake. The ammonia production in vitro was a function of grass CP content, extent of release into the soluble fraction and microbial utilisation. One effect of maturation was a brief period of ammonia surplus followed by insufficiency for microbial growth, but the rate and amounts of VFA produced were not affected by forage CP content or ammonia concentration. This suggests forage CP may have a greater effect on ruminant performance through provision of plant and microbial amino acids for absorption, than through microbial VFA production. 229 Maturation, including effects of leaf, stem and flower components had little effect on proportions of VFA. Rates of VF A production were similar for young, medium and mature ryegrass, possibly because all we re minced to a similar particle size distribution, but this will be similar to fora ge eaten by grazing cows . Chewing will have a significant impact on rate and extent of fe rmentation as well as clearance from the rumen. The grazing trials showed exceptional responses with lotus silage fed to cows. Future research would benefit cow nutrition if the reasons for the large responses to lotus could be explained. The maize silage supplement resulted in a moderate increase in milk production, but this would have been lower with pasture containing less nitrogen, typical of a normal summer. In contrast, mixed silages such as maize and sulla silages gave good responses and a more stable rumen environment than either maize or sulla silage fed with pasture. Cows preferred the mixed silages to a single silage, except for lotus. In sacco incubations suggested a less active fermentation with increasing proportion of maize silage in the diet. This may have important implications for provision of nutrients for cows but more studies are needed to understand effect of in vivo environment on digestion rate and nutrient release. The CNCPS model does provide a good method for diet evaluation, to reduce the numbers and enable a better design of cow trials. The model gives good predictions of milk production if cow energy balance is close to zero, as with the trials reported here. More investigations of model capability for use under grazing systems are necessary, especially using data from whole lactation trials. Long term, systems trials have the advantage of quantifying carry over effects associated with short trials. Ryegrass maturation data will form the foundation for an improved understanding of the NV of ryegrass pastures. The data do not include impacts of grass quality on intake, and this is a serious limi tation to use of the CNCPS model and to all studies of FV from pasture. However, th e effects of maturation on degradation of protein and fibre have been clearly defined and show a surprising resilience for VFA production by bacteria when forage N is limiting. Cow trials have given some useful information about silage supplementation but future work should provide mixed supplements to form an appropriate proportion of the diet, dependent on the quality of available pasture. 230 There have been some limitations to the work presented in this thesis, due to insufficient knowledge in early stages of this work and insufficient time for more measurements. It would have been worthwh ile to measure proportions of leaf, stem and flower from the pastures with increasing maturity, and this may be done in future. An additional useful measurement would be the energy required to mince grasses at different stages of maturity. These measurem ents may form the basis of future trials if intake is to be targeted as a topic for balancing pasture with supplements. Analyses focused on a 6% fractional ou tflow rate from the rumen and although some analyses used contrasting values (e .g.: 2%), this aspect of rumen function probably requires more consideration, especially when comparing fibre and protein. Outflow rates affect both intake and nutrient supply through effective degradability. The CNCPS model is sensitive to feed supply and although intakes are not well defined by the model, the issues of feed availability under New Zealand grazing systems need to be accommodated when the model is used. In vitro data presented here would be more valuable if it could be combined with microbial growth. Future in vitro studies with forages should incorporate microbial growth to give a more comprehensive and meaningful discussion of products of digestion. It is important that these factors be incorporated in the model to improve the rumen aspects, especially as DNA-based systems are showing promise for rapid and inexpensive measurements of bacterial numbers. Future research should investigate sulla characteristics, especially agronomic requirements, harvesting and chemical composition. This Mediterranean legume has good potential to provide high yields of high quality forage, and is being used by several commercial farmers. However, little information has been published and more trials should test the benefits of both the NSC and condensed tannins for dairy cow nutrition and performance. The information presented in this thesis showed that mixed forages (sulla, maize and lotus) are preferred by cows, to indivi dual forage supplements and farmlet trials over a whole lactation should investigate benefit of mixtures on cow nutrition, performance and sustainability. Use of an autumn calving herd would enable effects of ryegrass maturation to be avoided and au tumn pasture would provide a high quality diet for cows at peak lactation. Supplements would have a major role in winter when grass availability is low and could give ideal opportunities to compare contrasting silages. The CNCPS model might be used to balance diets in long term trials and 231 economic analyses would determine profitabil ity and sustainability of farm systems. These analyses apply to all farming systems where choices of diets are available. 232 List of publications derived from this PhD project Chaves, A.V.; Waghorn, G.C.; Brookes, I.M Wood ward, S.L. 2003. Empirical evaluation of the NRC dairy cattle model to predict performance of dairy cows fed pasture with corn and sulla silages. Proceedings of the VI International Symposium on the Nutrition of Herbivores : in press. Chaves, A.V.; Burke, J.L.; Waghorn, G.C.; Brookes, I.M.; Woodward, S.L. 2003. Empirical assessment of the CNCPS model to predict performance of dairy cows fed pasture with silages supplements. Proceedings of the New Zealand Society of Animal Production 63: 91-9 5. Chaves, A.V., Waghorn, G.C.; Brookes, I.M.; Hedderley, D. 2002. Digestion kinetics of ryegrass. Proceedings of the New Zealand Society of Animal Production 62: 157-1 62. Chaves, A.V.; Woodward, S.L.; Waghorn, G.C.; Brookes, I.M.; Holmes, C.W.; Laboyrie, P.G. 2002. Post-peak supplementation of pasture fed dairy cows with sulla and maize silages. Proceedings of New Zealand Grassland Association 64: 125-12 8. Chaves, A.V.; Waghorn, G.C.; Tavendale, M.H. 2002. A simplified method for lignin measurement in a range of forage species. Proceedings of New Zealand Grassland Association 64: 129-13 3. Chaves, A.V.; Waghorn, G.C.; Assis, A.G.; Oliveira, J.S. 2002. In sacco digestion kinetics of Panicum maximum cv. Tanzânia, P . maximum cv. Mombaça, and Oats ( Avena sativa L.) prepared by mincing fresh material. Proceedings of the XXXIX annual meeting of Brazilian Society of Animal Science In.: Ruminant nutrition session. Recife, PE, 2002 - Editor UFRPE. Chaves, A.V.; Waghorn, G.C.; Assis, A.G.; Oliveira, J.S.; Attwood, G. 2002. In vitro nutrient release from Panicum maximum cv. Tanzânia, P. maximum cv. Mombaça, and Oats ( Avena sativa L.) prepared by mincing fresh material. Proceedings of the XXXIX annual meeting of Brazilian Society of Animal Science In.: Ruminant nutrition session. Recife, PE, 2002 - Editor UFRPE. 233 Chaves, A.V.; Waghorn, G.C.; Woodward, S.L.; Brookes, I.M. 2002. Digestion kinetics of pasture and forage mixed rations prepared by mincing fresh material. Journal of Animal Science 8 0 (Suppl. 1)/ Journal of Dairy Science 85 (Suppl. 1): 396. Chaves, A.V.; Woodward, S.L.; Waghorn, G.C.; Brookes, I.M.; Holmes, C.W.; McNabb, W. 2002. Balancing diets for cows grazing pa sture post-peak lactation using forage mixed rations. Journal of Animal Science 80 (Suppl. 1)/ Journal of Dairy Science 85 (Suppl. 1): 204. Chaves, A.V.; Waghorn, G.C.; Brookes, I.M.; Burke, J.L. 2001. Digestion kinetics of mature grasses. Proceedings of the New Zealand Society of Animal Production 61: 8-12. Woodward, S.L.; Chaves, A.V.; Waghorn, G.C.; Laboyrie, P.G. 2002. Supplementing pasture-fed dairy cows with pasture silage, ma ize silage, Lotus silage or sulla silage in summer – does it increase production? Proceedings of New Zealand Grassland Association 64: 85-8 9. Burke, J.L; Waghorn, G.C.; Chaves, A.V. Invited paper 2002. Improving animal performance using forage based diets. Proceedings of the New Zealand Society of Animal Production 6 2 : 267-27 2. Woodward, S.L.; Chaves, A.V.; Grayling, L.; Waghorn, G.C. 2001. Specialist silages for increased milksolids production in late summer-autumn. Proceedings of the Ruakura Farmers’ Conference 53: 74-75. 234 APPENDICES 235 APPENDIX 1: I n vitro incubation. McDougall’s buffer Recipe for McDougall’s artificial saliva used as a buffer in the in vitro incubations. The buffer was prepared the day prior to incubations to ensure the components were dissolved prior to gassing with CO 2 on the day of incubation. One litre was sufficient for 3 x 24 containers with six standards. NaHCO 3 9.8 g/L Na2 HPO 4 .12H 2 O 9.3 g/L NaCl 0.47 g/L KCl 0.57 g/L CaCl 2 anhydrous 0.04 g/L MgCl 2 anhydrous 0.06g/L Reducing agent The reducing agent was cysteine sulphide immediately prepared before the incubations as follows. Mix in order: 0.315 g cysteine hydrochloride 48 mL water 2 mL 1N NaOH 0.315 g sodium sulphide (crystals to be rinsed in water and blotted dry to adjust weight) In vitro incubation procedure 1. Prepare forages freeze and weigh about 2.5g wet weight (0.5g DM) into bottles. 2. Place bottle in freezer after weighing on the previous day. Put caps on bottles in freezer. 3. Bottles are weighed empty without caps. When washing bottles never use detergent. Feed cows first thing in morning so you sample liquor 2-3 hours after onset of eating. 4. Prepare Excel sheet with bottle weights for data entry. APPENDICES 236 5. Start numbering microcentrifuge tubes. Day of incubation: 1. Turn incubator on previous day so the temperature is stable for incubation. 2. Place buffer in a bucket of hot water, wa rm to 40°C whilst gassing for 45 minutes before dispensing into bottles. 3. Put samples in incubator (up to 90 mi nutes before adding rumen inoculant). 4. The rumen samples, collected by hand from the rumen into a cheese cloth bag, were strained through cheesecloth into a warm thermos® flask using a funnel. The flask was filled to minimise exposure of the liquor to air and taken directly to the incubation room. 5. Make sufficient reducing agent for 0.5 mL/sample. 6. Add buffer to bottles by removing two at a time, then dispense 12 mL buffer and 0.5 mL reducing agent. Weigh and return bottles to incubator. This is the rate limiting step for incubations, and one bottle is al ways being gassed whilst other is having buffer/reducing agent and weighed. This require two people, one to collect bottles, weigh after addition of buffer and redu cing agent and the other to purge the bottles with CO 2 and add buffer and reducing agent. 7. Allow reducing agent to work for about 15 minutes whilst getting rumen contents. 8. Bring cow in to yards; put hot water in flask; take bucket, cheese cloth (Figure 1.1A below), thermos and funnel to cow; obtain about 4 kg contents from the cow and squeeze into thermos. 9. In sacco bags are normally placed in the ru men immediately after rumen contents have been taken. 10. Dispensing rumen liquor into incubation bo ttle requires 2 people; one to open and close bottles and the other to pipette liquor. 11. Calibrate pH meter; measure pH of rumen contents. 12. Pipette 3 mL rumen liquor into each bo ttle. Cap and return to incubator. NB: Cut tip off pipette to faci litate handling rumen contents; keep stirring rumen liquor whilst sampling. Depress pipette before i mmersing tip in liquor; sample 5 cm below surface. 13. Take 2 samples of rumen contents. One for ammonia analysis (1 mL + 15 µL concentrated HC1) and one for VFA dete rmination (1 mL of rumen contents). 14. Microcentrifuge tubes should be labelled pr ior to incubations. For each run there are 3 x 24 samples plus 6 standards, so about 156 tubes are needed for samples used to determine ammonia concentrations (76 for acidified media and 76 for storing the supernatant). A fu rther 32 are needed for VFA samples taken at 0, 6, 12 and 24 hours. Only two samples are taken from each sample type at each time (i.e. 3 bottles are sub-sampled into two tubes) but tubes are needed for the initial APPENDICES 237 centrifugation and for storage. All samples were stored at -20 oC prior to ammonia and VFA analysis. FIGURE 1.1A – The cheese cloth used to strain the rumen liquor has large holes enabling small particulate material pass into the liquor. From left to right, 2.5 kg roll of cheese cloth (stockinette), placement of cheese cloth in bucket to strain the liquor and the thermos flask to transport rumen liquor from cow to the incubator (bottles). APPENDICES 238 APPENDIX 2: Ammonia determination method (Chaney and Marbach 1962). The procedure is based on Chaney and Marbach (19 62) and is a colorimetric procedure based on a combination of reagen ts for the catalysed indophenol reaction for the determination of ammonia, which produces a stable blue colour. Solution one: 0.11M Phenol 10 g/L 0.17 mM Sodium nitroprusside (sodium ni trosopentacyanoferrate (III)) 0.05 g/L Dissolve in distilled water. Make about 100 mL at any one time and store in a brown bottle. Store reagent at 4 0 C and do not keep for more than a week. Solution two: 0.13 M Sodium hydroxide 5 g/L Sodium hypochlorite 0.42 g/ L (Commercial bleach 17mL/L) Dissolve in distilled water. Make about 100 mL at any one time and store in a brown bottle. Store reagent at 4 0 C and do not keep for more than one week. S TANDARDS Solution three: 60 mMol/L NH 4 Cl; pre-weigh a 250 mL beaker. Weigh 802.35 mg of NH 4 Cl. Add it to the beaker. Add about 200 mL of artificial (made with distilled water) and 5 mL of concentrated HCl. Dissolve and make up to 250 mL with artificial saliva. Record final weight. Stock; weight of NH 4 = 0.8374. Final weight = 250.31. MW of NH 4 Cl = 53.49 Actual concentration was 62.5 mMol L -1 NH4 Cl (or 3.35 mg g -1 ). Solution four: Dilution Buffer; Combine 245 mL of artificial saliva and 5 mL HCl. Upper detection method limit for the assay is 80 nMol. The standard curve needs to be in the range of 0-75 nMol NH 3 . Prepare standards gravimetrically. Store at –20 0 C. APPENDICES 239 0 nM NH3 (in 20 µL); 0 µL of stock NH4Cl (solution 3) and 4000µL of buffer 5 nM NH3 (in 20 µL); 17 µL of stock NH4Cl (solution 3) and 3983µL of buffer 10 nM NH3 (in 20 µL); 34 µL of stock NH4Cl (solution 3) and 3966µL of buffer 20 nM NH3 (in 20 µL); 67 µL of stock NH4Cl (solution 3) and 3933µL of buffer 30 nM NH3 (in 20 µL); 100 µL of stock NH4Cl (solution 3) and 3900µL of buffer 35 nM NH3 (in 20 µL); 117 µL of stock NH4Cl (solution 3) and 3883µL of buffer 40 nM NH3 (in 20 µL); 134 µL of stock NH4Cl (solution 3) and 3866µL of buffer 50 nM NH3 (in 20 µL); 167 µL of stock NH4Cl (solution 3) and 3833µL of buffer 55 nM NH3 (in 20 µL); 184 µL of stock NH4Cl (solution 3) and 3816µL of buffer 60 nM NH3 (in 20 µL); 200 µL of stock NH4Cl (solution 3) and 3800µL of buffer 70 nM NH3 (in 20 µL); 234 µL of stock NH4Cl (solution 3) and 3766µL of buffer 75 nM NH3 (in 20 µL); 250 µL of stock NH4Cl (solution 3) and 3750µL of buffer Predicted concentration and the amount of NH3 in each reaction when 20 μL of each standard is used. 0 nM/L and in 20 µL there is 0 nM (0 ng). 