Rapeseed meal as a supplementary protein for dairy cows on grass silage-based diet, with the emphasis on the Nordic AAT-PBV feed protein evaluation system Mikko Tuori University of Helsinki Department of Animal Science SF - 00710 Helsinki, Finland Academic Dissertation To he presented with thepermission of the Faculty ofAgriculture and Forestry of the University ofHelsinki, forpublic criticism in the Auditorium 82, Viikki, on November 27,1992 at 12 noon. Agric. Sei. Finl. 1 (1992) https://www.c-info.fi/en/info/?token=isvkAld0FAB-dFRG.levPX3vSLpFpe_NYPMwDEg.bLv_DzQ0GofK5efljSGIwMl4_yj3BEPV95XVFj6bM-XcImT4eOhQiIXnmWLx-O19XUevHJgsIetkYc2AAOdFqmc9ZXg-ec5WaQOgeJQ3enPEkIOsHDB9KPWvWfWHORtK8VDjYwL9L6QbfGFX0PuayXyAof9ZRGh7tFnaMeiX2o2tttb0kmLf8xwdlHjeuY0T3DrYUp71JsKZBov8mt7LzRGlN5wPDL7xuXk49uFmaypzuJHhSW3sZCGduPrvNDa877c 369 Acknowledgements The present experiments were conducted at the Department of Animal Science, Univer- sity of Helsinki, with the exception of one conducted in Maaninka at the North Savo Research Station of the Agricultural Research Centre. First and foremost, 1 would like to thankProfessor Unto Tulisalo for suggesting rapeseed meal as the topic of this doctoral dissertation, and for his support and encouragement dur- ing the work. Further, I am indebted to Professor Liisa Syrjälä-Qvist for suggesting that the work be carried out the Department of Animal Science, for supervising several of the experiments, and for her unfailing support during the study. My heartfelt thanks go to our former Director, now Director General, Professor Esko Poutiainen for supervising the first experiment. I am very grateful to Dr. Pekka Huhtanen for his advice and comments on the calcula- tions, to Professor Matti Näsi for his preliminary comments on the manuscript, to Mr. Veijo Viiva for his advice on statistics, to Professor Vappu Kossila and Dr. Tuomo Varvikko for their constructive criticism on the manuscript. Carrying out the experiments called for much work in the laboratory and with the ani- mals, for which I am indebted to the staffof the Department of Animal Science, too numer- ous to be mentioned here. My special thanks go to Ms. Marjatta Suvitie, and Mr. Kalle Rinne, and the staffofNorth Savo Research Station for carrying out one of the experiments. Ms. Tuija Niskanen, Mr. Mikko Maisi, Ms. Heli-Maria Ojanperä, Ms. Aila Asikainen and Ms. Taina Voutilainen, who assisted in the experiments while undergraduate students, also deserve to be thanked. Ms. Terttu Heikkilä, and Dr. Pekka Huhtanen have earned my gratitude by providing the milking trial data, and so has Mr. Vesa Toivonen for amino acids analyses and Ms. Aila Vanhatalo for determination of contents in the mobile bag tests at the Institute of Animal Production, Agricultural Research Center. I would like to thank Ms. Liisa Fellman-Paul for translating the mainbody oftext to Eng- lish and Dr. Andrew Root and his wife Taija, for checking the text of the tables and appen- dices etc. The financial support provided by Öljynpuristamo Oy, and the Agricultural Research Foun- dation of August Johannes and Aino Tiura is acknowledged with gratitude. Last, but far from least, my thanks to my wife Ritva, whose support and encouragement have helped me complete the work, and to my children Katri and Eeva for their patience during my long days at work. Mikko Tuori 27 November 1992 Agric. Sei. Finl. 1 (1992) Agric. Sei. Fin!. 1 (1992) 370 TABLE OF CONTENTS Acknowledgements 369 Abstract 375 1. Introduction 376 2. Review of literature 376 2.1 Rapeseed meal 376 2.1.1. Rapeseed production 376 2.1.1.1. Origins and cultivation of turnip rape 376 2.1.1.2. Cultivation of rapeseed in Finland 378 2.1.2. Compositionof rapeseed 379 2.1.2.1. Whole seed 379 2.1.2.2. Rapeseed meal 379 2.1.2.3. Rapeseed protein 380 2.1.2.4. Rapeseed fat 381 2.1.2.5. Carbohydrates of rapeseed meal 382 2.1.2.6. Minerals 382 2.1.2.7. Glucosinolates 382 2.1.2.7.1. Structure and classification ofrapeseed glucosinolates 382 2.1.2.7.2. Analysis of glucosinolates in RSM 384 2.1.2.7.3. Physiological effects ofglucosinolates 384 2.1.2.7.4. Glucosinolates in milk 385 2.1.2.8. Sinapine 386 2.1.2.9. Phytic acid 386 2.1.2.10. Tannins 386 2.2. Rapeseed meal as a protein supplement for dairy cows 387 2.3. New systems for evaluating feed protein in dairy cattle 387 2.3.1. AAT-PBV feed protein evaluation system 387 2.3.1.1. Effective degradation of feed protein 388 2.3.1.2. Microbial nitrogen contamination of feed residues in bag 388 2.3.1.3. Small particle loss from the bag 388 2.3.1.4. Fractional outflow rate from the rumen 388 2.3.1.5. Efficiency ofmicrobial protein synthesis 389 2.3.1.6. Amino acid content and digestibility of duodenal proteins 391 2.3.1.7. Utilization efficiency of amino acids for lactation in dairy cows 391 3. Objectives of the study 392 4. Material and methods 392 4.1. Experiment 1 393 4.1.1. Animals and management 393 4.1.2. Experimental design and treatments 393 4.1.3. Feeds 393 371 Agric. Sei. Finl. 1 (1992) 4. 1.4. Sampling and analytical methods 393 4.1.5. Statistical analysis 394 4.2. Experiment 2 394 4.2.1. Animals and management 394 4.2.2. Experimental design and treatments 394 4.2.3. Feeds 394 4.2.4. Sampling and analytical methods 395 4.2.5. Statistical analysis 395 4.3. Experiment 3 395 4.3.1. Experimental animals and management 395 4.3.2. Experimental design and treatments 395 4.3.3. Feeds 395 4.3.4. Sampling and analytical methods 396 4.3.5. Statistical analysis 396 4.4. Experiment 4 397 4.4.1. Animals and management 397 4.4.2. Experimental design and treatments 397 4.4.3. Feeds 397 4.4.4. Sampling and analytical methods 397 4.4.5. Statistical analysis 397 4.5. Experiment 5 398 4.5.1. Animals and management 398 4.5.2. Experimental design and treatments 398 4.5.3. Feeds 398 4.5.4. Sampling and analytical methods 398 4.5.5. Statistical analysis 399 4.6. Estimating the effect of RSM on the milk and protein yield 399 4.7. Comparison of AAT-PBV and DCP systems 399 5. Results and discussion 400 5.1. Composition and nutritional value of feeds 400 5.1.1. Chemical composition of feeds 400 5.1.2. Intestinal degradation of RSM and SBM as measured by mobile nylon bag technique... 402 5.1.3. Amino acid content ofRSM and SBM 402 5.1.4. Glucosinolate content of RSM 403 5.2. Effect ofprotein supplement on the digestibility of the diet 405 5.3. Effect ofproportion of RSM on milk production 405 5.3.1. Feed intake 405 5.3.2. Milk yield 405 5.4. Effect ofprotein protection on milk yield and protein content 410 5.5. Goitrincontent ofmilk 412 5.6. DCP and AAT protein evaluation systems in milk production 413 5.6.1. Comparison of DCP and AAT systems 413 5.6.2. Effect of corrections of AAT-PBV values of RSM and other feeds 414 5.6.3. Effect of corrections of AAT values on the utilization of feed protein 416 372 Agric. Sei. Finl. 1 (1992) 6. Conclusions 419 6.1. Glucosinolate content 419 6.2. Effect ofrapeseed meal on milk yield 419 6.3. Protein protection 419 6.4. Goitrincontent of milk 419 6.5. DCP and AAT 419 7. References 420 Selostus 430 Appendices 431 373 Agric. Sei. Finl. 1 (1992) Agric. Sd. Fint. 1 (1992) 374 Rapeseed meal as a supplementary protein for dairy cows on grass silage-based diet, with the emphasis on the Nordic AAT-PBV feed protein evaluation system Mikko Tuori Tuori, M. 1992. Rapeseed meal as a supplementary protein for dairy cows on grass silage-based diet, with the emphasis on the Nordic AAT-PBV feed protein evaluation system. Agric. Sei. Finl. 1: 367-439. (Univ. Helsinki, Dept. Anim. Sci., SF - 00710 Helsinki, Finland.) The effect ofrapeseed meal (RSM) supplementation on the performance of dairy cows on direct cut grass silage based diets was studied in five feeding trials. The proportion ofRSM varied from 0% to 33% in the concentrate mixture (the grain was an oat-barley mixture of 1:1). In one experiment the treatments were RSM and soybean meal (SBM), while in another experiment forage was either grass silage or bam dried hay cut at the same maturity. In addition, this was compared to the data of other trials in Finland dur- ing the last ten years, in which RSM supplementation had been used. Using this data the response in terms ofmilk yield to RSM supplementation was estimated. The util- ization of protein in milk production was estimated by the Nordic AAT-PBV protein evaluation system. During the experiments (1983-1990) the varieties of turnip rape were changed from high glucosinolate, containing single-zero, to low glucosinolate containing double- zero varieties, while the glucosinolate content was reduced from 40-50 pinoles to 14 pmoles per g of defatted meal. Heat-moisture treatment (™Öpex) further reduced the glucosinolate contentby half. By replacing grain with RSM in the concentratemixture with ad libitum silage feed- ing, the silage intake increased by 0.43 kg per kg increase in RSM on the basis of dry matter (DM) (non significant). The response in increased milk production was 0.77 kg in milk or 0.70 kg in energy corrected milk (ECM) yield (P<0.02), and in protein yield 27 g/d (P<0.01) per kg increase in RSM DM. Although the linear effect of the RSM level was significant, the effect on the milk yield was reduced when the level of RSM was over 12-16% of concentrate mixture. The protein content of milk increased by 0.07 g/kg per MJ increase in metabolizable energy intake (P<0.02). Heat-moisture treatment of RSM increased milk production significantly in one experiment (21.9 kg vs. 23.9 kg milk or 23.4 vs. 25.2 kg ECM/d), (P<0.03). In two other experiments heat treatment had no noticeable effect on milk yield. In comparing SBM with RSM on the same crude protein basis in the concentrate, no difference in milk yield was observed. The goitrin content ofthe milkwas reduced when the glucosinolate contentRSM, or the level of RSM in the diet, was reduced. With Öpex-treated double zero RSM, the milk contained less than 10 pg/1 (sensitivity ofanalysis 2 pg/1) goitrin. The utilization of AAT in milk production was also estimated using different con- stants in the calculations of AAT-PBV values of the feeds. When the proportion of AAT ofmicrobial origin increased, the coefficient ofvariation of the AAT utilization reduced. This is affected by correcting the microbial-N contamination of the in sacco analysis, lowering the estimate for the rumen outflow rate (k-value) from 0.08 to 0.03, and changing the estimate for the efficiency ofmicrobial protein synthesis (MPS). The best model was obtained using the method ofVoigt and Piatkowski (1991) for calcu- lating MPS. Key words: dairy cows, rapeseed meal, grass silage feeding, protein protection, protein evaluation systems, AAT/PBV 375 Agric. Sei. Fin!. 1 (1992) 1. Introduction In Finland milk production accounts for ca. 30% of the gross income from agriculture. The fact that most of the feeds can be produced on the farm, especially for dairy cattle on a grass silage-based diet, works in favour of dairy farming. Hence the proportion of forage in the total feed units averaged 55.4% in 1991 on milk recorded farms, the propor- tionof silage being 33.1%. Forage was supplement- ed by grain and protein concentrates, the most important of which were soybean meal and rape- seed meal. In 1991 the combined production ofoil- seed meals and oil cakes totalled 190 million kilograms, ofwhich 41% was of rapeseed and 59% of soybean. Rapeseed meal is given mainly to cattle, whereas soybean meal is mainly used in pig and poultry feeds, with only 20-25% used in cattle feed. Replacing a part of the soybean meal in cattle feed with rapeseed meal could increase the do- mestic cultivation of rape, thereby boosting the national self-sufficiency in cattle feeds. Rapeseed cultivation and production of rapeseed meal could increase in the future ifrapeseed oil were used as a diesel fuel. The production of feed protein for ruminants can be increased by using high amounts of nitrogen fer- tilizer in swards.This was demonstrated in the 1970 s by the so-called “Green line”project (Ettala and Lampila 1974, Ettala et al. 1974, Ettala et al. 1978). Grass silage of high crude protein content and high digestibility, supplemented with plain grain concentrate, will satisfy the requirement for Table I, World oilseed production. digestible crude protein even at high production levels. However, supplementing grain withprotein concentrates in a diet offorage with high crude pro- tein content has also increased milk production (Castle and Watson 1976, Gordon and Murray 1979). This indicates that the DCP system has its limitations, and new systems were developed during the 1970 s and 1980s. The new feed protein evaluation systems, for instance the Nordic AAT- PBV system (NKJ 1985), have divided the amino acid nitrogen absorbed in the duodenuminto thatof microbial origin and of feed origin, enabling the protein requirement of the dairy cow to be estima- ted more exactly than by using the DCP system. In the present study various rapeseed meals were used as supplements to the grass silage based-diet of dairy cows, and their effect on milk production and composition, as well as on the utilization of feed protein, was examined with the AAT-PBV system. 2. Review of literature 2.1 Rapeseed meal 2.1.1. Rapeseedproduction 2.1.1.1. Origins and cultivation of turnip rape Among the most commonly cultivated oilseed plants the turnip rape comes thirdafter soybean and linseed (Table 1). Since World War II it has become the most important oilseed plant in the 1962- 1972- 1982- 1986/87 1987/88 1988/89 1989/90 1990/91 1964 1974 1984 Soybeans 30.4 53.7 89.9 98.1 103.7 95.4 106.0 104.3 Cottonseed 21.2 24.9 28.8 27.6 31.6 32.6 31.2 33.3 Rapeseed 4.0 7.3 15.3 19.8 23.5 20.4 21.5 24.0 Sunflowerseed 7.2 10.9 16.8 18.8 21.0 20.6 21.7 22.3 Groundnuts 10.7 11.1 13.2 15,0 15.1 16.2 15.6 15.916.2 15.6 15.9 World» 79.9 121.3 178.2 194.4 209.5 202.7 211.2 216.1 » Soybeans, cottonseed, rapeseed, sunflowerseed, groundnuts, copra, palmkemels, linseed Amounts shown are million metric tons; adapted from Toepfer International (1990) 376 Agric. Sei. Finl. 1 (1992) temperate zone. Traditionally rapeseed was cultiv- ated in India, China and Japan. In Europe rapeseed was cultivated since the 14th century, although rape oil has only been used for cooking since the 1940 (Shahidi 1990a). Today China, Canada, India, France, Germany, Great Britain and Poland are important producers of rapeseed (Tables 2 and 3). According to the FAO statistics (FAO 1991) there was some rapeseed cultivation in forty coun- tries in 1990. The name “rape” in rapeseed was derived from the Latin word for turnip (rapum). Turnip, ruta- baga, cabbage, Brussels sprout, mustard and many other common vegetables are closely related to rape. Rape grows in low temperatures and tolerates humidity, thus thriving in such temperature zones where soybean and sunflower do not survive. Downey and Röbbelen (1989) studied the ori- gins of the oilseed plants of the Brassica family. Certain species of the Brassica family may well be among the earliestknown cultivated plant, as some vegetable forms of this family were commonly used as early as in the Neolithic age, and reference was made to oilseed rape and mustard in Indian Sanskrit texts from the 21st to the 16th century B.C. These plants and their medicinal qualities were also mentioned in Greek, Roman and Chinese texts from the 6th to the 3rd century B.C. In Europe rape cultivation did not begin until the early Middle Ages. The commercial cultivation ofrape began in the Netherlands as early as in the 16th century. Rape oil was traditionally used as an illuminant (lamp oil) and lubricant for steam engines, gaining a notable market share among food oils in the West only after World War 11, thanks to improved vari- eties and more efficient processing techniques (Downey and Röbbelen 1989). The major varieties of oilseed plants in the Bras- sica family are Brassica campestris or turnip rape, and B. napus or rape. The former originated in the high plateaus in Turkey, from where it spread two thousand years ago to cover an area from the islands in the western Atlantic Ocean to China and the East coast of Korea, from northern Norway to the Sahara and North India. Brassica napus is a cross between B. campestris and B. oleracea. The latter originates in the Mediterranean region, and it is generally accepted that B. napus originated in southern Europe (Downey and Röbbelen 1989). There are spring and winter varieties of both B. campestris and B. napus. The winter varieties tend to yield more, but their wintering characteristics are poorer than those of winter grain crops. Table 3. Rapeseed production in Europe. 1979- 1988 1989 1990 1981 France 871 2469 1803 2011 Germany 618 1640 2869 2157 Poland 434 1199 1586 1206 UK 274 1040 976 1231 Denmark 204 504 655 819 Czechoslovakia 165 380 387 380 Sweden 313 305 422 401 Finland 68 121 125 117 Europe, total 3203 8076 8261 8754 Amounts shown are 1000 metric tons; adapted from FAO Yearbook 1990 Table 2. Rapeseed production of some important producers. 1962- 1972- 1982- 1986/87 1987/88 1988/89 1989/90 1990/91 1964 1974 1984 Canada 0.21 1.22 2.78 3.79 3.85 4.31 3.10 3.26 PR China 1.04 1.20 4.72 5.88 6.61 5.04 5.44 6.55 India 1.19 1.92 2.65 2.61 3.37 4.02 3.80 4.00 Poland 0.28 0.49 0.63 1.30 1.19 1.18 1.58 1.20 EC-12 0.37 0.50 2.91 3.69 5.94 5.18 4.99 5.87 Amounts shown are million metric tons; adapted from Toepfer International (1990) 377 Agric. Sei. Finl. 1 (1992) Breeding of the presently cultivated turnip rape and rape varieties has been pursued especially in Canada. B. campestris was imported to Canada in 1936, and a few years later B. napus was brought from Argentina. Commercial rapeseed production began in 1942 in Canada, the objective being to produce lubricant for the Allied war machinery. The erucic acid content of rapeseed was high, as it is still today in the Asian countries, where rape oil containing between 22 and 60 per cent of erucic acid is still being produced (HEAR, or high erucic acid rapeseed). The first rapeseed varieties with reduced erucic acid content, yielding rape oil con- taining less than 5% erucic acid (zero variety, “single low” or “single zero” varieties), were bred in Canada in 1968. These varieties are also known as LEAR (low erucic acid rapeseed). The first varieties low in both erucic acid and glucosinolate were licensed in 1974 (the “double low” or “double zero” varieties), i.a. the “Tower” variety. The low glucosinolate character came from the Polish variety “Bronowski”. A double zero yellow seeded variety of turnip rape, “Candle”, was developed in 1976. Having a low fibre content it is called a triple low (or triple zero) variety (Shahidi 1990b). The brand name “Canola” was assumed in Ca- nada in 1979 for all double low varieties. Canola was defined as a rapeseed variety yielding oil con- taining less than 2 per cent erucic acid, from which a defatted rapeseed meal containing less than 30 pmoles/g glucosinolate can be produced. Glucosin- olates include, among others, four aliphatic gluco- sinolates: gluconapin, progoitrin, glucobrassicana- pin and napoleiferin (Shahidi 1990b). 2.1.1.2. Cultivationof rapeseed in Finland In practice, the cultivation of oilseed plants began in Finland in 1942 with the sowing of Argentinean linseed flax, although fiber flax and hemp had been traditionally grown (Valle 1953). To a small extent hemp had been grown for seed and oil production. Linseed flax was being studied from 1924 at the Agricultural Research Center and from 1939 at the Plant Breeding Institute of Hankkija Wholesale Cooperative, resulting in two Finnish linseed flax varieties, “Vaanila” and “Tikkurila” (Maa-ja metsätalousministeriö 1975). From 1947 to 1951 the annual linseed flax cultivation covered an area of 2,000 to 4,000 hectares, the seed harvest averaging only 652 kg per hectare (Valle 1953). The cultivation of winter turnip rape began towards the end of the 1940 s (Table 4), and by 1959 the area under winter turnip rape cultivation had increased to 18,600 hectares. By 1976 the cul- tivation of winter varieties had ceased in favour of spring varieties of rape and turnip rape (Pahkala and Sovero 1988). Their combined cultivation area was at its largest in 1988, totalling 82,900 hec- tares (80,300 spring turnip rape and 2,600 spring rape. The respective figures for 1991 were 59,300 and 1,700 hectares (National Board of Agriculture 1991a). The yield of spring turnip rape has aver- aged 1,500 kg/ha (Table 4), being 1,780 kg/ha in Table 4. Rapeseed production in Finland. 1948- 1956- 1961- 1966- 1971- 1976- 1981- 1986- 1955 1960 1965 1970 1975 1980 1985 1990 Area, 10.8 9.9 6.7 5.1 10.1 31.7 61.1 76.6 1000 ha Yield, 1210 1120 1140 1440 1430 1530 1470 1530 kg/ha Production, 12.5 11.3 7.5 7.6 14.6 47.8 88.3 115.4 IOOOt/year Adapted from National Bord ofAgriculture, Helsinki, Finland. 378 Agric. Sei. Fint. 1 (1992) 1990, when the corresponding figure for spring rape was 2,090 kg/ha (National Board of Agricul- ture 1991a). The first single zero varieties were brought to Finland in 1976, followed later by double zero varieties. In 1982 fifteen per cent of the spring tur- nip rape cultivated was of double zero varieties, whereas the corresponding share ofspring rape was as high as nearly 70 per cent. At that time turnip rape covered 89.5% of the total rapeseed cultiva- tion area. By 1988 the double zero varieties cov- ered some 70 per cent of all spring turnip rape (varieties “Kova” and “Valtti”), and almost all spring rape (“Topas”) (National Board of Agricul- ture 1991b). 2.1.2. Composition ofrapeseed 2.1.2.1. Whole seed The properties of the whole seed of the turnip rape have been reviewed (e.g. Röbbelen and Thies 1980, Bell 1984, Bertram et al. 1986, Henkel and Mosenthin 1989). The seeds ofrape and turnip rape are normally black, reddish brown, or occa- sionally yellow in colour, and round, with a diame- ter ranging from 1.5 to 3.2 mm. Winter rape has the largest seeds. The ripe rapeseed comprises an embryo (84 to 86 per cent of the dry weight) and a hull with a single cell layer of adhering aleuron as the only remains of the endosperm. The embryo consists mainly oftwo large cotyledons with an oil content of about 50 per cent. The cotyledons con- tain protein granules similar to those in the aleuron layer (Bengtsson et al. 1972). The proportion of the hull is between 13.0% and 18.7% of the dry weight of the rapeseed (Appelq- vist 1972,Bertram et al. 1986), or 27% to 30% of rapeseed meal after the extraction of oil. The black or brown seeds have thicker hulls than the yellow seeds. The hull still contains 9 to 13 per cent hexane-soluble fat and 15 to 18 per cent protein in addition to fibre, its main ingredient. The dark colour of the hull is mainly due to condensed polyphenols. Of the undesirable constituents of the seed, glucosinolate and sinapine are mainly found in the core, whereas tannins are in the hull (Ber- tram et al. 1986, Henkel and Mosenthin 1989). The oil content of the rapeseed varies from one species or variety to another. In winter rape it typic- ally ranges from 42 to 50 per cent and in spring rape from 37 to 47 per cent of the dry matter. The respective figures for turnip rape are 40 to 48 per cent and 36 to 46 per cent (Appelqvist 1972). After the extraction of fat the rapeseed meal contains roughly 40 per cent crude protein, depending on the species and variety. For instance, in 1969 in Ca- nada the protein content of the defatted meal (DM) varied from 33.0 to 47.9 per cent (Appelqvist 1972). 