Introduction
In organic dairy farming systems, grass, legumes and their ensiled conserves represent the main feedstuffs and major protein sources. In ruminant nutrition, ruminal microbes contribute considerably to the protein supply of the animal. For an optimal growth of rumen microbes and synthesis of microbial protein, a balanced and simultaneous ruminal protein and carbohydrate supply is crucialReference Nocek and Russell1. Carbohydrates can be divided into non-structural carbohydrates (starch, sugar and pectins), which are readily fermentable in the rumen, and structural carbohydrates (cell wall constituents consisting of cellulose, hemicellulose and lignin), which are potentially degradable in the rumen. Both determine the energy value of a feedstuff. From grass, legumes and their silages, rumen protein supply is usually higher than energy supply and, most notably, protein degradation occurs more rapidly than carbohydrate fermentation. Therefore, slow ruminal carbohydrate degradation rate in relation to rapid nitrogen release is claimed as one of the main factors for high environmental nitrogen losses, low microbial protein synthesis and low feed efficiencyReference Givens and Rulphin2.
Traditionally, protein-rich dairy diets are supplemented with cereal grains, which are high in starch and low in protein content. However, in sustainable agriculture systems, feeding grain must be viewed critically with regard to amount of forage utilization, intact nutrient cycles and the use of potential human food in animal feeding. Therefore, in the past decade, several studies examined alternatives to grain such as fodder beets, beet pulp, whole crop silages, maize silage or energy-rich by-productsReference Eriksson, Murphy, Ciszuk and Burstedt3–Reference Moran and Croke7. A nutritional consequence when substituting cereals is that starch is replaced by other carbohydrate sources.
The general aim of organic dairy farming is to reach a high level of self-sufficiency. This can be achieved by minimizing purchased concentrate use and focusing on milk production from forage. Therefore, on organic dairy farms, maize silage could be an attractive crop instead of cereal grains. Maize silage is a reliable forage supplement that can be grown in most grassland regions of moderate altitude. It supplies ruminally digestible fiber, but also noteworthy amounts of ruminally degradable starchReference Sniffen, O'Connor, Soest, Fox and Russell8. Therefore, the objective of this experiment was to evaluate whether grain and maize silage are interchangeable energy sources in grass–clover silage-based diets in terms of feed and nutrient intake, milk yield, milk constituents and feed efficiency.
Animals, Materials and Methods
Experimental design and animals
The experiment was conducted as a repeated change-over design at the organic farm of the Agricultural High School Ursprung in the province of Salzburg, Austria (570 m NN, 1250 mm precipitation, 8.5°C average annual temperature) from November 2005 to February 2006. The experiment was carried out with 13 Holstein Friesian dairy cows, housed in a cubicle housing system with Calan gates for individual feeding. Cows were allotted to two treatment groups (six cows in treatment G and seven cows in treatment M) according to milk yield, days in milk and number of lactations. At the beginning of the experiment, average (± standard deviation) milk yield, days in milk and number of lactations were 25.9±6.16 kg per day, 105±59.0 days in milk and 2.9±1.82, respectively. The experiment lasted for 15 weeks with a 3-week adaptation period to Calan gates and feeding regimen at the beginning of the experiment and a change-over period of 2 weeks in weeks 9 and 10.
Feeding regimen
In the experiment, two grass–clover silage-based rations, intended to be equal in dietary energy content, were compared. The chemical composition and nutritional values of the feeds are listed in Table 1. Grass–clover silage was a mixture of first and third cut and derived from approximately 50% grassland and 50% perennial clover–grass leys. Hay was harvested from second cut. Maize silage was harvested in mid-October, chopped to lengths of approximately 10 mm and ensiled as bales. The grain mixture consisted of equal amounts of barley, wheat, rye and maize, which represent commonly used energy sources in concentrates for dairy cows. Furthermore, 50 g of a commercial mineral and vitamin mixture were given to each cow daily.
Table 1. Chemical composition and calculated nutritive values of feeds (g kg−1 dry matter (DM) unless stated).