0.25 nM/L and in 20 µL there is 5 nM (273 ng). 0.5 nM/L and in 20 µL there is 10 nM (546 ng). 1.0nM/L and in 20 µL there is 20 nM (1075 ng). 1.5 nM/L and in 20 µL there is 30 nM (1605 ng). 1.75 nM/L and in 20 µL there is 35 nM (1878 ng). 2.0nM/L and in 20 µL there is 40 nM (2150 ng). 2.5 nM/L and in 20 µL there is 50 nM (2680 ng). 2.75 nM/L and in 20 µL there is 55 nM (2953 ng). 3.0nM/L and in 20 µL there is 60 nM (3209 ng). 3.5 nM/L and in 20 µL there is 70 nM (3755 ng). 3.75 nM/L and in 20 µL there is 75 nM (4012 ng). Add 20 μL of standard or unknown (in duplic ate) to each well in the plate. Add 100 μL of Solution 1 to each well. Add 100 μL of Solution 2 to each well. Leave on plate to incubate at room temperature for 30 minutes. Read absorbency at 625 nm using a spectrophotometer. APPENDICES 240 APPENDIX 3: VFA determination method. VFA concentration was determined by gas chromatography (GC) as described by Attwood et al. (1998 ). The GC used a nitroterephthalic acid modified polyethylene glycol column (DB – FFAP, 30m x 0.53mm x 1.0 μm film thickness; J and W scientific, Ca USA) attached to a Hewlett Packard 6890 series system. Helium was the carrier gas at a flow rate of 5mL/minute. The oven temperature started at 85 oC and held for 5 minutes before the next sample was injected. Peaks were de tected using a flame ionisation detector, identified by comparison with standards, an d integrated using HP chemstation software (version 4.02). The n-caproic acid (10mMol fin al concentration) was used as an initial standard. Basic procedure: 1. Label wide mouth, yellow lid vials and put on ice. 2. Defrost VFA samples, shake and spin for 15 minutes at 15000 rpm. 3. Add 100 µL of n-caproic acid to vials. 4. Add 100 µL of phosphoric acid to vials. 5. Add 1 mL of sample followed to vials and then screw on lid. Note it is important that minimal volatilisation occurs (i.e. replace eppendorf lid and screw on yellow vial lid as quickly as possible). 6. Shake vial to ensure good mix. Note: The n-caproic and phosphoric acids should be stored at 4 oC, and in large batches it may cool to room temperature and separate out. APPENDICES 241 APPENDIX 4: I n vitro data from Chapter 3. TABLE 4.1A - Net ammonia production from in vitro incubation of maturing ryegrass a. Data are mean values for in µMol NH 3 /mM plant N. Negative values indicate a net utilisation of ammonia present in rumen inoculum, with large values a consequence of low plant N concentrations. Run A Incubation time (h) 0 2 4 6 8 10 12 24 22d mowing date 1 0 19.67 30.05 65.18 66.50 38.00 64.24 72.62 31d mowing date 1 0 14.79 13.63 23.23 26.25 9.43 -1.03 -7.86 10d mowing date 2 0 25.84 34.52 38.24 27.51 79.53 1.16 24.90 Standard 112.2 294.0 Run B Incubation time (h) 0 2 4 6 8 10 12 24 53d mowing date 1 0 53.26 19.09 46.29 11.16 -24.86 -34.47 -4.62 32d mowing date 2 0 21.59 47.46 54.65 27.24 4.19 -28.09 23.62 22d mowing date 3 0 51.81 39.66 53.33 27.91 -1.05 -45.74 20.12 Standard 83.2 265.2 Run C Incubation time (h) 0 2 4 6 8 10 12 24 74d mowing date 1 0 58.18 -27.92 -54.51 -54.75 -53.25 -54.01 -39.06 53d mowing date 2 0 48.59 -25.00 -46.99 -46.54 -46.99 -45.96 -33.50 43d mowing date 3 0 23.49 -35.65 -53.65 -54.37 -54.25 -54.74 -34.43 Standard 169.7 508.9 Run D Incubation time (h) 0 2 4 6 8 10 12 24 88d mowing date 1 0 84.38 16.38 -38.32 -53.93 -52.49 -46.22 -6.56 67d mowing date 2 0 49.89 -13.81 -87.62 -89.42 -89.56 -83.41 -43.97 57d mowing date 3 0 99.68 20.38 -52.22 -62.35 -59.17 -57.65 -0.78 Standard 195.8 331.6 Run E Incubation time (h) 0 2 4 6 8 10 12 24 105d mowing date 1 0 138.0 -134.6 -290.7 -326.6 -320.7 -325.9 -235.8 84d mowing date 2 0 71.9 -133.6 -252.3 -297.5 -282.2 -293.1 -189.3 74d mowing date 3 0 190.0 -68.5 -217.7 -250.6 -230.6 -230.4 -183.5 Standard 273.8 736.0 a Runs A – E refer to in vitro incubation runs described in Chapter 3 (Table 3.1). APPENDICES 242 TABLE 4.2A - Concentrations (mMol/L) of volatile fatty acids production from in vitro incubation of runs A to E as desc ribed in section 3.3 – Chapter 3. Run A Time (h) acetate propionate Iso-but butyrate Iso-val valerate Total 22 days mowing date 1 0 13.4 4.4 1.8 0.2 0.3 0.5 20.8 6 53.2 18.2 9.5 0.8 1.0 1.5 84.2 12 80.5 27.7 13.0 1.3 1.8 2.4 126.7 24 92.6 30.1 15.0 1.2 1.7 2.4 143.0 31 days mowing date 1 0 14.5 3.1 1.8 0.0 0.3 0.5 20.3 6 54.7 20.6 8.6 0.6 0.7 1.4 86.6 12 62.8 24.3 14.3 0.8 1.0 1.9 105.2 24 98.2 35.0 14.3 1.1 1.3 2.2 152.1 10 days mowing date 2 0 14.5 3.7 2.1 0.0 0.4 0.6 21.2 6 48.0 16.8 7.8 0.7 0.9 1.4 75.5 12 66.5 22.7 10.3 0.8 0.9 1.6 102.7 24 91.6 32.1 14.0 1.4 1.5 2.4 143.0 Run B Time (h) acetate propionate Iso-but butyrate Iso-val valerate Total 53 days mowing date 1 0 12.0 4.2 1.5 0.5 0.7 0.9 19.8 6 40.8 11.4 8.3 0.5 0.6 1.2 62.8 12 73.3 20.9 12.7 0.7 0.6 1.7 109.9 24 86.0 24.0 12.2 0.9 0.9 1.8 125.8 32 days mowing date 2 0 14.3 4.7 2.3 0.4 0.5 0.7 22.8 6 41.0 11.5 8.4 0.5 0.7 1.3 63.4 12 70.2 18.8 12.0 0.6 0.6 1.7 104.0 24 94.6 24.6 13.1 1.3 1.2 2.2 137.0 22 days mowing date 3 0 14.7 3.1 1.9 0.0 0.4 0.5 20.5 6 45.3 12.5 9.8 0.6 0.8 1.5 70.5 12 69.8 20.0 12.2 0.7 0.7 1.8 105.2 24 87.9 26.7 14.2 1.1 1.1 2.3 133.3 APPENDICES 243 Run C Time (h) acetate propionate Iso-bu t butyrate Iso-val valerate Total 74 days mowing date 1 0 10.21 3.18 1.28 0.15 0.20 0.27 15.29 6 48.10 18.30 8.90 0.00 0.38 0.92 76.60 12 65.20 27.10 14.90 0.75 0.60 1.30 109.85 24 73.10 27.70 16.90 0.83 0.77 1.50 120.80 53 days mowing date 2 0 9.77 2.52 1.07 0.13 0.17 0.22 13.88 6 48.10 17.20 7.90 0.00 0.37 0.84 74.41 12 70.60 27.30 15.30 0.60 0.64 1.40 115.84 24 86.70 29.10 15.80 0.99 0.79 1.50 134.88 43 days mowing date 3 0 9.77 3.40 1.70 0.00 0.41 0.45 9.77 6 50.20 18.90 8.20 0.00 0.36 0.90 50.20 12 63.40 26.80 16.30 0.71 0.58 1.40 63.40 24 81.60 32.40 18.90 1.00 0.83 1.80 81.60 Run D Time (h) acetate propionate Iso-but butyrate Iso-val valerate Total 88 days mowing date 1 0 12.5 2.7 1.9 0.0 0.0 0.5 17.6 6 89.4 26.3 14.2 0.7 0.6 2.0 133.1 12 103.6 33.1 18.1 1.1 0.7 2.3 158.9 24 114.1 30.2 18.4 0.8 0.9 2.4 166.8 67 days mowing date 2 0 13.4 2.6 1.6 0.2 0.3 0.5 18.5 6 85.1 24.9 13.5 0.7 0.5 1.8 126.5 12 101.1 29.4 16.3 0.7 0.7 2.1 150.3 24 118.9 34.1 19.7 1.1 0.9 2.4 177.1 57 days mowing date 3 0 14.2 4.6 2.9 0.3 0.5 0.9 23.3 6 75.8 26.2 14.3 0.8 0.5 1.8 119.4 12 107.6 30.1 18.3 1.0 0.8 2.3 160.1 24 113.6 33.1 20.4 0.8 0.9 2.7 171.5 APPENDICES 244 Run E Time (h) acetate propionate Iso-but butyrate Iso-val valerate Total 105 days mowing date 1 0 11.6 5.8 2.9 0.3 0.5 0.8 21.9 6 71.9 18.8 12.9 0.0 0.7 1.8 106.1 12 112.0 28.8 19.1 0.0 0.9 2.4 163.2 24 122.0 32.0 21.1 1.0 1.0 2.6 179.7 84 days mowing date 2 0 11.6 3.3 2.1 0.2 0.4 0.6 18.2 6 75.6 21.1 13.8 0.0 0.6 1.8 112.9 12 112.0 29.8 19.1 0.9 0.9 2.4 165.1 24 120.0 33.7 21.6 1.0 1.0 2.6 179.9 74 days mowing date 3 0 11.6 4.3 2.6 0.3 0.5 0.7 20.1 6 69.7 20.0 13.3 0.0 0.6 1.7 105.3 12 106.3 31.3 21.2 0.9 0.9 2.5 163.0 24 130.0 37.7 24.5 1.3 1.1 2.9 197.5 Abbreviations: Iso-but, iso-butyrate; iso-val, iso-valerate. APPENDICES 245 APPENDIX 5: Equations used to describe relationships between height, herbage mass and nutritive characteristics of ryegrass due to maturation (age; days of re-growth). The use of single or double asterisk in this and other equations denotes estimates parameters significantly different from zero at 5 and 1% levels respectively. This terminology applies to all equations in this chapter. ns = non significant. Height (cm) = 5.59±5.29 ns + 0.53±0.19 * x age + 0.002±0.002 ns x age 2 (r 2 =0.92; C V=1 4.5% ; Root MSE=6.79; mean height=47 cm). HM = - 0.53±0.56 ns + 0.034±0.02 ns x age + 0.0002 ±0.00016 ns x age 2 (r 2 =0.81 ; CV=3 4.2% ; Root MSE=0.72 t/ha; mean HM=2.11 t/ha). HM = - 0.95±0.2 ** + 0.065± 0.004 ** x height (r 2 =0.91; CV=23.1%; Root MSE=0.49 t/ha; mean HM=2.11 t/ha). CP (g/100g in the DM) = 24.98 ± 1.68 ** - 0.29 ± 0.06 ** x age + 0.001 ± 0.0005 * x age 2 (r 2 =0.82 ; CV=16.82%; Root MSE=2.16; mean CP=12.82 g/100g in the DM). NSC (g/100g in the DM) = 5.51 ± 1.21 ** + 0.014 ± 0.04 ** x age - 0.001 ± 0.0004 ** x age 2 (r 2 =0.28 ; CV=16.84%; Root MSE=1.55; mean NSC=9.22 g/100g in the DM). Lipid (g/100g in the DM) = 4.25±0.15** - 0.02±0.002** x age (r 2 =0.73; CV=11.28%; Root MSE=0.35; mean Lipid =3.07g/100g in the DM). NDF (g/100g in the DM) = 46.58±1.94** - 0.016±0.07ns x age + 0.0016± 0.0006 * x age 2 (r 2 =0.80 ; C V=4.76% ; Root MSE=2.48; mean NDF=52.18g/100g in the DM). ADF (g/100g in the DM) = 25.50±1.68** + 0.013±0.06ns x age + 0.001±0.0005 * x age 2 (r 2 =0.79 ; C V=7.01% ; Root MSE=2.16; mean ADF=30.75g/100g in the DM). Ash (g/100g in the DM) = 13.53±0.47** - 0.068±0.007** x age (r 2 =0.76 ; CV=11.28%; Root MSE=1.08; mean Ash =9.56g/100g in the DM). ME (MJ ME/kg DM) = 11.61±0.41** + 0.009±0.015ns x age - 0.0003 ±0.0001* x age 2 (r 2 =0.73; CV=4.85%; Root MSE=0.52 ; mean ME=10.80 MJ ME/kg DM) APPENDICES 246 APPENDIX 6: In sacco data from Chapter 3. TABLE 6.1A – NIRS calibration estimates parameters for in sacco residues and lignin concentration in forages determined by wet chemical analyses. Means in g/100 g DM. Constituent N Mean SEC RSQ SECV 1-VR CP 120 10.30 0.54 0.99 0.68 0.99 NDF 138 70.13 2.54 0.93 3.00 0.91 ADF 136 46.80 1.55 0.95 1.89 0.92 Lignin 86 4.65 0.30 0.99 0.45 0.98 Where N = Number of samples SEC = Standard Error of Calibration (e.g. NDF 70.13±2.54) RSQ = R Squared SECV = Standard Error of Cross Validation (e.g. NDF 70.13±3.00) 1-VR = Cross Validation RSQ TABLE 6.2A - Composition of ryegrass (0 hour) and in sacco residues over the 72 hours digestion period. Data are for crude protein (CP), neutral detergent fibre (NDF) and acid detergent fibre (ADF). g CP/100 g DM Area Date Age CP initial 0 2 6 12 24 72 1 11/09 /2 00 0 22 23.7 19.2 18.1 16.2 13.0 7.4 9.3 2 5/10 /20 00 24 18.6 15.7 12.3 8.5 6.2 5.7 9.0 3 13/10 /2 00 0 22 18.5 16.0 12.5 8.9 11.5 6.6 10.6 Average 23 20.3 17.0 14.3 11.2 10.2 6.5 9.6 STDE V 3.0 2.0 3.3 4.3 3.6 0.8 0.8 1 5/10 /20 00 45 16.8 13.8 11.4 8.2 6.1 6.3 9.7 2 24/10 /2 00 0 43 16.7 9.7 8.3 7.4 5.1 6.5 9.6 3 3/11 /20 00 43 13.1 9.8 9.8 9.8 5.9 4.7 13.0 Average 44 15.5 11.1 9.8 8.4 5.7 5.8 10.8 STDE V 2.1 2.3 1.6 1.2 0.6 0.9 1.9 1 3/11 /20 00 74 12.6 9.0 8.8 7.6 6.5 4.1 11.3 2 17/11 /2 00 0 67 9.7 4.0 4.5 4.8 3.8 4.7 5.3 3 4/12 /20 00 74 8.9 2.1 2.2 4.0 3.3 3.8 5.4 Average 72 10.4 5.0 5.2 5.5 4.5 4.2 7.3 STDE V 1.9 3.6 3.4 1.9 1.7 0.5 3.4 Total average 46 15.4 11.0 9.8 8.4 6.8 5.5 9.2 Total STDE V 4.7 5.7 4.7 3.5 3.3 1.2 2.5 APPENDICES 247 g NDF/ 100g DM Area Date Age NDF initial 0 2 6 12 24 72 1 11/09 /2 00 0 22 42.7 55.5 58.6 66.6 69.9 79.6 76.6 2 5/10 /20 00 24 44.8 66.9 73.2 67.8 76.1 76.1 73.0 3 13/10 /2 00 0 22 49.5 64.6 64.3 66.2 73.8 78.3 69.4 Average 23 45.7 62.3 65.4 66.9 73.2 78.0 73.0 STDE V 3.5 6.0 7.4 0.8 3.1 1.8 3.6 1 5/10 /20 00 45 48.7 70.9 72.4 65.7 73.2 74.5 68.5 2 24/10 /2 00 0 43 47.1 72.3 75.4 67.1 75.2 75.6 69.8 3 3/11 /20 00 43 48.8 64.6 64.3 66.2 73.8 78.3 69.4 Average 44 48.2 69.3 70.7 66.3 74.1 76.2 69.2 STDE V 1.0 4.1 5.7 0.7 1.0 1.9 0.7 1 3/11 /20 00 74 51.8 64.9 65.9 69.9 74.0 79.7 73.5 2 17/11 /2 00 0 67 51.1 70.5 72.2 75.5 78.1 80.0 79.6 3 4/12 /20 00 74 56.7 75.9 77.9 78.6 80.0 79.8 80.6 Average 72 53.2 70.4 72.0 74.6 77.4 79.8 77.9 STDE V 3.1 5.5 6.0 4.4 3.1 0.2 3.9 Total average 46 49.0 67.3 69.4 69.3 74.9 78.0 73.4 Total STDE V 4.1 5.9 6.3 4.6 2.9 2.1 4.6 g ADF/ 100g DM Area Date Age ADF initial 0 2 6 12 24 72 1 11/09 /2 00 0 22 21.8 32.1 33.2 43.1 45.5 47.6 46.6 2 5/10 /20 00 24 30.1 38.0 40.6 40.3 44.4 44.9 46.4 3 13/10 /2 00 0 22 26.1 37.1 41.2 45.0 41.4 45.8 46.3 Average 23 26.0 35.8 38.3 42.8 43.7 46.1 46.4 STDE V 4.2 3.2 4.5 2.4 2.1 1.4 0.2 1 5/10 /20 00 45 26.4 38.3 39.5 37.9 42.6 43.7 43.9 2 24/10 /2 00 0 43 26.3 41.1 41.8 38.9 44.7 43.9 42.0 3 3/11 /20 00 43 28.2 38.0 38.3 40.3 44.6 47.2 41.5 Average 44 27.0 39.1 39.9 39.0 44.0 44.9 42.4 STDE V 1.1 1.7 1.8 1.2 1.2 2.0 1.3 1 3/11 /20 00 74 30.2 38.1 39.2 42.1 45.1 48.1 44.0 2 17/11 /2 00 0 67 30.9 42.9 43.5 46.3 48.7 49.5 47.7 3 4/12 /20 00 74 35.6 45.9 47.2 50.3 51.5 50.9 48.1 Average 72 32.2 42.3 43.3 46.2 48.4 49.5 46.6 STDE V 2.9 4.0 4.0 4.1 3.2 1.4 2.3 Total average 46 28.4 39.1 40.5 42.7 45.4 46.8 45.1 Total STDE V 3.9 3.9 3.8 4.0 3.0 2.5 2.4 STEDE V: standard deviation. APPENDICES 248 FIGURE 6.1A - Dry matter (DM) disappearance during in sacco incubations of ryegrass in different ages (days of re-growth) for areas 1, 2 and 3. Area 1 25 50 75 100 % DM d is appe ar an ce 22 days 31 days 45 days 53 days 74 days 88 days 105 days Area 2 25 50 75 100 % DM d is appe ar an ce 10 days 24 days 32 days 43 days 53 days 67 days 84 days Area 3 25 50 75 100 0 12 24 36 48 60 72 Incubation time (hours) % DM d is appe ar an ce 14 days 22 days 33 days 43 days 50 days 57 days 74 days APPENDICES 249 FIGURE 6.2A – Crude protein (CP) disappearance during in sacco incubations of ryegrass in different ages (days of re-growth) for the three areas. Area 1 45 70 95 % CP d is appe ar an ce 22 days 31 days 45 days 53 days 74 days 88 days 105 days Area 2 45 70 95 % CP d is appe ar an ce 10 days 24 days 32 days 43 days 53 days 67 days 84 days Area 3 45 70 95 0 12 24 36 48 60 