2.1.2.2. Rapeseed meal Industrially the oil can be removed from the rape- seed either by pressing or by pressing and extract- ing with hexane. Before pressing, the crushed rape- seed is, when necessary, first dried, then purified, flaked and cooked in a stacked cooker in a temper- ature between 75° and 85°C, usually for 20 to 40 minutes (Unger 1990). Continuous action expel- lers or screw presses are normally used for pressing most of the oil from the seed. The fat content of the remaining rapeseed cake varies between 15 and 18 per cent. The rapeseed cake is then subjected to an extraction process using hexane in a temperature of about 55°C. After extraction the rapeseed meal (RSM), which contains from 1.5% to 5% fat, is toasted at 100°C, dried and allowed to cool before storage (Carr 1989). Rapeseed meal differs from rapeseed in that most ofthe oil has been removed. The removal ofoil rad- ically alters the composition of rapeseed, as the remaining fractional ingredients are then concentrated. In terms of a normalproximate anal- ysis of feeds, roughly 12% of rapeseed meal con- sists of crude fibre, mainly in the form of the hulls (Bell and Jeffers 1976). The crude fat content of rapeseed meal depends on the quantity ofoil left in the meal. The crude fat content, on average 4.1% of DM, includes the 379 Agric. Sei. Finl. 1 (1992) remaining oil, the impurities removed from the oil during purifying, and other compounds, the largest group of which being phosphatides (gums) (Bell and Jeffers 1976). The crude protein content ofrapeseed meal (N x 6.25) is roughly 40% of the defatted meal. The crude protein content of the hull ranges from 12% to 16% depending on the proportion of remaining embryonic matter, whereas that of the embryonic matter is about 52% (Bell 1984). The average composition of Finnish rapeseed meal and cake is shown in Table 5. 2.1.2.3. Rapeseed protein The main proteins in the seed are water-soluble albumin and salt-soluble globulin. Albumins form a major part of the metabolically active, vitally important protein in the cell. Albumins can also function as storage proteins. Regarding their amino acid composition, albumins are better dietary than storage proteins, especially as far as dietary sulphur-containing amino acids are concerned (Norton 1989). The storage protein is located in Table 5. Composition of the Finnish rapeseed meal and cakes 1989". Rapeseed meal Rapeseed Crude fat Crude fat Crude fat cake < 5 % 5-7 % >7 % No ofsamples 13 23 9 31 In dry matter (%) Ash x 8.3 8.1 7.8 7.1 CV 5.0 23 2.3 5.2 Crude protein x 37.9 37.9 36.5 33.8 CV 5.2 4.4 2.9 5.8 Crude fat x 4.1 5.3 7.7 17.2 CV20.8 17.4 9.6 14.4 Crude fibre x 13.8 13.3 13.3 11.1 CV 6.2 7.9 7.4 8.0 "Mean and coefficient ofvariation (%) Adapted from Valtion maatalouskemian laitos (1989) 380 the so-called protein bodies or aleuron grains (Appelqvist 1972). The majority of rapeseed proteins are storage proteins without enzymatic activity. Between 18% and 28% ofthe crude protein in the seed consists of sedimentation coefficient 12 S, a globulin also known as crucipherin (Bhatty et al. 1968, Schwenke et al. 1973b). It is a neutral protein, sol- uble in sodium chloride, with a molecular weight of 300,000 to 360,000(Schwenke 1990). The 2 S pro- teins are water soluble, alkaline albumins known as napins. Their molecular weight ranges from 12,000 to 14,000. They account for about 40% of the pro- tein content of the seed (Schwenke et al. 1973b). Albumins are relatively rich in sulphurous amino acids (cystine 6.9%) and lysine (9.0%) (Schwenke et al. 1973a). A newly identified group of proteins, oleosines, make up ca. 20% ofthe pro- tein content of the rapeseed. They are believed to be efficient emulsifying agents in the dry seeds (Murphy et al. 1991). Rapeseed meal includes more amino acids (AA) containing sulphur and slightly less lysine than soy- Table 6. Essential amino acid composition ofrapeseed meal and rapeseed hulls compared with soybean meal (Bell 1984). In dry matter (%) In crude protein (N x 6.25) Rape- Soy- Rape- Soy- seed bean seed bean hulls 3' meal h) meal hulls meal meal Arginine 0.28 2.50 3.26 2.49 6.11 6.44 Histidine 0.11 1.15 1.21 0.98 2.81 2.40 Isoleucine 0.45 1.63 2.37 4.00 3.98 4.69 Leucine 0.62 2.85 3.79 5.51 6.97 7.49 Lysine 0.66 2.45 3.14 5.87 5.98 6.22 Methionine trace 0.73 0.71 trace 1.78 1.40 Cystine 0.24 0.50 0.33 2.13 1.23 0.65 Tryptophfan 0.05 0.48 0.61 0.40 1.16 1.20 Phenylalanine 0.43 1.64 2.43 3.82 4.01 4.80 Valine 0.67 2.09 2.53 5.96 5.11 5.00 Threonine 0.69 1.84 1.92 6.13 4.50 3.80 Protein 11.25 40.95 50.57 100 100 100 (Nx6.25, %) a) Finlayson 1974, b) Clandinin et al. (1981), Sarwar et al. 1981 Agric. Sei. Finl. 1 (1992) bean meal (Table 6). The amino acids of Finnish rapeseed meal have been found to contain 5.4% to 5.9% lysine, 4.3% to 4.4% threonine, 2.3% to 2.5% methionine and 2.3% to 2.5% cystine (percentage of crude protein) (Näsi 1991, Näsi and Siljander- Rasi 1991). 2.1.2.4. Rapeseed fat The rapeseed fats are predominantly, more than 90%, triglycerides. In addition, there are less than 4% phospholipids, the most important one being lecithin (Table 7). Table 7. Fat composition of low erucic rapeseed (Zadernowski & Sosulski 1978). Triglycerides Diglycerides Free fatty acids Sterol esters 92.1 % 1.2 0.5 1.1 0.6 3.6 0.9 Sterols Phospholipids Glycolipids The fatty acid composition of rapeseed fats depends primarily on the variety (Table 8). For- merly varieties high on erucic acid, with erucic acid contents ranging from 45% to 50% of all fatty acids, were cultivated. In the 1960 s rapeseed oil caused an unusually high incidence of cardiac defects in laboratory animals, presumably due to erucic acid. As this phenomenon was not observed in humans, the nutritional risk may have been exag- gerated (Ackman 1990). The EEC Commission set the upper limit for erucic acid content at 15% in 1976, the present limit of 5% having been intro- duced in 1979. Low erucic acid rapeseed, i.e. single zero varieties, have an erucic acid content of less than 2% of all fatty acids (Ward et al. 1985). In Finland erucic acid content has been below 1% for several years. Rapeseed oil contains less linoleic acid, but more a-linolenic acid than soybean oil. Linoleic acid, also known as Vitamin F, tends to reduce the blood cholesterol level. The proportion of linoleic acid is higher in low erucic acid rapeseed varieties. A high linolenic acid content is an undesirable trait, as this Table 8. Comparison ofmajor fatty acids in some edible vegetable oils of commerce (w/w% fatty acids). Fatty acids HEAR* LEARb Candlec Tower1 Soybean Com Safflower Sunflower Peanut Olive Linseed 14:0 - 0.04 0.05 0.05 0.1 - 0.1 - 0.1 - - 16:0 4 3.48 3.55 3.88 10.8 11.4 6.5 6.2 10.0 11.0 5.5 18:0 1 1.50 1.38 1.56 4.0 1.9 2.3 4.7 2.3 2.2 4.3 20:0 1 0.42 0.43 0.50 ... . ... 22:0 <1 0.27 0.20 0.28 ... . . . . Total saturated 6 5.71 5.61 6.46 15.1 13.3 10.4 10.8 17.8 13.5 9.8 16:1 - 0.22 0.28 0.29 0.2 - 0.4 - 0.1 0.8 18:1 15 61.65 55.58 64.02 23.8 25.3 12.2 20.4 47.1 75.8 21.1 20:1 10 1.38 1.78 1.24 0.2 - - - 1.4 0.3 22:1 45 0.44 1.63 0.08 ....... Total 70 63.69 59.27 65.80 24.3 25.3 12.6 20.4 48.6 77.1 21.1 monounsaturated 18:2n-6 14 19.69 21.87 18.79 53.3 60.7 77.4 68.8 33.6 8.3 13.3 18:3n-3 9 10.65 12.99 8.59 7.1 0.7 0.4 - - 0.6 55.7 Total 13 30.34 34.86 27.61 60.6 61.4 77.8 68.8 33.6 8.8 69.0 polyunsaturated Adapted from Ackman (1990) " HEAR =high erucic acid rapeseed; b LEAR = low erucic acid rapeseed, B. campestris var. Torch ■ B. campestris var. Candle; d B. napus var. Tower 381 Agric. Sei. Finl. 1 (1992) acid oxidates easily, giving the oil an unpleasant odour. The oleic acid content ofrapeseed oil is not- ably higher than that ofsoybean, the former resemb- ling olive oil in this respect. Oleic acid has assumed increasing significance in recent studies on nutri- tion (Ackman 1990). 2.1.2.5. Carbohydrates of rapeseed meal RSM comprises about 50% crude fibre and frac- tions ofnitrogen-free extract. Hulled RSM contains roughly one third carbohydrate, one half protein (% N x 6.25), and various other minor constituents. There are relatively small quantities of monosac- charides and disaccharides, starch being almost non-existent. The composition of carbohydrates in hulledrapeseed meal is 14.5% pectin, 7% cellulose residues, 4.5% amyloid, 2% arabinan, and 1%arab- inogalactan (Bell 1984). Hulled rapeseed meal also contains protein (52%), lignin (2.6%), sinapine (2.4%), ash (3.7%), phytates (2%) and glucosino- lates (1%). Dry RSM matter contains about 10% soluble sugars. Theander and Åman (1976) reported 6.8% to 7.5% sucrose, 2.4% stachyose, 0.3% raffinose and 0.2 - 0.5% fructose in a moisture-free defatted turnip rape meal, a- galactocides of sucrose, containing raffinose and stachyose, cannot be absorbed by the human digest- ive system, as it lacks the enzyme a-galactosidase required for their hydrolysis. This has been observed to cause flatulence in humans and animals (Naczk and Shahidi 1990). Hull mass amounts to about 16% ofthe weight of the full seed, and 30% of the oil-free RSM. The hulls contain 34% N-free extract, 44% crude fibre, 14.5% pentosans, 32% cellulose, 3.8% sugars, 12% to 24% lignin, 6% to 12% polyphenols and 1.5% tannins (Bell 1984). Lignin content is lower in triple zero rapeseed varieties. Theander et al. (1977) reported 7.9% lignin in the hulls of yellow- hulled turnip rapeseed. The digestibility of hull energy has been found to be very low in pigs, rang- ing from 0% (brown hulls) and 30% (yellow hulls) (Bell and Shires 1982). 2.1.2.6. Minerals The mineral content of RSM was reviewedby Bell (1984). RSM dry matter contained 7.4% to 7.6% ash, 0.68% calcium and 1.2%to 2% phosphorus, of which nearly two thirds was phytin-bound phosphorus. In addition, the RSM dry matter con- tained 1.3%potassium, 0.64% magnesium, 0.8% to 1.7% sulphur and only 0.026% sodium (Summers et al. 1983). The phytin content of the rapeseed calls for careful attention when formulating a diet, as is the case with cereal grain and oilseed meal in general. Availability of RSM phosphorus to young chickens does not exceed 30% (Summers et al. 1983). 2.1.2.7. Glucosinolates 2.1.2.7.1. Structure and classification of rapeseed glucosinolates The presence of glucosinolates and phenolic cho- line esters in rapeseed limits its use as a feed for many animal species. Glucosinolates are an integ- ral group of sulphur-containing anions present in nature, the structure of which was studied by Ettlinger and Lunden (1954). More than 100 dif- ferent glucosinolates are known today (Sorensen 1990). The basic structure of glucosinolates is as fol- lows: /S-C# Hno5 R-C 3 Glucosinolates can be classified as alkenyleglu- cosinolates and indolylglucosinolates on the basis of their side chain. In the former the R is straight- chained, in the latter it is heterocyclic. Glucosino- lates are hydrophilic and strongly acidic com- pounds, isolated and handled as salts. The glucosin- olates of the species Brassica are derived biosyn- thetically from three amino acids: methionine, phenylalanine and tryptophan. About 30 different glucosinolates have been found in the seed ofBras- 382 Agric. Sei. Fint. 1(1992) sica napus (SORENSEN 1990). Glucosinolates num- bered 1 to 15 in Table 9 can be derived from methi- onine, 14 to 22 from phenylalanine, and 23 to 27 from tryptophan (Sorensen 1990). The following glucosinolates present in B. napus Table 9. Selected glucosinolates important for the quality of rapeseed and other cruciferous corps (Sorensen 1990). No. Trivial names Semisystematic names Glucosinolates derived from methionine: 1 Sinigrin Allylglucosinolate 2 CE Gluconapin But-3-enylglucosinolate 3CE Glucobrassicanapin Pent-4-enylglucosinolate 4 CE Progoitrin (2R)-2-Hydroxybut-3- enylglucosinolate 5 Epiprogoitrin (2S)-2-Hydroxybut-3- enylglucosinolate 6 CE Napoleiferin (2R)-2-Hydroxypent-4- enylglucosinolate 7 Glucoibervirin 3-Methylthiopropylglucosinolate 8 Glucoerucin 4-Methylthiobutylglucosinolate 9 Glucoberteroin 5-Methylthiopentylglucosinolate 10 Glucoiberin 3-Methylsulphinylpropylglucosinolate 11 Glucoraphanin 4-Methylsulphinylbutylglucosinolate 12 Glucoalyssin 5-Methylsulphinylpentylglucosinolate 13 Glucoraphenin 4-Methylsulphinylbut-3- enylglucosinolate 14 Glucocheirolin 3-Methylsulphonylpropyl- glucosinolate 15 Olucoerysolin 4-Methylsulphonylbutylglucosinolate Glucosinolates derived from phenylalanine: 16 Glucotropaeolin Benzylglucosinolate 17 Gluconasturtiin Phenethylglucosinolate 18 Glucobarbarin 2-Hydroxy-2-phenylethyl- glucosinolate 19 Glucolepigramin m-Hydroxybenzylglucosinolate 20 Sinalbin p-Hydroxybenzylglucosinolate 21 Glucolimnanthin m-Methoxybenzylglucosinolate 22 Glucoaubrietin p-Methoxybenzylglucosinolate Indol-3-ylmethylglucosinolatesbiosyntetically derived from typto- phan: 23 E Glucobrassicin Indol-3-ylmethylglucosinolate 24 Neoglucobrassicin N-Methoxyindol-3-ylmethyl- glucosinolate 25 Sulphoglucobrassicin N-Sulphoindol-3-ylmethyl- glucosinolate 26 E 4-Hydroxygluco- 4-Hydroxyindol-3-ylmethyl- brassicin glucosinolate 27 4-Melhoxygluco- 4-Methoxyindol-3-ylmethyl- brassicin glucosinolate C: Glucosinolates standard by Canola Council E: Glucosinolates standard by EEC are quantitatively the most prominent (Sang and Salisbury 1988): I Gluconapin II Progoitrin 111 Glucobrassicanapin IV Napoleiferin V Glucobrassicin VI 4-hydroxyglucobrassicin VII Sinigrin VIII Gluconasturtiin In addition to the four alkenylglucosinolates (gluconapin, progoitrin, glucobrassicanapin and napoleiferin) defined in the Canola standard, the quantities of the indolylglucosinolates glucobrassi- cin and 4-hydroxyglucobrassicin are included in the total glucosinolate content as defined in the EEC standard (EEC 210/55, ref. Schnug and Haneklaus 1988). After 1991 the total glucosino- late content of the rapeseed produced in the EEC countries may not exceed 20 pmoles/g in air-dry seed. In the double zero varieties the quantity of alkenylglucosinolate is reduced, resulting in a relat- ive increase in indolylglucosinolate content (mainly 4-hydroxyglucobrassicin) (Table 10). Glucosinolates are not toxic as such, but become so during enzymatic degradation. The release of myrosinase enzymes (thioglucoside glucohydro- lase, EC 3.2.1), present in plant cells and synthe- Table 10. Content of glucosinolates in some rapeseed culti- vars (Sang & Salisbury 1988). Glucosinolate Brassica campestris Brassica napus Candle Torpe Oro Regent (Canada) (Sweden) (Canada) (Canada) Glucosinolates (percentage of total amount) Gluconapin 26 31 12 18 Progoitrin 34 25 68 54 Glucobrassicanapin 21 26 6 3 Napoleiferin 7 6 5 2 Glucobrassicin tr tr tr 4tr tr tr 4 4 - hydroxyglucobrassicin? 7 5 17 Sinigrin 0 0 2 2 Gluconasturtiin 4 6 2 tr4 6 2 tr Total glucosinolates 52 42 111 41 (pmoles/g defatted meal) 383 Agric. Sei. Fint. 1 (1992) sized by microbes in the digestive tract, e.g. as rapeseed tissue degrades, causes disintegration of glucosinolates. The products of this disintegration are glucose and aglucon (Röbbelen and Thies 1980): /SH R-C. +Glucose 3 In a neutral environment aglucon releases sul- phates, and isothiocyanate (trivial name mustard oil) is generated: R - N = C = S + HSO.4 In a mildly acidic environment, or when coming into contact with ferrous ion, aglucons generate nitriles and elemental sulphur: R-C-N+S Isothiocyanates carrying a hydroxyl group in the (3-position form spontaneously cyclic compounds, oxazolidine-2-thiones. R-CH-CH -N=C=S—� CH, —NH i2i2 iOH R- CH C=S In a neutral or mildly alkaline environment iso- thiocyanates of indolylglucosinolate origin form thiocyanate ions (Table 11). Table 11. Principal degradation products of the rapeseed glucosinolates (Henkel & Mosenthin 1989). Glucosinolate Primary degradation Secundary degradation products products Progoitrin Goitrin pH-7: nitriles (traces) pH<7: nitriles and isothiocyanates Gluconapin lsothiocyanate Glucobras- lsothiocyanate sicanapin Napoleiferin 5-allyl-2-thiooxazolidon Glucobrassicin ph>7: thiocyanate or pH<7: nitriles 3-hydroxymethylindole 384 2. 1.2.7.2. Analysis ofglucosinolates in RSM Enzymatic degradation of glucosinolates by myro- sinases produces substances, the identification and measuring of which can be utilized in assessing glucosinolate content. These substances include isothiocyanates, goitrin (5-vinyl-oxazolidine-2- thione), nitriles and thiocyanate. The total gluco- sinolate content can be determined from the released glucose (McGregor et al. 1983). Far more sophisticated methods for analysing glucosinolates have been developed during the last two decades, enabling their classification according to the following criteria. Single glucosinolates can be identified by gas chromatography using tri-methylsilylised desul- phoglucosinolates (TMS derivative) (Underhill and Kirkland 1971, Thies 1976). This is also the reference method of the EEC. Reversed-phase li- quid high performance chromatography has made it possible to identify all known glucosinolates either as non-degraded glucosinolates (Helboe et al. 1980) or as desulphoglucosinolates (Sang and Truscott 1984). The coloured complex formed by a glucosinolate and palladium ion can be measured by means of spectrophotometry - a fast and sensitive method for determining total glucosinolate content (Thies 1982,Möller et al. 1985). The X-ray fluorescence method was developed recently for determining total glucosinolate content by means of measuring the sulphur content of an organic specimen. The method is based on the close relationship of the glu- cosinolate and sulphur contents of the rapeseed (Schnug 1987). Near infrared reflectance analysis is not precise enough, not even for screening pur- poses (McGregor 1990). 2.1.2.7.3. Physiological effects ofglucosinolates The degradation products of glucosinolates have various physiological effects, mainly affecting the thyroid. Degradation of glucosinolates is due to the release of myrosinase from degrading rapeseed tissue. The enzyme is also synthesized by intestinal Agric. Sei. Fint. 1 (1992) bacteria in the gastrointestinal tract. Hence, inacti- vating the enzyme, e.g. by means of heating rape- seed products, will not eliminate the harmful effect ofglucosinolates (Marangos and Hill 1974). Degrading glucosinolates produce isothiocya- nate, which decreases thyroxine synthesis by means ofreducing the amount of iodine used in the synthesis. This deficiency can be eliminated by adding iodine to the feed (Anke 1980). Goitrin causes a more significant functional disorder by preventing the oxidation of iodide into chemically active iodine and its binding with aromatic tyrosine (Gmelin 1969, Bergner and Schmidt 1972, Fen- wick and Heaney 1983). The increased supply of iodine cannot eliminate the deficiency. Instead, the thyroid gland begins to grow, thereby increasing the production and availability of tyrosine. The goitrogenic effect can be counteracted by adding trace elements to the diet (Menzel 1983). Glucosinolates are very prone to affect the pay- ability of feeds in monogastrics. Henkel and Kall- weit (1989, ref. Henkel and Mosenthin 1989), for example, gave chicks a feed mixture containing 20% single zero rapeseed meal (B. campestris). After a few days the birds refused to feed on it. The glucosinolate content of new double zero varieties is so low that such rapeseed meal may be given as the sole protein supplement feed (20% to 25% of the diet) to chicks and pigs (Classen et al. 1991, Campbell and Slominski 1991). In dairy cows glucosinolate-content affects the palatability of rapeseed meal less, and the meal of double zero varieties does not affect feed intake, not even in rel- atively high concentrations (Emanuelson 1989). In monogastrics glucosinolates induce enlarge- ment or cirrhosis of the liver, retarded fertility, and thyroidal enlargement, as well as frailty of legs in chicks (Hill 1979). These symptoms have been significantly reduced by using double zero rape- seed meal (Campbell and Slominski 1991), but its detrimental effect on the fertility of sows remains to some extent, depending on glucosinolate content (Etienne et al. 1991). The double zero rapeseed products have also had a negative effect on bovine fertility (Emanuelson 1989). 2.1.2.7.4. Glucosinolates in milk When feeds containing glucosinolates are given to cows, hydrolysis products are also secreted in the milk. Virtanen et al. (1959) and Virtanen (1963) reported that 0.05% of the goitrin (5-vinyl-oxazoli- done-2-thione) present in the feed was secreted into the milk. Consequently Bachmann et al. (1985) found that 0.1% ofthe goitrin in B. napus rapeseed meal was secreted into the milk. In Finland, Arstila et al. (1969) reported goitrin contents ranging from 35 to 100 pg per litre of milk. They assumed that the goitrin was derived from weeds of the Cruciferae family. The follow- ing year milk samples from the same area were found to be free of goitrin, presumably due to chemical weedkillers used on the farms. In a Cana- dian study of six western and three eastern dairies the goitrin content of milk did not exceed 2 pg per litre (Benns et al. 1979). In Switzerland, Bach- mann et al. (1985) measured goitrin contents rang- ing from 37 to 707 pg per litre, when 0.1 to 1.0kg as B. campestris rapeseed meal was given daily. The extracted rapeseed meal contained 46 pM goi- trinper gram, with a total glucosinolate-content of 77 pg per gram. In Finland the goitrin content of dairy milk was measured in 1983 from 224 samples from various regions in the country. The goitrin content exceeded the minimum discernible level of 2 pg per litre in only 19 samples, averaging 6.4 pg goitrin per litre in these samples (range 2to 31 pg/1) (Kauramaa 1983). The role of milk in endemic goitre has been the subject of much discussion. Peltola (1960) gave rats milk collected “from a moderately severe goi- tre endemic area” (Orimattila, Finland) and milk from “a non-goitrous district” (Porvoo, Finland). The milk from the “goitrous area” induced enlarge- ment of the thyroid gland, which could not be pre- vented by supplemental iodine. This result could not, however, be obtained by Virtanen (1963). Krusius and Peltola (1966) found that 0.1 pg/d goitrin induced nearly significant, and 0.5 pg/d sig- nificant, thyroid growth in rats. That means a ratio of 0.5 to 2.3 pg/d/kg metabolic live weight and 12 385 Agric. Sei. Finl. 1 (1992) to 59 pg/d per 75 kg. Ivarsson and Nilsson (1973) gave rats milk from cows that had been feeding on concentrates containing none, 4.2% or 8.1% RSM (69 pM isothiocyanate and 81 pM goitrin per gof DM). In rats thyromegaly could be prevented by increasing the intake of iodine, but the combined effect of goitrin and other goitrogenic substances could not be excluded. There is relatively little information about the quantity at which goitrin becomes harmful to man. Arstila et al. (1969) stated that 200 pg/d will suffice to inhibit the nor- mal accumulation of radioiodine in the thyroid gland, and this dose may also induce enlargement of the thyroid gland. 2.1.2.8. Sinapine Sinapine is a choline ester of sinapine acid (3,5- dimethoxy-4-hydroxycinnamic acid). Sinapine is a cation, and in the rapeseed it is combined with the glucosinolate sinalbin (p-hydroxybenzylglucosino- late), thus substituting potassium. The sinapine content of rapeseed varies between 0.78 and 1.33 per cent (Appelqvist 1972), being lower in B. cam- pestris than in B. napus (Mueller et al. 1978). In an alkaline environment sinapine is hydrolysed to sinapic acid and choline. Contrary to the free cho- line present in feeds, or choline chloride or hydro- lyzed sinapine, which are absorbed in the small intestine, bound choline is not absorbed in that manner (Goh et al. 1979). In the large intestine the intestinal bacteria converted sinapine to trimethyl- amine (Hobson-Frohock et al. 1973, March and MacMillan 1978), which, when absorbed, is con- verted to trimethyloxide by the liver and kidneys, and secreted. Certain species ofpoultry (those lay- ing brown eggs) have a genetic defect, resulting in an insufficient level of the TMA-oxydase activity for all the trimethylamine to be degraded, the remaining trimethylamine being deposited in the egg, affecting its flavour (Pearson et al. 1979). TMA-oxydase synthesis may also be adversely affected by free goitrin and tannin (Fenwick et al. 1981, Goh et al. 1985). Similarly, a fishy smell has been observed in milk (Andersen 1985). 