1 Not determined.
2 Utilizable crude protein in the duodenum.
Both diets were based upon ad libitum allowance of grass–clover silage, which was fed at 5:00 and 15:30. At 14:00, approximately 1.5 kg hay was provided for each cow. All feeds were removed and weighed before the next feeding. Treatment group G received 1 kg of the grain mixture and treatment group E 2.1 kg maize silage on a dry matter (DM) basis. Grain mixture and maize silage were fed on top of grass–clover silage in two equal portions at morning and evening feeding. Additionally, cows in treatment G were fed 1 kg hay in addition to grass–clover silage at morning feeding to decrease grass–clover silage intake. It was assumed that the grain mixture and the maize silage were consumed completely. Refusals of morning-fed hay in treatment G were manually collected and weighed before removing grass–clover silage. In the present paper, maize silage, the grain mixture and morning-fed hay are referred to as supplements. In both treatments, cows exceeding a daily milk production of 17 kg were fed a purchased, pelleted concentrate mixture at a rate of 1 kg DM per 2 kg additional milk yield via an automatic feeding station.
Data collection and analytical procedures
Individual forage intake was recorded during four 6-day recording periods in weeks 4, 8, 11 and 15 of the experiment using Calan gates. Grass–clover silage was offered twice a day to ensure refusals in amounts of 5–10%. Orts were weighed before next feeding. Intake of the concentrate mixture was recorded daily during the whole experiment. In each recording period, DM content of all used feeds and feed refusals (grass–clover silage and hay) was determined at least three times by oven drying at 105°C for 24 h. Furthermore, in each recording period, one sample of each dietary component and feed refusal (for each treatment separately), pooled over 3 days, was collected and analyzed using methods described by ALVA9 and VDLUFA10. Contents of utilizable crude protein in the duodenum (uCP) and energy were calculated according to GfE11. Cows were milked twice daily at 6:00 and 16:30 in a herring milking parlor and milk yields were recorded automatically at each milking. In each recording period, three pooled samples from morning and evening milk were collected from each cow and analyzed for fat, protein, lactose, urea and somatic cell counts. Between recording periods, individual milk samples were collected once a week. Body weights and body condition scores (BCSs) (1–5 points, 0.25 interval, 1: very lean, 5: very fat) were determined once in each recording period.
Statistical analysis
Data were analyzed with the statistical program package SAS12 according to the model developed by Kaps and LambersonReference Kaps and Lamberson13. However, only parameters which were statistically significant were included in the final reduced model. Statistical differences were considered to be significant when P<0.05 and trending towards significance when P<0.10.
where Y ijklmn=observation on cow j with treatment i, order of treatment k and period l, μ=overall mean, αi=fixed effect of treatment i (G, M), βk=effect of order k (I, II) of applying treatments, γl=fixed effect of period l (1, 2, 3, 4), δm=continuous covariate day of experiment m, SUB(β)jk=random effect of cow j within order k, εijklmn=random residual error.
Results and Discussion
Several studies have examined the supplementation of protein-rich forages (pasture, grass silage and legume silage) with maize silageReference Mitani, Takahashi, Ueda, Nakatsuji, Kondo and Okubo14–Reference Stockdale16. Other trials have compared different carbohydrate sources in dairy cow dietsReference Eriksson, Murphy, Ciszuk and Burstedt3, Reference Mogensen and Kristensen4, Reference Castillo, Kebreab, Beever, Barbi, Sutton, Kirby and France17–Reference Visser and Hindle19. However, surveys comparing grain and maize silage as supplements are rare. In the present study, cows were fed grass–clover silage-based diets, which were supplemented either with 1 kg of a grain mixture plus 0.5 kg hay (treatment G) or 2.1 kg maize silage (treatment M) on a DM basis. Dietary energy content was similar in both treatments (6.4 and 6.3 MJ net energy for lactation (NEL) kg−1 DM in treatments G and M, respectively). Generally, it is assumed that at low levels of supplementation, the energy source has only little effect on feed intake, milk yield and milk compositionReference Leaver, Garnsworthy and Wiseman20. Hence, most studies cited in the present paper used much higher supplementation rates than the study at hand.