72 Incubation time (hours) % CP d is appe ar an ce 14 days 22 days 33 days 43 days 50 days 57 days 74 days APPENDICES 250 FIGURE 6.3A - Neutral detergent fibre (NDF) disappearance over time during in sacco incubations of areas 1, 2 and 3. Area 1 0 25 50 75 100 % NDF d is appe ar an ce 22 days 31 days 45 days 53 days 74 days 88 days 105 days Area 2 0 25 50 75 100 % NDF d is appe ar an ce 10 days 24 days 32 days 43 days 53 days 67 days 84 days Area 3 0 25 50 75 100 0 12 24 36 48 60 72 Incubation time (hours) % NDF d is appe ar an ce 14 days 22 days 33 days 43 days 50 days 57 days 74 days APPENDICES 251 FIGURE 6.4A - Acid detergent fibre (ADF) disappearance over time during in sacco incubations of areas 1, 2 and 3. Area 1 0 25 50 75 100 % A DF d is appe ar an ce 22 days 31 days 45 days 53 days 74 days 88 days 105 days Area 2 0 25 50 75 100 % A DF d is appe ar an ce 10 days 24 days 32 days 43 days 53 days 67 days 84 days Area 3 0 25 50 75 100 0 12 24 36 48 60 72 Incubation time (hours) % A DF d is appe ar an ce 14 days 22 days 33 days 43 days 50 days 57 days 74 days APPENDICES 252 APPENDIX 7: I n vitro data from Chapter 4. TABLE 7.1A - Net ammonia production from in vitro incubation from five mature grasses. Data are mean values for in µMol NH 3 /mM plant N. Negative values indicate a net utilisation of ammonia present in rumen inoculum, with large values a consequence of low plant N concentrations. Time h 0 2 6 8 12 24 36 48 Perennial ryegrass Leaf 0 54.4 83.1 95.8 114.2 43.1 47.6 73.3 Stem 0 39.6 85.9 103.9 103.9 38.1 16.2 59.4 Flower 0 38.0 123.8 111.7 111.4 -31.5 -28.8 -22.4 Tall fescue Time h 0 2 6 8 12 24 36 48 Leaf 0 74.0 51.3 30.4 -69.7 -70.1 -62.4 -78.0 Stem 0 17.8 153.0 156.5 153.6 152.4 153.8 149.7 Flower 0 57.9 -99.5 145.4 147.8 119.7 118.8 116.1 Yorkshire fog Time h 0 2 4 6 8 10 12 24 48 Leaf 0 31.4 51.5 70.2 72.6 48.2 26.6 21.9 -0.6 Stem 0 37.7 41.1 34.8 12.8 48.6 -57.3 191.2 179.2 Flower 0 80.0 136.0 186.7 148.9 177.6 253.7 -49.3 -68.4 Phalaris Time h 0 2 6 8 12 24 36 48 Leaf 0 31.3 84.6 99.0 99.9 118.6 137.9 124.3 Stem 0 89.4 122.3 90.0 63.0 109.4 -99.9 -14.9 Flower 0 87.7 86.5 79.9 44.6 18.0 103.4 110.0 Paspalum Time h 0 2 6 8 12 24 36 48 Leaf 0 31.3 14.9 17.8 175.0 124.0 125.6 -59.6 Stem 0 53.0 223.7 123.4 260.3 262.9 261.7 250.9 Flower 0 81.4 119.0 203.2 147.1 -68.9 126.8 107.6 APPENDICES 253 TABLE 7.2A - Concentrations (mMol/g DM) of volatile fatty acids production from in vitro incubation from five mature grasses as described in section 4.4.5 – Chapter 4. Forage Component Time Acetate Propionate Butyrate Minor Total P. ryegrass Leaf 0 0.39 0.07 0.04 0.03 0.53 P. ryegrass Leaf 6 0.64 0.18 0.08 0.03 0.92 P. ryegrass Leaf 12 1.38 0.34 0.18 0.08 1.98 P. ryegrass Leaf 24 1.62 0.22 0.21 0.07 2.12 P. ryegrass Leaf 48 1.03 0.27 0.13 0.19 1.63 P. ryegrass Stem 0 0.37 0.07 0.04 0.02 0.50 P. ryegrass Stem 6 1.31 0.18 0.16 0.03 1.68 P. ryegrass Stem 12 0.83 0.12 0.14 0.00 1.09 P. ryegrass Stem 24 0.95 0.22 0.20 0.07 1.44 P. ryegrass Stem 48 0.79 0.21 0.18 0.22 1.39 P. ryegrass Flower 0 0.36 0.06 0.03 0.01 0.47 P. ryegrass Flower 6 0.91 0.24 0.17 0.07 1.40 P. ryegrass Flower 12 1.19 0.34 0.28 0.10 1.91 P. ryegrass Flower 24 1.76 0.41 0.32 0.14 2.64 P. ryegrass Flower 48 0.75 0.24 0.28 0.27 1.53 Tall fescue Leaf 0 0.65 0.14 0.09 0.05 0.93 Tall fescue Leaf 6 1.49 0.31 0.23 0.06 2.08 Tall fescue Leaf 12 2.25 0.54 0.32 0.08 3.18 Tall fescue Leaf 24 2.71 0.76 0.39 0.10 3.97 Tall fescue Leaf 48 2.35 0.65 0.33 0.08 3.40 Tall fescue Stem 0 0.61 0.12 0.08 0.03 0.83 Tall fescue Stem 6 1.10 0.33 0.20 0.02 1.65 Tall fescue Stem 12 1.54 0.47 0.29 0.05 2.35 Tall fescue Stem 24 1.65 0.42 0.42 0.15 2.64 Tall fescue Stem 48 1.81 0.96 0.56 0.18 3.52 Tall fescue Flower 0 0.52 0.10 0.07 0.02 0.71 Tall fescue Flower 6 1.06 0.42 0.26 0.04 1.78 Tall fescue Flower 12 1.88 0.79 0.48 0.09 3.24 Tall fescue Flower 24 2.19 0.89 0.53 0.11 3.73 Tall fescue Flower 48 2.31 0.96 0.62 0.13 4.01 Yorkshire fog Leaf 0 0.37 0.09 0.06 0.03 0.54 Yorkshire fog Leaf 6 0.77 0.15 0.14 0.03 1.08 Yorkshire fog Leaf 12 0.93 0.16 0.08 0.02 1.19 Yorkshire fog Leaf 24 1.91 0.53 0.25 0.08 2.78 Yorkshire fog Leaf 48 1.84 0.56 0.25 0.12 2.78 Yorkshire fog Stem 0 0.32 0.08 0.04 0.02 0.46 Yorkshire fog Stem 6 0.33 0.08 0.04 0.03 0.48 Yorkshire fog Stem 12 0.92 0.29 0.17 0.04 1.42 Yorkshire fog Stem 24 1.03 0.23 0.21 0.06 1.54 Yorkshire fog Stem 48 1.99 0.36 0.37 0.17 2.90 Yorkshire fog Flower 0 0.43 0.10 0.06 0.02 0.61 Yorkshire fog Flower 6 0.98 0.22 0.19 0.05 1.45 Yorkshire fog Flower 12 1.34 0.28 0.23 0.08 1.92 Yorkshire fog Flower 24 0.84 0.17 0.22 0.09 1.32 Yorkshire fog Flower 48 0.77 0.17 0.17 0.21 1.32 APPENDICES 254 Phalaris Leaf 0 0.52 0.05 0.05 0.04 0.65 Phalaris Leaf 6 0.94 0.19 0.16 0.09 1.38 Phalaris Leaf 12 1.30 0.27 0.23 0.10 1.90 Phalaris Leaf 24 1.76 0.32 0.28 0.19 2.54 Phalaris Leaf 48 1.54 0.26 0.26 0.16 2.21 Phalaris Stem 0 0.52 0.05 0.05 0.04 0.66 Phalaris Stem 6 0.83 0.18 0.12 0.04 1.17 Phalaris Stem 12 1.90 0.16 0.19 0.06 2.31 Phalaris Stem 24 1.40 0.19 0.15 0.00 1.74 Phalaris Stem 48 1.26 0.44 0.29 0.25 2.24 Phalaris Flower 0 0.55 0.05 0.05 0.04 0.69 Phalaris Flower 6 1.13 0.35 0.22 0.08 1.78 Phalaris Flower 12 1.50 0.24 0.26 0.14 2.14 Phalaris Flower 24 2.16 0.29 0.34 0.18 2.98 Phalaris Flower 48 1.58 0.25 0.27 0.15 2.25 Paspalum Leaf 0 0.42 0.09 0.07 0.04 0.62 Paspalum Leaf 6 1.01 0.21 0.16 0.05 1.43 Paspalum Leaf 12 1.23 0.25 0.19 0.07 1.74 Paspalum Leaf 24 0.79 0.14 0.12 0.02 1.08 Paspalum Leaf 48 0.66 0.27 0.17 0.20 1.31 Paspalum Stem 0 0.46 0.10 0.07 0.02 0.65 Paspalum Stem 6 0.98 0.27 0.22 0.05 1.52 Paspalum Stem 12 1.56 0.37 0.27 0.05 2.25 Paspalum Stem 24 1.76 0.56 0.34 0.07 2.72 Paspalum Stem 48 2.07 0.99 0.46 0.15 3.66 Paspalum Flower 0 0.39 0.10 0.05 0.03 0.58 Paspalum Flower 6 0.41 0.25 0.20 0.07 0.92 Paspalum Flower 12 0.46 0.27 0.22 0.07 1.02 Paspalum Flower 24 0.50 0.40 0.19 0.07 1.16 Paspalum Flower 48 0.62 0.22 0.19 0.20 1.23 APPENDICES 255 APPENDIX 8: In sacco data from Chapter 6. TABLE 8.1A – Dry matter (DM) degradation characteristics (% of total DM) from fresh minced forages incubated in dacron bags in the rumen from individual fistulated cows fed full pasture and four silages supplements as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h -1 ), and effective degradability (E) which takes in account the effect of passage from the rumen1 . Cow Diet In sacco A B C k E 6% E 8% 5 272 Pasture Pasture 46 38 16 0.067 63 60 7915 Pasture Pasture 47 36 17 0.077 64 61 5756 PM Pasture 40 40 20 0.057 62 60 7912 PM Pasture 38 41 21 0.079 67 64 3343 PMS Pasture 47 40 13 0.066 64 61 3792 PMS Pasture 45 39 16 0.071 64 61 3788 PS Pasture 42 42 16 0.065 65 62 7926 PS Pasture 41 39 21 0.078 65 62 5774 PSM Pasture 43 38 19 0.086 65 62 7920 PSM Pasture 43 42 15 0.058 64 61 5756 PM PM 46 36 19 0.047 62 59 7912 PM PM 43 36 20 0.062 62 59 3343 PMS PMS 43 41 16 0.052 62 59 3792 PMS PMS 39 47 14 0.045 59 56 3788 PS PS 46 38 16 0.083 68 65 7926 PS PS 46 36 18 0.085 67 65 5774 PSM PSM 42 39 19 0.073 63 61 7920 PSM PSM 46 38 16 0.045 62 59 Average 44 39 17 0.067 64 61 Model P 0.67 0.18 0.01 Forages P 0.67 0.063 0.004 Cow/diet P 0.31 0.17 Cow P 0.51 0.019 r2 0.17 0.89 0.98 1 Passage rate set at 0.06 h -1 and 0.08 h -1 . P: assessing goodness of fit for the overal l model and tests (Forages: effect of diet incubated; Cow/diet: diet effects on forages in sacco and Cow P: cow effects. APPENDICES 256 TABLE 8.2A – Crude protein (CP) degradation characteristics (% of total DM) from fresh minced forages incubated in dacron bags in the rumen from individual fistulated cows fed full pasture and four silages supplements as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h -1 ), and effective degradability (E) which takes in account the effect of passage from the rumen1 . Cow Diet In sacco A B C k E 2% E 5% E 8% 5 272 Pasture Pasture 56 37 7 0.104 87 81 77 7915 Pasture Pasture 53 39 7 0.132 90 85 81 5756 PM Pasture 54 35 11 0.136 87 82 78 7912 PM Pasture 60 29 11 0.141 81 77 74 3343 PMS Pasture 57 36 7 0.115 87 81 77 3792 PMS Pasture 51 41 8 0.129 91 85 81 3788 PS Pasture 58 33 9 0.156 85 81 78 7926 PS Pasture 56 35 9 0.148 87 82 79 5774 PSM Pasture 58 32 10 0.169 84 80 78 7920 PSM Pasture 61 32 7 0.098 83 77 74 5756 PM PM 62 27 10 0.093 85 80 77 7912 PM PM 55 34 11 0.135 84 80 76 3343 PMS PMS 60 32 8 0.099 86 81 77 3792 PMS PMS 62 32 6 0.062 86 79 76 3788 PS PS 67 28 5 0.080 89 84 81 7926 PS PS 65 28 6 0.093 89 84 81 5774 PSM PSM 65 26 9 0.124 88 84 81 7920 PSM PSM 66 26 8 0.071 86 81 78 Average 59 32 8 0.116 86 81 78 Model P 0.0009 0.15 0.0675 Forages P 0.0009 0.0464 0.0204 Cow/diet P 0.17 0.52 Cow P 0.9 0.13 r2 0.74 0.91 0.94 Abbreviations see Table 7.1A. APPENDICES 257 TABLE 8.3A Neutral detergent fibre (NDF) degradation characteristics (% of total DM) from fresh minced forages incubated in da cron bags in the rumen from individual fistulated cows fed full pasture and four silages supplements as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h -1 ), and effective degradability (E) which takes in account the effect of passage from the rumen1 . Cow Diet In sacco A B C k E 6% E 8% 5 2 7 2 Pasture Pasture 26 48 26 0.056 44 41 7915 Pasture Pasture 26 46 28 0.062 45 41 5756 PM Pasture 17 53 30 0.039 42 38 7912 PM Pasture 13 52 34 0.062 48 44 3343 PMS Pasture 26 53 21 0.055 46 43 3792 PMS Pasture 23 50 28 0.063 46 43 3788 PS Pasture 20 55 25 0.049 46 42 7926 PS Pasture 18 48 34 0.060 45 42 5774 PSM Pasture 22 49 30 0.063 46 42 7920 PSM Pasture 20 56 24 0.044 45 41 5756 PM PM 14 54 32 0.035 34 31 7912 PM PM 14 51 35 0.043 35 32 3343 PMS PMS 15 60 25 0.036 37 33 3792 PMS PMS 8 71 21 0.030 32 28 3788 PS PS 26 47 28 0.064 50 46 7926 PS PS 28 42 30 0.056 48 45 5774 PSM PSM 10 56 34 0.058 38 34 7920 PSM PSM 17 57 27 0.033 37 33 Average 26 48 26 0.056 49 46 Model P 0.0001 0.0696 0.0803 Forages P <.0001 0.0152 0.0248 Cow/diet P 0.0038 0.3919 0.3389 Cow P 0.3947 0.1884 r2 0.94 0.94 0.94 Abbreviations see Table 7.1A. APPENDICES 258 TABLE 8.4A Acid detergent fibre (ADF) degradation characteristics (% of total DM) from fresh minced forages incubated in dacron ba gs in the rumen from individual fistulated cows fed full pasture and four silages supplements as defined by soluble (A), degradable insoluble (B), undegradable residue (C = 100 – A – B) as well as fractional disappearance rate (k, h -1 ), and effective degradability (E) which takes in account the effect of passage from the rumen1 . Cow Diet In sacco A B C k E 6% E 8% 5 2 7 2 Pasture Pasture 26 48 26 0.051 43 40 7915 Pasture Pasture 28 44 29 0.056 42 39 5756 PM Pasture 17 54 29 0.034 41 37 7912 PM Pasture 13 52 35 0.059 47 43 3343 PMS Pasture 26 52 22 0.053 45 42 3792 PMS Pasture 25 47 28 0.056 44 41 3788 PS Pasture 20 54 26 0.048 45 41 7926 PS Pasture 18 47 36 0.056 44 40 5774 PSM Pasture 22 47 31 0.062 45 42 7920 PSM Pasture 20 56 24 0.041 44 40 5756 PM PM 14 57 29 0.026 31 28 7912 PM PM 12 51 37 0.039 32 29 3343 PMS PMS 15 59 26 0.035 37 33 3792 PMS PMS 9 72 19 0.026 31 27 3788 PS PS 24 47 29 0.070 49 46 7926 PS PS 26 42 31 0.062 48 45 5774 PSM PSM 13 52 35 0.059 39 35 7920 PSM PSM 18 55 27 0.031 37 33 Average 19 52 29 0.048 42 38 Model P 0.0001 0.1367 0.0290 Forages P 0.0002 0.0556 0.0085 Cow/diet P 0.0009 0.2384 0.1666 Cow P 0.3019 0.0825 r2 0.94 0.91 0.960 Abbreviations see Table 7.1A. APPENDICES 259 TABLE 8.5A – Particle size distribution of pasture, sulla and maize silages dry matter for in sacco incubations indicated by sieve aperture size either retaining or enabling material to pass. > 2mm 0.25-1mm Fine Pasture 35 40 25 Sulla silage 18 35 47 Maize silage 23 23 55 Average 25 33 42 STDE V 9 9 15 STDE V: standard deviation. APPENDICES 260 FIGURE 8.1A - Mean rumen fluid volatile fatty acids concentrations and acetate: propionate (A: P) ratios for cows fed pastur e and pasture plus maize and sulla silages. Data on the X axis are times after AM feeding. 0 40 80 120 0 3 6 9 12 Time after feeding (hours) A ce tate (m Mol /L) Pasture PMS PSM PS PM 0 8 15 23 30 0 3 6 9 12 Time after feeding (hours) Pro pio na te (m Mol /L) Pasture PMS PSM PS PM 0 6 12 18 24 0 3 6 9 12 n-butyr ate (m Mol /L) 0 5 10 0 3 6 9 12 Min or (m Mol /L) 0 40 80 120 160 200 0 3 6 9 12 Time after feeding (hours) Tota l VFA (m Mol /L) Pasture PMS PSM PS PM 0 2 4 6 0 3 6 9 12 Time after feeding (hours) Rati o A :P Pasture PMS PSM PS PM APPENDICES 261 APPENDIX 9: A simplified method for lignin measurement in a range of forage species1 . 9 . 1A - Abstract Lignin is the prime factor infl uencing the digestibility of plant cell wall material. As concentrations of lignin increase, digestibil ity, intake, and animal performance usually decrease. Presented is a simplified acid detergent lignin procedure which has been used to determine lignin concentration from a wide range of forages and also ryegrass at different stages of maturation. Forages used in this study included grasses, legumes, herbs and conserved material, with lignin co ncentration ranging from 2.02 to 21.1% of the DM. Legumes tended to have higher values than grasses, and ryegrass maturation was not accompanied by increased lignin co ncentration at 53 days of age. These results will be incorporated into a NIRS meth od for determining forage quality, and used in a dairy nutrition model to assist in ration formulation for dairy cows. Keywords : fibre; forages; lignin; analytical method. Short title : Lignin measurements in forages. 