2.1.2.9. Phytic acid Thompson (1990) has produced a comprehensive review of phytic acid. It is a myo-inositol (1,2,3,4,5,6-hexakis-dihydrogen phosphate) by structure. Phytic acid acts as the primary reservoir ofphosphorus, binding between 60 and 90 per cent of the phosphorus present in the rapeseed. Phytic acid also binds cations, and protects the plant from oxidative defects and molding during storage by binding zinc. Phytic acid is normally present in rapeseed as salts of Ca, Mg and K. The phytic acid content ofextracted B. campestris or B. napus rape- seed meal varies between 3.7% and 4.8%, averag- ing 4.4% of the defatted meal (Anjou et al. 1977). Phytic acidhas a detrimental effect on the utiliza- tion of zinc, although the high fibre content is also inversely related to the utilization of minerals (Nwokolo and Bragg 1977). On the other hand, Summers and Leeson (1985) found no difference between soybean or rapeseed meal in the utilization of minerals by chickens on normal diets. Phytic acids also bind proteins, but no effect on in vivo digestibility has been demonstrated. As a nutrient phytic acid is believed to reduce the incidence of cancer of the colon in man. 2.1.2.10. Tannins Kozlowska et al. (1990) studied the tannins present in rapeseed. Tannins are complex phenolic compounds with a molecular weight ranging from 500 to 3000. The tannin content of extracted rape- seed meal is 2 to 3 per cent (Fenwick and Hoggan 1976). Containing, for instance, tannic acid, tan- nins are bitter in taste and therefore obviously have, together with glucosinolates, a detrimental effect on palatability. The high tannin content of certain varieties of sorghum reduces the utilization of pro- tein in poultry, but a similar combined effect of tannin and protein has not been observed in the rapeseed (Kozlowska et al. 1990). 386 Agric. Sei. Finl. 1 (1992) 2.2. Rapeseed meal as a protein supplement for dairy cows The use of rapeseed meal in dairy cattle was reviewed e.g. by Thomke (1981) and Hill (1991). In the past the use of rapeseed meal or cakes as a protein supplement was limited due to the high glu- cosinolate content of single zero varieties of rapeseed. The addition ofbetween 0.5 and 1.5kg of RSM or cakes to the diet of dairy cows had no noticeable effect on milk production, compared to other protein supplements (Bunger et al. 1940, Poijärvi 1944, Jarl 1951, Nordfeldt 1958, Asplund and McElroy 1961,Lindell and Knuts- son 1976,Lindell 1976,Laarveld and Christen- sen 1976, Syrjälä-Qvist et al. 1982). Reduced palatability (Nordfeldt 1958) or milk yield was observed in some experiments, especially at high RSM levels (20 - 30% in the concentrate) (Asplund and McElroy 1961, Ingalls et al. 1968, Waldern 1973). Low glucosinolate double zero rapeseed cultivars have made RSM equal to other vegetable supple- ments in the diet of dairy cows (Ingalls and Sharma 1975, Sharma et al. 1977, Papas et al. 1978, Sanchez and Claypool 1983, Murphy et al. 1985, Vincent and Hill 1988, Vincent et al. 1990). Long term studies have produced some evidence of lowered fertility due to 0-RSM (Lindell 1976, Lindell and Knutsson 1976), and with 00-RSM on primiparous cows (Ahlin et al. 1985,Emanuel- son 1989, Emanuelson et al. 1991). This negative effect on primiparous cows was confounded with the type of rapeseed, as the glucosinolate content was highest in the beginning of the experiment, being reduced to one half in the course of the three- year experiment (Emanuelson 1989). 2.3. New systems for evaluating feed protein in dairy cattle The protein evaluation of ruminant feeds has tradi- tionally been based on digestible crude protein. During the last ten years many new systems have been reported, all taking into account the protein metabolism of both the rumen and the animal. Alderman (1987) has listed six new systems for evaluating feed protein: i. The British RDP/UDP system (rumen degraded and rumen undegraded protein) of Roy et al. (1977), subsequently published (1980) and revised by ARC (1984), which describes feeds in terms of truly absorbed amino nitrogen (TAAN). It is to be converted to metabolizable protein for practical purposes (MP=6.25 TAAN) (Webster 1987, 1992). ii. The French PDI system (protein digested in the intestine) (INRA 1978, Vérité et al. 1979, Vérité and Peyraud 1989). iii. The Swiss API (absorbable protein in the intes- tine), developed from the PDI system (Bickel and Landis 1987). iv. The German system of measuring the duodenal flow of crude protein (Ausschuss fur Bedarfsnor- men 1986, ref. Aldermann 1987, Rohr 1987). PiATKOwsKi et al. (1990) and Voigt and Piatkow- ski (1991) have corrected the efficiency of micro- bial protein synthesis by taking into account the negative effect of the proportion ofUDP in the feed protein. v. The Nordic 1 AAT-PBV system (absorbable amino acids in the duodenum; protein balance in the rumen) (Madsen 1985). vi. The American AP system (absorbed protein) (NRC 1986). 2.3.1. AAT-PBVfeedprotein evaluation system The AAT measures intraduodenally true absorb- able amino acids, which originate in the microbial protein synthesized in the rumen and the unde- graded feed protein (UDP) in the rumen. PB V is the balance between rumen degraded protein (RDP) and the protein requirement of microbes in the rumen (Madsen 1985, Hvelplund and Madsen 1990). The system for calculating the AAT-PBV values for feeds is presented in Appendix 1. 387 Agric. Sei. Finl. 1 (1992) ' The Nordic countries include Scandinavia and Finland 2.3.1.1. Effective degradation of feed protein Degradation of feed protein is measured using the in sacco method (Mehrez and orskov 1978), bearing in mind the recommendations of the CEC- EAAP workshop (Oldham 1987) and the Nordic Working Group (Madsen 1985, Lindberg 1985, Hvelplund 1990). The effective protein degrada- tion (EPD) is calculated according to a system devised either by Orskov and McDonald (1979), McDonald (1981) or Kristensen et al. (1982). The method reported by Orskov and McDonald (1979) is presented in Appendix 1. The correct EPD value, determined with the in sacco method, is always only an estimate, as it is affected by many factors in the feed and therumen, e.g., microbial nitrogen contamination of feed resi- dues in incubated samples, loss of particles from the bag, lower microbial density in the bag than outside it in the rumen, and the value used to calcu- late the rate of outflow from the rumen. The in sacco method is not suitable for feeds containing highly soluble protein, ifall soluble protein is not degraded in the rumen. The in vitro method, whereby inhibitors are used to stop the fermenta- tion, may be useful, but it requires automated labor- atory equipment (Broderick 1987, Broderick and Clayton 1992). 2.3.1.2. Microbial nitrogen contamination of feed residues in bag Feed residues in the bag after incubation include contaminated microbes, and the degradation values obtained are lower than true values (Varvikko and Lindberg 1985, Varvikko 1986). In Denmark a stomacher machine was used for washing bags to loosen microbes from the feed residue (Merry and McAllan 1983,Hvelplund and Möller 1987). In France, Michalet-Doreau and Ould-Bah (1989) developed a correction formula, whereby microbial nitrogen contamination is regressed against the nitrogen and NDF contents of the sample. On the other hand, after determining the nitrogen bound by NDF Lindberg (1988) calculated that the EPD of grass is constant and independent of the growth stage or level ofnitrogen fertilization. 2.3.1.3. Small particle loss from the bag In finely ground samples the loss of nitrogen in small particles is a problem, even a severe one, especially severe in oats, ranging from 30% to 70% ofthe total protein content. Weisbjerg et al. (1990) designed a correction formula for degradation val- ues adjusted for particle loss. Without any correc- tion for the loss of N in particles from the bag, this part will have to be included into fraction a, which is protein degraded rapidly in the rumen. In such feeds the EPD value will be excessively high and the AAT value excessively low. For a 60% loss of oat particles the AAT value will be ca. 30% too low. For oil meals the problem is less severe, with particle loss seldom exceeding 15%. Making corrections for particle loss is not devoid of problems, the underlying assumption being that protein in small particles is degraded at the same rate than the protein remaining in the bag, which may not always be true. The pore size of the bag fabric cannot be reduced, because intrusion of bacteria into the bag would be reduced. According to Meyer and Mackie (1986) the optimal pore size was 53 pm, and even then the concentration ofmicrobes was only 60% to 80% of that outside bag in the rumen. With cellulolytic bacteria the effect was even greater. Huhtanen and Khalili (1992) reached the same conclusion after measuring enzyme activity in the bag and in the particles present in the rumen. Much higher car- boxymethylcellulase and xylanase activity was measured in particles in the free rumen contents than in feed particles within the nylon bag. In some feeds protein is fully soluble, and then the EPD value is 100% and UDP 0%.This need not be correct, because some of the peptides have been observed to flow the duodenum (Chen et al. 1987). 2.3.1.4. Fractional outflowrate from the rumen In the Nordic system the outflow rate of particles 388 Agric. Sei. Finl. 1 (1992) from the rumen is constant, k = 0.08 (Hvelplund and Madsen 1990). In France the k-value is also constant, i.e. 0.06 for all feeds, but in Britain it var- ies from 0.02 to 0.08 depending on the feeding level (Vérité and Peyraud 1989,ARC 1984). The outflowrate for particles ofconcentrate has usually been reported to be higher than that of roughage particles (Colucci et al. 1982, Colucci et al. 1990). At maintenance level the outflow rate of particles has been measured as 3 to 4 per cent, ris- ing to 6 to 9 per cent at higher feeding levels (Ganev et al. 1979, Lindberg 1982, Elinam and ORSK.OV 1984, Colucci et al. 1982, Hadjipanayi- otou et al. 1988, Colucci et al. 1990). Markers, for instance straw or protein feed treated with neutral detergent and mortanded with chromium, were used in the studies quoted. Such treatment removes the soluble fraction of the marker carrier, and it behaves as indigestible feed particles. The outflow rate of indigestible NDF has been found to be much higher than that of digest- ible NDF (Tamminga et al. 1989, Huhtanen and Khalili 1991). This can be confirmed by means of the rumen evacuation technique, when the average rumen content can be accurately determined, and the ratio between daily faecal output of cell wall components and those contained by the rumen gives an indication of the average rate of passage of cell wall components from the rumen (Robinson et al. 1987, Tamminga et al. 1989). The nitrogen flowing into the duodenum com- prises ammonium nitrogen, with an outflow rate similar to that of the liquid phase, undegraded feed nitrogen with an outflow rate similar to that of NDF, and microbial nitrogen, which in the liquid phase is partly free and partly attached to feed particles. The outflow rate of nitrogen is between that of the liquid phase and that of the solid phase according to Tamminga et al. (1989). They reported an increase in the outflow rate of total-N from 0.038 to 0.060 when the feed intake of dairy cows increased from 6.0 to 24.0 kg per day. The average outflowrate of NDF, which is close to that of undegraded feed protein, varied from 0.019 to 0.025 in dairy cows in two trials. Huhtanen and Khalili (1992) reported outflow rates of NDF ranging from 0.016 to 0.017 in cattle on a diet of grass silage. The in sacco method also gives too low degrada- tion values for the potential degradable fraction in the bag (Tamminga et al. 1989), probably due to lower microbial enzyme activity in the bag than outside it (Meyer and Mackie 1986, Huhtanen and Khalili 1992). This, together with the overes- timated outflowrate for UDP, leads to underestim- ates of the effective degradation of feed protein in the rumen. 2.3.1.5. Efficiency of microbialprotein synthesis The most important factor for the new protein estim- ation systems is estimating the synthesizing effi- ciency of microbial protein in the rumen, as micro- bial protein accounts for 40% to 80% of the total amino acid nitrogen (AAN) entering the small intestine (Smith 1975, Sniffen and Robinson 1987). All systems are based on constant factors related to the feed energy released in therumen, but they differ in terms of how the energy released from the feed for use by microbes is estimated. The AAT-PBV system relates the energy released in the rumen to the total digestible carbohydrate content (DCHO), the British system relates it to the fer- mentable organic matter in the rumen (FOM), or total ME, whereas the French system relates it to the total digestible organic matter content less the sum ofthe ether extract, bag UDP and fermentation products. The experimental data shows great variations in the efficiency of microbial protein synthesis between diets. For instance, the AAT-PBV system is based on a value of20 grams of microbial amino- nitrogen per kilogram of DCHO. The figure was arrived at in an experiment where the microbial efficiency on a concentrate-rich diet was 60% lower than on a roughage-rich diet (20 g vs. 32 g AAN/kg DCHO) (Hvelplund and Madsen 1985). The flow of microbial nitrogen was measured from the difference between the total NAN content and the undegraded feed nitrogen, measured by the in 389 Agric. Sei. Finl. 1 (1992) sacco method. The DAPA (2,6-diaminopimelic acid) system was also used to measure the flow of microbial nitrogen, and this gave approximately 10 per cent higher efficiencyvalues for protein synthe- sis with concentrate-rich diets compared to the dif- ference method (Hvelplund and Madsen 1985). Tamminga (1981) measured an average effi- ciency valueof microbial protein synthesis of32 N grams per kilogram of fermentable carbohydrates in the rumen (FOM), with a variation of 25 to 38 grams. Hagemeister et al. 1980 reported value of 35 g microbial N synthesis per kg FOM. In a review by ARC (1984) the microbial efficiency was 26.8 grams microbial N/kg fermented organic matter (FOM) with a grass silage diet, whereas the corresponding figure for silage diet with concen- trate was 36 grams. The value of 32 grams was taken as a constant for this system. Microbes in the rumen either exist free in liquid, or are attached to small feed particles, most of the microbial mass being of a particle-associated popu- lation and microorganisms associated with proto- zoa (Olubobokun et al. 1988). It is therefore important to know how the markers behave with microbes in different pools. Protozoa contain no diaminopimelic acid, and the microbes in particles contain less DAPA and ribonucleic acid (RNA) than the microbes in the liquid phase (Merry and McAllan 1983, Craig et al. 1987).There is no dif- ference in the purin content between particle and liquid associated microbes, but the ratio of purin over total nitrogen is lower in protozoa than in bac- teria(Firkins et al. 1987a). The use of these mark- ers will result in excessively low microbial protein flow rates. Firkins (1987b) observed a 10% lower nitrogen flow rate in duodenalbacteria using purins as markers, compared to 15N. Increasing the feeding level will also increase the efficiency of microbial protein synthesis. This effect is minimal in growing cattle (Firkins 1987b, Glenn et al. 1989,Rooke et al. 1992),but in dairy cattle the effect of feeding level is really significant. Increasing the feed intake from 6.0 kg to 17.4 kg organic matter (OM) per day resulted in an increase in bacterial nitrogen yield at a growing rate 17.4-34.8 g bacterial N/kg OM apparently digested in the rumen (OMADR) (Robinson et al. 1985). The authors suggested that an increase in the quantity of fiber escaping the rumen, associated with increased feed intake, stimulates bacterial escape by washing out particles that are at an earlier stage of digestion with more bacteria attached. As the retention time of particles in the rumen decreases, the number of attached bacteria may increase rapidly. This may also explain the high efficiency values of 40 to 60 g microbial N/kg OMADR in dairy cows with a high feed intake (Rode and Satter 1988, McCarthy et al. 1989, Klusmeyer et al. 1990,Cameron et al. 1991). Fermented feeds, such as grass silage, contain fermentation products, that are not utilized as energy by microbes. Only about 25% of the energy derived from lactic acid is utilized by microbes (Demeyer and Van Nevel 1979). On the other hand, Chamberlain (1987) observed that ca. 50% of the energy released by fermenting proteins is util- ized by microbes. The new systems do not take into account these factors. Nousiainen (1992) pre- sented results on how this DCFIO-correction affects the EPD and AAT values of silage. He also calculated the true indigestible duodenal protein as a function of the crude protein content of silage, and used a double pool model (ARC 1984). AAT- values, calculated with the system, increased from 71 to 83 g/kg silage DM, when crude fibre content increased from 26% to 32% of DM. The corrected AAT-values increased less, from 72 to 78 g/kg DM, respectively. The results reported by Jaakkola et al. (1991) showed that microbial nitrogen yield decreased when the quantity of fermentation products increased. However, compared to silage, the micro- bial yield was lower with hay (Jaakkola and Huh- tanen 1990). There are inexplicable variations in the efficiency of microbial protein synthesis. For instance, Jaakkola et al. (1991) measured a mean efficiency of 39 g MN/kg OMADR in growing steers. Jaakkola and Huhtanen (1992a) used growing steers again, obtaining a mean efficiency of only 15 g MN/kg OMADR. In both trials the 390 Agric. Sei. Finl. 1 (1992) same methods and markers were used, and the feed was in both cases direct cut grass silage. These results showed that many related factors affecting microbial synthesis in the rumen remain unexplained. 2.3.1.6. Amino acid content and digestibility of duodenal proteins The proteins flowing through the duodenum con- sist of microbial, undegraded feed protein and endogenous protein. According to many published reports the proportion of amino acids in microbial protein is ca. 80% (ref. Hvelplund and Madsen 1990), and this value has been adopted by several systems (ARC 1980, NRC 1985, Vérité and Pey- raud 1989). In the AAT-PBV system the propor- tion of amino acid in the microbialprotein is lower, being 70%, based on the studies of Hvelplund and Möller (1980). Later Hvelplund (1986) estim- ated the proportion of amino acids to be between 0.62 and 0.72 in 49 isolated bacterial samples, and he still considers 70% to be the correct value. That value is also used in Germany (Rohr 1987). The proportion of amino acids in undegraded protein is estimated as 100% by many systems (ARC 1980, NRC 1985, Vérité and Peyraud 1989). The AAT-PBV system uses estimates of 85% for concentrates and 65% for roughage. This is consistent with the data in eighteen published papers collected by Hvelplund (1986). The digestibility of the duodenal amino acids varies in different systems: 80% (NRC 1985), 85% (ARC 1984) and 90% (Rohr 1987). In the French system digestibility of microbial amino acid pro- tein is 80%, and that of undegraded feed protein varies from 55% to 90% (Vérité and Peyraud 1989). In the AAT-PBV system the rate of true digestibility is 85% for amino acids of microbial origin and 82% for those of feed origin (Madsen 1985). The true digestibility of undegraded feed protein can be determined by using the duodenal bag (or mobile bag) technique, whereby feed samples incubated in the rumen and abomasum are passed through the small intestine in small nylon bags (Hvelplund 1985, Rae and Smithard 1985, Voigt et al. 1985, Varvikko and Vanhatalo 1988, Vérité and Peyraud 1989, Hvelplund and Madsen 1990). If TU means truly indigestible feed protein, the digestibility of UDP equals (UDP- TU)/UDP (Vérité and Peyraud 1989, Hvelplund and Madsen 1990). This method was adopted by the AAT-PBV system (Hvelplund 1990). TU can be determined from an intact feed sample using the mobile bag technique for concentrates. Preincuba- tion in the rumen is necessary for forage, as it increases digestion of feed protein in the intestines (Hvelplund et al. 1992). 2.3.1.7. Utilization efficiency of amino acids for lactation in dairy cows The utilization efficiency of amino acids is affected by the potential performance of the animal, the quantity of energy and AA supplied, and the bal- ance of AA in the feed (Rulquin and Journet 1987, Oldham 1987a). The theoretical maximum rate of 82% for efficiency would apply only when an ideal mix of AA is supplied, and the AA input places the first limit on performance (Oldham 1987a). The optimal values for efficiency obtained at normal feeding levels of energy and protein are, however, ca. 20% lower, (Rulquin and Journet 1987). The British protein evaluation system has adopted the value of 0.80 for the utilization effi- ciency of truly absorbed amino acids (ARC 1984), which may be too high for practical applications (Oldham 1987b). Kaufmann (1977) estimated the utilization rate to be 70%, and Piatkowski et al.(1990) suggested the same. The French system adopted the value of 64% for utilization efficiency of PDI in lactating cows, as values ranging from 0.58 to 0.69 were obtained in 24 trials(Vérité and Peyraud 1989). The AAT-PBV system adopted a variable rate of utilization efficiency ranging from 0.75 to 0.80 (Madsen 1985). In a practical feeding trial a lower efficiency rate was obtained, and there have been great variations within a single year and on one site (Kristensen et al. 1985, Oldham 391 Agric. Sei. Fin!. 1 (1992) 1987a). The protein requirement for milk production can be calculatedby dividing the milk protein yield by the efficiency of protein utilization. In the French system the protein requirement is 48 g of protein digested in the intestine per kilogram of milk (31/0.64). The recommendation for AAT require- ment in the Nordic system is 37-42 g AAT/kg ECM (Hvelplund 1990) corresponding an efficiency of 0.74-0.84. 3. Objectives of the study In Finland milk production is mainly based on forage and grain feeding. Unwilted grass silage usually accounts for one third of the total net energy intake (fattening feed units) in recorded herds. Grain is supplemented by protein concen- trates, soybean meal (SBM) and rapeseed meal being the most important ones. Rapeseed (mainly of turnip rape, Brassica campestris L.) is the only important oilseed plant that thrives in Finland. The present study examined how milk yield and quality is affected by various quantities of rapeseed meal in the diet, and the different treatments ofrapeseed. The feed protein evaluation in dairy cattle, which in Finland is still based on digestible crude protein, is (or will be) based on the AAT-PBV system in neighbouring Scandinavia. The second objective of this study was to compare the DCP and AAT-PBV systems using data ofFinnish milk production trials as material, from which various parameters describ- ing the utilization of feed protein were calculated. 