9.2A - Introduction 9.2.1A - Importance of lignin for ruminants Lignin is a major anti-nutritional componen t of grasses. Lignin limits cell wall (fibre) digestion by providing a physical barrier to microbial attack and the concentration of both fibre and lignin incr eases as plants mature (Van Soest, 1978 ; Chaves et al., 2002). Ruminants can digest the cellulose and hemicellulose components of fibre but the lignin inhibits the rate and extent of digestion especially when the proportion of lignin in fibre begins to increase. Lignin precursors also have anti-microbial properties (Jung and Fahey, 1983). However it benefits plants by providing mechanical support, water impermeabi lity and protection from insects. The consequences of lignification are summarised in Table 9.1A. 1 Previous published in the Proceedings of New Zealand Grassland Association , 2002, 129-13 3. Best paper poster award at New Zealand Grasslands Conference 2002. APPENDICES 262 Abundant data show negative correlations between the lignin concentration in plants and both dry matter (DM) and fibre digestibility (Jung et al., 1997). Lignin is more prevalent in grass stem than leaf and absent from legume leaves. In contrast to grasses, there are minimal ferulate cross linkages between lignin and hemicellulose in legume stems (Hatfield et al. , 1999), so the lignin concentration is less detrimental to nutritive value of legumes, compared to grasses. This was demonstrated by more rapid degradation of legume fibre than grass fibre (Jung et al., 1997). A good understanding of fibre, its composition, degradation and effects on energy availability is essential for mainta ining a nutrient supply to high producing animals. The rate of digestion, breakdown by chewing and clearance from the rumen determine feed intake, and rumen capacity is likely to limit future increases in productivity from ryegrass based pasture (Wag horn, 2002). Flowering is associated with significant reductions in intake. There are several chemical classifications of fibre, usually based on sequential removal of constituents by boiling in detergen ts or acids. Plant residues remaining after boiling in neutral detergent are not availabl e for ruminant absorption unless they are degraded by microbial fermentation. The neutral detergent fibre (NDF; the residue after boiling in a neutral detergent) comprises hemicellulose, cellulose, lignin and ash, commonly referred to as the cell wall or fibre fr action of plants. Digestion of NDF in an acid detergent removes hemicellulose leaving the acid detergent fibre (ADF) fraction, and a further digestion in 12 M sulphuric acid removes cellulose leaving lignin and ash. Lignin is determined after the mineral (ash) fraction is measured by ashing. Measuring lignin is complicated by the extensive cross linkages with cellulose and hemicellulose and the insolubility of this polymer, (Hatfield et al., 1994). A variety of methods have been developed to estimate lignin with the acid detergent lignin (ADL) procedure (Van Soest, 1967) usually employed for forage analysis. This paper presents a method, which was adapted from Goering and Van Soest (1970) to simplify lignin measurements, together with data from a wide range of forages and also ryegrass of increasing maturity. Th ese data will be incorporated into a NIRS (Near Infra Red Spectrometer) method for determining forage quality, enabling improved predictions of animal performance and formulation of mixed forage rations for dairy cows to avoid current limitations of rumen fill. APPENDICES 263 9.3A – Material and methods 9.3.1A - Measurement procedure Lignin is usually determined as part of a sequential extraction comprising three refluxing and filtering steps; firstly with neutral detergent followed by acid detergent and lastly 12 M H 2 SO 4 digestion to leave lignin and ash residues. The standard ADL determination uses about 1 g dried material, requires condensing and refluxing apparatus including 600 mL Berzelius beakers for boiling and 50 ml fritted glass crucibles (coarse porosity) to retain residues after re fluxing. The main modifications of the ADL procedure of Goering and Van Soest (1970), presented here, are the removal of the neutral detergent fibre step and a reduction in the quantity of material required for analysis. Screw capped culture tubes (16 x 100 mm) have been used for digestion in both acid detergent and 12 M H 2 SO 4 to obtain an acid-insolub le residue lignin and ash residue. Glass microfibre filters (Watman 55 mm GF/C) were used for collecting fibre residues from the digestion tubes after the final digestion and for ashing. The modified technique is presented in Table 9.2A. This technique was tested to determine variability and precision and used to determine lignin concentration in a wide range of forage samples. 9.3.1.1A - Repeatability Replicate analysis is the primary means of ev aluating data variability or precision. The coefficient of variation (CV) was used as a measure of repeatability: CV (%) = (SD/µ) x 100 Where SD = sample standard deviation and µ = mean of replicate analyses. Precision was monitored by a reference standard (ryegrass/white clover pasture) analyzed with every batch of 20 forage samples. Standard deviation (SD) was calculated from repeated analysis of control samples and duplicates of the forage samples. The modified ADL method has been used to determine the lignin concentration of a range of forage types and also ryegrass of increasing (21 to 105 days) maturity. These data were required for model predic tion of forage mixed rations for high producing dairy cows. APPENDICES 264 9 . 4A - Results The technique developed for lignin measur ements (Table 9.2A) resulted in good repeatability for a wide range of forage types (Table 9.3A). In most cases the CV was less than 5%. It is however important to be aware that the condensed tannins present in some forages have been shown in this laboratory to comprise part of the lignin fraction after acid digestion. True lignin will be calculated from 12 M sulphuric acid residue less the values for ash and tannins determined by a separate analysis. The forages used for the lignin assay (T able 9.3A) are used in farming systems and were part of a systematic study of feeding values by Burke et al. (2000 ) and Chaves et al. (2001 ). All grasses except those in the “maturation” trial were leafy and vegetative, and had similar concentrations of lignin (about 3% of DM) except for cocksfoot and paspalum. Lignin concentratio n in clover and lucerne were higher than most of grasses but the values for Lotus species were elevated by the presence of tannins. Most surprising were the high values for lignin in chicory and plantain, both of which have very low fibre concentrations (Burke et al., 2000). The concentration of lignin in ryegrass increased as the grass aged, and was strongly correlated with the decline in organic matter digestibility (Figure 9.1A; r 2 = 0.90). Lignin and NDF concentration increased simult aneously, but the relative rate of change showed lignin increased more rapidly than NDF. Mature ryegrass was digested much more slowly by rumen mi croflora after mincing ( in sacco) than young ryegrass (Figure 9.1A; Chaves et al., 2002). 9 . 5A - Discussion The data presented here show that a rapid lignin assay, based on fewer steps and smaller quantities of samples and reagents (compared to Goering and Van Soest, 1970) can give good repeatability. Analys is of diverse forage samples confirms the generally lower values in grasses than legumes, show higher values when ensiled and slow increases in concentration with ryegrass maturation. Although the increases in lignin and NDF concentration in ryegrass correspond to slower in sacco degradation (Burke et al., 2000; Chaves et al., 2001), there was no increase in lignin concentration for the first 53 days of re-growth (Table 9.3A) when most grazing would take place. This suggests that either structural changes in the non lignin fibre components were responsible for slow er degradation or that cross linkages between lignin and hemicellulose were responsible (Figure 9.2A). These changes cannot be detected by analyses of lignin concentration. Lignin concentration is not APPENDICES 265 the sole determinant of fibre degr adation rate of ryegrass (Inoue et al., 1989) and the linkages with fibre constituents (Ralph et al., 1995; Wilson and Hatfield, 1997) have a major impact on fibre degradation. Other factors affecting fibre degradation include the extent of chewing (Waghorn, 2002), diet composition, especially inclusion of readily fermentable substrates, level of feeding and rumen pH. Future improvements in the nutritive value of ryegrass must also reduce the extent or effectiveness of cross linkages between lignin, hemicellulose and cellulose in manner that enables plants to retain their structural integrity. These changes may in volve biotechnology and the identification of the brown midrib mutant of corn, with substantial higher digestibility (Barrière et al., 1992 ), provides an incentive for investigation. Casler (1997) demonstrated that a 20 year plant breeding effort could theoretically result in a new cultivar (for most perennial forage species) with an in vitro digestibility up to 10% higher than the original population. The simplified procedure for lignin analysis and future prediction by NIRS (Corson et al., 1999) will improve estimation of feeding value for fresh and ensiled forages. The principal driver for this procedure was a need for including lignin values in the CNCPS (Cornell dairy nutrition) model to assist in ration formulation for dairy cows. APPENDICES 266 TABLE 9.1A - Consequences of lignification. o Required by the plant to maintain structural integrity o Lignin can be a significant component of grass leaves, stems, sheaths and legume stems o Lignin concentration is higher in C4 grasses than C3 grasses o Lignin concentration is higher in legumes than grasses o Lignin concentrations can be dependent upon analytical technique, with Klason lignin values being 2 - 4 times as high as acid detergent lignin for grasses and 30 - 50% higher for legumes o Lignification is especially intense in sc hlerenchyma and vascular tissues which form long indigestible fibres o Lignin is not degradable by bacteria o Slows rate of intake, digestion and increases the need for chewing and salivation o Lignin is impermeable to bacteria; cell wall rupture is a prerequisite to digestion, from the cell lumen o Limited rupture during eating limi ts initial rate of degradation o Continual damage to cell walls by chewing increases surface area of polysaccharide for colonisation by fibrolytic bacteria o Passage from the rumen (clearance) is a prerequisite for further intake and non degradable cell walls must be chewed to a size able to pass a 1 or 2 mm sieve aperture for sheep and cattle respectively o Within feed types (e.g. temperate grasses) lignin concentration is poorly correlated with digestibility ( in vivo and/or in vitro) o Linkages via ferulic and p-coumeric acids between core lignin and cell wall polysaccharides appear to have a major impact on feeding value and are more important than concentration per se. o Dietary lignin is determined by proporti ons of legume and grass, maturity and selection of leaf, stem and sheath constituents APPENDICES 267 o Limitations attributable to lignin apply to all forage fed animals o The nutritive value of tropical grasses are further reduced by urinary nitrogen losses associated with excretion of metabolised phenolics o The impact of lignin can be especially severe when animals have been selected for high genetic merit under good feeding regimens (e.g. total mixed ratio) and then fed pasture o Dairy cattle are affected to a large extent by lignification, because the demands for milk production exceed capacity to provide nutrients from pasture. Consequences include significant bodyweight loss, failure to ovulate and rapid declines in productivity o Solutions include selecting animals with more effective chewing, larger rumen capacity, ability to allow larger particles to escape from the rumen o Breeders may modify lignin to enable easier cell rupture, less binding to polysaccharides, fewer lignified tissues or identify genotypes with brittle fibres more easily broken to pass from the rumen o Lignin is the principal limitation to production from forages APPENDICES 268 TABLE 9.2A - Modified procedures for the determination of acid detergent lignin (ADL). (Procedure to obtain acid deterg ent fibre (ADF) for ADL analysis) 1. Dry 16 mL glass tubes at 100 oC for at least an hour. 2. Weigh tubes to 4 or 5 decimal places in groups of 10 at a time 2 . 3. Dry finely ground samples at 60 oC and weigh about 250 mg into weighed tubes. (Grinding was carried out in a Wiley mill with a 1 mm screen (Mertens, 1992a)). 4. Add 10 mL acid detergent fibre solution (2% cetyl trimethylammonium bromide in 1 litre 0.5M H 2 SO 4 ) to the sample. Put cap on tube. Vortex. 5. Reflux over a steady heat (water bath) for one hour at 95 - 100 oC, vortex every 10 minutes. 6. Centrifuge for 10 minutes at 3000 rpm an d remove supernatant by suction. 7. Add 15 mL hot distilled water to residue; vortex. Centrifuge for 10 minutes at 3000 rpm and discard supernatant. Repeat this step 3 times. 8. Add 15 mL acetone to the residue. Vortex . Centrifuge for 10 minutes at 3000 rpm and discard supernatant. Repe at this step 1 more time. 9. Evaporate residual acetone in water bath at 60 oC. 10. Dry tubes at 90 oC overnight. 11. Weigh tubes and residue using hot weighing technique to 4 or 5 decimal places, calculate ADF (Mertens, 1992a). 2 When 24 tubes (or GMF) were removed from 105 oC in a single batch, weight of first tube did not change over the time the 24 th tube were weighed (~5 minutes) so can remove more than 10 at once. 2. Include 8 additional GMF blanks with each batch of 22 forage samples to determine ashing losses from GMF. APPENDICES 269 (ADL procedure) A protective mask MUST be worn when handling 12M H 2 SO 4 in this assay. 12. Label glass microfibre filters (GMF) with marker pen (both sides), dry at 100 oC over night and weigh 10 tubes at a time 1 to 5 decimal places. 13. Add 1.5 mL 12 M H 2 SO 4 to tubes containing residues (in fume cupboard) and digest at 30 oC for 60 minutes mixing carefully every 10 minutes. 14. Following digestion the acid-insoluble residue was collected by filtration using 45 mm Buchner funnels with pre-weighed 55 mm Whatman GF/C glass microfibre filters. An extensive washing with wate r and a final acetone rinse (twice) was used prior to drying the samples overnight at 100 oC. 15. Weigh the filters and residue to 5 decimal places. 16. Ash at 450 oC for 6 hours. Weigh GMF with ash to 5 decimal places. 17. ADL was determined as the difference in weight of the residue before APPENDICES 270 TABLE 9.3A - Comparative ADL concentration (% of DM) of different feeds. Feed ADL 1 % CV % Fresh grasses Lolium perenne (Perennial ryegrass AR1 2 ) 2.92 1.14 Holcus lanatus (Yorkshire fog) 3.13 2.20 Lolium perenne (Perennial ryegrass Nil Endophyte) 3.38 2.23 Bromus willdenowii (Prairie grass) 3.78 0.92 Festuca arundinacea (Tall fescue) 3.78 2.97 Pennisetum clandestinum (Kikuyu) 3.82 2.05 Dactylis glomerata (Cocksfoot) 5.13 2.72 Paspalum dilatatum (Paspalum) 6.85 2.52 Legumes and herbs Trifolium repens (White clover) 5.87 2.04 Medicago sativa (Lucerne) 6.12 3.34 Trifolium pratense (Red Clover) 6.23 12.53 Cichorium intybus (Chicory) 6.97 6.96 Lotus corniculatus (Birdsfoot trefoil) 7.22 3 4.40 Lotus pedunculatus (Lotus major) 16.97 3 3.48 Plantago lanceolata (Plantain) 21.10 0.72 Conserved Zea maize (Maize grain) 2.02 2.72 Avena sativa (Oat silage) 4.29 4.03 Lolium perenne (Perennial ryegrass silage) 4.31 4.56 Zea maize (Maize silage) 4.36 3.87 Zea maize (Maize silage) 4.95 1.31 Medicago sativa (Lucerne silage) 7.33 0.06 Medicago sativa (Lucerne hay) 8.04 0.40 Perennial ryegrass maturation Perennial ryegrass 21 days 2.65 4.54 Perennial ryegrass 31 days 2.38 5.94 Perennial ryegrass 53 days 2.51 3.10 Perennial ryegrass 74 days 2.63 0.27 Perennial ryegrass 88 days 3.02 Perennial ryegrass 105 days 4.35 Pasture reference standard 5.34 3.78 Pasture reference standard 5.78 2.33 1 Mean of values. 2 AR1, AgResearch cultivar with endophyte ( Neotyphodium lolii) selection for high in peramine but low in other key alkaloids. 3 Lignin includes condensed tannins. APPENDICES 271 FIGURE 9.1A - Relationships between lignin, estima ted organic matter digestibility (% OMD) of maturing perennial ryegrass. OMD and lignin concentration in the dry matter (DM) (r 2 = 0.90) and in sacco DM disappearance of ryegrass harvested at five maturation ages (P > 0.05). 0 25 50 75 0 12 24 36 48 60 72 Incubation time (hours) % DM r em ai ni ng 22 days 31 days 45 days 53 days 88 days %OM D = 94.3 - 7.2*Lignin 50 60 70 80 2 2.5 3 3.5 4 4.5 Lignin concentration (% of DM) % Or ga ni c m atte r di ge sti bili ty R yegrass becoming mature APPENDICES 272 FIGURE 9.2A - Lignin-carbohydrate complex showing ferulate crosslinkages. O H O H O O O H O O OR O H RO O O OO O O H OH O O O O O R' O O O O O R' O O O H O H O H RO CH 3 CH 3 CH 3 Adapted from Chesson (1988 ). Lignin Ferulic cross-linking Polysaccharide (cellulose/hemicellulose) Bibliography AAC 1990. Australian Agricultural Council. Standing Committee on Agriculture, Ruminants Subcommitee . Feeding standards for Australian livestock: ruminants . CSIRO, Melbourne, Australia. Adesogan, A.T., Givens, D.I. Owen, E. 2000. Measuring chemical composition and nutritive value in forages. In Field and laboratory methods for grassland and animal production research (ed. -t-Mannetje, L. and Jones, R. M.), pp. 263-2 78. CABI Publishing, Wallingford, UK. AFRC 1992. Nutritive requirements for ruminants animals: protein. Nutrition Abstracts and Reviews 62: 787-83 5. AFRC 1993. Energy and protein requirements of ruminants . CAB International, Wallingford, UK. Agnew, R.E. and Newbold, J.R. 2002. Nutritional standards for dairy cattle . http://www.bsas.org.uk/publs/reports/FinalDairy.pdf Akin, D.E. 1979. Microscopic evaluation of forage digestion by rumen microorganisms-a review. Journal of Animal Science 48: 701-7 10. Alderman, G., France, J. Kebreab, E. 2001. A critique of the Cornell Net Carbohydrate and Protein System with emphasis on dairy cattle. 3. The requirements model. Journal of Animal and Feed Sciences 10: 361-383. Annison, E.F., Lindsay, D.B. Nolan, J.V. 2002. Digestion and metabolism. In Sheep nutrition (ed. Freer, M. and Dove, H.), pp. 95-118. CSIRO Publishing, Canberra, Australia. AOAC 1990. O f ficial methods of analysis of the Association of Official Analytical Chemists ('15th'). Association of Official Analytical Chemists, Arlington, VA. ARC 1980. T he nutrient requirements of ruminant livestock . Commonwealth Agricultural Bureaux, Slough, England. 273 Armstrong, D.G. 1964. Evaluation of artificially dried grass as a source of energy for sheep. II. The energy value of cocksfoot, timothy and two strains of rye-grass at varying stages of maturity. Journal of Agricultural Science, Cambridge 62: 399-41 6. Armstrong, D.G. 1982. Digestion and utilization of energy. In " Nutritional limits to animal production from pastures ", Eds. Hacker, J. B., Commonwealth Agriculture Bureaux, pp. 225-2 44. Attwood, G.T., Klieve, A.V., Ouwerkerk, D. Patel, B.K.C. 1998. Ammonia-hyperproducing bacteria from New Zealand ruminants. Applied and environmental microbiology 64: 1796 –1 804. Auldist, M.J., van der Poel, W., Laboyrie, P.G. Prosser, C.G. 1998. Influence of pasture allowance on the composition and cheese-yielding potential of milk. Proceedings of the New Zealand Grassland Association 60: 199– 202. Baldwin, R.L. 1995. Modeling ruminant digestion and metabolism ('First'). Chapman & Hall, London, England. Baldwin, R.L., Thornley, J.H.M. Beever, D.E. 1987. Metabolism of the lactating cow. II. Digestive elementary of a mechanistic model. Journal of Dairy Research 54: 107- 1 31. Bargo, F., Muller, L.D., Delahoy, J.E. Cassidy, T.W. 2002. Milk response to concentrate supplementation of high producing dairy cows grazing at two pasture allowances. Journal of Dairy Science 85: 1777 -1 79 2. Bargo, F., Muller, L.D., Kolver, E.S. Delaho y, J.E. 2003. Production and digestion of supplemented dairy cows on pasture. Journal of Dairy Science 86: 1-42. Barrell, L.G. 2000. Preparation of fresh forages for incubation and prediction of nutritive value. Bachelor of applied science (Honours) degree, Massey University, Palmerston North, New Zealand, p. Pages. Barrell, L.G., Burke, J.L., Waghorn, G.C., Attwood, G.T. Brookes, I.M. 2000. Preparation of fresh forages for incubation and prediction of nutritive value. Proceedings of the New Zealand Society of Animal Production 60: 5-8. Barrière, Y., Traineau, R., Emile, J.C. Hébert, Y. 1992. Variation and covariation of silage maize digestibility estimated from digestion trials with sheep. Euphytica 59: 61-72. 274 Bateman, H.G., Clark, J.H., Patton, R.A., P eel, C.J. Schwab, C.G. 2001. Accuracy and precision of computer models to predict passage of crude protein and amino acids to the duodenum of lactating cows. Journal of Dairy Science 84: 649-66 4. Beauchemin, K.A., Yang, W.Z. Rode, L.M. 2003. Effects of particle size of alfalfa-based dairy cow diets on chewing activity, ruminal fermentation, and milk production. Journal of Dairy Science 86: 630–64 3. Beever, D.E. and Cottrill, B.R. 1994. Protein systems for feeding ruminant livestock: a European assessment. Journal of Dairy Science 77: 2031 -2 04 3. Bell, C.C. and Keene, B. 1996. Subtropical grasses in Bay of Plenty dairy pastures. Proceedings of the New Zealand Grassland Association 57: 55-5 8. Berchielli, T.T., Furlan, C.L., Aroeira, L.J.M., Nogueira, J.R. Ruggieri, A.C. 2000. Estimation of the dry matter intake of lactating co ws in intensive grazing Coastcross grass (Cynodon dactylon (L.) Pers cv. Coastcross). Acta Scientiarum 22: 843-8 49. Berzaghi, P., Herbein, J.H. Pola n, C.E. 1996. Intake, site, and extent of nutrient digestion of lactating cows grazing pasture. Journal of Dairy Science 79: 1581 -15 89. Blackwell, M.B. and Keogh, R.G. 1999. Endophyte toxins and performance of spring- calving dairy cows in Northland. In " Ryegrass endophyte: an essential New Zealand symbiosis ", Eds. Woodfield, D. R. and Matthew, C., New Zealand Grassland Association, Grassland Research and Practice Series No. 7 pp. 45-50. Blair, L. and Cummins, K.A. 1983. Microb ial nitrogen contamination of feedstuffs contained in polyester bag. Journal of Dairy Science 66: 189-. Broderick, G.A. 1987. Determination of protein degradation rates using a rumen in vitro system containing inhibitors of microbial nitrogen metabolism. British journal of nutrition 58: 463-47 5. Broderick, G.A. 1995. Desirable characterist ics of forage legumes for improving protein utilization in ruminants. Journal of Animal Science 73: 2760 -2 77 3. Broderick, G.A. and Albrecht, K.A. 1997. Ruminal in vitro degradation of protein in tannin-free and tannin containi ng forage legume species. Crop Science 37: 1884 - 1 891. 275 Broderick, G.A., Koegel, R.G. , Walgenbach, R.P. Kraus, T.J. 2002. Ryegrass or alfalfa silage as the dietary forage for lactating dairy cows. Journal of Dairy Science 85: 1894 –1 901. Brookes, I.M. 2003. Feed planning for dairy farms: the basic concepts. In "D airy 3 ", Eds. Brookes, I. M., Massey University, 1 pp. 185-19 1. Brouk, M. and Belyea, R. 1993. Chewing activity and digestive responses of cows fed alfalfa forages. Journal of Dairy Science 76: 175-18 2. Bryant, A.M. and Trigg, T.E. 1982. The nutrition of the grazing dairy cow in early lactation. In Dairy production from pasture. (ed. Macmillan, K. L. and Taufa, V. K.), pp. 185-20 7. New Zealand Society of Animal Production, Hamilton, New Zealand. Burke, J.L. 2003. Economic use of complementary feeds in dairy grazing systems. In "D airy 3 ", Eds. Brookes, I. M., Massey University, pp. 153-163. Burke, J.L., Waghorn, G.C. Brookes, I.M. 2002a. An evaluation of sulla (Hedysarum coronarium) with pasture, white clover and lucerne for lambs. Proceedings of the New Zealand Society of Animal Production 62: 152-15 6. Burke, J.L., Waghorn, G.C., Brookes, I.M., Attwood, G.T. Kolver, E.S. 2000. Formulating total mixed rations from forages – defining the digestion kinetics of contrasting species. Proceedings of the New Zealand Society of Animal Production 60: 9-14. Burke, J.L., Waghorn, G.C. Chaves, A.V. 2002 b. Improving animal performance using forage-based diets. Proceedings of the New Zealand Society of Animal Production 62: 267-27 2. Butler, B.M. and Hoogendoorn, C.J. 1987. Pa sture quality and animal performance over late spring and summer. Proceedings of the New Zealand Society of Animal Production 47: 31-33. Calsamiglia, S., Stern, M.D. Bach, A. 2000. Enzymatic and microbial-cell preparation techniques for predicting rumen degradation and postruminal availability of protein. In Forage evaluation in ruminant nutrition (ed. Givens, D. I., Owen, E., Axford, R. F. E. and Omed, H. M.), pp. 259-27 9. CABI Publishing, Wallingford, UK. 276 Carlson, I.T., Oram, R.N. Surprenant, J. 1996. Reed canarygrass and other phalaris species. In Cool-season forage grasses (ed. Moser, L. E., Buxton, D. R. and Casler, M. D.), pp. 569-6 04. American Society of Agronomy, Inc. Crop Science Society of America, Inc. Soil Science Society of America, Inc., Madison, WI (USA). Carran, R.A. 2002. Nitrous oxide, its form ation and place in farm green house gas balances. Proceedings of the New Zealand Society of Animal Production 62: 211- 2 12. Carruthers, V.R., Neil, P.G. Dalley, D.E. 199 7. Effect of altering the non-structural: structural carbohydrate ratio in a pastur e diet on milk production and ruminal metabolites in cow in early lactation. Animal science 64: 393-4 02. Casler, M.D. 1997. Breeding for improved forage quality: potentials and problems. In " Proceedings of the XVIII international grassland conference ", Eds. pp. 323-3 30. Chaney, A.L. and Marbach, E.P. 1962. Modified reagents for determination of urea and ammonia. Clinical chemistry 8: 130-1 32. Chaves, A.V., Waghorn, G.C., Brookes, I.M. Burke, J.L. 2001. Digestion kinetics of mature grasses. Proceedings of the New Zealand Society of Animal Production 61: 8-12. Chaves, A.V., Waghorn, G.C., Brookes, I.M. Hedderley, D. 2002. Digestion kinetics of ryegrass. Proceedings of the New Zealand Society of Animal Production 62: 157-1 62. Cherney, D.J.R., Cherney, J.H. Lucey, R.F. 1993. In vitro digestion kinetics and quality of perennial grasses as influenced by forage maturity. Journal of Dairy Science 76: 790- 7 97. Cherney, D.J.R. and Mertens, D.R. 1998. Modelling grass utilization by dairy cattle. In Grass for dairy cattle (ed. Cherney, J. H. and Cherney, D. J. R.), pp. 351-37 1. CAB International, Cambridge, UK. Cherney, D.J.R., Patterson, J.A. Lemenager, R.P. 1990. Influence of in situ bag rinsing technique on determination of dry matter disappearance. Journal of Dairy Science 73: 391-39 7. Chesson, A. 1988. Lignin-polysaccharide co mplexes of the plant cell wall and their effect on microbial degradation in the rumen. Animal Feed Science and Technology 21: 219-22 8. 277 Chilibroste, P., Tamminga, S. , Boer, H., Gibb, M.J. den Dikken, G. 2000. Duration of regrowth of ryegrass (Lolium perenne) effects on grazing behavior, intake, rumen fill, and fermentation of lactating dairy cows. Journal of Dairy Science 83: 984-99 5. Clark, D.A. 1995. Summer milk - pasture and crops. In " Proceedings of the Ruakura farmers conference ", Eds. Lomas, J., Brebner Print Ltd, 78 pp. 10-16. Clark, D.A., Harris, S.L., Thom, E.R., Waugh, C.D., Copeman, P.J.A. Napper, A.R. 1997a. A comparison of Barkant turnips and Superc how sorghum for summer milk production. Proceedings of the New Zealand Grassland Association 59: 157–16 1. Clark, D.A., Matthew, C. Crush, J.R. 2001. More feed for New Zealand dairy systems. Proceedings of the New Zealand Grassland Association 63: 283-28 8. Clark, D.A., Penno, J.W. Neil, P.G. 1997b. Nutr itional merits and problems of pasture. In Milk composition, production and biotechnology (ed. Welch, R. A. S., Burns, D. J. W., Davis, S. R., Popay, A. I. and Prosser, C. G.), pp. 397-418. CAB International, Hamilton, New Zealand. Clark, D.A., Thom, E.R., Waugh, C.D. Burggraaf , V.T. 1999. Milk production from perennial ryegrass pastures containing different levels of endophyte. Proceedings of the New Zealand Society of Animal Production 59: 258-259. Cohen, D.C. and Doyle, P.T. 2001. Effect of sample preparation on in situ estimates of protein degradability for white clover herbages. Australian Journal of Experimental Agriculture 41: 619-6 24. Corson, D.C., Waghorn, G.C., Ulyatt, M.J. Lee, J. 1999. NIRS: Forage analysis and livestock feeding. Proceedings of the New Zealand Grasslands Association 61: 127- 1 32. Cowan, R.T. and Lowe, K.F. 1998. Tropic al and subtropical grass management and qualitu. In Grass for dairy cattle (ed. Cherney, J. H. and Cherney, D. J. R.), pp. 101- 135. CAB International, Cambridge, UK. CSIRO 1990. Feeding standards for Austra lian livestock. Ruminants. CSIRO Publications, East Melbourne, Australia. Czerkawski, J.W. and Breckenridge, G. 1977 . Design and development of a long-term rumen simulation technique (Rusitec). British journal of nutrition 38: 371-38 9. 278 Dado, R.G. and Allen, M.S. 1994. Variat ion in and relationships among feeding, chewing, and drinking variab les for lactating dairy cows. Journal of Dairy Science 77: 132-14 4. Dado, R.G. and Allen, M.S. 1995. Intake limitations, feeding behavior, and rumen function of cows challenged with rumen fill from dietary fiber or inert bulk. Journal of Dairy Science 78: 118–1 33. Dalley, D.E., Roche, J.R., Grainger, C. Moate, P.J. 1999. Dry matter intake, nutrient selection and milk production of dairy cows grazing rainfed perennial pastures at different herbage allowances in spring. Australian Journal of Experimental Agriculture 39: 923-9 31. Davies, L.J. and Hunt, B.J. 1983. Growth of tropical grass introductions in mixed swards with ryegrass and clover under mowing. New Zealand Journal of Agricultural Research 26: 415-42 2. Delaby, L., Peyraud, J.L. Delagarde, R. 2001. Effect of the level of concentrate supplementation, herbage allowance and milk yield at turn-out on the performance of dairy cows in mid lactation at grazing. Animal science 73: 171–18 1. Dellow, D.W., Waghorn, G.C., Moseley, G. Re id, C.S.W. unpublished. Effect of forage species and level of intake on the breakd own of food particles in the rumen of sheep fed fresh white clover or fresh perennial ryegrass at two stages of growth. 