4. Material and methods The experimental data of experiments of 1 to 5, including the materials and methods, is presented here. The data concerning experiments 6 to 9 has been reported elsewhere (See Table 12). The data of experiments 1 to 7 included feed intake and milk production with calculated DCP and AAT intakes, whereas the AAT intake was excluded in experi- ments 8 and 9 (RSM vs. control). In all experiments unwilted grass silage was given ad libitum (in one experiment silage or hay cut at the same growth stage was given). The quant- ity of concentrate was fixed at 6-8.5 kg per day, except for two trials, in which the concentrate was given according to milk yield (Exp. 2 and 6). Table 12. List of the experiments, ofwhich the data was used in this study. No. of Research Number of Treatments Source Exp. Station Years Cows Diets Experimental design 1 Viikki 11 1983-84 24 4 Continuous, comparison period of 11 weeks RSM levels, treament ofRSM 2 1984-85 24 2 Continuous, comparison period of 12weeks Treatment ofRSM 1 3 1987-88 10 5 Two 5x5 Latin squares, 25-day's periods RSM levels, treatment of RSM 4a Maaninka41 1987-88 8 4 Two 4x4 Latin squares, 4-week's period RSM-levels, silage 2,3 4b Maaninka4' 1987-88 8 4 Two 4x4 Latin squares, 4-week's period RSM-levels, hay 2,3 5 Viikki" 1989-90 14 7 Cyclic change-over, 4 periods of4 weeks RSM-and SBM-levels and treatments 6 Jokioinen3 ' 1986-87 30 6 Continuous, comparison period of 8 weeks RSM-levels, silage growth stage 4 7 Viikki" 1987-88 12 6 Cyclic change-over, 4 periods of4 weeks Treatment ofRSM 5 8 Suitia 2' 1987-88 20 4 4x4 Latin square, 4-week period RSM-level 6 9 Viikki" 1990-91 16 8 Cyclic change-over, 4 periods of4 weeks RSM-level 7 Research station: 11 University of Helsinki, Viikki Experimental Farm; 21 University of Helsinki, Suitia Experimental Farm; 31 Agricultural Research Center of Finland, Department of Animal Production; 41 Agricultural Research Center of Finland, North Savo Experimental Station Sources: 1) Tuori & Syrjälä-Qvist (1987); 2) Tuori & Syrjälä-Qvist (1988); 3) Suvitie & Rinne (1988); 4) Heikkilä et ai, (1989); 5) Huhtanen (1991): 2 experimental groups; 6) Huhtanen et ai. (1991): 2 experimental groups; 7)Huhtanen (1992, unpublished) 392 Agric. Sei. Finl. 1 (1992) 4.1. Experiment 1 4.1.1. Animals and management Twenty-four Finnish Ayrshire cows calving in the autumn were used in a continuous milk production trial. Sixteen cows were multiparous and eight pri- miparous. The cows were housed in a cowshed with individual stalls. Feed was given individually twice a day, and orts were collected and weighed once a day. The feed was accessible for seven hours daily. Unwilted grass silage was given in sufficient quantities to ensure orts averaging 5% to 10% of the mass of the silage, with intakes measured and recorded. Concentrate was offered at a fixed rate: 7 kg per day during lactation weeks 1 to 14, and 3.5 kg per day during weeks 15 to 28. The daily ration ofhay was 2 kg. The cows were milked twice daily commencing at 6 a.m. and 4 p.m. The milk was weighed and the figures were recorded with a True- test milkometer five days a week. The cows were weighed every second week. This report includes feed intake and production data for lactation weeks 4 to 14only. 4.1.2. Experimental design and treatments A randomized block design with four treatments was used. The cows were taken to the trial a few days after calving, being assigned to six blocks according to calving date and age, and then assign- ed within each block randomly to four treatments. The experiment covered lactation weeks 1 to 28, with the exception of the last blocks, which ended at about lactation week 20. One heifer died during the trial in her 19th week of lactation. The treat- ments consisted of diets including three different protein ratios in the concentrate and one group with treated protein. The control diet consisted of a concentrate including only a grain mixture, where- as the other diets consisted of a concentrate with 23% or 37% protein supplement ofwhich 88% was rapeseed meal. The smallerprotein supplement was either untreated or heat-treated. 4.1.3. Feeds The feeds used in the experiment were direct cut silage, mainly of the first grass crop of the season, containing timothy, meadow fescue and red clover (ratio 60:20:20). The herbage was harvested with a precision harvester, and 5 litres of formic acid per tonne of silage was applied at harvesting. The silage was stored in concrete tower silos. The hay consisted mainly of timothy. A pelleted protein supplement consisting of 88% either untreated turnip rapeseed meal (Brassica campestris L., var. “Span” and “Torch”, both single low varieties), or dry heat-treated rapeseed meal {Brassica napus L.), 8% dried brewers grain and 4% molasses. The control group was given the same amount of molasses with grain as a supple- ment. The dry heat-treatment was performed by Öljynpuristamo Oy, Helsinki. The total glucosinolate content of untreatedRSM was 43 pmol/g of fat-free matter, and of treated RSM 8 pmol/g fat-free matter, measured by gas- liquid-gas chromatography (GLC). The grain mixture of oat and barley (ratio 1:1) was preserved with propionic acid and rolled before feeding. The cows were also given daily 250 grams of a mineral mixture containing 17.5% Ca, 7.8% P and 6.0% Mg. 4.1.4. Samplingand analytical methods The silage samples taken weekly were dried at 103°C for 24 hours to determine their dry matter contents. The dry matter content was corrected for volatile losses by adding 41% of lactic acid, 89% of volatile fatty acids (VFA) and 100% of ammonia (Porter et al. 1984). The samples taken for establis- hing the fermentationpattern of silage, and for the in sacco determination, were kept frozen until analy- sed. For analysis the samples were pooled to seven batches according to the date and origin of the silage. Samples were taken from every new batch of hay and grain brought into the bam. For proximate (Weende) analysis samples were dried in a vacuum at 50°C and then milled through a 1 mm screen. The 393 Agric. Sei. Finl. 1 (1992) in vitro digestibility of silage and hay was measured (Tilley and Terry 1963), and standard procedures were followed in the proximate analysis. Samples of fresh silage were analysed for reducing sugars (Somogyi 1945, with modifications of Salo 1965), lactic acid (Barker and Summerson 1941), volatile fatty acids (Huida 1973), ammonia nitrogen (McCullough 1967), and for total and water soluble Kjeldahl nitrogen. Degradation of protein was determined in a sheep in sacco (Mehrez and ORSKOV 1977, Setälä 1983). The incubation times of concentrates were 2,5, 8 and 24 hours, and an additional 48 hours for silage and hay. The sheep was given a mixture of silage, hay and barley (ratio 45:45:10 onDM basis). The glucosinolate content of RSM was measured at the Department of General Chemistry of the University of Helsinki, using the GLC method described by Heaney and Fenwick (1980), with minor modifications as per Hase et al. (1988). The digestibility values for concentrate feeds in calculating feed values were derived from feed tables (Salo et al. 1982).Energy corrected milk was calculated according to Sjaunja et al. (1990). Milk samples were taken once a week (evening and morning) and analysed at the laboratory of Valio Finnish Co-operative Dairies’ Association for fat and protein with the infra red (IR) technique on a Milko-Scan 300 (Foss Electric, Denmark). Milk urea was measured using the enzymatic colo- rimetric method with a KONE CD analyzer (Raja- mäki and Rauramaa 1984). Goitrin content (L-5- vinyloxazolidine-2-thione) was measured at the laboratory of Valio using high performance liquid chromatography (HPLC) (Benns et al. 1979) and a reverse-phase column (Rauramaa 1983). 4.1.5. Statistical analysis A fixed model of least squares variance analysis was used to calculate the corrected means for intake and production data (Harvey 1966): yjk = p + diet+block + e..k The WSYS statistical software developed by Vilva (1989) was used in all calculations. 4.2. Experiment 2 4.2.1. Animals and management Twenty-four autumn-calving Finnish Ayrshire cows, twelve primiparous and twelve multiparous, were used in a continuous milk production trial. A standardization period of 4 weeks preceded the test period of twelve weeks. At the start of the trial the time since calving averaged 38 days. Concentrate was offered at therate of 0.3 kg per kilogram of fat- corrected milk. Unwilted grass silage ad libitum and about 2 kg hay were also available daily. The cows were housed as in experiment 1, except that milk yield was measured with a True-Test milko- meter every day. They were weighed once a fort- night, and twice at the beginning and end ofperiod. 4.2.2. Experimental design andtreatments A 2x2 factorial randomized block design was used, the factors being the treatment of RSM and bovine primiparity vs. multiparity. First the cows were assigned into five groups on the basis of calving date and age, after which they were randomly assigned to one oftwo treatments consisting of dif- ferent diets. The control diet consisted of a grain mixture including 17% turnip rapeseed meal and the test diet of the same amount of heat-moisture treated turnip rapeseed meal. 4.2.3. Feeds Direct cut silage was harvested partly with a flail and partly with a precision harvester into a tower silo, and 5 litres offormic acid per tonne was added to it. The grass was predominantly timothy, includ- ing some meadowfescue and red clover. A mixture of rolled barley and oat (ratio 1:1) was used as grain. The oat was dried, and the barley was preserv- ed with propionic acid. The RSM was of a single zero variety (Emma). The heat treatment of RSM (™Öpex-treatment at the plant of Öljynpuristamo Oy) was carried out in a pressure chamber in elevat- ed temperature and moisture. The total glucosino- 394 Agric. Sei. Fin!. 1 (1992) late content was 49 pmoles in untreated, and 15 pmoles in treated RSM, as measured with the GLC method by the Department of General Chemistry, University of Helsinki. 4.2.4. Samplingand analytical methods Feed and milk samples were taken as described in experiment 1. For analysis the weekly silage and hay samples were pooled over four-week periods. Every two samples of concentrate were pooled for analysis. The feed was analysed, and the in sacco degradable protein was determined as in experi- ment 1. The in sacco determinations were carried out in a sheep feeding on silage, hay, barley and RSM (dry matter ratio 40:40:10:10). 4.2.5. Statistical analysis The milk production data was analysed in terms of covariance, using the least squares variance ana- lysis (Harvey 1966): y... =a+T.+ A. + b. *P... + e...J ijk i j 1 ijk ijk where T; is treatment, A j =primiparity vs. multi- parity of the cow; and P jjk is the production rate of the cow during the standardization period. The effects of the block and interaction between age and treatment were not significant, and they were excluded from the model. WSYS statistical soft- ware was used for the calculations (Vilva 1989). Feed intake was analysed without covariance. The least squares corrected means were calculated from production data (Appendix 3), but uncorrected data was used in comparing the protein systems. 4.3. Experiment 3 4.3.1. Experimental animals and management Ten autumn-calving Ayrshire cows were used in a changeover trial of five periods, each lasting 25 days. Five of the cows were primiparous and five were multiparous. At the beginning of the experi- ment the average time since calving was 50 days (8 to 84 days). The cows were individually fed in a cowshed, having free access to silage at all times. Concentrate was fed at the fixed rate of 8 kg per day. In addition, 250 g ofa mineral mixture (17.0% Ca, 8.0% P, 6.0% Mg, 6.0% Na) was given daily. No hay was offered. The cows were weighed on two consecutive days in the beginning, in the middle and at the end of each period, always before the afternoon feeding. The change in live weight was calculated by regression. 4.3.2. Experimental design and treatments An experimental design of two balanced 5x5 Latin squares was used. Multiparous cows were assigned to one squareand primiparous cows to another. The treatments consisted of different ratios or treat- ments of rape seed meal in the diet. The control treatment (A) consisted of a diet devoid ofRSM, in which a mixture of barley and oat (ratio 1:1) was given as concentrate. The grain was partially re- placed by varying quantities of heat-moisture trea- ted (Öpex) double zero RSM (var. “Esko”), in treatment B by 12%, and in treatment C by 24%. Treatment D contained 24% untreated single zero RSM (var. “Emma”), and treatment E 24% untrea- ted, glucosinolate-reduced double zero RSM treated with FeSO..4 4.3.3. Feeds The direct cut silage came from swards of first cut timothy and meadow fescue. It was harvested by precision harvester, treated with 5 litres of formic acid (80%) per tonne of grass, and stored in bunker silos. The double zero RSM in treatments B and C was heat and moisture treatedby the same method as in experiment 2 (Öpex treatment at the plant of Öljyn- puristamo Oy). In treatment D the single zero RSM was not heat-treated, and in treatment E, instead of the heat and moisture treatment, the double zero RSM was treated with ferrosulphate to increase the breakdown of glucosinolates. The RSM of single zero variety “Emma” was not quite pure, however. 395 Agric. Sei. Finl. 1 (1992) as some double zero RSM of variety “Kova” was mixed with it. The total glucosinolate content, meas- ured by high performance liquid gas chromato- graphy (HPLC), was 28, 13 and 7 pmol/g fat-free meal respectively for the single zero variety “Emma”, heat treated double zero variety “Esko” and glucosinolate-reduced double zero “Esko”. 4.3.4. Samplingand analytical methods The weekly samples ofsilage were analysed for dry matter and pH, and then pooled over each experi- mental period. Samples of concentrates (grain, RSM) were taken in the last week of each test period and analysed separately. In addition to the analyses described under experiments 1 and 2, neutral detergent fiber (NDF), acid detergent fiber (ADF) and acid detergent lignin (ADL) were deter- mined according to Goering and Van Soest (1970). Acid insoluble ash was determined from feeds and faeces with 2-N HCL according to Van Keulen and Young (1977). The glucosinolate con- tents of various rapeseed meals were measured by HPLC according to Björkqvist and Hash (1988) at the Department of Chemistry, Helsinki University of Technology. Degradation of protein in sacco was determined using a cow, which was given 2 kg of hay, 6 kg of concentrate, and silage ad libitum daily. The concentrate contained oat, barley and RSM. The incubation times of concentrates were 3,6, 12, 24 and 48 hours, with an additional72 hours for silage. For all feeds the loss due to washing off, and for concentrates the loss ofparticles containing protein in the bag were determined. Particle loss was taken as the difference between washing loss and soluble loss. Soluble loss of crude protein was determined by taking a sample ofca. 1 g, incubating it in water for one hour at 39°C, centrifuging and washing it three times, after which the amount of nitrogen in the precipitate was determined(Huhtanen 1991). Milk samples were collected from four consecu- tive milkings during the last two weeks of each trial period. Milk fat, protein and lactose contents were determined by the Valio laboratory using the IR technique on a MilkoScan 605. The pooled samples of the last week of each trial period were analysed for urea (Rajamäki and Rauramaa 1984) by the Valio laboratory, and for goitrin (Benns et al. 1979, Rauramaa 1983) by the laboratory of Viljavuus- palvelu Oy. The goitrin content was measured from the samples of primiparous cows only. The apparent digestibility of the diets was meas- ured using acid-insoluble acid as marker (Van Keulen and Young 1977). During the last five days of each trial period faeces samples were taken twice a day from the rectum of each cow. The indi- vidual samples were pooled into one sample per period, and kept frozen until analysed upon being dried at 100°C for 72 hours and milled through a 1.5 mm screen. 4.3.5. Statistical analysis The production data was analysed in terms of least squares variance (Harvey 1966)using WSYS stat- istical software (Vilva 1989) and following the model: yijldm =M + si +cj(s i) + p k +Tl+ (s*p)ik + (S*Di, + ejjklm where S., C, P. and T, are the effects of thei’ y k 1 square, the cows in the squares, the period and the treatment, respectively. The residual degree offree- dom was 24. The effect of the treatment was further analysed using contrasts (Snedecor and Cochran 1980); Cl linear effect of heat-moisture treated double zero variety RSM C 2 quadratic effect of heat-moisture treated double zero variety RSM C 3 effect ofvariety (zero variety dietD versus dou- ble zero variety diet E). The effect of FeS04 treatment was confounded. C 4 effect of heat treatment of RSM: heat treated RSM on diet C versus untreated RSM on diet E; the effect FeS0 4treatment confounded. 396 Agric. Sei. Finl. 1 (1992) I. 4.4. Experiment 4 4.4.1. Animals and management Sixteen autumn-calving Ayrshire cows were in a changeover trial covering four periods. The length of each period was 28 days. Eight of the cows were primiparous and the other eight multiparous. The average time lapse since calving to the beginning of the experiment was 89 days. The multiparous cows had calved 4.9 times on average. The cows were housed and individually fed in a cowshed with separate stalls, where they had free access to forage for 12 hours a day. They were given hay or silage at 6 a.m. and 1 p.m, and concentrates at 8 a.m. and 12 noon. Forage (hay or silage) was available ad libi- tum, whereas concentrates were rationed to 7.5 kg daily throughout the experiment. The daily diet was supplement with 250 grams of mineral mixture (See exp. 3 for composition). The consumption of feed was measured every day, and the milk yield was measured at every milking. The cows were weighed on two consecutive days at the beginning and at the end of each period. 4.4.2. Experimental design and treatments The experimental design was four 4x4 Latin squares. Two squares, one with primiparous and another with multiparous cows, were feeding on silage, while the cows in the other two squares were feeding on hay. There were four treatments within each square: 0,8, 16 or 24 per cent of RSM added to a concentrate containing mainly oat and barley. 4.4.3. Feeds Direct cut silage and hay (a mixture of 20% timothy, 20% meadow fescue and 60% couch grass) were obtained from a sward cut for the first time in the season. In the spring the sward had been fertilized with 112 kg nitrogen, 22 kg phosphorus and 45 kg potassium per hectare. The grass was harvested with flail harvester, adding 4.9 litres of AIV-11-solution (80% formic acid) per tonne. The barn-dried hay was cut at the same time as the silage was harvested. The double zero variety “Esko” was used as RSM. 4.4.4. Samplingand analytical methods Samples of silage and hay were taken weekly and the dry matter of both, in addition to the pH value of silage, were determined. Samples for feed ana- lyses were taken during the last week of each period and analysed in the same manner as in expe- riment 3. Faeces samples were subjected to proxi- mate and fiber analysis. Milk samples were taken once a week at morning and evening milkings and analysed at the laboratory of Lapinlahti Dairy for fat, protein and lactose, in addition to employing the NIR technique and Seralyzer reflection photo- metry for urea analysis. Milk samples taken during the last week of each period were analysed goitrin (Benns et al. 1979,Rauramaa 1983) by the Valio laboratory. The energy content of feed, faeces and milk samples were measured with an adiabatic process on a Parr 1241 adiabatic bomb calorimeter. The energy values ofmilk samples were determined by soaking a filter paper in ca. 3.5 g of the sample and igniting it in the bomb. Ruminal degradation of protein was determined in sacco using a cow. The cow was given grass silage, hay and concentrate. The incubation times ofall feeds were 3,6,12,24,48 and 72 hours. Wash- ing off and particle loss from the bag were also determined. The apparent digestibility of the diets was meas- ured using acid insoluble ash as a marker (Van Keulen and Young 1977). Faeces were collected from the rectum twice a day during the last five days of each test period. 4.4.5. Statistical analysis Production data was analysed for RSM levels by means of the least squares variance analysis (Har- vey 1966) with the general linearprogram of SAS (1989) following the model: 397 Agric. Sei. Finl. 1 (1992) y, )k.m„ =H+F,+v Pk + T, + + (p*p).k + (F*T)ffl + (A*P)jk + (F*A*P)ijk + (F*A*T) jj, + C (F ) + C (A.) + C (F*A )+ emv v m v y m v \y ijkltnn where F., A,, Pk, T, and Cm signify the effects of forage (silage vs. hay), the age of the cow (primi- parous vs. multiparous), period, treatment and the animal’s position in the square, respectively. The residual degree of freedom was 24. The effect of the treatment was further partitioned into the effects of orthogonal contrasts (linear, quadratic and cubic). Table 13. Composition of concentrate mixtures in the experiment 5. Treatment Control RSMI RSM2 TRSM SEMI SBM2 TSBM2 Barley 100.0 44.0 38.0 44.0 45.8 41.6 45.8 Oats 100.0 44.0 38.0 44.0 45.8 41.6 45.8 RSM 12.0 24.012.0 24.0 TRSM 12.0 SBM 8.4 16.8 TSBM 8.4 RSM = rapeseed meal (00-var.), SBM = soybean meal, T = Öpex-treated meal 4.5. Experiment 5 4.5.1. Animals and management A changeover experiment of four test periods using fourteen autumn-calving Ayrshire cows. The length of each period was 28 days. Seven of the cows were primiparous, and the other seven multiparous. The time since calving averaged 53 days at the beginning of trial. The cows were housed as in experiment 3. Grass silage was avail- able ad libitum, the concentrate being rationed to 8 kg per day during the whole experiment. The diet was supplemented daily with 250 g of a mineral mixture containing 16.8% Ca, 7.7% P, 6.0% Mg and 6.9% Na. The feeds were freely accessible at all times. The cows were weighed on two consecu- tive days at the beginning, in the middle and at the end of the period. The change in weight was calcu- lated by regression. 4.5.2. Experimental design and treatments A cyclic changeover design was used in the experi- ment (Davis and Hall 1969), involving seven treatments, fourteen cows and four periods. The treatments consisted of diets including various pro- portions of SBM, treated SBM (TSBM), RSM, or treated RSM (TRSM) in the concentrate mixture (Table 13). 4.5.3. Feeds Direct cut silage, containing predominantly timothy and meadow fescue, was harvested from swards for the first time in the season. The swards has been fertilized with 70 kg nitrogen, 30 kg phosphorus and 80 kg potassium per hectare. The silage was harvested with a fine chopper, and 4 to 5 litres of AIV-II solution (80% formic acid) was added to each tonne of silage as a preservative. SBM and RSM were extracted at Öljynpuristamo Oy. The soybeans were imported from USA, the rapeseed was of the Finnish double zero variety “Kova”. Both SBM and RSM were either untreated after extraction, or heat-moisture treated (Öpex- method by Öljynpuristamo Oy). The total quantity of glucosinolates, determined by HPLC, was 31 pmoles in seeds, 13 pinoles in RSM, and 8 pmoles in the treated RSM per gramof fat-free matter. 4.5.4. Sampling and analytical methods The feeds were sampled and analysed as in experi- ment 3. Silage and hay were analysed for each period, the samples of concentrates were pooled for periods 1 and 2 and again for 3 and 4. The glucosi- nolate contents were determined using HPLC (Björkvist and Hase 1988) at the laboratory of Öljynpuristamo Oy. Amino acid content of rape- seed and soybean meals was analyzed using a gas chromatographic method(Näsi and Huida 1982). 398 Agric. Sei. Finl. 1 (1992) Ruminal degradability of the feeds was measured in sacco using one heifer. The heifer was given grass silage, hay and 1.5 kg concentrate daily. The incubation times were 3,6, 12, 24,48 and 72 hours for concentrates, with an additional 96 hours for silage and hay. The loss through washing off the bag was determinedfor all feeds, as well as particle loss for concentrates. Intestinal disappearance of protein was measured for intact rapeseed and soybean meals by the mobile nylon bag technique (Hvelplund 1985, Varvikko and Vanhatalo 1988). Approximately 1.1 g offeed was weighed in 3.5 x 5.0 cm heat sea- led polyester bags. The pore size of the cloth was 16 pm and the open surface 5% of the area. The bags were introduced through the T-cannula into the proximal duodenum. Once excreted, the bags were machine washed for 50 minutes at 40°C. Ini- tial incubations, averaging a batch of six bags each, were carried out in four heifers and one bull. After drying in 60°C all the residues of one batch (one incubation) were combined, milled through a 0.8 mm sieve and analyzed fornitrogen. Milk samples were taken over 2 milkings and pooled to one sample in the third and fourth week of each test period. Milk fat, protein and lactose was determined by the IR-technique and urea was determined once in each period by the Valio labor- atory. The apparent digestibility of the diets was meas- ured in multiparous cows, using acid insoluble ash (AIA) as a marker. The collection and treatment of faeces followed the method used in experiment 3. 