26. Demment, M.W., Peyraud, J.-L. Laca, E.A. 1995. Herbage intake at grazing a modelling approach. In " Recent developments in the nutrition of herbivores. Proceedings of the IVth international symposium on the nutrition of herbivores ", Eds. Journet, M., Grenet, E., Farce, M.-H., Thériez, M. and Demarquilly, C., INRA Editions, pp. 121-141. Dijkstra, J. and France, J. 1994. Modelling and methodology in animal science. In " IVth international workshop on modelling nutrient utilization in farm animals ", Eds. Danfaer, A. and Lescoat, P., National Institute of Animal Science, pp. 9-18. Dijkstra, J., Neal, H.D.C., Beever, D.E. France, J. 1992. Simulation of nutrient digestion, absorption and outflow in the rumen: model description. Journal of Nutrition 122: 2239 -2 256. Dixon, R.M. and Stockdale, C.R. 1999. Asso ciative effects between forages and grains: consequences for feed utilisation. Australian Journal of Agricultural Research 50: 757-77 3. 279 Draper, N.R. and Smith, H. 1981. Applied regression analysis ('2nd'). John Wiley & Sons, New York, NY. Dulphy, J.P., Faverdin, P. Jarrige, R. 1989. Feed intake: the fill unit systems. In Ruminant nutrition. Recommended allowances and feed tables (ed. Jarrige, R.), pp. 61-71. INRA, Paris, France. Dynes, R.A., Henry, D.A. Masters, D.G. 2003. Characterising forages for ruminant feeding. Asian-Australasian Journal of Animal Sciences 16: 116-12 3. Easton, H.S. 1999. Endophyte in New Zealand ryegrass pastures, an overview. In " Ryegrass endophyte: an essential New Zealand symbiosis ", Eds. Woodfield, D. R. and Matthew, C., New Zealand Grassland Association, Grassland Research and Practice Series No. 7 pp. 1-9. Easton, H.S., Baird, D.B., Camaron, N.E., Kerr, G.A., Norriss, M. Stewart, A.V. 2001. Perennial ryegrass cultivars: herbage yield in multi-site plot trials. Proceedings of the New Zealand Grassland Association 63: 183– 18 8. Easton, H.S. and Couchman, J.N. 1999. Ryegrass endophyte and cattle growth in Northland. In " Ryegrass endophyte: an essential New Zealand symbiosis ", Eds. Woodfield, D. R. and Matthew, C., New Ze aland Grassland Association, Grassland Research and Practice Series No. 7 pp. 57-62. Elizalde, J.C., Merchen, N.R. Faulkner, D.B. 1999. In situ dry matter and crude protein degradation of fresh forages during the spring growth. Journal of Dairy Science 82: 1978 -1 990. Emanuele, S. and Staples, C.R. 1988. Effect of forage particle size an in situ digestion kinetics. Journal of Dairy Science 71: 1947 -1 954. Englyst, H.N., Trowell, H., Southgate, D.A. T. Cummings, J.H. 1987. Dietary fiber and resistant starch. American journal of clinical nutrition 46: 873-87 4. Forbes, J.M. 1986. T he voluntary food intake of farm animals . Butterworth, London, England. Forbes, J.M. 2000. Physiological and metabolic aspects of feed intake control. In Farm animal metabolism and nutrition (ed. D’Mello, J. P. F.), pp. 319-3 33. CABI Publishing, Edinburgh, UK. 280 Fox, D.G. and Barry, M.C. 1994. Using whole animal models to evaluate and refine cattle diets under widely varying conditions. In " IVth international workshop on modelling nutrient utilisation in farm animals ", Eds. Danfaer, A. and Lescoat, P., National Institute of Animal Science, pp. 143-162. Fox, D.G., Sniffen, C.J., O'Connor, J.D., Russell, J.B. Van Soest, P.J. 1992. A net carbohydrate and protein system for evaluati ng cattle diets: III. Cattle requirements and diet adequacy. Journal of Animal Science 70: 3578 -35 96. Fox, D.G., Tylutki, T.P., Tedeschi, L.O., Van Amburgh, M.E., Chase, L.E., Pell, A.N., Overton, T.R., Russell, J.B. 2003. The net carbohydrate and protein system for evaluating herd nutrition and nutrient excretion - CNCPS version 5.0 – Model documentation (CNCPS manual). http://www.cncps.cornell.edu/cncps/download/CNCPSManual.pdf France, J. and Siddons, R.C. 1993. Volatile fatty acid production. In Quantitative aspects of ruminant digestion and metabolism (ed. International, C.), pp. 107-121. Forbes, J.M.; France, J., Wallingford, UK. Fulkerson, W.J., Slack, K., Hennessy, D.W. Ho ugh, G.M. 1998. Nutrients in ryegrass (Lolium spp.), white clover (Trifolium repens) and kikuyu (Pennisetum clandestinum) pastures in relation to season and stage of regrowth in a subtropical environment. Australian Journal of Experimental Agriculture 38: 227–2 40. Gill, M., Beever, D.E. France, J. 1989. Biochemical bases needed for the mathematical representation of whole animal metabolism. Nutrition abstract reviews 2: 181-20 0. Givens, D.I. and Deaville, E.R. 1999. The current and future role of near infrared reflectance spectroscopy in animal nutrition: a review. Australian Journal of Agricultural Research 50: 1131 -1 14 5. Goering, H.K. and Van Soest, P.J. 1970. Forage fiber analyses (apparatus, reagents, procedures, and some applications) . Agriculture Research Service USDA, Washington, D.C., USA. Grainger, C. and McGowan, A.A. 1982. The significance of pre-calving nutrition of the dairy cow. In " Dairy production from pasture", Eds. Macmillan, K. L. and Taufa, V. K., New Zealand and Australian societies of animal production, pp. 135-171. Gray, D.I. and Lockhart, J.C. 1996. Management of summer feed deficits on three high- performing dairy farms in the Manawatu. Proceedings of the New Zealand Grassland Association 57: 9-13. 281 Hacker, J.B. and Minson, D.J. 1981. The digestibility of plant parts. Herbage abstracts 51: 459– 48 2. Hall, M.B. 1998. Making nutritional sens e of nonstructural carbohydrates. In " 9th Annual Florida Ruminant Nutrition Symposium ", Eds. pp. 108-12 1. Hamilton-Manns, M. and Crothers, R. 1999. Farmer experience of perennial ryegrass endophyte on a Manawatu dairy farm. In " Ryegrass endophyte: an essential New Zealand symbiosis ", Eds. Woodfield, D. R. and Matthew, C., New Zealand Grassland Association, Grassland Research and Practice Series No. 7 pp. 51-56. Hanna, W.W., Monson, W.G. Burton, G.W. 1973. Histological examination of fresh forage leaves after in vitro digestion. Crop Science 13: 98-10 2. Harris, S.L., Clark, D.A., Auldist, M.J., Waug h, C.D. Laboyrie, P.G. 1997a. Optimum white clover content for dairy pastures. Proceedings of the New Zealand Grassland Association 59: 29–3 3. Harris, S.L., Clark, D.A. Jansen, E.B.L. 1997b. Optimum white clover content for milk production. Proceedings of the New Zealand Society of Animal Production 57: 169- 1 71. Harris, S.L., Clark, D.A. Laboyrie, P.J. 1998a. Birdsfoot trefoil – an alternative legume for New Zealand dairy pastures. Proceedings of the New Zealand Grassland Association 60: 99–1 03. Harris, S.L., Clark, D.A., Waugh, C.D., Coperman, P.J.A. Napper, A.R. 1998b. Use of ‘Barkant’ turnips and ‘Superchow’ sorg hum to increase summer-autumn milk production. Proceedings of the New Zealand Society of Animal Production 58: 121- 1 24. Hatfield, R.D., Jung, H.J.G., Ralph, J., Buxton, D.R. Weimer, P.J. 1994. A comparison of the insoluble residues produced by klason lignin and acid detergent lignin procedures. Journal Science Food Agriculture 65: 51-5 8. Hatfield, R.D., Ralph, J. Grabber, J.H. 1999. Cell wall structural foundations: molecular basis for improving forage digestibilities. Crop Science 39: 27-3 7. Hodgson, J. 1990. Grazing management: science into practice . Longman scientific and technical, Essex, England. 282 Hodgson, J. and Brookes, I.M. 1999. Nutrition of grazing animals. In New Zealand pasture and crop science (ed. White, J. and Hodgson, J.), pp. 117-13 2. Oxford University Press, Auckland, N.Z. Hodgson, J., Matthews, P.N.P., Matthew, C. R.J., L. 1999. Pasture measurement. In New Zealand pasture and crop science (ed. White, J. and Hodgson, J.), pp. 59-65. Oxford University Press, Auckland, N.Z. Hoffman, P.C., Combs, D.K. Casler, M.D. 1998. Performance of lactating dairy cows fed alfalfa silage or perennial ryegrass silage. Journal of Dairy Science 81: 162–16 8. Holmes, C.W. 1987. Pastures for dairy cows. In Livestock feeding on pasture (ed. Nicol, A. M.), pp. 133-14 3. New Zealand Society of Animal Production, Hamilton, New Zealand. Holmes, C.W., Brookes, I.M., Garrick, D.J., Mackenzie, D.D.S., Parkinson, T.J. Wilson, G.F. 2002. Milk production from pasture . Massey University, Palmerston North, New Zealand. Hoogendoorn, C.J. 1986. Studies on the effects of graz ing regime on sward and dairy cow performance . Massey University, Palmerston North, New Zealand, p. Pages. Hoover, W.H., Crooker, B. A. Sniffen, C.J. 1976. Effects of differential solid-liquid removal rates on protozoal numbers in cont inuous culture of rumen contents. Journal of Animal Science 43: 528-5 34. Hristov, A. and Broderick, G.A. 1994. In vitro determination of ruminal protein degradability using [15N]-ammonia to correct for microbial nitrogen uptake. Journal of Animal Science 72: 1344 -1 35 4. Huntington, G.B. and Archibeque, S.L. 1999. Practical aspects of urea and ammonia metabolism in ruminants. In " Proceedings of the American Society of Animal Science ", Eds. Journal of Animal Science, pp. 1-11. Huntington, J.A. and Givens, D.I. 1995. The in situ technique for studying the rumen degradation of feeds: a review of the procedure. Nutrition Abstracts and Reviews 65: 65–9 3. 283 Hvelplund, T. and Weisbjerg, M.R. 2000. In situ techniques for the estimation of protein degradability and postrumen availability. In Forage evaluation in ruminant nutrition (ed. Givens, D. I., Owen, E., Axford, R. F. E. and Omed, H. M.), pp. 233-2 58. CABI Publishing, Inoue, T., Brookes, I.M. Barry, T. 1989. Effe cts of selection for shear strength on the voluntary intake and digestion of perennial ryegrass fed to sheep. Proceedings of the New Zealand Society of Animal Production 49: 221-2 24. INRA 1989. Ruminant nutrition. Recommended allowances and feed tables . Libbey Eurotex Eurotext, Paris, France. Ishler, V., Heinrichs, J. Varga, G. 1996. From feed to milk: understanding rumen function . http://www.das.psu.edu/dcn/catnut/PDF/rumen422.pdf Jacobs,J.L., Rigby, S.E., McKenzie, F.R., Ward, G.N., Kearney, G. Effect of lock up and harvest dates on dairy pasture dry matter yield and quality for silage in south- western Victoria. Australian Journal of Experimental Agriculture 38: 131-1 38. John, A. and Ulyatt, M.J. 1987. Importance of dry matter content to voluntary intake of fresh grass forages. Proceedings of the New Zealand Society of Animal Production 47: 13-1 6. Johnston, J.D. and Shivas, I.C. 1999. Formulating the fourth diet. In " Biotechnology in the feed industry : proceedings of Alltech's fifteenth annual symposium ", Eds. Lyons, T. P. and Jacques, K. A., Nottingham University Press, pp. 321-341. Jones, D.I.H. and Theodorou, M.K. 2000. Enzyme techniques for estimation digestibility. In Forage evaluation in ruminant nutrition (ed. Givens, D. I., Owen, E., Axford, R. F. E. and Omed, H. M.), pp. 155-173. CABI Publishing, Wallington, UK. Jung, H.G. and Fahey, G.C. 1983. Nutritiona l implications of phenolic monomers and lignin: a review. Journal of Animal Science 57: 206-21 9. Jung, H.G., Mertens, D.R. Payne, A.J. 1997. Correlation of acid detergent lignin and klason lignin with digestibility of forage dry matter and neutral detergent fiber. Journal of Dairy Science 80: 1622 -1 62 8. Kennedy, P.M. and Murphy, M.R. 1988. The nutritional implications of differential passage of particles through the ruminant alimentary tract. Nutrition research reviews 1: 189-20 8. 284 Kohn, R.A., Kalscheur, K.F. Hanigan, M. 1998. Evaluation of models for balancing the protein requirements of dairy cows. Journal of Dairy Science 81: 3402 -3 41 4. Kolver, E.S., Barry, M.C., Penno, J.W. Muller, L.D. 1996. Evaluation of the Cornell Net Carbohydrate and Protein System for dairy cows fed pasture-based diets. Proceedings of the New Zealand Society of Animal Production 56: 251-254. Kolver, E.S. and de Veth, M.J. 2002. Prediction of ruminal pH from pasture-based diets. Journal of Dairy Science 85: 1255 -1 26 6. Kolver, E.S. and Muller, L.D. 1998. Performance and nutrient intake of high producing Holstein cows consuming pasture or a total mixed ration. Journal of Dairy Science 81: 1403 -1 411. Kolver, E.S., Muller, L.D., Varga, G.A. Cassidy, T.J. 1998. Synchronization of ruminal degradation of supplemental carbohydrate with pasture nitrogen in lactating dairy cows. Journal of Dairy Science 81: 2017 -20 28. Kolver, E.S., Penno, J.W., MacDonald, K.A. , McGrath, J.M. Carter, W. 1999. Mowing pasture to improve milk production. In " Proceedings of the Ruakura farmers conference ", Eds. Media, M., Times Print, 51 pp. 95-96. Kolver, E.S., Roche, J.R., de Veth, M.J., Thorne , P.L. Napper, A.R. 2002. Total mixed rations versus pasture diets: Evidence for a genotype x diet interaction in dairy cow performance. Proceedings of the New Zealand Society of Animal Production 62: 246-25 1. Kolver, E.S., Roche, J.R., Miller, D. Densley, R. 2001. Maize silage for dairy cows. Proceedings of the New Zealand Grassland Association 63: 195-20 1. Koong, L.J., Baldwin, R.L., Ulyatt, M.J. Charlesworth, T.J. 1975. Iterative computation of metabolic flux and stoichiometric parame ters for alternate pathways in rumen fermentation. Computer programs in biomedicine 4: 209-213. Korte, C.J. 1982. Grazing management of pere nnial ryegrass white clover pasture in late spring. Proceedings of the New Zealand Grassland Association 43: 80-8 4. Lapierre, H. and Lobley, G.E. 2001. Nitrogen recycling in the ruminant: a review. Journal of Dairy Science 84: E223– E23 6. LIC 2002. Livestock Improvement Corporation: Dairy statistics 2001-2002 . Livestock Improvement Corporation Limited, Hamilton, New Zealand. 285 López, S., France, J., Dhanoa, M.S., Mould, F. Dijkstra, J. 1999. Comparison of mathematical models to describe disa ppearance curves obtained using the polyester bag technique for incubating feeds in the rumen. Journal of Animal Science 77: 1875 -1 888. MacDonald, K.A. 1999. Determining how to make inputs increase your economic farm surplus. In " Proceedings of the Ruakura farmers conference ", Eds. Media, M., Times Print, 51 pp. 78-87. MacDonald, K.A., Penno, J.W., Nicholas, P.K., Lile, J.A., Coulter, M. Lancaster, J.A.S. 2001. Farms systems - impact of stocki ng rate on dairy farm efficiency. Proceedings of the New Zealand Grassland Association 63: 223-2 27. Macmillan, K.L. 1997. Why don't cows cycle? In " Proceedings of the Ruakura farmers conference ", Eds. Lomas, J., WordWise and Thom, E. R., Brebner Print Ltd, 49 pp. 52- 62. Madsen, J. and Hvelplund, T. 1994. Prediction of in situ protein degradability in the rumen. Results of a European ringtest. Livestock Production Science 39: 201-21 2. Mambrini, M. and Peyraud, J.L. 1994. Mean retention time in digestive tract and digestion of fresh perennial ryegrass by lactating dairy cows: influence of grass maturity and comparison with a maize silage diet. Reproduction Nutrition Development 34: 9-23. Marotti, D.M., Chapman, D.F., Cosgrove, G.P., Parsons, A.J. Egan, A.R. 2002. New opportunities to improve forage based production systems. Proceedings of the New Zealand Society of Animal Production 62: 273-277. Martz, F.A. and Belyea, R.L. 1986. Role of particle size and forage quality in digestion and passage by cattle and sheep. Journal of Dairy Science 69: 1996 -2 00 8. Mathers, J.C. and Aitchison, E.M. 1981. Direct estimation of the extent of contamination of food residues by microbial matter after incubation within synthetic fibre bags in the rumen. Journal of Agricultural Science, Cambridge 96: 691. Matthews, P.N.P., Harrington, K.C. Hampton, J.G. 1999. Management of grazing systems. In New Zealand pasture and crop science (ed. White, J. and Hodgson, J.), pp. 153-174. Oxford University Press, Auckland, N.Z. Mayer, D.G. and Butler, D.G. 1993. Statistical validation. Ecological Modelling 68: 21-3 2. 286 McDonald, I.M. 1981. A revised model for the estimation of protein degradability in rumen. Journal of Agricultural Science, Cambridge 96: 251-2 52. McGrath, D.F., Dawson, J.E., Thomson, N.A. Simons, H.P. 1998. More summer milk - the opportunities identified. In " Proceedings of the Ruakura farmers conference ", Eds. Lomas, J., Brebner Print Ltd, 50 pp. 85-94. McNabb, W.C., Waghorn, G.C., Peters, J.S. Barry, T.N. 1996. The effect of condensed tannins in Lotus pedunculatus on the solubi lization and degradation of ribulose-1,5- bisphosphate carboxylase (Rubisco) protei n in the rumen and the sites of Rubisco digestion. British journal of nutrition 76: 535-54 9. Mehrez, A.Z. and Ørskov, E.R. 1977. The use of a Dracon bag technique to determine rate of degradation of protein and energy in the rumen. Journal of Agricultural Science, Cambridge 88: 645-65 0. Menneer, J., Ledgard, S., McLay, C. Silvester, W. 2001. What impact does dairy cow pugging have on clover N2 fixation and long-term farm production? Proceedings of the New Zealand Grassland Association 63: 63-67. Mertens, D.R. 1992a. Critical conditions in determining detergent fibers. In " Proceedings of National Forage Testing Association (NFT A) Forage Analysis Workshop ", Eds. pp. c1-c8. Mertens, D.R. 1992b. Nonstructural and structural carbohydrates. In Large Dairy Herd Management (ed. Van Horn, H. H. and Wilcox, C. J.), pp. 219-23 5. ADSA, Champaign, IL. Mertens, D.R. 1994. Regulation of forage intake. In Forage quality, evaluation and utilization (ed. Fahey, G. C., Moller, S., Merten s, D. R. and Collins, M.), pp. 450-49 3. American Society of Agronomy, Inc. Crop Science Society of America, Inc. Soil Science Society of America, Inc., Madison, WI. Mertens, D.R. 1997. Creating a system for meeting the fiber requirements of dairy cows. Journal of Dairy Science 80: 1463 -1 48 1. Messman, M.A., Weiss, W.P. Albrecht, K.A. 1996. In situ disappearance of individual proteins and nitrogen from legume forages containing varying amounts of tannins. Journal of Dairy Science 79: 1430 -1 43 5. 287 Michalet-Doreau, B. and Ould-Bah, M.Y. 1992. In vitro and in sacco methods for the estimation of dietary nitrogen degradability in the rumen: a review. Animal Feed Science and Technology 40: 57-86. Min, B.R., Barry, T.N., Attwood, G.T. McNa bb, W.C. 2003. The effect of condensed tannins on the nutrition and health of ruminants fed fresh temperate forages: a review. Animal Feed Science and Technology 106: 3-19. Minson, D.J. 1990. Forage in ruminant nutrition . Academic Press, Inc., San Diego, California - USA. Minson, D.J. and Wilson, J.R. 1980. Comparative digestibilit y of tropical and temperate forage - a contrast betw een grasses and legumes. Journal of the Australian Institute of Agricultural Science 46: 247-2 49. Minson, D.J. and Wilson, J.R. 1994. Prediction of intake as an element of forage quality. In Forage quality, evaluation and utilization (ed. Fahey, G. C., Moller, S., Mertens, D. R. and Collins, M.), pp. 533-56 3. American Society of Agronomy, Inc. Crop Science Society of America, Inc. Soil Science Society of America, Inc., Madison, WI. Mitchell, P.L. 1997. Misuse of regressi on for empirical validation of models. A gricultural Systems 54: 313-32 6. Moller, S. 1997. An evaluation of major nutrients in dairy pasture in New Zealand and their effects on milk production and herd reproductive performance . Doctor of Philosophy, Massey University, Palmerston North, p. Pages. Moller, S.N., Parker, W.J. Edwards, N.J. 1996. Within-year variation in pasture quality has implications for dairy cow nutrition. Proceedings of the New Zealand Grassland Association 57: 173-17 7. Moore, K.J. and Hatfield, R.D. 1994. Carbohydrates and forage quality. In Forage quality, evaluation and utilization (ed. Fahey, G. C., Moller, S., Mertens, D. R. and Collins, M.), pp. 229-28 0. American Society of Agronomy, Inc. Crop Science Society of America, Inc. Soil Science Society of America, Inc., Madison, WI. Mould, F.L. and Ørskov, E.R. 1984. Manipulation of rumen fluid pH and its influence on cellulolysis in sacco, dry matter degradation and the rumen microflora of sheep offered either hay or concentrate. Animal Feed Science and Technology 10: 1-14. 288 Murphy, M.R. 1990. Chemical and physical properties of lignocellulose that limit intake and digestion. In Microbial and plant opportunities to improve lignocellulose utilization by ruminants (ed. Akin, D. E., Ljungdahl, L. G., Wilson, J. R. and Harris, P. J.), pp. 17-22. Murphy, M.R., Baldwin, R.L. Koong, L.J. 1982. Estimation of stoichiometric parameters for rumen fermentation of roughage and concentrate diets. Journal of Animal Science 55: 411-42 1. Niezen, J.H., Waghorn, G.C., Lyons, T.B. Corson, D.C. 1998. The potential benefits of ensiling the forage legume sulla compared with pasture. Proceedings of the New Zealand Grassland Association 60: 105-10 9. Nocek, J.E. 1985. Evaluation of specific vari ables affecting in situ estimates of ruminal dry matter and protein digestion. Journal of Animal Science 60: 1347 -13 58. Nocek, J.E. 1988. In situ and other methods to estimate ruminal protein and energy digestibility: a review. Journal of Dairy Science 71: 2051 -2 06 9. Nocek, J.E. 1997. Bovine acidos is: implications on lameness. Journal of Dairy Science 80: 1005 – 1028. Nocek, J.E. and English, J.E. 1986. In situ degradation kinetics: evaluation of rate determination procedure. Journal of Dairy Science 69: 77-87. Nocek, J.E. and Grant, A.L. 1987. Characteriza tion of in situ nitrogen and fiber digestion and bacterial nitrogen contamination of hay crop forages preserved at different dry matter percentages. Journal of Dairy Science 64: 552-56 4. Nocek, J.E. and Russell, J.B. 1988. Protein and energy as an integrated system. Relationship of ruminal protein and carboh ydrate availability to microbial synthesis and milk production. Journal of Dairy Science 71: 2070 -2 107. Nozière, P. and Michalet-Doreau, B. 2000. In sacco methods. In Farm animal metabolism and nutrition (ed. D’Mello, J. P. F.), pp. 233-25 3. CABI Publishing, Edinburg, UK. NRC 2001. National Research Council: Nutrient requirement of dairy cattle ('7th revised edition'). National Academy Press, Washington, DC. 289 O ' Connor, J.D., Sniffen, C.J. , Fox, D.G. Chalupa, W. 1993. A net carbohydrate and protein system for evaluating cattle diets: IV. Predicting amino acid adequacy. Journal of Animal Science 71: 1298 -1 311. Olubobokun, J.A., Craig, W.M. Pond, K.R. 1990. Effects of mastication and microbial contamination on ruminal in situ forage disappearance. Journal of Animal Science 68: 3371. Ørskov, E.R. 1999. Supplement strategies for ruminants and management of feeding to maximize utilization of roughages. Preventive Veterinary Medicine 38: 179-18 5. Ørskov, E.R. 2000. The in situ technique for the estimation of forage degradability in ruminants. In Forage evaluation in ruminant nutrition (ed. Givens, D. I., Owen, E., Axford, R. F. E. and Omed, H. M.), pp. 175-21 3. CABI Publishing, Ørskov, E.R. and McDonald, I. 1970. The estimation of protein degradability in the rumen from incubation measurements weighted according to rate of passage. Journal of Agricultural Science, Cambridge 92: 499-50 3. Penno, J.W. 2002. T he response by grazing dairy cows to supplementary feeds . Ph.D., Massey University, Palmerston North, p. 305. Penno, J.W., Holmes, C.W., MacDonald, K.A. Walsh, B.J. 1998. The effect of stage of lactation and season on milksolids resp onse to supplementary feeding of dairy cows. Proceedings of the New Zealand Society of Animal Production 58: 113-11 6. Peyraud, J.L., Delaby, L., Delagarde, R. Marquis, B. 1998. Effect of energy supplementation and herbage allowance on herbage intake by grazing dairy cows. Effet de l'apport de concentre energetique et des quantites d'herbe offerte sur l'ingestion des vaches laitieres au paturage. Rencontres Recherches Ruminants 5: 217-2 20. Phillips, C.J.C. 1988. The use of conserved forage as a supplement for grazing dairy cows. Grass and forage science 43: 215-23 0. Pinares, C. and Holmes, C.W. 1996. Effects of feeding silage and extending lactation on the pastoral dairy system. Proceedings of the New Zealand Society of Animal Production 56: 239-2 41. 290 Pitt, R.E., Van Kessel, J.S., Fox, D.G., Pell, A.N. , Barry, M.C. Van Soest, P.J. 1996. Prediction of ruminal volatile fatty acids and pH within the net carbohydrate and protein system. Journal of Animal Science 74: 226-2 44. Pond, K.R., Ellis, W.C. Akin, D.E. 1984. Ingestive mastication and fragmentation of forages. Journal of Animal Science 58: 1567 -15 74. Poppi, D.P. 1996. Prediction of food inta ke in ruminants from analyses of food composition. Australian Journal of Agricultural Research 47: 489-50 4. Poppi, D.P., Hendricksen, R.E. Minson, D.J. 1985. The relative resistance to escape of leaf and stem particles from the rumen of cattle and sheep. Journal of Agricultural Science, Cambridge 105: Poppi, D.P., Hughes, T.P. L’Huillier, P.J. 1987. Intake of pasture by grazing ruminants. In Livestock feeding on pasture (ed. Nicol, A. M.), pp. 55-63. New Zealand Society of Animal Production, Hamilton, New Zealand. Ralph, J., Grabber, J.H. Hatfield, R.D. 1995. Lignin-ferulate cross-links in grasses: active incorporation of ferulate polysaccharide esters into ryegrass lignins. Carbohydrates research 275: 167-1 78. Reeves III, J.B. 2000. Use of near infrared reflectance spectroscopy. In Farm animal metabolism and nutrition (ed. D’Mello, J. P. F.), pp. 185-20 7. CABI Publishing, Edinburgh, UK. Reid, C.S.W. 1984. The progress of solid feed residues through the rumino-reticulum: the ins and outs of particles. In " Ruminant physiology: concepts and consequences ", Eds. Baker, S. K., Gawthorne, J. M., Mackintosh, J. B. and Purser, D. B., University of Western Australia, pp. 79-84. Reis, R.B. and Combs, D.K. 2000. Effects of corn processing and supplemental hay on rumen environment and lactation performance of dairy cows grazing grass-legume pasture. Journal of Dairy Science 83: 2529 -2 538. Robinson, P.H. and McQueen, R.E. 1997. Influence of level of concentrate allocation and fermentability of forage fiber on chewing behavior and production of dairy cows. Journal of Dairy Science 80: 681-6 91. 291 Russell, J.B., O'Connor, J.D., Fox, D.G., Van Soest, P.J. Sniffen, C.J. 1992. A net carbohydrate and protein system for ev aluating cattle diets: I. Ruminal fermentation. Journal of Animal Science 70: 3551-3 56 1. Russell, J.B. and Strobel, H.J. 1993. Microbial energetics. In Quantitative aspects of ruminant digestion and metabolism (ed. Forbes, J. M. and France, J.), pp. 164-186. CAB International, Rys, G.T., Smith, N. Slay, M.W. 1988. Alternative forage species in Hawkes Bay. Proceedings of the Agronomy Society of New Zealand 18: 75-8 0. SAS 2001. T he SAS system for Windows . In (Ed.), (Vol. pp. Cary, NC, USA: SAS Institute. Satter, L.D. and Slyter, L.L. 1974. Effect of ammonia concentration on rumen microbial protein production in vitro. British journal of nutrition 32: 199-20 8. Sauvant, D. 1991. The use of modelling to predict animal responses to diet. In " Proceedings Ralston Purina International Scientific Advisory Board Meeting ", Eds. pp. Shabi, Z., Arieli, A., Bruckental, I., Aharoni, Y., Zamwel, S., Bor, A. Tagari, H. 1998. Effect of the synchronization of the degradation of dietary crude protein and organic matter and feeding frequency on ruminal fermentation and flow of digesta in the abomasum of dairy cows. Journal of Dairy Science 81: 1991 -20 00. Shaw, R.J., Thomson, N.A., McGrath, D.F. Dawson, J.E. 1997. Mora summer milk - an on farm demonstration of research principles. Proceedings of the New Zealand Grassland Association 59: 149–1 55. Sheaffer, C.C., Seguin, P. Cuomo, G.J. 1998. Sward characteristics and management effects on cool-season gr ass forage quality. In Grass for dairy cattle (ed. Cherney, J. H. and Cherney, D. J. R.), pp. 73-100. CAB International, Cambridge, UK. Simons, H.P., Thomson, N.A., McGrath, D.F. Dawson, J.E. 1998. More Summer Milk – a further investigation into improving summer milk production. Proceedings of the New Zealand Grassland Association 60: 27–3 1. Smoler, E., Rook, A.J., Sutton, J.D. Beever , D.E. 1998. Prediction of milk protein concentration from elements of the metabolizable protein system. Journal of Dairy Science 81: 1619 -1 623. 