4.5.5. Statistical analysis The following formula for statistical analysis was used: y ttUm = H + B,+ Cj(S ä) + p k + T, + (B*P) jk + (B*T)a + where Ef, C, Pk and T, are the effects of the block (viz. primiparity vs. multiparity), the cow in the block, the period and the treatment, respectively. The effect of the treatment was partitioned to the linear effect of protein level (Cl), quadratic effect of protein level (C2), effect of SBM vs. RSM (C3) and effect ofRSM treatment (C4). 4.6. Estimating the effect of RSM on the milk and protein yield In each experiment there was a calculated deviation of the RSM group from the 0-RSM group for RSM intake and milk production. These deviations were used as measurements when calculating linear and non-linear regressions between RSM intake and production. Non-linear regression was calculated using the Gauss-Newton weighted least square technique (SAS 1989). Only RSM trials with vary- ing proportions of RSM were included (Experi- ments 1, 3-6, 8-9). The effect of treatment on protein protection of RSM was studied in experiments 1-3 and 5, in which a protected RSM group was included. 4.7. Comparison ofAAT-PBV and DCP systems Milk production experiments 1 to 8 were used as data for calculating the parameters, and their coef- ficient of variation, associated with the utilization of protein in milk production: i. Linear regression between protein intake and production. ii. Multiple regression, where protein production was explained in terms ofprotein intake and cor- rected ME-intake. The corrected ME was estim- ated by subtracting from the actual ME intake the intake adjusted for estimated regression, where ME intake was explained in terms ofpro- tein intake. iii. Feed protein utilization in milk production; pro- teinproduction by feed protein for production (g milk protein/g feed protein). Feed protein con- version in milk production: feed protein for pro- duction by milk production (g feed protein/kg ECM). 399 Agric. Sei. Finl. 1 (1992) 5. Results and discussion 5.1. Composition and nutritional value of feeds 5.1.1. Chemical composition offeeds The chemical composition and energy and protein value of the feeds used in experiments 1 to 5 are is presented in Appendix 2, and Tables 14 and 15 Table 14. The average composition of the feeds (Exp. I -5). contains a summary. The crude protein content of silage was fairly high in all experiments (mean 17.2%, variation 16.3% to 18.5% in DM), while the crude protein content of RSM ranged from 32.7 to 39.1 per cent. The fat content of RSM was higher in experiments 1 to 4, ca. 9% to 7% ofDM, than in the fifth experiment, where it was 5% to 6% of DM. The mean degradation parameters and AAT- PBV values of different feeds in experiments 1 to 5 No. of DM, % In DM, % analysis Ash Crude Ether Crude NDF" ADF" ADL" protein extract fibre Silage 28 22.9 7.1 17.2 5.6 28.1 50.8 29.2 1.9 Hay 11 85.1 6.8 8.8 2.2 34.7 70.4 36.9 2.7 Hay(early cut) 4 87.5 9.0 18.0 2.4 29.5 58.6 30.6 2.4 Barley 14 82.5 2.6 11.8 2.8 5.1 20.5 5.4 0.8 Oat 14 87.1 3.2 12.7 6.3 10.7 26.5 11.7 2.3 RSM 15 88.7 7.6 36.0 8.0 12.7 27.5 19.0 8.6 TRSM(Opex)21 13 88.1 7.8 34.9 8.3 13.0 27.1 18.5 7.9 SBM 2 87.9 7.4 49.5 3.0 7.6 13.8 8.0 0.8 TSBM(Opex)2 » 2 89.2 6.5 50.1 3.8 8.2 14.6 8.3 0.4 " NDF, ADF and ADL are determined only in Exp, 3-5. 2) heat-moisture treatment for protein protection. Average fermentation quality of silage: pH 3.94; (percentage in DM): lactic acid 4.8, sugars 4.2, acetic acid 1.7, butyric acid 0.03; percentage of total N: soluble N 52.8; ammoniura-N 4.1 Feeds fed in different experiments: Silage, barley, oats and RSM in all experiments; hay in Exp. 1,2 and 5; early cut hay in Exp. 4, RSM(Öpex) in Exp. 1,2,3,4,5; SBM and SBM(Öpex) in Exp. 5. Table 15. The average energy and protein values of the feeds (Exp. 1-5). Silage Hay Hay Barley Oats RSM RSM SBM SBM (early cut) Öpex Öpex per kg DM: FFU 0.754 0.561 0.692 1.166 1.038 1.020 1.014 1.056 1.069 ME,MJ 10.64 9.18 10.14 13.62 12.34 12.12 12.07 12.46 12.58 NEL, MJ 5.92 5.19 5.90 7.99 7.20 7.26 7.23 7.48 7.56 DCP, g 124 53 128 88 102 299 290 445 451 AAT, g 76 75 80 103 72 141 157 183 242 PBV, g 40 -42 42 -49 9 149 115 223 146 FFU = fattening feed unit (0.7 kg starch), ME metabolizable energy according to MAFF (1975), NEL net energy in lactation according to VAN ES (1978), DCP digestible crude protein, AAT= absorbable amino acids in the duodenum, PBV = protein balance in the rumen. Calculation of AAT/PBV-values: ramen outflow rate (k-value) 0.08, EPD-values calculated according to ORSKOV and McDonald (1979) and values for roughages are corrected formicrobial N contamination in bag residues (Michalet-Doreau & Ould-Bah 1989), true digestibility ofUDP (TD.UDP) is constant 0.82. 400 Agric. Sei. Fin!. 1 (1992) are presented in Table 16. The mean values of de- gradation at different incubation times of feeds in experiments 3 to 5 are given in Appendix 2. The calculation methods of EPD and AAT-PBV values follow the system in general (See footnote in Table 15), except that there is no correction for degrada- tion according to particle loss from the bag during incubation (Hvelplund and Madsen 1990). The average degradability of the silage protein was low (Table 16). In the first two experiments the EPD of silage was only 59% to 68% (with micro- bial nitrogen adjustment), but rose to between 81% and 87% in the other experiments. The effect of heat treatment on oilseed meals varied (Table 17). In the first experiment the EPD value ofdry-heated RSM was only 11%, while that of the untreated RSM was 50%. In experiment 3 there was only a slight difference in EPD between the untreated and heat-treatedRSM. In experiment 4 treatedand untreatedRSM were not compared, as only Öpex-treated RSM was used. That RSM was relatively highly degradable. In experiments 2 and Table 16. The average values for protein degradationparameters and AAT-PBV values of the different feeds (Exp. 1-5). Degradation parameters % AAT, AAT, AAT, PBV a be Micr.N EPD, % MPS, UDP, total g/kg DM corr. g/kg DM g/kg DM g/kg DM Silage 43.9 45.7 0.088 8.6 74.9 53.3 22.5 75.8 40.2 Hay 21.5 48.9 0.098 14.3 61.5 57.4 17.8 75.2 -42.0 Hay,e 44.2 41.6 0.092 8.1 74.2 54.8 24.7 79.6 41.6 Barley 23.3 71.4 0.150 69.0 77.3 25.4 102.7 -48.5 Oat 67.7 25.6 0.218 86.3 60.0 11.9 71.9 8.7 RSM 19.5 75.7 0.086 57.8 35.4 105.2 140.6 149.4 RSM-Öpex 6.1 82.6 0.099 50.5 35.9 121.3 157.1 114.6 SBM 20.0 82.7 0.065 57.2 35.6 147.7 183.3 223.3 SBM-Öpex 9.3 100.4 0.037 40.8 35.0 206.9 241.9 145.7 Calculation of AAT/PBV-values: see footnote in Table 15. Table 17.Protein degradation parameters and AAT-PBV values ofRSM in the different experiments. Exp. Treatment Degradation parameters % AAT, AAT, AAT, PBV no. abc EPD, % MPS, UDP, total g/kg DM g/kg DM g/kg DM g/kg DM 1 RSM-0 Untreated 19.8 78.1 0.050 49.8 34.8 123.8 158.5 117.5 RSM-00 Heated 4.3 71.2 0.008 10.4 32.3 253.1 285.4 -12.3 2 RSM-0 Untreated 22.5 72.4 0.080 58.7 36.3 104.0 140.3 151.2 RSM-0 Öpex 1.8 72.1 0.085 38.9 36.3 152.0 188.3 77.6 3 RSM-00 FeS0 4 4.6 88.8 0.115 56.9 34.6 100.1 134.7 131.4 RSM-0 Untreated 7.4 84.9 0.118 58.0 35.6 97.8 133.4 133.8 RSM-00 Öpex 3.2 95.6 0.103 50.7 35.5 114.2 149.8 108.9 4 RSM-00 Öpex 12.7 79.3 0.147 64.2 36.2 81.8 118.0 149.0 5 RSM-00 Untreated 29.8 65.4 0.095 65.3 35.4 94.4 129.8 195.8 RSM-00 Öpex 13.2 83.2 0.059 48.2 35.4 136.8 172.2 123.2 See footnotes in Tables 15 and 16 401 Agric. Sei. Finl. 1 (1992) 5 EPD of untreated RSM was 59% and 65%, and those of Öpex-treated RSM were 39% and 48% respectively. 5.1.2. Intestinal degradation ofRSM and SBM as measured by mobile nylon bag technique The disappearance in the intestines of intact samples of RSM and SBM in experiment 5 was measured using the mobile nylon bag technique (Table 18). Feeds were not treated with pepsin- HCI, as it has a negligible effect on the disap- pearance offeed crude protein (CP) in the intestine (Varvikko and Vanhatalo 1991, Vanhatalo and Aronen 1991). Measured with this technique, the heat-moisture treatment had no effect on the intestinal degradation of dry matter and crude pro- tein, a finding also reported by Vanhatalo and Aronen (1991). The estimated true digestibility of UDP [TD.UDP = (UDP-TU)/UDP] was calculated using predetermined EPD values and the values for indi- gestible nitrogen of feeds determined in the present study. The true digestibility of UDP, as estimated with this technique, agreed with theresults ofVan- hatalo and Aronen (1991), when TD.UDP of SBM and RSM ofthe same varieties and treatments were measured with the bag technique after 10 hours’ incubation in the rumen. Here estimates (based on the measured TU value) oftrue digestibil- ity values of UDP were lower for RSM and higher Table 18. Disappearance of DM and CP of intact RSM and SBM during intestinal incubation in mobile nylon bag (Exp. 5). EPD(%) Intestinal Calculated in the disappearance (%) TD.UDP rumen DM CP RSM-untreated 65.3 67.6 89.1 68.6 RSM-Öpex 48.2 68.6 89.1 79.0 SBM-untreated 57.1 85.3 97.1 93.2 SBM-Öpex 40.8 83.4 97.4 95.6 5.E.1)1.971.17 11 5 animals, 4 feeds, 19 observations, resid. d.f. 11(1 miss- ing obs.) 402 for SBM than using 7% as the TU value, as pro- posed in the AAT-PBV system (Hvelplund 1990). Microbial-N contamination of the mobile bag resi- dues was not taken into account in this estimate, but it was lower for concentrate feeds than for fibrous feeds (Varvikko and Vanhatalo 1990). 5.1.3. Amino acidcontent ofRSM and SBM The amino acid contents of RSM and SBM were measured in experiment 5 (Table 19). After heat and moisture treatment the amino acid content of both RSM and SBM was reduced, but more mar- kedly in RSM than in SBM. Especially the lysine, histidine and arginine contents had decreased by more than 15% in RSM, the corresponding reduction being only 4% to 7% in SBM. Earlier experiences concerning these have been variable: Näsi and Siljander-Rasi (1991) found heat and moisture treatment to have a minor effect on lysine content, whereas Näsi et al. (1985) found a noti- ceable reduction in lysine and the available lysine content of dry-heat treated RSM. Rae et al. (1983) Table 19. Amino acid composition of the RSM and SBM (both untreated and Öpex-heat-moisture treated ) in experi- ment 5. g/l 6g N Rapeseed meal Soybean meal Normal Öpex Normal Öpex Lysine 5.7 4.8 6.7 6.2 Histidine 2.4 2.0 2.3 2.2 Arginine 6.2 5.3 7.6 7.1 Asparticacid 7.9 7.4 11.4 11.0 Threonine 4.9 4.6 4.2 4.1 Serine 4.7 4.4 5.2 5.2 Glutamic acid 17.2 15.5 18.3 17.9 Proline 5.8 5.9 5.2 5.1 Glycine 5.5 4.8 4.1 4.5 Alanine 4.6 4.3 4.5 4.4 Valine 5.4 5.2 4.8 4.6 Isoleucine 4.7 4.6 4.8 4.7 Leucine 7.5 7.0 7.9 7.8 Tyrosine 3.4 3.2 4.0 3.9 Methionine 1.5 1.4 0.8 0.7 Phenylalanine 4.5 4.2 5.5 5.4 Total 91.6 84.0 96.9 94.3 Agric. Sei. Fint. 1(1992) reported that treating RSM with formaldehyde (1.2 g FA/100 g CP) reduced lysine content by 29%. Formaldehyde treatment with 0.4 to 0.8 g FA per 100 g CP decreased the lysine content only by 2% to 6% (Setälä and Syrjälä-Qvist 1984/85). 5.1.4. Glucosinolatecontent ofRSM A considerable reduction in glucosinolate content (Table 20) was observed between single-low (Experiments 1 and 2) and double-low varieties of rapeseed (Exp. 3, 4 and 5). The total quantity of glucosinolates was reduced from 40 - 50 pinoles per gram of fat-free material in experiments 1 and 2 to 14pmoles per gram in experiment 5. However, even the largest quantity of glucosinolates, 49 pmoles in exp. 2, is quite moderate compared to values ranging from 100 to 205 pmoles/g for single-low canola meal (Shahidi 1990b). In 1987 (Exp. 3) the single-low RSM was found to contain less glucosinolates than the canola stand- ard of 30 pg alkenylglucosinolates per one gram of defattedmeal. That was due to the contamination of the single-low rapeseed with double-low varieties. The rapeseed meal produced in 1989 (Exp. 5) ful- filled the EEC standard of 20 pmoles per gram of seed (included the indolylglucosinolates 4- hydroxyglucobrassicin and glucobrassicin) (Table 21). Since the cultivation of the new turnip rape- seed variety, “Kulta”, began in Finland, the gluco- sinolate content has dropped to below 10 pmoles per gram of seed (Vilkki 1991). The main glucosinolates of the double-low turnip rapeseed are progoitrin, glucobrassicanapin, glu- conapin and 4-hydroxyglucobrassicin. The propor- tion of the last one has increased, while total quant- ity of glucosinolates has decreased. Sang and Salisbury (1988) measured doubled proportions of that glucosinolate in double-low rapeseed com- pared to the single zero varieties (Table 10). Having been crushed and extracted, the rapeseed meal containedabout half the quantity of glucosin- olates found in intact seeds, as calculated from defatted dry matter. The heat and moisture treat- ment of RSM further reduced the quantity to a half (Tables 20 and 21). Especially the proportion of 4- hydroxyglucobrassicin had decreased during the processing, whereas the proportion of progoitrin had increased (Table 21). The analytical methods used do not, however, include measurements of the degraded products of glucosinolates. The presence of moisture during heating was found to be an essential factor in reducing the quantity of gluco- sinolates (Reynolds and Young 1964, Appelqvist and Josefsson 1967, Belzile et al. 1963, Eapen et al. 1968,Shahidi andNACZK 1990). Table 20. Glucosinolate content (pmoles/g defatted meal) of the rapeseed meal in different experiments. Exp. Origin GNA GBN PRO NAP Others Total no. 1 RSMO(I9B3, B. camp.) 10 21 10 2 43 TRSM(I9B3, B. napus) 3 2 3 0 8 2 RSMO(I9B4, B. camp.) 14 11 21 3 49 TRSMO(Öpex)(I9B4, B. camp.) 3 4 7 1 15 3 RSMO(I9B7, B. camp.) 8 8 10 2 28 RSMOO(Öpex)(l9B7, B. camp.) 4 3 4 1 1 13 RSMOO(FeSO 4 ) (1987, B. camp.) 2 1.5 2 1.5 17 5 RSMOO(I9B9, B. camp.) 1.4 3.4 5.6 0.8 2.3 13.5 RSMOO(Öpex)(l9B9, B. camp.) 1.5 1.6 3.2 0.4 0.8 7.5 TRSM (1983, B.napus) was ofDanish origin, all others were ofFinnish origin. Treatments: 1983:dry-heated, 1984-1989 heat-moisture treated in a pressurized chamber; FeS0 4 =reduced glucosinolates by FeS04 ; GNA = gluconapin, GBN = glucobrassicanapin, PRO= progoitrin, NAP =napoleiferin 403 Agric. Sei. Finl. 1 (1992) Table 21. Glucosinolate content ofRSM in details from experiment 5. pmoles/g DM pmoles/g defatted matter Seed RSM TRSM Seed RSM TRSM Gluconapin 3.0 1.4 1.5 5.0 1.5 1.6 Glucobrassicanapin 3.7 3.4 1.6 6.2 3.6 1.7 Progoitrin 5.6 5.6 3.2 9.3 5.9 3.4 Napoleiferin 0.9 0.8 0.4 1.5 0.8 0.4 Gluconasturtiin 0.2 0.2 0.1 0.3 0.2 0.1 Glucobrassicin 0.1 0.05 0 0.2 0.05 0 Neoglucobrassicin 0.03 0 0 0.05 0 0 4-hydroxyglucobrassicin 3.0 0.5 0.06 5.0 0.5 0.06 SI" 0.8 0.7 0.4 1.3 0.7 0.4 52 1.0 0.4 0 1.7 0.4 0 53 0.2 0.2 0.1 0.3 0.2 0.1 Others 0.3 0.2 0.1 0.5 0.2 0.1 Total 18.83 13.45 7.46 31.35 14.05 7.86 -indolyl glucosinolates 3.13 0.55 0.06 5.22 0.58 0.06 11 STS3 are glucosinolates with S-containing R-moiety Table 22. The effect ofRSM on the digestibility of the diet. RSM-level in the concentrate (%) S.E. Significance ofcontrasts 0 8 12 16 24 Lin. Quadr. Cubic Experiment 3 OM 76.6 - 77.2 - 74.7 1.21 0.73 0.22 CP 72.4 74.5 - 74.4 1.13 Burgess & 10.0-24.6 0.27 9.6 0.04 0.10" Nicholson 1984 11.3 SBM 14.2-17.3 0.06 6.8 0.29 -0.03 Mayne & Gordon 14.2-20.7 0.09 5.7 0.17 0.00 1985 14.2-24.1 0.10 5.8 0.16 0.03 12.7 SBM 12.7-20.7 0.13 8.5 0.24 0.07 Murphy et. ai. RSM 12.7-20.4 0.12 7.3 0.19 0.06 1985 12.8 SBM 19.3-25.0 0.05 4.2 0.02 Gordon & Unsworth 1986 15.2 SBM 18.4-25.1 0.10 4.2 0.06 0.03 Peoples & Gordon 1989 20.0 RSM 13.5-16.0 0.24 14.8 0.32 0.00 Heikkilä et al. 13.5-18.0 0.11 9.6 0.27 -0.09 1989 16.0 " 13.5-16.0 0.28 14.0 0.24 0.40 13.5-18.0 0.13 9.1 0.22 0.22 15.7 SBM 12.2-21.0 0.19 8.1 0.14 0.07 Cody et ai. 1990 18.7 SBM.FM 17.5-22.9 0.13 7.4 0.13 0.08 Small & Gordon 1990 14.2 RSM 12.7-16.2 0.23 15.0 0.32 0.06 Huhtanen et ai. 1991 15.5 RSM 13.3-17.1 0.52 17.8 0.03 -0.10 Huhtanen 1992 14.2-17.8 0.37 22.4 0.22 0.08 (unpublished) 15.8-19.6 0.35 14.7 0.10 0.23 16.5-20.2 0.36 15.1 0.14 0.10 FM = fish meal, GNC= groundnut cubes, RSM =rapeseed meal, SBM =soybean meal; 11 total DM intake 406 Agric. Sei. Finl. 1 (1992) Table 24. Responses of intake when increasing rape- seed meal or crude protein content in the concentrateor diet (n=24). Mean s.d. Min. Max. per kg increase of RSM DM Forage 0.15 0.43 -0.58 1.45 Total intake 0.34 0.40 -0.35 1.59 per 10 g increase of concentrateCP (in DM) Forage 0.05 0.13 -0.18 0.40 Total intake 0.11 0.12 -0.111 0.44 per 10 g increase of diet CP (in DM) Forage 0.11 0.32 -0.45 0.96 Total intake 0.27 0.30 -0.27 1.06 production parameters were calculated (Table 25). The non-linear regressions are shown in Table 26 and Figures 1-4. The determinationcoefficients for nonlinear equations were calculated from the reg- ressions between the actual and regressed variable Y’s (Table 26). The effect of RSM on the milk and protein yield deviate only slightly from the linearity, as can be seen from the determination of coefficients in Tables 25 and 26. This is due to the maximum quantities ofRSM having been reasonable, the lar- gest being only 1.7kg DM per day. The reactions in terms of milk yield to the intake of increased quant- ities of RSM can be calculated from the exponen- Table 25. Linear regressions estimating the response in milk and protein yield and content: Y= a + b |X 1 (X, = change in intake or content of the diet). Dependent Independent a b R 2 S.E. Significance variable Y variable X (P-value) Amilk yield (kg/d) ARSM (kg DM/d) 0.168 0.77 22.4 0.56 0.020 AECM(kg/d) ARSM " 0.235 0.70 23.5 0.49 0.016 Aprotein yield (g/d) ARSM " 6.75 29.2 31.1 17.1 0.005 Amilk yield (kg/d) ACPc (g/kg DM) -0.01 0.28 33.3 0.52 0.003 AECM yield (kg/d) ACPc " 0.32 0.19 19.7 0.51 0.030 Aprotein yield (g/d) ACPc " 11.3 7.9 24.5 17.9 0.014 Amilk yield (kg/d) ACPd " -0.06 0.74 33.7 0.52 0.003 AECM yield (kg/d) ACPd " 0.33 0.48 18.1 0.51 0.038 Aprotein yield (g/d) ACP d " 11.9 19.3 22.0 18.2 0.021 Aprotein yield (g/d) ACPd (gCP/d) 7.0 0.108 41.0 15.8 (k =0.08) 73.2 82.8 60.3 88.1 65.3 48.2 57.1 40.8 EPDi)2) (k = 0.08) 73.2 82.8 50.8 68.6 55.9 38.0 50.8 32.5 EPDi)2) (ic =o.03) 84.5 92.4 73.2 81.4 73.9 61.6 73.2 59.5 AATI (system; k=0.08) PBVI 73 72 11l 71 130 23 196 172 183 242 223 146-34 55 -47 123 AAT3 (k=0.08) PBV3 73 72 120 90 155 199 205 271 -4 159 84 192 104-34 55 -60 AAT7 (k=0.08) PBV7 76 73 123 95 167 207 219 279 169 90-40 52 -66 -13 140 72 AATI2 (k=0.03) PBVI2 72 65 104 83 122 4 204 149 147 192 270 213-31 68 -39 153 AATI7 (k=0.03) PBVI7 103 94 145 116 147 20 -106 -52 161 173 175 218 224 170-85 113 AAT 18 (k=0.03) PBVIB 93 138 162 169 209 -4 183 139 243 193 74 85 103 -28 43 -29 " EPD with microbial-N correction for roughage (Michlet-Doreau & Ould-Bah 1989) 2> Particle loss correction (Weisbjerg et al. 1990) Explanations for the AAT values (see also Table 32): - Efficiency of microbial protein synthesis = 20 g AAN/kg DCHO (AATI, AAT3, AAT7, AATI2), 30 g/kg (AATI7), (185-1.31 *UDP%)*DOM (AAT 18); - Particle loss correction for concentrates(AAT3, AAT7, AATI2, AATI7, AATIB) - DCHO correction (AAT7, AATI2, AAT 17) 415 Agric. Sd. Finl. 1 (1992) untreated RSM. AAT value decreased by 28 % (Table 31). Higher values for outflow rate were measured by using mortanded straw marker or pro- tein supplements, which produce more rumen in- digestible particles. Lower values were measured for NDF, which presented the outflow of protein better (Tamminga et al. 1989). Using corrected digestible carbohydrate values obtained by calculating the microbial protein syn- thesis had only minor effect on the AAT-values of the feeds. In silage the AAT-value remained unchanged, when increased energy for microbes from rumen degradable protein was balanced with the loss of energy in fermentation acids. Table 31 also contains AAT-values calculated with a higher valuefor the efficiency ofmicrobial protein synthe- sis (30 g AA-N/kg DCHO), or calculated with the formula devised by Voigt and Fiatkowski (1991). 5.6.3. Effect ofcorrections ofAA T values on the utilization offeedprotein Average protein utilization and its coefficient of variation was calculated after the above mentioned corrections or adjustments to AAT-values were made (Tables 32 and 34). In Table 33 milk protein yield has been regressed against the protein and energy intake. Energy intake (ME) is corrected by eliminating the effect of protein intake. The correction for microbial-N contamination with the k-value of 0.08 gave roughage similar AAT values to those obtained with the k-value of 0.03 without microbial-N correction. The variation in protein utilization also stayed at the same level (Table 32). Both particle loss correction and that of DCHO reduced the variation in utilizationparameters. The Table 32. Mean values and variation coefficients for utilization of feed protein using different correction for AAT (yij = p + experiment + n=34). Feed Assumptions for calculating the AAT Diet Utilization of Protein con- EPD-% feed protein version rateprotein k-value TD.UDP Micr.N Particle DCHO- ofrough- corr. loss corr. age2) corr. g/kg ECM Mean CV4) Mean CV corr. DCP 0.432 8.03 72.0 7.80 AATI" 0.08 0.82 AAT2 0,03 75.2 0.763 4.15 40.4 4.45 75.7 0.767 4.24 40.1 4.54 + AAT3 0.08 AAT4 0.03 + t + 70,1 0.693 3.57 44.4 3.87 70.5 0.697 3.67 44.1 3.95 AATS 0.08 AAT6 0.03 + + 75.20.716 3.9143.0 4.24 + 75.70.719 3.3542.7 4.27 AAT7 0.08 AATB 0.03 + + + 70.20.659 3.32 46,7 3.72 + + 70.5 0,661 3.5746.6 4.03 AAT9 0.08 TU=.O7 3) + AATIO 0.08 “ + 75.20.809 5.1038.2 5.30 + 70.20.678 4.1445.5 4.59i 11 AATI is assumed to be according to the AAT-PBV system; 2) for concentrate feeds k = 0.08; 31 TD.UDP = (UDP - TU)/UDP; 4) CV =coefficient of variation (%) Corrections for AAT: Microbial-N correction for roughage (Michalet-Doreau & Ould-Bah 1989);particle loss correction for concentrate feeds (Weisbjerg et al. 1990); DCHO-corr. for all feeds: DCHO + 0.50 * RDP - 0.75 * lactic acid - VFA, (Nousiainen 1992). 