292 Sniffen, C.J., O'Connor, J.D., Van Soest, P.J., Fox, D.G. Russell, J.B. 1992. A net carbohydrate and protein system for evaluating cattle diets: II. Carbohydrate and protein availability. Journal of Animal Science 70: 3562 -3 577. Stockdale, C.R. 1995. Maize silage as a supp lement for pasture-fed dairy cows in early and late lactation. Australian Journal of Experimental Agriculture 35: 19-26. Stockdale, C.R. 1996. Substitution and production responses when lactating dairy cows graze a white clover pasture supplemented with maize silage. Australian Journal of Experimental Agriculture 36: 771-77 6. Stockdale, C.R. 1997. Supplements improve th e production of dairy cows grazing either white clover or paspalum-dominan t pastures in late lactation. Australian Journal of Experimental Agriculture 37: 295– 30 2. Stockdale, C.R. 1999a. Effects of cereal grain, lupins–cereal grain or hay supplements on the intake and performance of grazing dairy cows. Australian Journal of Experimental Agriculture 39: 811– 81 7. Stockdale, C.R. 1999 b. Effects of season and time since defoliation on the nutritive characteristics of three irrigated perennial pasture species in northern Victoria 1. Energy, protein and fibre. Australian Journal of Experimental Agriculture 39: 555-5 65. Stockdale, C.R. 2000a. Differences in body co ndition and body size affect the responses of grazing dairy cows to high-energy supplements in early lactation. Australian Journal of Experimental Agriculture 40: 903–9 11. Stockdale, C.R. 2000b. Levels of pasture substitution when concentrates are fed to grazing dairy cows in northern Victoria. Australian Journal of Experimental Agriculture 40: 913–9 21. Stockdale, C.R. and Dellow, D.W. 1995. The productivity of lactating cows grazing white clover and supplemented with maize silage. Australian Journal of Agricultural Research 45: 1205 -1 21 7. St-Pierre, N.R. 2001. Integrating quantitative findings from multiple studies using a mixed model methodology. Journal of Dairy Science 84: 741-75 5. St-Pierre, N.R. 2003. Reassessment of bias es in predicted nitrogen flows to the duodenum by NRC 2001. Journal of Dairy Science 86: 344–35 0. 293 St-Pierre, N.R. and Thraen, C.S. 1999. Animal grouping strategies, sources of variation, and economic factors affecting nutrient balance on dairy farms. Journal of Dairy Science 82: 72-8 3. Suksombat, W., Holmes, C.W. Wilson, G.F. 1994. Effects of herbage allowance and a high protein supplement on performance of dairy cows grazing on autumn-winter pastures. Proceedings of the New Zealand Society of Animal Production 54: 83-86. Terrill, T.H., Rowan, A.M., Douglas, G.B. Ba rry, T.N. 1992. Determination of extractable and bound condensed tannin concentration in forage plants, protein concentrate meals and cereal grains. Journal of the Science of Food and Agriculture 58: 321-32 9. Thom, E.R., Clark, D.A., Van Vught, V.T., Wa ugh, C.D. 1998. Pasture species and drought impact on milk yield 1. Milk yield responses in the Waikato. Proceedings of the New Zealand Grassland Association 60 : 39– 44. Thom, E.R., Clark, D.A. Waugh, C.D. 1999. Endophyte and dairy production in New Zealand: experience at the Dairyi ng Research Corporation. In " Ryegrass endophyte: an essential New Zealand symbiosis ", Eds. Woodfield, D. R. and Matthew, C., New Zealand Grassland Association, Grassland Research and Practice Series No. 7 pp. 39-44. Thomson, N.A. 1996. Irrigation and pasture quality. Proceedings of Ruakura Farmers Conference 48: 58-66. Thomson, N.A., Davis, K.L., McGrath, J.M., Hainsworth, R.J. Clough, J.C. 1998. Impact of feeding maize silage on dairy production. Proceedings of the New Zealand Grassland Association 60: 51-55. Tilley, J.M.A. and Terry, R.A. 1963. A two stage technique for the in vitro digestion of forage crops. Journal of the British Grassland Society 18: 104–1 11. Towers, N.R. 1986. Facial eczema - problems and successes in control. Proceedings of the New Zealand Grassland Association 47: 121-1 27. Tylutki, T.P. and Fox, D.G. 1997. Application of the Cornell nutrient management planning system: optimizing herd nutrition. In "Proceedings of the Cornell nutrition conference ", Eds. Cornell University, pp. 54-60. 294 Ulyatt, M.J. 1973. The feeding value of herbage. In Chemistry and biochemistry of herbage (ed. Butler, D. G. and Bailey, L. D. ), pp. 131-17 8. Academic Press, Londin, UK. Ulyatt, M.J. 1981. The feeding value of herbage: Can it be improved? New Zealand Agricultural Science 15: 200-20 5. Ulyatt, M.J. 1982. Plant fibre and regulation of digestion in the ruminant. In " Dietary fibre in human and animal nutrition ", Eds. Wallace, G. and Bell, L., The Royal Society of New Zealand, pp. 103-10 7. Ulyatt, M.J. 1984. Pasture composition and animal production. In " Ruminant physiology: concepts and consequences ", Eds. Baker, S. K., Gawthorne, J. M., Mackintosh, J. B. and Purser, D. B., University of Western Australia, pp. 195-203. Ulyatt, M.J., Dellow, D.W., John, A., Reid, C.S.W. Waghorn, G.C. 1986. The contribution of chewing during eating and rumination to the clearance of digesta from the ruminoreticulum. In Control of digestion and metabolism in ruminants (ed. Milligan, L. P., Grovum, W. L. and Dobson, A.), pp. 498-5 15. Prentice-Hall, Englewood Cliffs, NJ-USA. Ulyatt, M.J., Thomson, D.J., Beever, D.E., Evans, R.T. Haines, M.J. 1988. The digestion of perennial ryegrass (Lolium perenne cv. Melle) and white clover (Trifolium repens cv. Blanca) by grazing cattle. British journal of nutrition 60: 137-14 9. Ulyatt, M.J. and Waghorn, G.C. 1993. Limitation s to high levels of dairy production from New Zealand pastures. In Improving the quality and intake of pasture-based diets for lactating dairy cows (ed. Edwards, N. J. and Parker, W. J.), pp. 11-32. Massey University, Palmerston North, New Zealand. Van Soest, P.J. 1967. Development of a comprehensive system of feed analyses and its application to forages. Journal of Animal Science 26: 119-12 8. Van Soest, P.J. 1978. Preharvest factors influencing quality of conserved forage. Journal of Animal Science 47: 712-72 0. Van Soest, P.J. 1994. Nutritional Ecology of the Ruminant ('2nd edition'). Cornell University Press, Ithaca, NY. 295 Van Vuuren, A.M., Van der Koelen, C.J. Vroons-De Bruin, J. 1993. Ryegrass versus corn starch or beet pulp fiber diet effects on digestion and intestinal amino acids in dairy cows. Journal of Dairy Science 76: 2692 -27 00. Vérité, R. and Journet, M. 1970. Influence de la teneur en ean et da la deshydratation de I’herbe sur sa valeur alimenta ire pour less vaches laitieres. Annales de zootechnie 19: 255-26 8. Verkerk, G., Stevens, J. McKay, B. 2000. Management for sustainable reproduction. In " Proceedings of the Ruakura farmers conference ", Eds. Lomas, J., Brebner Print, 52 pp. 53-65. Waghorn, G.C. 1986. Changes in rumen digesta of cows during a restricted feeding period when offered fresh red clover, lucerne, or lucerne hay. New Zealand Journal of Agricultural Research 29: 233-24 1. Waghorn, G.C. 2002. Can forages match concentrate diets for dairy production? Proceedings of the New Zealand Society of Animal Production 62: 261-266. Waghorn, G.C., Adams, N.R. Woodfield, D.R. 2002. Deleterious substances in grazed pastures. In Sheep nutrition (ed. Freer, M. and Dove , H.), pp. 333-35 6. CSIRO Publishing, Canberra, Australia. Waghorn, G.C. and Barry, T.N. 1987. Pasture as a nutrient source. In Livestock feeding on pasture (ed. Nicol, A. M.), pp. 21-38. New Zealand Society of Animal Production, Hamilton, New Zealand. Waghorn, G.C. and Caradus, J.R. 1994. Screening white clover cultivars for improved nutritive value - development of a method. Proceedings of the New Zealand Grassland Association 56: 49-53. Waghorn, G.C., Douglas, G.B., Niezen, J.H. , McNabb, W.C. Foote, A.G. 1998. Forages with condensed tannins - their management and nutritive value for ruminants. Proceedings of the New Zealand Grassland Association 60: 89-9 8. Waghorn, G.C. and Jones, W.T. 1989. Bloat in cattle, 46. Potential of dock (Rumex obstusfolius) as an antibloat agent for cattle. New Zealand Journal of Agricultural Research 32: 227-23 5. 296 Waghorn, G.C. and McNabb, W.C. 2003. Co nsequences of plant phenolic compounds for productivity and health of ruminants. Proceedings of the Nutrition Society 62: 383-39 2. Waghorn, G.C., Reed, J.D. Ndlovu, L.R. 1997. Condensed tannins and herbivore nutrition. In " X VIII International Grassland Congress ", Eds. Buchanan-Smith, J. G., Bailey, L. D. and McCaughey, P., III pp. 153-166. Waghorn, G.C., Reid, C.S.W., Ulyatt, M.J. John, A. 1986. Feed comminution, particle composition and distribution between the four compartments of the stomach in sheep fed chaffed lucerne hay at two feeding frequencies and intake levels. Journal of Agricultural Science, Cambridge 106: 287-296. Waghorn, G.C. and Shelton, I.D. 1995. Effect of condensed tannins in Lotus pedunculatus on the nutritive value of ryegrass (Lolium perenne) fed to sheep. Journal of Agricultural Science, Cambridge 125: 291-297. Waghorn, G.C., Shelton, I.D., McNabb, W.C. McCutcheon, S.W. 1994. Effects of condensed tannins in Lotus pedunculatus on its nutritive value for sheep 2. Nitrogenous aspects. Journal of Agricultural Science, Cambridge 123: 109-11 9. Waghorn, G.C., Shelton, I.D. Thomas, V.J. 1989. Particle breakdown and rumen digestion of fresh ryegrass ( Lolium perenne L.) and lucerne ( Medicago sativa L.) fed to cows during a restricted feeding period. British journal of nutrition 61: 409-42 3. Waite, R., Johnson, M.J. Armstrong, D.G. 1964. Evaluation of artificially dried grass as a source of energy for sheep. I. The effect of stage of maturity on the apparent digestibility of rye-grass, cooksfoot and timothy. Journal of Agricultural Science, Cambridge 62: 391-39 8. Wales, W.J., Dellow, D.W. Doyle, P.T. 1999a. Degradabilities of dry matter and crude protein from perennial herbage and supple ments used in dairy production systems in Victoria. Australian Journal of Experimental Agriculture 39: 645–6 56. Wales, W.J., Doyle, P.T., Stockdale, C.R. Dellow, D.W. 1999b. Effects of variations in herbage mass, allowance, and level of supplement on nutrient intake and milk production of dairy cows in spring and summer. Australian Journal of Experimental Agriculture 39: 119-1 30. 297 Wales, W.J., Williams, Y.J. Doyle, P.T. 2001. Effect of grain supplementation and the provision of chemical or physical fibre on marginal milk prod uction responses of cows grazing perennial ryegrass pastures. Australian Journal of Experimental Agriculture 41: 465-4 71. Walker, G.P., Stockdale, C.R., Wales, W.J., Doyle, P.T. Dellow, D.W. 2001. Effect of level of grain supplementation on milk production responses of dairy cows in mid–late lactation when grazing irrigated pastures high in paspalum (Paspalum dilatatum Poir.). Australian Journal of Experimental Agriculture 41: 1-11. Wang, Y., Douglas, G.B., Waghorn, G.C., Barry, T.N. Foote, A.G. 1996a. Effect of condensed tannins in Lotus corniculatus upon lactation performance in ewes. Journal of Agricultural Science, Cambridge 126: 353-362. Wang, Y., Waghorn, G.C., McNabb, W.C., Barry, T.N., Hedley, M.J. Shelton, I.D. 1996b. Effect of condensed tannins in Lotus corniculatus upon the digestion of methionin and cysteine in the small intestine of sheep. Journal of Agricultural Science, Cambridge 127: 413-4 21. Wattiaux, M.A., Mertens, D.R. Satter, L.D. 1991. Effect of source and amount of fiber on kinetics of digestion and specific gravity of forage particles in the rumen. Journal of Dairy Science 74: 3872 -38 83. Weakley, D.C., Stem, M.D. Satter, L.D. 1983. Factors affecting disappearance of feedstuffs from bags suspended in the rumen. Journal of Animal Science 56: 493- 5 07. Weimer, P.J. 1996. Why don’t ruminal bacteria digest cellulose faster? Journal of Dairy Science 79: 1496 -1 502. Weimer, P.J., Waghorn, G.C., Odt, C.L. Mertens, D.R. 1999. Effect of diet on population of three species of ruminal cellulolytic bacteria in lactating dairy cows. Journal of Dairy Science 82: 122-1 34. Weiss, W.P. 1994. Estimation of digestibil ity of forages by laboratory methods. In Forage quality, evaluation and utilization (ed. Fahey, G. C., Moller, S., Mertens, D. R. and Collins, M.), pp. 644-68 1. American Society of Agronomy, Crop Science Society of America and the Soil Science Society of America, Madison, Wisconsin. 298 Wilman, D. and Agiegba, J.E. 1982. The effects of clover variety, cutting interval and nitrogen application on herbage yields , proportions and heights in perennial ryegrass-white clover swards. Grass and forage science 37: 1-13. Wilson, G.F. and Moller, S. 1993. Imbalances in New Zealand pastures for dairy cows. In Improving the quality and intake of pasture-based diets for lactating dairy cows (ed. Edwards, N. J. and Parker, W. J.), pp. 59-70. Department of Agricultural and Horticultural Systems Management, Massey University, Palmerston North, New Zealand. Wilson, G.F., Moller, S., Parker, W.J. Hodgson, J. 1995. Seasonal differences in pasture composition and nutritional implications. In " Dairyfarming Annual ", Eds. Massey University, 47 pp. 46-48. Wilson, J.R. 1991. Plant structures, their digestive and physical breakdown. In " Recent advances on the nutrition of herbivores ", Eds. Ho, Y. W. and al., e., Malaysian Society of Animal Production, pp. 207-216. Wilson, J.R. 1993. Organization of forage plant tissue. In Forage cell wall structure and digestibility (ed. Jung, H. G., Buxton, D. R., Hatfield, R. D. and Ralph, J.), pp. 1-32. USDA-Agricultural Research Service and the U.S. Dairy Forage Research Center, Madison, Wisconsin. Wilson, J.R., Deinum, B. Engels, F.M. 1991. Temperature effects on anatomy and digestibility of leaf and stem of tr opical and temperate forage species. Netherlands Journal of Agricultural Science 39: 31-48. Wilson, J.R. and Hatfield, R.D. 1997. Structural and chemical changes of cell wall types during stem development: consequenc es for fibre degradation by rumen microflora. Australian Journal of Agricultural Research 48: 165-18 0. Wilson, J.R. and Mertens, D.R. 1995. Cell wall accessibility and cell structure limitations to microbial digestion of forage. Crop Science 35: 251-2 59. Woodward, S.L., Auldist, M.J., Laboyrie, P.J. Jansen, E.B.L. 1999. Effect of Lotus corniculatus and condensed tannins on milk yield and milk composition of dairy cows. Proceedings of the New Zealand Society of Animal Production 59: 152-15 5. 299 Woodward, S.L., Chaves, A.V., Waghorn, G.C. Laboyrie, P.J. 2002. Supplementing pasture-fed dairy cows with pasture silage, maize silage, Lotus silage or sulla silage in summer– does it increase production? Proceedings of the New Zealand Grassland Association 64: 85-8 9. Woodward, S.L., Waghorn, G.C., Ulyatt, M.J. Lassey, K.R. 2001. Early indications that feeding Lotus will reduce meth ane emissions from ruminants. Proceedings of the New Zealand Society of Animal Production 61: 23-2 6. 300