416 Agnc. Sei. Fint. 1(1992) Table 33. Standard error and coefficient ofvariation ofmilk protein yield: (protein yield). = a + experiment +b,*(protein intake). + b2 *(corrected ME intake). +e.. R 2Protein intake S.E. CV of estimate DCP 16.4 2.2491.6 AATI AAT2 AAT3 AAT4 AATS AAT6 AAT7 AATB AAT9 AAT 10 82.8 18.5 2.54 81.8 18.8 2.58 84.9 18.3 2.51 84.3 18.6 2.55 85.1 16.8 2.35 84.1 17.0 2.33 2.2587.4 16.4 16.9 2.3286.5 82.2 19.4 2.65 86,7 17.7 2.43 effect of changing the value of TD.UDP was unexpected. The constant value of 0.82 resulted in lower variations than with the estimated value of the equation TD.UDP = (UDP-TU)ZUDP, where TU had the fixed value of 0.07 (Table 32). The value of TU should probably be determined for dif- ferent feeds as Hvelplund et al. (1992) have pro- posed, otherwise a constant value for TD.UDP is preferable. The variation was further reduced as the value of protein synthesis efficiency increased, in addition to all the corrections (Table 34). The rumen out- flow rate of 0.03 was used for all feeds. This rate is higher or close to the outflow rate ofNDF or rough- age particles reported in many studies (Mäkelä 1956,Setälä 1983,Tamminga et al. 1989, Huhta- nen and Khalili 1991). The efficiency of microbial protein synthesis per energy unit varies greatly (Hvelplund and Mad- sen 1985), especially by rising at a directratio with increased feeding levels (Robinson et al. 1985, Sniffen et al. 1987). The production level in the present data was moderate (22.5 kg milk/day, intake 16.1 kg DM/day). The proportion of AAT, calculated from the microbial mass, increased from 79% to 83% oftotal AAT intake when the value of efficiency was changed from 20 to 30 g microbial amino-N per kg DCHO (with k = 0.03). This may be too high a value for protein of microbial origin, although in some studies involving dairy cows with high milk yields, the proportion has been 70 to 80 per cent (Klusmeyer et al. 1990, Ferlay et al. 1992). All the same, the variation in protein utiliza- tion decreased. Voigt and Piatkowski (1990, 1991) found a Table 34. Mean values and variation coefficients for utilization of feed protein using different efficiencies ofmicrobial protein synthesis (y. = p + experiment +e.; n=34). Feed Efficiency of k-value AAT Utilization g protein/kg ECM protein micr. protein (all intake of feed protein Mean CV synthesis, g feeds) (g/d) Mean CV amino-N/kg DCHO AAT7 20 0.08 1311 0.659 3.32 46.7 3.72 AATU 18 0.03 1190 0.869 3.29 35.4 4.07 AATI2 20 0.03 1289 0.776 2.94 39.7 3.62 AATI3 22 0.03 1388 0.701 2.70 43.9 3.29 AATI4 24 0.03 1487 0.639 2.51 48.1 3.03 AATIS 26 0.03 1585 0.588 2.38 52.3 2.84 AATI6 28 0.03 1684 0.544 2.28 56.6 2.69 AATI7 30 0.03 1783 0.506 2.22 60.8 2.58 AATIB" 0.03 1448 0.663 2.05 46.3 2.43 Corrections: microbial-N for roughage, particle loss for concentrate feeds, DCHO for all feeds 11 Efficiency ofmicrobial protein synthesis: (185-I.3I*UDP%)*DOM (Voigt and Piatkowski 1991) 417 Agric. Sei. Finl. 1 (1992) negative correlation between the proportion of UDP and the efficiency of microbial protein syn- thesis in the rumen. This may be caused by a decrease in the energy released in the rumen, or a decrease in the availability of amino acids and pep- tides to microbes. In the present study the dietary intake ofAAT was calculated using the formula for microbial protein synthetized (MPS) ofVoigt and Piatkowski (1991): MPS = (185-1.31 * UDP%) * DOM where DOM refers to digestible organic matter (kg). The proportion of amino acid nitrogen was taken as 0.75 and digestibility as 0.85. For UPD the respective values were taken from the Nordic system. Variation coefficients for protein (AATIB) utilization are shown in Table 34. The residual variance of the utilization of protein was significantly different (P<0.01) between AATI (Nordic system) and AATIB (Rostock). However, when AAT intake was calculated using individual AAT 18 values for feeds, the variations in the Nordic and Rostock systems differed less. This may indicate that the protein values of the feeds are not quite additive, and a more correct pro- tein intake value can be calculated from the DOM and UDP values of the total diet. This would be a disadvantage to practical diet formulation. The Rostock method is less sensitive to varia- tions of feed protein degradation. The results of some experiments in the present study agree with that finding, e.g., when reducing the protein degra- dation of RSM had, on average, no effect on milk or protein yield. With the Nordic method the AAT value of hay was equal or even higher than that of grass silage, regardless of the higher digestibility of organic matter and the crude protein content of silage. The Rostock method gave higher protein values for silage than for hay. The values ofprotein feeds were also higher than those obtained with the Nordic method (Table 31). The AAT system gave carbohydrate concentrates very high values compared to DCP values. The AAT value of barley was 40 to 50 per cent higher than that of good quality grass silage, and the value of hay was higher than that of silage cut earlier. Jaakkola and Huhtanen (1992) have showed that the flow of non-ammonium nitrogen (NAN) remained almost equal when the proportion of con- centrate (barley plus RSM) increased from 25 to 75% in the diet in terms of dry matter. In that study forage consisted either of direct cut silage or hay cut at the same maturity. Calculated on the basis of AAT, the AAT increase should have been 37% according to the concentrate. When the microbial AAT flow was calculated according to Voigt and Piatkowski (1991), the increase was almost the same as measured, i.e. 3 to 4 per cent with an increased proportion of the concentrate. The results of the present study (experiment 4), where silage and hay cut at same maturity were compared, no difference between the two kinds of roughage in terms of milk yield was observed, and the calculated AAT consumption per kg ECM was 3 grams higher on a diet of hay than on silage. A shortage of amino acids was evident, as replacing some of the concentrate with RSM increased the milk production on both diets. The fermentation characteristics of silage affect microbial protein synthesis. Good fermentation properties increase microbial protein synthesis, compensating for the greater amounts of bypass protein of dried forage (Jaakkola et al. 1991). In experiment 6 (Heikkilä et al., unpublished) silage cut at two different growth stages were compared. Calculated AAT values were equal for both lots of silage (EPD values were determined), yet milk production was 7.7% higher, and ME intake was 5.5% higher with the earlier lot of silage added to the diet than with the later one. Milk yield correlated much higher with CP or DCP concentra- tions than AAT concentrations in the diet. There are considerable advantages in using the new protein systems compared to DCP in that the former can be developed far easier. Although these present studies show that on the diets commonly used in Finland, i.e. high quality grass silage and hay with grain-based concentrate, there are some problems associated with the AAT-system, but they can be overcome applying new knowledge. 418 Agric. Sei. Finl. 1 (1992) However, more research is needed to develop a more precise method of estimating the microbial protein synthesis in therumen, which appears to be the most important protein source for dairy cows. 6. Conclusions 6.1. Glucosinolate content In the course of the study the types of RSM were changed from single zero to double zero rapeseed varieties, whereby their glucosinolate contents were reduced from 40-50 to 15 pmoles per gram of defatted meal. Heat-moisture treatment further reduced the glucosinolate content by half. The glu- cosinolate content of the latest Finnish turnip rape varieties is less than 10 pinoles per gram of defat- ted meal. 6.2. Effect of rapeseed meal on milk yield The inclusion of RSM in the concentrate to supple- ment the staple diet of grass silage of dairy cows (silage ad libitum) increased the average milk yield by 0.7 kg ECM per kg RSM dry matter. The average protein yield increased by 27 grams per kg RSM dry matter. These increases were statistically significant. Milk yield increased when the ratio of RSM in the concentrate was raised to between 12% and 16 %. Increasing the ratio of RSM further to 24% had a minor effect on milk yield. Although the additional RSM seemed to reduce the fat content significantly in one trial, a similar effect was not observed in the other trials. The effect on protein content of the milk was not significant. The effect RSM had on milk yield and quality was attributed to the increase in energy supply and the specific protein effects. RSM and SBM were of equal value when 12% or 24% of RSM in the concentrate were substituted with SBM with comparable protein contents. 6.3. Protein protection Protecting the RSM protein by heat-moisture treat- ment reduced the effective rumen degradation (EPD) by 6 to 20 %-units. In experiment 2, where EPD of RSM was reduced 20 %-units by the treat- ment, milk yield was improved significantly by heat-moisture treatment (21.9 vs. 23.9 kg milk/d). In two other experiments, where EPD of RSM was reduced 6 or 17 %-units, there was no effect on milk yield. In experiment 1 EPD of heated RSM was 39%-units lower than that ofuntreated with no positive effect on milk yield. In that experiment heat treatment seemed to be too severe, judging by thereduced quantity of available lysine. One expla- nation for the different effect of treatment of RSM could be the level of glucosinolates. In experiment 2 the original level of glucosinolates in RSM was high, and was reduced by treatment. In the experi- ments 3 and 5 the glucosinolate content in RSM was lower, and presumably further reduction by the heat treatment did not give any advantage. 6.4. Goitrin content of milk Changing from single zero to doublezero varieties of RSM reduced the goitrin content of milk. Heat- moisture treating the RSM resulted in a further notable reduction in its goitrin content. For cows fed on heat-moisture treated RSM made of the “Kulta” variety, with a glucosinolate content of 2.5 pmoles per gram defatted meal, the goitrin content was only 3.5 to 6.4 pgrams per litre. Such a low goitrin content of milk should not cause any risk of thyroid problems in people. 6.5. DCP and AAT The applicable parameters for the utilization DCP and AAT were calculated from the average data of the feeding trials involving 34 groups ofcows. The observations related to trials where a staple diet of silage was supplemented with concentrate of varying protein contents. In relating protein yield to the energy and protein supply, DCP was better than uncorrected AAT, whereas the protein utilization varied less with the AAT system than with DCP. The variation in protein utilization was further reduced when AAT was corrected in such a way 419 Agric. Sei. Finl. 1 (1992) that the rumen-degradability of protein increased, or the microbial synthesis became more efficient, or both. It would indicate that as far as cows with high milk yield are concerned, the AAT system exaggerates the role of rumen undegradable feed protein in the supply of absorbable protein. As a result, hay was given a better protein value than silage, and the calculated utilization of AAT was poorer with a dietofhay than with one of silage. As a consequence, the current calculation method leads to AAT variable requirements ofAAT accord- ing to dietary variations. Using the method of Voigt and Piatkowski (1991) in estimating MPS, which gave the lowest coefficient of variation, the AAT values of silage, oat and rapeseed and soybean meals were increased, whereas AAT value of barley was reduced compared to AAT values of the system. These changes are supported by the present results. The present feeding trial data is limited to predo- minantly silage-based feeding. The results of making corrections to AAT systems indicate trends only, and more basic study is needed especially conseming the efficiency of microbialprotein synt- hesis with high producing dairy cows. The AAT-PBV system, like the other new sys- tems, has many advantages compared to DCP, which is at the end of its development. With the new protein systems there is more opportunity to fiillfill the requirements of diminish nitrogen out- put into the environment, and it is possible to derive new knowledge concerning the protein metabolism in the ruminant. References Ackman, R.G. 1990. Canola fatty acids - an ideal mixture for health, nutrition, and food use. In: Shadidi, F. (ed.). Canola and rapeseed: production, chemistry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 81-98. Ahlin, K.-Å., Emanuelson, M., Edqvist, L.-E., Larsson, K. & Wiktorsson, H. 1985. Rapeseed products as feed for dairy cows. Preliminary results from longterm study. In: Sorensen, H. (ed.). Advances in the production and utilization of cruciferous crops. In series: World Crops: Production, utilization, description. Vol. 11: 222-229. Ahrar, M. & Schingoethe, D.J. 1979. Heat-treated soybean meal as a protein supplement for lactating cows. J. Dairy Sci. 62: 932-940. Alderman, G. 1987. Comparison of rations calculated in the different systems. In: Jarrige, R. & Alderman, G. (eds.). Feed evaluation and protein requirement systems for ruminants. Proc. CEC seminar, Brussels. CEC L- -2985 Luxembourg, p. 283-297. Andersen, P.E. 1985. Double low rapeseed meal in diets to dairy cows. In: Sorensen, H. (ed.), Advances in the pro- duction and utilization of cruciferous crops. In series: World Crops: Production, utilization, description, Vol. 11:218-221. Anjou,K., Lönnerdal, 8., Uppström, B. & Åman, P. 1977. Composition of seeds from some Brassica cultivars. Swedish J. Agric. Res. 7: 169-178. Anke, M., Schwarz, S., Hennig, A., Groppel, 8., Grön, M., Zenker, G. & Glös, S. 1980. Der Einfluss zusätzlicher Zink- und Jodgaben auf rapsexraktionsbe- dingte Schäden beim Schwein. Mb. Veterinärmed. 35: 90-94. Appelqvist,L.-Å. 1972. Chemical constituents ofrapeseed. In: Appelqvist, L.-Å. & Ohlson, R. (eds.). Rapeseed. Cultivation, composition, processing and utilization. Elsevier Pubi. Co. Amsterdam 1972. p. 123-173. Appelqvist,L.A. & Josefsson, E. 1967. Method for quanti- tative determination of isothiocyanates and oxazolidinet- hiones in digests ofseed meals ofrape and turnip rape. J. Sci. Food Agric. 18:510-519. ARC 1980, The nutrient requirements ofruminant livestock. Technical review by an Agricultural Research Council working party. Commonwealth Agricultural Bureaux, Famham Royal, UK. 351 p. ARC 1984. The nutrient requirements ofruminant livestock. Supplement No I. Technical review by an Agricultural Research Council working party. Commonwealth Agri- cultural Bureaux, Famham Royal, UK. 45 p. Arstila, A., Krusius, F.-E. & Peltola, P. 1969. Studies on the transfer of thio-oxazolidone-type goitrogens into cow’s milk in goitre endemic districts of Finland and in experimental conditions. Acta Endocr. 60: 712-718. Ashes, J.R., Mangan, J.L. & Sidhu, G.S. 1984. Nutritional availability of amino acids from protein cross-linked to protect against degradation in the nimen. Br. J. Nutr. 52: 239-247. Asplund,M. & McElroy, L.W. 1961.Effects ofhigh levels of rapeseed oil meal in dairy cow rations on milk pro- 420 Agric. Sei. Fint. 1 (1992) duction, J. Dairy Sei. 44: 2338 (abstr.). Bachmann, M., Theus, R., LOthy, J. & Schlatter, C. 1985. Vorkommen von goitrogenen Stoffen in Milch. 1. Mitteilung: Übergang von Goitrin in die Milch von Kiihen hei Verfutterung von Rapsextraktionsschrot. Zeitschr. Lebensm. Unters. Forsch.lBl: 375-378. Barker, S.B. & Summerson, W.H. 1941. The colorimetric determination of lactic acid in biological materials. J. Biol. Chem. 138: 537-554. Bell, J.M. 1984. Nutrients and toxicants in rapeseed meal: a review. J. Anim. Sci. 58: 996-1010. & Jeffers, H.F. 1976. Variability in the chemical compo- sition ofrapeseed meal. Can. J. Anim. Sci. 56; 269-273. & Shires, A. 1982. Composition and digestibility by pigs ofhull fractions from rapeseed cultivars with yellow or brown seed coats. Can. J. Anim. Sci. 62: 557-565. Belzile, R.J., Bellk, J.M. & Wetter, L.R. 1963. Growth depressing factors in rapeseed oil meal. V. The effects of myrosinase activity on the toxicity of the meal. Can. J. Anim. Sci. 43: 169-173. Ben-Ghedalia, D., McMeniman, N.P. & Armstrong, D.G. 1978. The effect ofpartially replacing urea nitrogen with protein N on N capture in the rumen ofsheep fed a puri- fied diet. Br. J. Nutr. 39: 37-44. Bengtsson,L., Hofsten, von, A. & Lööf, B. 1972. Botany of rapeseed. In: Appelqvist, L.-A. & Ohlson, R. (eds.). Rapeseed. Cultivation, composition, processing and util- ization. Elsevier Pubi. Co. Amsterdam, p. 36-48. Benns, G., L’Abbé, M.R. & Lawrence, J.F. 1979. High- pressure liquid chromatography detection of the antithy- roid compound 5-vinyloxazolidine-2-thione in milk. J. Agric. Food Chem. 27: 426-428. Bergner, H. & Schmidt, W. 1972. Untersuchungen zur Wirkung von Rapsextraktionsschrot auf das KUkenwachstum. 3. Mitteilung. Die Wirkung von ges- taffelten Rapsextraktionsschrotgaben auf den 131- Ein- bau in die Kilkenschilddruse. Arch. Tieremährg. 22: 41- 47. Bertilsson, J. 1991 . Different protein levels to dairy cows - effect of reduced protein degradability and lower nitro- gen supply on animal performance. 6th Int. Symp. Pro- tein Metabolism and Nutrition. Heming, Denmark, p. 309-311. Bertram, H.J., Brauer, D., Lembke, K.G., Hatje, G., Hen- kel, H., Kley, G., Lange, D., Lennerts, L. & Röbbe- len, G. 1986. Raps auf neuen Wegen: 00-Sorten in ZOchtung, Anbau und Verwendung. Verlag Th. Mann. Gelsenkirchen-Buer, 98 p. Bhatty, R.S., McKenzie, S.L. & Finlayson, A.J. 1968. The proteins ofrapeseed (Brassica napus L.) soluble in salt solutions. Can. J. Biochem. 46: 1191-1197. Bickel, H. & Landis, J. 1987. Present situation ofprotein evaluation forruminants in Switzerland. In: Jarrige, R. & Alderman, G. (eds.). Feed evaluation and protein requi- rement systems forruminants. Proc. CEC seminar, Brus- sels. CEC L-2985 Luxembourg, p. 41-45. Björkqvist, B. & Hase, A. 1988. Separation and determi- nation of intact glucosinolates in rapeseed by high-per- formance liquid chromatography. J. Chromatogr. 435: 501-507. Broderick, G.A. 1987. Determination of protein degrada- tion rates using a rumen in vitro system containing inhi- bitors ofmicrobial nitrogen metabolism. Br. J. Nutr. 58: 463-475. & Clayton, M.K. 1992.Rumen protein degradation rates estimated by non-linear regression analysis of Michae- lis- Menten in vitro data. Br. J. Nutr. 67; 27-42. BONQER, H., Fissmer, E, & Reising, F. 1940. FUtterungsversuche mit Rapsriickständen an Milchkuhen. Zeitschr. Tieremähr. Futtermittelkde. 4: 183-200. Burgess, P.L. & Nicholson, J.W. 1984. Protein levels in grass silage-based total mixed rations for dairy cows in rnidlactation. Can. J. Anim. Sci. 64: 435-442. Cameron, M.R., Klusmeyer, T.H., Lynch, G.L., Clark, J.H. & Nelson, D.R. 1991. Effects ofurea and starch on rumen fermentation, nutrient passage to the duodenum, and performance ofcows. J. Dairy Sci. 74: 1321-1336. Campbell, L.D. & Slominski, B.A. 1991.Nutritive quality of low-glucosinolate Canola meal for laying hens. Proc. Bht Int. Rapeseed Congr., Saskatoon, Canada. Vol. 2: 442-447. Carr, R. 1989, Processing of oilseed. In: Röbbelen, G.R. et al. (eds.). Oil crops of the world: Their breeding and uti- lization. McGraw-Hill Pubi. Co, New York. p. 226-259. Castle, M.E. & Watson, J.N. 1976. Silage and milk pro- duction. A comparison between barley and groundnut cake as supplements to silage ofhigh digestibility. J. Br. Grassld. Soc. 31: 191-195. & Watson, J.N. 1984. Silage and milk production: a comparison between concentrates containing different amounts of protected protein as supplements for silage ofhigh digestibility. Grass and Forage Sci. 39: 93-99. Chalupa, W.V. 1992. A model to estimate rumen escape protein and amino acid composition. Proc. Distiller Feed Confer., Cine. Ohio, USA. Vol. 47: 87-97. Chamberlain, D.G. 1987. The silage fermentation in rela- tion to the utilization of nutrients in the rumen. Process Biochemistry, April 1987: 60-63. —, Martin, P.A. & Robertson, S. 1989. Optimizing com- pound feed use in dairy cows with high intakes ofsilage. In: Haresign, W & Cole, D.J.A. (eds.). Recent Advances in Animal Nutrition. Butterworths. p. 175-193. Chen, G., Sniffen, C.J. & Russell, J.B. 1987. Concentra- tion and estimated flow of peptides from the rumen of dairy cattle: effects ofprotein quantity, protein solubil- ity, and feeding frequency. J. Dairy Sci. 70: 983-992. Clandinin, D.R., Robblee, A.R., Slinger, S.J. & Bell, J.M. 1981.Composition ofcanola meal. In: Canola meal for livestock and poultry. Pubi. 59: 8-11. Canola Council of Canada, Winnipeg. Classen, H.L., Bell, J.M. & Clark, W.D. 1991. Nutritio- 421 Agric. Sei. Finl. 1 (1992) nai value of very low glucosinolate Canola meal for broiler chickens. Proc. Bth Int. Rapeseed Congr., Saska- toon, Canada. Vol 2: 390-395. Cody, R.F., Murphy, J.J. & Morgan, DJ. 1990. Effect of supplementary crude protein level and degradability in grass silage-based diets on performance of dairy cows, and digestibility and abomasal nitrogen flow in sheep. Anim. Prod. 51: 235-244. Colucci, P.E., Chase, L.E. & Van Soest, P.J. 1982. Feed intake, apparent diet digestibility, and rate ofparticulate passage in dairy cattle. J. Dairy Sci. 65: 1445-1456. —, Macleod, G.K., Grovum, W.L., McMillan, I. & Barney, D.J. 1990. Digesta kinetics in sheep and cattle fed diets with different forage to concentrate ratios at high and low intakes, J. Dairy Sci. 73: 2143-2156. Craig, W.M., Brown, D.R., Broderick, G.A. & Ricker, D.B. 1987. Post-prandial compositional changes of fluid- and particle-associated ruminal microorganisms. J. Anim. Sci. 65: 1042-1048. Davis, A.W. & Flall, W.B. 1969. Cyclic change-over designs. Biometrika 56: 283-293. Demeyer, D. & Van Nevel, C. 1979. Protein fermentation and growth by rumen microbes. Ann. Rech. Vet. 10: 277-279. Downey, R.K. & Röbbelen, G. 1989.Brassica species. In: Röbbelen, G.R. et al. (eds.). Oil crops of the world: Their breeding and utilization. McGraw-Hill Pubi.Co. New York. p. 339-362. Eapen, K.E., Tape, N.W. & Sims, R.P. 1968. New process for the production of better quality rapeseed oil and meal. I. Effect ofheat treatments on enzyme destruction and color ofrapeseed oil. J. Am. Oil Chem. Soc. 45: 194- 196, Elimam, M.E. & ORSKOV, E.R. 1984. Factors affecting the outflow of protein supplements from the rumen. 1. Fee- ding level. Anim. Prod. 38: 45-51. Emanuelson, M. 1989. Rapeseed products of double low cultivars to dairy cows: Effects of long-term feeding and studies on rumen metabolism. Swedish Univ. Agric. Sci. Dept. Anim. Nutr. Managem. Report 189, 182p. —, Ahlin, K.-Å. & Wiktorson, H. 1991. Rapeseed pro- ducts of 00-cultivarsto dairy cows. Effects of long-term feeding on animal performances. Proc. Bth Int. Rapeseed Cong., Saskatoon, Canada. Vol. 2: 430-435. Emery, R.S. 1978. Feeding for increased milk protein. J. Dairy Sci. 61: 825-828. Etienne, M., Dourmad, J.Y., Obidzinski, W., Evrard, J. & Vachot, C. 1991. Effects of low glucosinolate rapeseed meal in sow pregnancy diet. Proc. Bth Int. Rapeseed Congr., Saskatoon, Canada. Vol. 2: 376-381. Ettala, E. 1976. Factors affecting the composition ofmilk. I. Effect of energy and protein levels in grass silage- and pasture-based diets. Ann. Agric. Fenn. 15: 182-195. &Lampila, M. 1974. Konserverat gräs som energi- och proteinkälla for mjölkkor. NJF:s grovfodersymposium, Randers. Husdjur 45: F: 1-11. —, Lampila, M. & Rissanen, H. 1978. Effect of concent- rate feeding level in grass silage-based diets on milk production of dairy cows. Ann. Agric. Perm. 17: 175- 185. —, Takala, M. & Lampila, M. 1974. Typpilannoitustasot lypsylehmien säilörehuruokinnassa. Kehittyvä Maata- lous 18: 51-59. Ettlinger, M.G. & Lundeen, A.J. 1954. The structure of sinigrin and sinalbin; an enzymatic rearrangement. J. Amer. Chem. Soc. 78: 4172. Faldet, M.A. & Sätter, L.D. 1991. Feeding heat-treated full fat soybeans to cows in early lactation. J. Dairy Sci. 74: 3047-3054. FAO 1991. FAO yearbook Production 1990, vol 44. FAO Statistics Series no. 99, Rome, 1991. 379 p. Fenwick, G.R. & Fleaney, R.K. 1983. Glucosinolates and their breakdown products in cruciferous crops, foods and feedingstuffs. Food Chem. 11; 249-271. —, Pearson, A.W., Greenwood, N.M. & Butler, E.G. 1981. Rapeseed meal tannins and egg taint. Anim. Feed Sci. Technol. 6:421-431. Fenwick, R.G. & Hoggan,FI. A. 1976. The tannin content of rapeseed meals. Br. Poult. Sci. 17: 59-62. Ferlay, A., Legay, F., Bauchart, D., Poncet, C. & Doreau, M. 1992. Effect of a supply ofraw or extruded rapeseeds on digestion in dairy cows. J. Anim. Sci. 70: 915-923. Finlayson, J. 1974. The amino acid composition of rape- seed hulls. Can. J. Anim. Sci, 54: 495-496. Firkins, J.L., Berger, L.L., Merchen, N.R., Fahey, G.C. & Mulvaney, R.L. 1987a. Ruminal nitrogen metabolism in steers as affected by feed intake and dietary urea concentration. J. Dairy Sci. 70: 2302-2311. —, Lewis, S.M., Montgomery, L., Berger, L.L., Merchen, N.R. & Fahey, G.C. Jr. 1987b. Effects offeed intake and dietary urea concentration on ruminal dilu- tion rate and efficiency of bacterial growth in steers. J. Dairy Sci. 70: 2312- 2321. Ganev, G,, ORSKOV, E.R. & Smart, R. 1979. The effect of roughage or concentrate feeding and rumen retention time on total degradation of protein in the rumen. J. Agric. Sci. 93: 651-656. Garnsworthy, P.C. 1989. The interaction between dietary fibre level and protein degradability in dairy cows. Anim. Prod. 48: 271-281. Glenn, 8.P., Varga, G.A., Huntington, G.B. & Waldo, D.R. 1989. Duodenal nutrient flow and digestibility in Holstein steers fed formaldehyde- and formic acid-treat- ed alfalfa or orchardgrass silage at two intakes. J. Anim. Sci. 67: 513-528. Gmelin, R. 1969. Glucosinolate (Senfolglucocide). Präp. Pharm. 5: 33-41. Goering, H.K. & Van Soest, P.J. 1970. Forage fibre ana- lysis (Apparatus, reagents, procedures and some applica- tions). USDA Agricultural Handbook, No. 379, Was- hington, DC. 20 p. 422 Agric. Sei. Fin!. 1(1992) Ooh, Y.K., Mueller, M.M., Clandinin, D.R. Robblee, A.R. 1979. The effects of choline and sinapine bisulfate in a laying ration on the incidence of fishy odor in eggs from brown-shelled egg layers. Can. J. Anim. Sci. 59: 545-549. —, Robblee, A.R. & Clandinin, D.R. 1985. Influence of glucosinolates and free oxazolidinethione in a laying diet containing a constantamount of sinapine on the thyroid size and hepatic trimethylamine oxidase activity of brown-egg layers. Can. J. Anim. Sci. 65: 921-927. Gordon, F.J. 1979. The effect of protein content of the supplement for dairy cows with access ad libitum to high digestibility, wilted grass silage. Anim. Prod. 28: 183- 189. 1980. The effect of silage type on the performance of lactating cows and the response to high levels ofprotein in the supplement. Anim. Prod. 30: 29-37. 1984. The effect of level ofconcentrate supplementation given with grass silage during the winter on the total lactation performance of autumn-calving dairy cows. J. Agric. Sci. 102: 163-179. 1987. The influence of the system of silage harvesting and feeding and the use of protected protein on milk production. Grass and Forage Sci. 42: 9-19, & McMurray, C.H. 1979.The optimum level ofprotein in the supplement for dairy cows with access to grass silage. Anim. Prod. 29; 283-291. &Unsworth, E.F. 1986.The effects ofsilage harvesting system and supplementation of silage-based diets by protein and methionine hydroxy analogue on the perform- ance of lactating cows. Grass and Forage Sci.4l: 1-8. Hadjipanayiotou, M., Koumas, A., Georghiades, E. & Hadjidemetriou, D, 1988. Studies on degradation and outflow rate ofprotein supplements in the rumen of dry and lactating Chios ewes and Damascus goats. Anim. Prod. 46: 243-248. Haoemeister, H., LOppino, W. & Kaufmann, W. 1980. Microbial protein synthesis and digestion in the high- yielding dairy cow. In; Haresign, W. (ed.). Recent advances in animal nutrition -1980, Nutr. Conf. Feed Man. 14th. p. 67-84. Harvey, W.R, 1966.Least-square analysis of data with un- equal subclass numbers. ARS 20-8, Agric. Res. Service, U.S. Dept. Agric. 157 p. Hase, A., Johansson, M.-L. & Viljava, T.-R. 1988. Sources oferror in the analysis of glucosinolates by gas liquid chromatography. J. Am. Oil Chem. Soc. 65:647- 651. Heaney, R.K. & Fenwick, G.R. 1980. The analysis of glu- cosinolates in Brassica species using gas chromatog- raphy. Direct determination of the thiocyanate ion pre- cursors, glucobrassicin and neoglucobrassicin. J. Sci. Food Agric. 31:593-599. &Fenwick, G.R. 1985. Brassica vegetables - a major source of glucosinolates in the human diet. In: Sorensen, H. (ed.). Advances in the production and utilization of cruciferous crops. In series: World Crops: Production, utilization, description, Vol. 11: 40-49. Heikkilä, T., Väätäinen, H., Lampila, M. & Toivonen, V. 1989. Säilöntäaineet säilörehun valmistuksessa. Säilön- täaine-ja väkirehukoe. MTTK, KEL/ERA. Toimintaker- tomus jatutkimustuloksia 1989, Jokioinen,Finland, p. 5- 7. Helboe, P., Olsen. O. & Sorensen, H. 1980. Separation of glucosinolates by high performance liquid chromato- graphy. J. Chromatogr. 197: 199-205. Henkel, H. & Mosenthin, R. 1989.Rapssaat und Rapspro- dukte in der Tieremährung. Übersicht. Tieremährg. 17: 139-190. Hill, R. 1979. A review of the toxic effects of rapeseed meals with observations on meal from improved varie- ties. Br. Vet. J, 135: 3-16, 1991. Rapeseed meal in the diets of ruminants. Nutr. Abstr. Rev. (Series B): 139-155. Hobson-Frohock, A., Land, D.G., Griffiths, H.S. & Cur- tis, R.F. 1973. Egg taints: association with trimethyla- mine. Nature 243: 304-305. Holter, J.8., Byrne, J.A. & Schwab, C.G. 1982. Crude protein for high milk production. J. Dairy Sci. 65: 1175- 1188. Huhtanen, P. 1991.The response to replacement of barley with wheat bran and treatment of rapeseed meal in the diets of dairy cows given grass silage ad libitum, Acta Agric. Scand. 41:415- 426. & Khalili, H. 1991.Sucrose supplements in cattle given grass silage-based diet. 3. Rumen pool size and digestion kinetics. Animal Feed Sci. Technol. 33; 275-287. & Khalili, H. 1992.The effect of sucrose supplements on particle-associated carboxymethylcellulase (EC 3.2.1.4) and xylanase (EC 3.2.1.8) activities in cattle given grass-silage-based diet. Br. J. Nutr. 67: 245-255. —, Khalili, H. & Näsi, M. 1991. A comparison ofuntrea- ted and formaldehyde-treated barley distiller’s solubles and rapeseed meal as protein supplements in dairy cows given grass silage ad libitum. J. Agric. Sci. Finl. 63: 455- 463. Huida, L. 1973. Quantitative determination of volatile fatty acids from rumen sample and silage by gas-liquid chro- matography. J. Scient. Agric. Soc. Finl. 45: 483-488. Hvelplund, T. 1985. Digestibility of rumen microbial pro- tein and undegraded dietary protein estimated in the small intestine of sheep and by in sacco procedure. Acta Agric. Scand. Suppl. 25: 132-144. 1986. The influence of diet on nitrogen and amino acid content ofmixed nimen bacteria. Acta Agric. Scand. 36: 325- 331. 1990. The AAT-PBV protein evalution system. Notes from a Nordic meeting concerning the AAT-PBV pro- tein evaluation system, held at Tune Landboskole Den- mark, 3rd-sth December 1990. & Madsen, J. 1985. Amino acid passage to the small intestine in dairy cows compared with estimates of 423 Agric. Sei. Fint. 1 (1992) microbial protein and undegraded dietary protein from analysis on the feed. Acta Agric. Scand. Suppl. 25: 21- 36. & Madsen, J. 1990. A study of the quantitative nitrogen metabolism in the gastro- intestinal tract, and the resul- tant new protein evaluation system for ruminants. The AAT- PBV system. Diss. Inst. Anim. Sci. Royal Vet. Agric. Univ. Copenhagen. 62 p. & Möller,E. 1987. Proteinvurdering af graes og klover ved forste slaet. XVIII. NJF congres, Arhus, 5 p. & Möller, P.D. 1980. Fodringens indflydelse på vom- bakteriernes kemiske saramensaetning. Medd. nro 310, Statens Husdyrbrugsforsog, Copenhagen. & Weisbjerg, M.R., Hvelplund, T. & Andersen, L.S. 1992. Estimation of the true digestibility of rumen undegraded dietary protein in the small intestine of ruminants by the mobile bag technique. Acta Agric. Scand. Sect. A, Anim. Sci. 42: 34-39. Ingalls, J.R. & Sharma, H.R. 1975. Feeding Bronowski, Span and commercial rapeseed meals with or without addition of molasses or flavor in rations of lactating cows. Can. J. Anim. Sci. 55: 721-729. —, Seale, M.E. & McKirdy, J.A. 1968. Effect ofrapeseed meal and urea on ad libitum consumption of grain rations by dairy cows. Can. J. Anim. Sci. 48: 437-442. INRA 1978. Alimentation des ruminants. Institut National de la Recherche Agronomique ofFrance, INRA Publica- tions, Versailles. 597 p. Iwarsson, K. & Nilsson, P.O. 1973. Rape seed meal as a protein supplement for dairy cows. 11. Investigations in rats on the goitrogenic properties ofmilk from cows fed rapeseed meal. Acta Vet. Scand. 14: 595-609. Jaakkola, S. & Huhtanen, P. 1990. Nitrogen digestion and rumen fermentation in cattle given silage or dried grass with three levels of concentrate. Proc. 9th Silage Conf. Faculty of Agric. Univ. Newcastle upon Tyne, UK. p. 116-117. & Huhtanen, P. 1992a. The effect of grass maturity on nitrogen metabolism in the rumen of cattle receiving silage- based diet. Proc. 14th Gen. Meet. European Grassland Federation, Lahti, Finland, p. 588-590. Huhtanen, P. 1992b. The effects of the forage preserva- tion method and the proportion of concentrate on nitro- gen digestion and rumen fermentation in cattle. Grass For. Sci. 1992(in press). —, Huhtanen, P. & Hissa, K. 1991. The effect ofcell wall degrading enzymes or formic acid on fermentation qual- ity and on digestion of grass silage by cattle. Grass For. Sci. 46: 75- 87. Jarl, F. 1951. Utfodringsfbrsök med svenskt rapsmjöl till mjölkkor. Kungl. Lantbrukshögskolan och Statens Lantbruksfbrsök. Statens Husdjursforsök, meddelande nr 45: 1-42. Kaim, M., Neumark, H., Folman, Y. & Kaufmann, W. 1987. The effect of two concentrations of dietaryprotein and of formaldehyde-treated soya-bean meal on the per- formance ofhigh- yielding dairy cows. Anim. Prod. 44: 333-345. Kaufmann, W. 1977. Calculation of the protein require- ments for dairy cows according to measurements of N metabolism. In: Protein metabolism and nutrition. Eur. Assoc. Anim. Prod. Pubi. 22; 126-129. Wageningen 1977. & Luffing, W. 1979.Zum Einfluss von Protected protein und HMM-Ca auf die Leistung von Milchkuhen. Zeitschr. Tierphysiol. Tieremährg. Futtermittelkde. 41: 202-217. Klusmeyer, T.H., McCarthy, R.D. Jr., Clark, J.H. & Nelson, D.R. 1990.Effect of source and amount ofpro- tein on ruminal fermentation and passage ofnutrients to the small intestine of lactating cows. J. Dairy Sci. 73: 3526-3537. Kozlowska, H., Naczk, M., Shahidi, F. & Zadernowski, R. 1990. Phenolic acids and tannins in rapeseed and Canola. In: Shahidi, F. (ed.). Canola and rapeseed: pro- duction, chemistry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 193-210. Kristensen, E.S., Möller, P.D. & Hvelplund, T. 1982. Estimation of the effective protein degradability in the rumen ofcows using the nylon bag technique combined with the outflow rate. Acta Agric. Scand. 32: 123-127. Kristensen, V.F., Munksgaard, L. & Andersen, P.E. 1985. Protein value of grass silage for dairy cows. Acta Agric. Scand. Suppl. 25: 145-154. Krusius, F.-E. & Peltola, P. 1966. The goitrogenic effect ofnaturally occuring L-5-vinyl- and L-5- phenyl-2-thio- oxazolidone in rats. Acta Endocr. 53: 342-352. Laarveld, B. & Christensen, D.A. 1976.Rapeseed meal in complete feeds for dairy cows. J. Dairy Sci. 59: 1929- 1935. Laird, R. Leggate, A.T. & Castle, M.E. 1979. The effect of supplementary protein on the performance of dairy cows offered grass silage ad libitum. Anim. Prod. 29: 151-156. Lindberg, J.E. 1982. Ruminal flow rate of soya-bean meal, rapeseed meal and cottonseed meal in cows fed at main- tenance and at three times maintenance. J. Agric. Sci. 98: 689-691. 1985. Estimation of rumen degradability of feed proteins with the in sacco technique and various in vitro methods: a review. Acta Agric. Scand. Suppl. 25: 64-97. 1988. Influence ofcutting time and N fertilization on the nutritive value of timothy. 2. Estimates ofrumen degra- dability of nitrogenous compounds. Swedish J. Agric. Res. 18: 85-89. Lindell, L. 1976.Rapeseed meal in rations for dairy cows. 2. Comparison of two levels of rapeseed meal. Swedish J. Agric. Res. 6: 65-71. & Knutsson, P.-G. 1976. Rapeseed meal in rations for dairy cows. 1. Comparison of three levels of rapeseed meal. Swedish J. Agric, Res. 6: 55-63. Maa- ja metsätalousministeriö 1975. Öljykasvitoimikunnan 424 Agric. Sei. Finl. 1 (1992) mietintö. Komiteamietintö 1975:63. Helsinki, 109 p. Madsen, J. 1985. The basis for the proposed Nordic protein evaluation system forruminants. The AAT-PBV system. Acta Agric. Scand. Suppl. 25: 9-20. MAFF 1975, Energy allowance and feeding systems for ruminants. Tech. 8u11.33. 79 P. London. Mäkelä, A. 1956. Studies on the question ofbulk in the nut- rition of farm animals with special reference to cattle. Acta Agralia Fennica 85: 1-130. Marangos, A. & Hill,R. 1974. The hydrolysis and absorp- tion of thioglucosides ofrapeseed meal. Proc. Nutr. Soc. 33: 90A (abstr). March, B.E. & MacMillan, C. 1978. Trimethylamine pro- duction in the caeca and small intestine as a cause of fishy taints in eggs. Poultry Sci. 58: 93-98. Mayne, C.S. & Gordon, F.J. 1984. The effect of type of concentrate and level of concentrate feeding on milk production. Anim. Prod. 39: 65-76. & Gordon, F.J. 1985. The effect of concentrate-to- forage ratio on the milk-yield response to supplementary protein. Anim. Prod. 41: 269-279. McCarthy, R.D., Klusmeyer, T.H., Vicini, J.L., Clark, J.H. & Nelson, D.R. 1989. Effects of source of protein and carbohydrate on ruminal fermentation and passage of nutrients to the small intestine of lactating cows. J. Dairy Sci. 72: 2002-2016. McCullough, H. 1967. Determination of ammonia in whole blood by a direct colorimetric method. Clin. Chem. Acta. 17: 297-304. McDonald, I. 1981. A revised model for the estimation of protein degradability in the rumen. J. Agric. Sci. 96: 251-252. McGregor, D.l. 1990. Application of near infrared to the analysis of oil, protein, chlorophyll, and glucosinolates in Canola/rapeseed. In: Shadidi, F. (ed.). Canola and rapeseed: production, chemistry, nutrition and proces- sing technology. Van Nostrand Reinhold, New York. p. 221-231. —, Mullin, W.J. & Fenwick, G.R. 1983. Review ofanaly- sis of glucosinolates: analytical methodology for deter- mining glucosinolate composition and content. J. Assoc. Offic. Anal, Chem. 66; 825-849. Mehrez, A.Z. & ORSKOV, E.R. 1977. A study of the artifi- cial fibre bag technique for determining the digestibility of feeds in the rumen. J. Agric. Sci. 88: 645-650. Menzel, E. 1983. Untersuchungen iiber die schildrilsenvergrössemde (strumigene) Wirkung raps- haltiger Rationen in Abhängigkeit von der Rapssorte, dem Glucosinolatgehalt und Futterungszusätzen. Diss. Christian- Albrechts-Univ. Kiel. 143 p. Merry, R.J. & McAllan, A.B. 1983. A comparison of the chemical composition ofmixed bacteria harvested from the liquid and solid fractions of rumen digesta. Br. J. Nutr. 50: 701-709. Meyer, J.H. & Mackie, R.I. 1986. Microbiological evalua- tion of the intraruminal in sacculus digestion technique. Appi. Environm. Microbiol, 51: 622-629. Michalet-Doreau, B. & Ould-Bah, M.Y, 1989. Estima- tion of the extent ofbacterial contamination in bag resi- dues and its influence on in sacco measurements of forage nitrogen degradation in rumen. XVI International Grassland Congress, Nice, France, 1989: 909-910. Mdller, P., Plöger, A. & Sorensen, H. 1985.Quantitative analysis of total glucosinolate content in concentrated extracts from double low rapeseed by the Pd- glucosino- late complex method. In: Sorensen, H. (ed.). Advances in the production and utilization of cruciferous crops. In series: World Crops: Production, utilization, description, Vol. 11: 97-110. Morgan, D.J. 1985. The effect of formalin-treated soya bean meal upon the performance of lactating cows. Anim. Prod, 41: 33-42. Mueller, M.M., Ryl, E.8., Fenton, T. & Clandinin, D.R. 1978. Cultivar and growing location differences on the sinapine content ofrapeseed. Can. J. Anim. Sci. 58: 579- 583. Murphy, J.J., Gleeson, P.A. & Morgan, D.J. 1985. Effect ofprotein source in the concentrate on the performance ofcows offered grass silage ad-libitum. lr. J. Agric. Res. 24: 151-159. Murphy, D.J., Cummins, 1.,Edwards, E.W., Hills, M.J. & Li, M. 1991. Oleosins - a new class of abundant seed protein in rapeseed. Proc. Bth Int. Rapeseed Congr., Sas- katoon, Canada. Vol 3: 922-927. Naczk, M. Shahidi, F. 1990. Carbohydrates of canola and rapeseed. In: Shadidi, F. (ed.). Canola and rapeseed: pro- duction, chemistry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 211-220. Näsi, M. 1991. Digestibility and protein utilization respon- ses ofsoyabean and rape seed meal to physical and enzy- matic treatments in diets for growing pigs. J. Agric. Sci. Finl. 63: 465-474. & Huida, L. 1982. Digestibility of amino acids in pig diets containing Eurolysine bacterial protein or Pekilo protein, with special reference to a gas chromatographic method used in amino acid determination. J. Agric. Soc.Finl. 54; 279-285. & Siljander-Rasi, H. 1991. Effects of thermal proces- sing on digestibility and protein utilization of rapeseed meal of medium and low glucosinolate type in diets for growing pigs. J. Agric, Sci. Finl. 63: 475-482. —, Alaviuhkola, T. & Suomi, K. 1985. Rapeseed meal of low- and high-glucosinolate type fed to growing-finis- hing pigs, J. Agric. Sci. Finl. 57: 263-269. National board of agriculture 1991a. Monthly review of agricultural statistics 10/1991. Helsinki. 31 p. National board of agriculture. 1991b. Yearbook of farm sta- tistics 1990. Official statistics of Finland, Agriculture and forestry 1991:2. Helsinki. 254 p. NCR 1985. Ruminant nitrogen usage. National Research Council (U.S.) Subcommittee on Nitrogen Usage in Ruminants. US National Academy of Sciences, Was- 425 Agric. Sei. Fint. 1 (1992) hington DC. 138 p. NKJ 1985. Protein evaluation forruminants. Proc. NKJ-NJF Seminar no. 72, Copenhagen. Acta Agric. Scand. Suppl. 25. 222 p. Nordfeldt, S. 1958. Smältbarhetsförsök och utfodringsfbr- sök med rapsmjöl och senapsmjöl med låg resp. hög fett- halt till mjölkkor. Kungl. Lantbrukshögskolan och Sta- tens Lantbruksforsök. Statens Husdjursforsök, medde- lande nr 66: 1-38. Norton, G. 1989. Nature and biosynthesis of storage pro- teins. In: Röbbelen, G.R. et al. (eds.). Oil crops of the world: Their breeding and utilization. McGraw-Flill Publ.Co. New York. p. 165-191. Nousiainen, J. 1992. Säilörehun ja heinän AAT- ja PBV- arvojen laskeminen NIR- ja titrausanalytiikan avulla. Suomen Maatal. Tiet. Seuran Tiedote 16: 143-150. Nwokolo, E.N. & Bragg, D.B. 1977. Influence of phytic acid and crude fibre on the availability ofminerals from four protein supplements in growing chicks. Can. J. Anim. Sci. 57; 475-477. Oldham, J.D. 1987. Towards a European standard method for assessing protein degradability. Circular in con- nection with EAAP-EEC work shop. Brussels. 4 p. 1987a. Efficiencies ofamino acid utilisation. In: Jarrige, R. & Alderman, G. (eds.). Feed evaluation and protein requirement systems for ruminants. Proc. CEC seminar, Brussels, CEC L-2985 Luxembourg, p. 171- 186. —1987 b. Testing and implementing the modem systems: UK. In: Jarrige, R. & Alderman, G. (eds.). Feed evalua- tion and protein requirement s ystems for mminants. Proc. CEC seminar, Brussels, CEC L-2985 Luxem- bourg. p. 269-281. Olubobokun, J.A., Graig, W.M. & Nipper, W.A. 1988. Characteristics ofprotozoal and bacterial fractions from microorganisms associated with ruminal fluid or par- ticles. J. Anim. Sci. 66: 2701-2710. ORSKOV, E.R. & McDonald, 1.1979.The estimation ofpro- tein degradability in the rumen from incubation meas- urements weighted according torate ofpassage. J. Agric. Sci. 92: 499-503. —, Hughes-Jones, M. & McDonald, I. 1980. Degradabi- lity ofprotein supplements and utilization of undegraded protein by high-producing dairy cows. In: Haresign, W. (ed.). Recent advances in animal nutrition -1980, Nutr. Conf, Feed Man. 14th. p. 85-98. —, Reid, G.W. & McDonald, I. 1981. The effects ofpro- tein degradability and food intake on milk yield and composition in cows in early lactation. Br. J. Nutr. 45: 547-555. Pahkala, K. & Sovero, M. 1988.The cultivation and breed- ing of oilseed crops in Finland. Ann. Agric. Fenn. 27: 199-207. Papas, A.M., Ames, S.R., Cook, R.M., Sniffen, C.J., Polan C.E. & Chase, L. 1984. Production responses of dairy cows fed diets supplemented with ammonium salts of iso C-4 and C-5 acids. J. Dairy Sci. 67: 276-293. Papas, A., Ingalls, J.R. & Cansfield, P. 1978. Effects of Tower and 1821 rapeseed mealsand Tower gum on milk yield, milk composition and blood parameters of lacta- ting dairy cows. Can. J. Anim. Sci, 58: 671-679. Pearson, A.W., Butler, E.J., Curtis, R.F., Fenwick, G.R., Hobson-Frohock, A. & Land, D.G. 1979. Effect of rapeseed meal on trimethylamine metabolism in the domestic fowl in relation to egg taint. J. Sci. Food Agric. 30: 799-804. Peirce-Sandner, S 8., Papas, A.M., Rogers, J.A., Swee- ney, T.F., Cummins, K.A., Conrad, H.R. & Muller, L.D. 1985. Supplementation of dairy cow diets with ammonium salts ofvolatile fatty acids. J. Dairy Sci. 68: 2895-2907. Peltola, P. 1960. Goitrogenic effect ofcow’s milk from the goitre district of Finland. Acta Endocr. 34: 121-128. Peoples, A.C. & Gordon, F.J, 1989.The influence ofwilt- ing and season of silage harvest and the fat and protein concentration of the supplement on milk production and food utlization by lactating cattle. Anim. Prod. 48: 305- 317. Piatkowski, 8., Gurtler, H. & Voigt, J. 1990. Grundzilge der Wiederkäuer-Emährung. Gustav Fischer Verlag Jena 1990,232 p. Poijärvi, I. 1944. Inhemska linfrö-, vallmo- och rapskakors fodervärde. Praktisk Försöksverksamhet 1,7: 4-6. Porter, M.G., Patterson, D.C., Steen, R.W. Gordon, F.J. 1984. Determination of dry matter and gross energy of grass silage. In: Gordon, F.J. & Unsworth, E.F. (eds.). Seventh silage conference 1984,The Queen’s University ofBelfast, p. 89- 90. Rae, R.C, & Smithard, R.R. 1985. Estimation of true nitro- gen digestibility in cattle by a modified nylon bag technique. Proc. Nutr. Soc. 44: 116A. —, Ingalls J.R. & McKirdy, J.A. 1983. Response of dairy cows to formaldehyde-treated canola meal during early lactation. Can. J. Anim. Sci. 63: 905-915. Rajamäki, S. & Rauramaa, A. 1984. The automated deter- mination ofurea in milk. Finn. Chem. Lett. 2:47-48. Rauramaa, A. 1983. Content of L-5-vinyl-2-thio-oxazoli- done in milk. Finnish Chemistry (Kemia-kemi) 10, no. 11: 964 (abstr.). Rees, P. & Rowlinson, P. 1983.The effects of soya protein degradability upon the production and food intake of dairy cows offered grass silage ad libitum. Anim. Prod. 36: 502 (abstr.). Reynolds, J.R. & Young, C.G. 1964. Effect of seed prepa- ration on efficiency and oil quality in filtration extraction ofrapeseed. J. Am. Oil Chem. Soc. 41: 63-65. Röbbelen, G. & Thies, W. 1980. Biosynthesis of seed oil and breeding for improved oil quality of rapeseed. In: Tsunoda, S. et al. (eds,). Brassica crops and wild allies. Biology and breeding. Japan Sci. Soc. Press, Tokyo, p. 253-283. Robinson, P.FI. & Kennelly, J.J. 1988. Influence of intake of rumen undegradable protein on milk production of 426 Agric. Sei. Fin!. 1(1992) late lactation Holstein cows. J. Dairy Sci. 71: 2135-2142. Sniffen, C.J. & Van Soest, P.J. 1985. Influence of level of feed intake on digestion and bacterial yield in the forestomachs of dairy cattle. Can. J, Anim. Sci. 65: 437- 444. Tamminga, S. & Van Vuuren, A.M. 1987. Influence of declining level of feed intake and varying the proportion of starch in the concentrate on rumen ingesta quantity, composition and kinetics of ingesta turnover in dairy cows. Livest. Prod. Sci. 17:37-62. Rode, L.M. & Sätter, L.D. 1988. Effect of amount and length ofalfalfa hay in diets containing barley or corn on site of digestion and rumen microbial protein synthesis in dairy cows. Can. J. Anim. Sci. 68: 445-454. Rohr, K. 1987. Present situation of the modem protein sys- tems: Germany. In: Jarrige, R. & Alderman, G. (eds.). Feed evaluation and protein requirement systems for ruminants. Proc. CEC seminar, Brussels. CEC L-2985 Luxembourg, p. 25-27. Rooke, J.A., Rymer, C., Maya, F.M. & Armstrong, D.G. 1992. Effect of including barley or molassed sugar beet feed in grass silage diets on their digestion by cattle and sheep. J. Sci. Food Agric. 58: 475-483. Roy, J.H.8., Balch, C.C., Miller, E.L., orskov, E.R. & Smith, R.H. 1977. Calculations of the N-requirements for ruminants from nitrogen metabolism studies. In: Pro- tein metabolism and nutrition. Eur. Assoc. Anim, Prod. (EAAP), Pubi. 22: 126-129. Rulquin, H. & Journet, M. 1987. Efficiency of utilization of amino acids (AA) for lactation in ruminants. In: Jar- rige, R. & Alderman, G. (eds.). Feed evaluation and pro- tein requirement systems for ruminants. Proc. CEC seminar, Brussels. CEC L-2985 Luxembourg, p. 213- 223. Salo, M.-L. 1965. Determination of carbohydrate fractions in animal foods and faeces. Acta Agr. Perm. 105: 1-102. —, Tuori, M. & Kiiskinen, T. 1982. Rehutaulukot ja ruo- kintanormit. Helsinki. 70 p. Sanchez, J.M. & Claypool, D.W. 1983. Canola meal as a protein supplement in dairy rations. J. Dairy Sci. 66: 80- 85. Sang, J.P. & Salisbury, P.A. 1988. Glucosinolate profiles of international rapeseed lines. J. Sci, Food Agric. 45: 225-261. & Truscott, RJ. 1984. Liquid chromatographic deter- mination of glucosinolates in rapeseed as desulphoglu- cosinolates. J. Assoc. Offic. Anal. Chem. 67: 829-833. Sarwar, G., Bell, J.M., Sharby, T.F. & Jones, J.D. 1981. Nutritional evaluation ofmeals and meal fractions deri- ved from rape and mustard seed. Can. J. Anim. Sci. 61: 719-733. SAS 1989. SAS Institute Inc. SAS/STAT User's guide, ver- sion 6, fourth edition, vol. 2. Gary, NorthCarolina, USA. 846 p. Schnug,E. 1987.Eine Methode zurschnellen und einfachen Bestimmung des Gesamtglucosinolatgehaltes in Griinmasse und Samen von Kruziferen durch die quanti- tative Analyse enzymatisch freisetzbaren Sulfates. Fat Sci. Technol. 89: 438-442. & Haneklaus, S. 1988. Theoretical principles for the indirect determination of the total glucosinolate content in rapeseed and meal quantifying the sulphur concentra- tion via X-ray fluorescence (X-RF method). J. Sci. Food Agric. 45: 243-254. Schwenke, K.D, 1990, Structural studies on native and che- mically modified storage proteins from rapeseed (Bras- sica napus L.) and related plant proteins. Nahrung 34: 225-240. —, Raab, 8., Uhlig, J., Tkocz, H., Behlke, J., Böttoer, M. & Freimuth, U. 1973a. fiber Samenproteine. 3. Mitt. Isolierung und Charakterisierung der Albumine aus Son- nenblumen- und Rapssaraen. Nahrung 17: 791-809. —, Simo, B. & Raab, R. 1973b. Über Samenproteine. 2. Mitt. Fraktionenverteilung der Proteine aus Rapssamen. Nahrung 17: 579-586. Setälä, J. 1983. The nylon bag technique in the determina- tion of ruminal feed protein degradation. J. Scient. Agric, Soc, Finl, 55: 1-78. & Syrjälä-Qvist, L. 1984/85. Degradation ofcmde pro- tein and quality of undegradable protein in untreated or formaldehyde-treated rapeseed meal. Anim. Feed Sci. Technol. 12: 19-27. Shahidi, F. 1990a. Production and Canola: Global pro- duction and distribution. In: Shahidi, F. (ed.). Canola and rapeseed: production, chemistry, nutrition and pro- cessing technology. Van Nostrand Reinhold, New York, p. 3-13. —1990 b. North American production of Canola. In: Sha- hidi, F. (ed.). Canola and rapeseed: production, che- mistry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 15-23. & Naczk, M. 1990, Removal of glucosinolates and other antinutrients from Canola and rapeseed by methanol/ammonia processing. In: Shahidi, F. (ed.). Canola and rapeseed: production, chemistry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 291-306. Sharma, H.R., Ingalls, J.R. & McKjrdy, J.A. 1977. Effects of feeding a high level ofTower rapeseed meal in dairy rations on feed intake and milk production. Can, J. Anim. Sci. 57: 653-662. Sjaunja, L.0., Baevre, L., Junkkarinen, L., Pedersen. J. & Setälä, J. 1990. ANordic proposal for an energy cor- rected milk (ECM) formula. 27th Session International Committee of Recording and Productivity ofMilk Ani- mal. Paris, France, p. 156-157. Sloan, B.K. Rowlinson, P. & Armstrong, D.G. 1988.The influence of a formulated excess of rumen degradable protein or undegradable protein on milk production in dairy cows in early lactation. Anim. Prod. 46: 13-22. Small, J.C. & Gordon, F.J. 1990. A comparison of the res- ponses by lactating cows giwen grass silage to changes 427 Agric. Sei. Fint. 1 (1992) in the degradability or quantity ofprotein offered in the supplement. Anim. Prod. 50: 391-398. Smith, R.H. 1975. Nitrogen metabolism in the rumen and the composition and nutritive value of nitrogen com- pounds entering the duodenum. In: McDonald, I.W. & Warner, A.C.I, (eds.). Digestion and metabolism in the ruminant. Proc. 4th Int. Symp. Ruminant Physiol. Syd- ney, Australia, p. 399-415. Snedecor, G.W. & Cochran, W.G. 1980. Statistical meth- ods. Seventh edition. The lowa State University Press, Ames, lowa, USA. 507 p. Sniffen, C.J. & Robinson, P.H. 1987. Microbial growth and flow as influenced by dietary manipulations. J. Dairy sci. 70: 425-441. Somogyi, M. 1945. A new reagent for the determination of sugars. J. Biol. Chem. 160: 61-68. Sorensen, FI. 1990. Glucosinolates: Structure-Properties- Function. In: Shadidi, F. (ed.). Canolaand rapeseed: pro- duction, chemistry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 149-172. Spörndly, E. 1986. Effects of diet to dairy cows on milk protein content. Rep. 159. Swed. Univ. Agric. Sci. Upp- sala, Sweden. 24 p. 1989, Effects of diet on milk composition and yield of dairy cows with special emphasis on milk protein con- tent. Swed. J. Agric. Res. 19: 99-106. Summers, J.D. & Leeson, S. 1985. Mineral profile of Canola and soybean meal. Can, J. Anim. Sci. 65: 913- 919. —, Bong, D.L. & Leeson, S. 1983. Sodium, potassium and phosphorus in canola and soybean meal. Nutr. Rep. Int. 28; 955-963. Suvitie, M. & Rinne, K. 1988. Effect of supplementing grain concentratewith rapeseed meal on milkproduction and milk composition with dairy cows on grass silage or hay based feeding. Polish-Finnish seminar on “Nutritive value and utilization of rape seeds protein and other plant protein in animal feeding”, 13.-14. 12.1988, Jab- lonna, Warsaw. Mimeogr. 14p. Syrjälä-Qvist, L., Tuori, M. & Setälä, J. 1982. Rape seed meal as a protein source for high-production dairy cows on grass silage- and hay-based feeding. J. Scient. Agric, Soc. Finl. 54: 145-153. Syrjälä, L., Poutiainen, E. & Koskela, V.-H. 1978. Unt- reated and formaldehyde treated skimmilk powder as a protein supplement for dairy cows. J. Scient. Agric. Soc. Finl. 50: 155-165. Tamminga, S. 1981.Effect of the roughage/concentrate ratio on nitrogen entering the small intestine of dairy cows. Neth. J. Agric, Sci. 29: 273-283. —, Robinson, P.H., Vogt, M. & Boer, H. 1989. Rumen ingesta kinetics of cell wall components in dairy cows. Animal Feed Sci. Technol. 25: 89-98. Tesfa, A.T. 1992. Rapeseed oil in ruminants diet: its effect on rumen metabolism and animal performance. Diss, Univ. Helsinki, Dept. Anim. Sci. 53 p. —, Tuori, M. & Syrjälä-Qvist, L. 1991. Replacement of grain by tallow or rapeseed oil in the diet of lactating cows and its effects on milk yield and composition. Finn. J. Dairy Sci. 49: 54-64. Theander, O. & Åman, P. 1976.Low-molecular carbohyd- rates in rapeseed and turnip rapeseed meals. Swed. J. Agric, Res. 6: 81- 85. —, Åman, P., Miksche, G.E. & Yasuda, S. 1977, Carbo- hydrates, polyphenols, and lignin in seed hulls of differ- ent colors from turnip rapeseed. J. Agric. Food Chem. 25: 270-273. Thies, W. 1976. Quantitative gas liquid chromatography of glucosinolates in a microliter scale. Fette Seifen Anstrichmittel 78: 231-234. - 1982. Complex-formation between glucosinolates and tetrachloropalladate (II) and its utilization in plant breed- ing. Fette Seifen Anstrichmittel. 84; 338-342. Thomas, C. & Rae, R.C. 1988. Concentrate supplementa- tion of silage for dairy cows. In: Gamsworthy, P.C. (ed.). Nutrition and lactation in the dairy cow. Butterworths. p. 327-354. Thomas, P.C., Chamberlain, D.G., Robertson, S., Sha- moon, S.A. & Watson, J.N. 1984. The effect of the source and level of protein supplement on milk pro- duction in dairy cows given silage diets. In: Gordon, F.J. & Unsworth, E.F. (eds.). Seventh silage conference. Silage production & utilization. Summary of papers, p. 45-46. Thomke, S. 1981.Review ofrapeseed meal in animal nutri- tion; ruminant animals. J. Am. Oil Chem. Soc. 58: 805- 810. Thompson, L. 1990. Phytates in Canola/rapeseed. In: Sha- didi, F. (ed.). Canola and rapeseed: production, chem- istry, nutrition and processing technology. Van Nostrand Reinhold, New York. p. 173-192. Thuen, E. & Vik-Mo, L, 1985. Comparison ofthree systems for protein evaluation on the basis ofproduction experi- ments in lactating dairy cows. Acta Agric. Scand. Suppl. 25; 155-162. Tilley, J.M.A. & Terry, R.A. 1963. A two-stage technique for in vitro digestion of forage crops. J. Br. Grassl. Soc. 18: 104-111. Toepher International 1990. Statistische Informationen zum Getreide- und Futtermittelmarkt. Toepher International, Hamburg, p. 50-52. Tuori, M. & Syrjälä-Qvist, L. 1987. Heat-moisture treated rapeseed meal as a protein supplement for dairy cows. Proc. 7th Int. Rapeseed Congr., Poznan, Poland. Vol 7: 1836. & Syrjälä-Qvist, L. 1988. Effect of supplementing grain concentratewith rapeseed meal on the digestibility and N- utilization with dairy cows on grass silage or hay based feeding. Polish-Finnish seminar on “Nutritive value and utilization of rape seeds protein and other plant protein in animal feeding”, 13.-14.12.1988, Jab- lonna, Warsaw. Mimeogr 3 p. 428 Agric. Sei. Finl. 1 (1992) Underhill, E.W. & Kirkland, D.F, 1971. Gas chromatog- raphy of trimehylsilylderivatives of glucosinolates. J. Chromatogr. 57: 47-54. Unger, E.H. 1990. Commercial processing of Canola and rapeseed: crushing and oil extraction. In: Shahidi, F. (ed.). Canola and rapeseed: production, chemistry, nutri- tion and processing technology. Van Nostrand Reinhold, New York. p. 235-249. Valle, O. 1953. Kotimainen kasviöljyn tuotanto: Öljykas- vien viljelymme nykyinen vaihe. [Present status of the growing oil plants in Finland, English abstr.]. J. Industr. Chem. 10: 147- 150. Valtion maatalouskemian laitos. 1989. Reports of the State Institute of Agricultural Chemistry with English Sum- mary. Vantaa, Helsinki, 103p. Van Es, AJ.H. 1978.Feed evaluation for ruminants. I. The systems in use from May onwards in the Netherlands. Livest. Prod. Sci. 5: 331-345. Van Keulen, J. & Young, B.A. 1977. Evaluation of acid- insoluble ash as a natural marker in ruminant digestibil- itystudies. J. Anim. Sci. 44: 282-287. Vanhatalo, A. & Aronen, I. 1991. The ruminal and intes- tinal degradation estimates of either untreated or physi- cally treated rapeseed meals measured by nylon bag techniques. Proc. Bth International rapeseed congress “Rapeseed in a changing world”, Saskatoon, Canada, p. 424-429. Varvikko, T. 1986. Evaluation of rumen-undegraded feed protein. Diss. Dept. Anim. Husbandry, Agric. Res. Centre, Jokioinen, Finland. 28 p. & Lindberg, J.E. 1985. Estimation ofmicrobial nitrogen in nylon-bag residues by feed 15N dilution, Br. J. Nutr. 54: 473-481. & Vanhatalo, A. 1988. Intestinal nitrogen degradation in sacco of hay and grass silage. Proc. VI World Conf. Anim. Prod. 1988, Helsinki, Finland, p. 359. & Vanhatalo, A. 1990. The effect of differing types of cloth and of contamination by non-feed nitrogen on intestinal digestion estimates using porous synthetic- fibre bags in a cow. Br. J. Nutr. 63: 221-229. & Vanhatalo, A. 1991. Intestinal nitrogen degradation of hay and grass silage estimated by the mobile bag technique. World Rev. Anim. Prod. 26: 73-76. Vérité, R. & Journet, M. 1977. Utilisation des tourteaux traités au formol par les vaches laitiéres. 11. Effets sur la production laitiére du traitement des tourteaux et du niveau d’apport azoté au début de la lactation. Ann. Zoo- tech. 26: 183-205. & Peyraud, J.-L. 1989. Protein: the PDI systems. In: Jar- rige, R. (ed.). Ruminant nutrition. Recommended allo- wances & feed tables. INRA, Paris, 1989.p. 33-47. —, Journet, M. & Jarrige,R. 1979. A new system for the protein feeding of ruminants: the PDI system. Livest. Prod. Sci. 6: 349-367. Vik-Mo, L. 1985. Merits ofnew protein evaluation princip- les applied to Norwegian dairy diets. Acta Agric. Scand. Suppl, 25:200-211. Vilkki, J. 1991. Kulta-kevätrypsi. Maatalouden tutkimus- keskus, Tiedote 3/91, 22 p. Vilva, V. 1989. WSYS statistical program. Dept. Anim. Breed. Univ. Helsinki. Vincent, I, & Hill, R. 1988. Low-glucosinolate rapeseed meal as a protein source for milk production. Anim. Prod. 46: 505-506 (abstr.). Vincent, 1.C., Hill, R. & Campling, R.C. 1990. A note on the use of rapeseed, sunflower and soya-bean meals as protein sources in compound foods for milking cattle. Anim. Prod. 50: 541-543. Virtanen, A. I. 1963. An introduction to studies on factors in food- and fodder-plants which disturb the function of the thyroid gland, the transfer of these factors to milkby the cow, and the alleged goitogecic properties of milk. In: Virtanen, A.I. (ed.). Final report on investigations on the alleged goitrogenic properties ofmilk. Biochemical Inst., Helsinki, p. 1-26. —, Kreula, M. & Kiesvaara, M. 1959. On the transfer of L-5-vinyl-2-thiooxazolidone from the rumen to the milk. Acta Chem. Scand. 13: 1043-1044. Voigt, J. & Piatkowski, B. 1991.Stickstoffumsatz im Ver- dauungstrakt von Milchkuhen. 42nd Ann. Meet. EAAP, Berlin, Germany, p. 442. —, Piatkowski, 8., Engelmann, H. & Rudolph, E. 1985. Measurement of the postruminal digestibility of crude protein by the bag technique in cows. Arch. Tieremähr. 35: 555-562. Volden, H., Harstad, O.M. & Henne, M. 1992. Praktisk utproving av nytt proteinvurderingssystem. Forsok på landbrukskoler. In: Husdyrforsoksmotet, Norges landb- rukshogskole. SFFL, Faginfo no. 13 1992, Ås, Norway, p. 567-582. Waldern, D.E. 1973. Rapeseed meal versus soybean meal as the only protein supplement for lactating cows fed com silage roughage rations. Can. J. Anim. Sci. 53; 107- 112. Ward, J.T., Basford, W.D., Hawkins, J.H. & Holliday, J.M. 1985. Oilseed rape. Farming Press LTD,lpswich. 298 p. Webster, A.J.F. 1987. Metabolizable protein - the U.K. approach. In: Jarrige, R. & Alderman, G. (eds.). Feed evaluation and protein requirement systems for ruminants. Proc. CEC seminar, Bmssels. CEC L-2985 Luxembourg, p. 47-53. Webster, J. 1992. Metabolizable protein: the new UK fee- ding system forruminants. Int. Milling Flour Feed 1992, May: 22-23. Weisbjerg, M.R., Bhargava,P.K., Hvelplund, T. & Mad- sen, J. 1990. Anvendelse af nedbrydningsprofiler i fodermiddelvurderingen (Summary: Use of degradation curves in feed evaluation). Beretn. 679 fra Statens Hus- dyrbrugsforsog. 33 p. Zadernowski, R. & Sosulski, F. 1978. Composition oftotal lipids in rapeseed. J. Am. Oil Chem. Soc. 55: 870-872. 429 Agric. Sei. Fin!. 1 (1992) SELOSTUS Rypsirouhe lypsylehmien valkuaisrehuna säilörehuvaltaisella ruokinnalla Mikko Tuori Helsingin yliopisto Lypsylehmien väkirehun täydentämistä rypsirouheella tut- kittiin viidessä ruokintakokeessa tuoresäilörehuun perustu- valla ruokinnalla. Rypsirouheen osuus väkirehuseoksessa vaihteli 0-33 % loppuosan ollessa ohran ja kauran seosta. Yhdessä kokeessa oli mukana samat koejäsenet sekä rypsi- rouheesta että soijarouheesta, ja toisessa kokeessa säilörehu tai samalla kasvuasteella korjattu latokuivattu heinä. Lisäksi tehtiin laskelmia rypsirouheen vaikutuksesta maitotuotok- seen aineistosta, johon oli kerätty Suomessa viime vuosina tehdyt rypsivertailukokeet. Valkuaisen hyväksikäyttöä tar- kasteltiin pohjoismaisen AAT-PBV -valkuaisjärjestelmän mukaan. Tutkimusaikana rypsilajikkeet muuttuivat runsaasti glu- kosinolaatteja (40-50 pmoolia per g rasvatonta rouhetta) sisältävistä 0-lajikkeista 00-lajikkeisiin, joiden glukosino- laattipitoisuus aleni noin kolmannekseen. Rypsirouheen lämpökäsittely Öpex-menetelmällä alensi rouheen glukosi- nolaattipitoisuuden edelleen noin puoleen. Kun viljaa korvattiin rypsirouheella (maks. 1.7 kg KA/d) säilörehun ollessa ad libitum, säilörehun syönti lisääntyi keskimäärin 0.43 kg KA per kg rypsirouheen kuiva-aineen lisäystä (n.s.). Vaikutus maitotuotokseen oli keskimäärin 0.77 kg (P<0.02) ja energiakorjattuun maitotuotokseen (ECM) 0.70 kg per kg rypsirouheen kuiva-aineen lisäystä (P<0.02). Rypsirouheen sisällyttäminen väkirehuun 12-16 % lisäsi tuotosta, mutta osuuden noustessa edelleen 24 %:iin vaikutus oli vähäinen. Maidon valkuaispitoisuuteen vaikutti merkitsevästi muuntokelpoisen energian saannin lisäys, val- kuaispitoisuus nousi 0.07 g/kg per MJ ME:n lisäystä (P<0,02). Rypsirouheen Öpex-käsittely (kostea kuumennus) lisäsi maitotuotosta merkitsevästi yhdessä kokeessa (21.9 vs. 23.9 kg maitoa tai 23.4 vs. 25.2 kg ECM/d), (P<0,03). Kahdessa muussa kokeessa käsittelyllä ei ollut vaikutusta maitotuo- tokseen. Soijarouheella jarypsirouheella ei ollut eroa vaiku- tuksessa maitotuotokseen, kun rouheita syötettiin sama määrä raakavalkuaisena mitattuna. Maidon goitriinipitoisuus aleni rypsilajikkeiden glukosi- nolaattipitoisuuden jarypsirouheen määrän alentuessa. Käy- tettäessä Öpex-käsiteltyjä 00-rypsirouheita maidon goitriini- pitoisuus oli enää alle 10 pg/1 (analyysin herkkyysraja 2 Pg/1)- AAT:n hyväksikäyttöä maidontuotannossa estimoitiin käyttämällä erilaisia vakioita AAT-PBV -arvojen laskemi- sessa. Hyväksikäytön vaihtelukerroin aleni, kun laskennalli- sesti lisättiin mikrobivalkuaisen osuutta AAT:stä. Tähän suuntaanvaikuttavat mikrobi-N kontaminaation huomioimi- nen in sacco määrityksessä, pötsin virtausvakion alentami- nen o.oB:sta o.o3:een ja mikrobivalkuaissynteesin tehokkuu- den lisääminen. Paras malli saatiin laskemalla mikrobival- kuaissynteesin tehokkuus Voigt ja PiATKOWSKin (1991) mukaan. 430 Agric. Sei. Fin!. 1 (1992) AAT-PBV feed protein evaluation system, calculation of feed values and requirements AAT =MCP * aam * daam + UDP * aaf * daaf MCP =l79* DCHO (or MCP =20 g microbial amino-N per kg DCHO) UDP =CP * EPD PBV =RDP - MCP where AAT = absorbable amino acid protein of the feed PBV = ruminal protein balance of the feed MCP = microbial crude protein synthetized in the rumen CP = crude protein content of the feed UDP = rumen undegraded crude protein of the feed EPD = effective degradation of feed crude protein in the rumen aa = proportion ofamino acid protein in the microbial protein (aam) or in the undegraded feed protein (aaf) daaf = digestibility of undegraded feed amino acids daam = digestibility ofmicrobial amino acids DCHO = digestible carbohydrates of feed (dig. crude fibre + dig. nitrogen free extracts) RDP = EPD *CP (= rumen degraded dietary crude protein) Constants: Proportion ofamino acid protein in absorbable crude protein; - 0.70 in microbial crude protein - 0.85 in undegraded concentratecrude protein - 0.65 in undegraded roughage crude protein Digestibility ofamino acids (AA): -0.85 for AA from microbial crude protein -0.82 for AA from undegraded feed crude protein (UDP) Digestibility ofUDP can be estimated by using the equation ofHVELPLUND & MADSEN (1990): TD =(UDP-TU)AJDP where TD = true digestibility ofUDP in the small intestine TU = fraction of true indigestible crude protein of the feed TU should be determined by using the mobile nylon bag technique, and if not determined, value of0.05-0.07 can be used. Calculation ofEPD value for feeds according to orskov & McDonald (1979): p =a +b*(l- e"*') EPD =a + b � c/(c + k) where p =measured bag degradation ofprotein at time t a = rapid degraded fraction ofprotein b = slow degraded fraction c = degradation rate of fraction b k =fractional outflow rate EDP = effective degradation ofprotein 431 Appendix 1 Recommendations ofAAT values for dairy cattle: Requirements ofAAT are from Madsen (1985) and later revised by Nordic protein group (Hvelplund 1990). Requirement for maintenance is 3.25 g AAT * W07 5. For the milk production the recommendation was 45 g/kg ECM (Madsen 1985),now 40 g/ECM (energy corrected milk yield, Sjaunja et al. 1990), variation 37-42 g. At production levels below 25 kg ECM per day a decrease of3-4 g AAT per kg ECM may be justified((Hvelplund 1990). In Denmark the AAT-concentration of the diet can be expressed as 97 g AAT/feed unit for high producing, decreasing to about 90 g AAT/FU for low producing cows. In Norway AAT-recommendation (Volden et al. 1992) is: AAT, g/kg ECM = (40 * ECM + 0.2 * ECM2 )/ECM =40 + O.2*ECM Additional requirement for pregnancy is 60, 100 and 172 g AAT/d at the 7th, Bth and 9th month ofpregnancy. The ideal PBV value should be 0, but minimum value of -200 g/d for high producing and -300 g/d for low producing cows can be accepted. As maximum PBV-value 900 g/d is accepted in Denmark ((Hvelplund 1990) but there is no recommenda- tion in Norway (Volden et al. 1992). 432 Appendix 1. Composition of feeds in experiments 1 and 2 DM In dry matter (%) (%) Ash Crude Ether Crude protein extract fibre Experiment 1 Silage 24.9 7.4 16.3 5.6 28.3 Hay 83.8 6.3 7.7 2.2 35.4 Barley 75.6 2.5 10.1 2.3 5.8 Oat 82.4 3.5 11.7 5.4 12.2 RSM 88.6 7.4 35.4 9.3 12.8 RSM-heated 86.3 7.5 40.5 4.5 14.6 Experiment 2 Silage 22.5 6,6 16.8 6.4 30.3 Hay 83.5 6.5 9.6 2.1 34.3 Barley 75.1 2.3 11.4 2.3 5.2 Oat 87.4 3.1 13.3 5.5 10.8 RSMO 88.0 8.0 36.2 6.7 12,5 RSMO-Öpex 88.0 8.0 35.7 6.9 13.0 RSM =turnip rapeseed meal (0-var,), RSM-heated = dry heated rapeseed meal; RSM-Opex =heat-moisture treated turnip rapeseed meal (0-var.) Composition of feeds in experiment 3 DM In dry matter (%) (%) Ash Crude Ether Crude NDF ADF ADL protein extract fibre Silage 20.9 6.8 18.5 5.4 27.3 49.7 29.4 1.7 Barley 87.3 3.0 12.7 3.7 5.2 20.9 5.5 0.4 Oat 88.6 3.2 11.5 7.0 10.8 27.3 12.4 2.2 RSMOO-Öpex 88.2 8.0 33.3 10.7 12.1 27.2 17.4 7.4 RSMO 88.5 7.7 33.4 10.2 12.9 28.2 18.9 8.6 RSMOO-FeSO4 88.8 8.1 33.3 11.5 12.1 25.7 16.5 6.9 RSMO =turnip rapeseed meal (0-var.); RSMOO-Öpex =heat-moisture treated turnip rapeseed meal (00-var.); RSMOO-FeSO4 =turnip rapeseed meal (00-var.), glucosinolate content is reduced by ferrosulphate treatment. Composition of feeds in experiment 4. DM In dry matter (%) (%) Ash Crude Ether Crude NDF ADF ADL protein extract fibre Silage Hay 21.9 6.6 16.6 5.3 87.5 9.0 18.0 2.4 87.0 2,4 11.7 2,6 88.7 2.7 12.9 7.0 87.2 8.0 32.7 9.4 31.0 56.2 33.0 2.2 29.5 58.6 30.6 2.4 4.9 19.9 5.2 0.8 9.3 23.5 10.3 1.9 13.7 25.6 17.9 7.0 Barley Oat RSMOO-Öpex RSM-Öpex = heat-moisture treated turnip rapeseed meal (00-variety); hay is cut at the same maturity as silage 433 Appendix 2. Composition of feeds in experiment 5 DM In dry matter (%) (%) Ash Crude Ether Crude NDF ADF ADL protein extract fibre Silage 24.4 7.9 17.8 5.3 23.5 46.5 25.2 2.0 Hay 87.9 7.5 9.1 2.4 34.3 70.4 36.9 2.7 Barley 87.5 2.2 13.0 3.3 4.6 20.9 5.5 1.3 Oat 88.3 3.4 13.9 6.6 10.2 28.7 12.5 2.8 RSMOO 89.7 7.3 39.1 5.1 13.2 28.0 20.3 9.5 RSMOO-Opex 88.9 7.3 37.9 6.4 13.0 28.5 20.2 9.5 SBM 87.9 7.4 49.5 3.0 7.6 13.8 8.0 0.8 SBM-Öpex 89.2 6.5 50.1 3.8 8.2 14.6 8.3 0.4 RSM =turnip rapeseed meal (00-var.); RSM-Öpex =heat-moisture treated turnip rapeseed meal (00-var.); SBM =soybean meal Energy values of the feeds in experiments 1-5 Exp. 1 Exp. 2 Exp. 3 Exp. 4b Exp. 5 FFU ME NEL FFU ME NEL FFU ME NEL FFU ME NEL FFU ME NEL Silage 0.730 10.30 5.67 0.742 10.47 5.79 0.735 10.39 5.78 0.771 10.88 6.04 0.791 11.18 6.32 Hay 0.519 8.77 4.94 0.554 9.07 5.12 0.692 10.14 5.90 0.611 9.69 5.51 Barley 1.166 13.61 7.97 1.163 13.59 7.96 1.172 13.69 8.06 1.173 13.70 8.04 1.158 13.53 7.92 Oat 0.950 11.53 6.61 1.036 12.32 7.17 1.080 12.59 7.38 1.072 12.78 7.51 1.053 12.54 7.35 RSM 1.088 12.96 7.57 0.983 11.82 7.03 1.063 12.40 7.48 0.960 11.59 6.84 RMS-Öpex" 1.065 12.56 7.61 0.984 11.82 7.03 1.072 12.51 7.57 1.019 12.15 7.29 0.981 11.81 7.01 RSM-FeS04 1.082 12.63 7.67 SBM 1.056 12.46 7.48 SBM-Öpex 1.069 12.58 7.56 '•ln experiment 1 dry heat-treated RSM Average rumen degradability of the feed protein in experiments 3-5. No of Water Washing Incubation time, hrs analysis soluble loss 3 63 6 12 24 48 72 Silage 13 60.6 63.8 70.1 81.1 86.2 90.0 91.3 Hay 1 47.3 48.3 50.4 62.0 79.1 80.9 85.8 Hay(e) 4 56.3 54.8 59.8 73.5 79.8 85.2 85.8 Barley 9 15.3 41.9 56.0 66.4 88.3 91.9 95.4 96.3 Oat 9 14.7 80.0 84.3 88.1 93.6 94.6 94.9 95.0 RSM 12 11.2 18.5 36.0 47.3 74.0 88.4 92.2 92.8 RSMtr 11 9.7 14.7 31.8 46.2 70.7 84.7 91.3 92.4 SBM 2 13.4 24.5 40.8 40.2 60.3 93.2 99.1 99.3 SBMtr 2 8.7 20.0 23.0 26.3 43.8 66.6 97.9 99.2 434 Appendix 2. Feed intake and milk production in trial 1 (means of least squares of the lactation weeks of4-14) Treatments Significance Control RSMI RSM2 TRSMI S.E. (P-value) DM intake (kg/d) Forage 9.92 10.15 9.90 10.21 0.97 0.92 Grain RSM Total 5.73 4.62 3.49 4.30 0.22 1.40 2.10 1.25 0.11 15.87 16.39 15.70 15.98 1.12 0.78 Milk production Milk (kg/d) 24.36 26.78 23.89 26.68 2.98 0.24 ECM (kg/d) Fat (g/d) 26.06 26.96 24.91 27.18 2.57 0.43 1176 1173 1102 1182 126 0.66 Protein (g/d) 752 797 736 815 66 0.18 Milk composition Fat (g/kg) 44.2 4.0 0.2748.4 44.3 46.0 Protein (g/kg) 30.9 30.2 30.9 30.6 1.8 0.92 2.9