Feed and nutrient intake
Maize silage and grain differ in their nutritional characteristics, especially in their contents of starch and fiber (Table 1). Generally in maize silage, the B1 fraction (starch) of the CNCPS (Cornell net carbohydrate and protein system) is lower as compared with cereal grain (40 versus 70% of carbohydrates) and the B2 fraction (hemicellulose and cellulose) higher (45 versus 20% of carbohydrates)Reference Gruber, Stögmuller, Taferner, Haberl, Maierhofer, Steiner, Steinwidder, Schauer and Knaus21, Reference Taferner22.
Daily forage dry matter intake (DMI) (grass–clover silage, hay and maize silage) was significantly higher in treatment M (15.7 versus 14.6 kg DM). Grass–clover silage intake was 0.5 kg DM lower in treatment M as compared with treatment G (Table 2). However, in the present study, DMI and energy content of supplements were different and no data for a treatment group without supplementation were available. Hence, with the present results, no reliable statement for substitution rates can be made. Intake of the concentrate mixture averaged 3.0 kg DM in both treatments. Therefore, percentage of dietary forage intake was rather high (P<0.01) at 79% of total daily DMI in treatment G and 85% in treatment M, taking into account an average milk yield of 23.2 kg per day.
Table 2. Effect of maize silage versus grain-mixture on daily feed and nutrient intake.
No difference between treatments was observed for total daily DMI, estimated intake of NEL and uCP (Table 2). Contrary to uCP intake, intake of CP was significantly lower in treatment M, which resulted in different dietary CP concentrations (15.7 and 14.9% of DM in treatments G and M, respectively, P<0.01). RNB is an estimated value for the protein supply of the ruminal microbes and is calculated as CP minus uCP divided by 6.2511. RNB was significantly lower in treatment M (29 versus 40 g), which reflected the lower CP intake in treatment M (Table 2).
In the study in question, protein balance, defined as uCP intake as percentage of requirements for maintenance and observed milk production11, was markedly positive (+13%) in both treatments. Reasons for the positive protein balance could be the following: provision of grass–clover silage ad libitum for the whole day, a slight overestimation of the uCP feed value as well as discrepancies in the estimation of the protein and energy requirements of the GfE11.
Overall hay intake was, according to plan, significantly higher in treatment G (1.2 versus 0.7 kg DMI); however, intake was considerably lower in both groups (6 and 4% of total daily DMI in treatments G and M, respectively) than initially planned and expected. Although treatment G was supplemented with hay at morning feeding, daily intake of crude fiber, NDForg and ADForg was on average 3–4% higher in treatment M, which was statistically significant (Table 2). This reflected mainly differences in the fiber content of the two energy supplements (fiber intake from maize silage was 50% higher than fiber intake from grain mixture plus hay consumed at morning feeding).
Milk production and feed efficiency
Several authors have dealt with the effect of different carbohydrate supplements on milk yield and milk composition. Inconsistency in the results is often attributed to differences in the source of carbohydrate, the quantity of supplementation and the forage quality. Generally, it is assumed that good roughage quality is likely to increase substitution rates and reduce production responses to supplementsReference Bargo, Muller, Kolver and Delahoy23, Reference Faverdin, Dulphy, Coulon, Verite, Garel, Rouel and Marquis24.
Results of the present study concerning milk performance and feed efficiency are shown in Tables 3 and 4, respectively. In the present study, daily milk yield was unaffected by energy supplementation, which is in agreement with results of Castillo et al. Reference Castillo, Kebreab, Beever, Barbi, Sutton, Kirby and France17 and Visser et al. Reference Visser, Van der Togt and Tamminga25, who fed concentrate mixtures with different structural and non-structural carbohydrates and did not observe any difference in DMI, energy intake or milk yield. Contrary to this, Moran and CrokeReference Moran and Croke7 fed different amounts of maize silage or wheat to pasture-fed dairy cows at the same level of supplemental energy and observed numerically higher milk yields when feeding wheat. However, this might have resulted from the trend for higher supplemental ME intake and lower calculated pasture substitution in the wheat-fed treatment.
Table 3. Effect of maize silage versus grain-mixture on daily milk yield and milk composition.
Table 4. Effect of maize silage versus grain-mixture on feed utilization.
1 Concentrate mixture and grain mixture.
2 N in milk in % of N intake.
3 (NEL intakeforage −NEL requirementmaintenance)/NEL requirement kg−1 milk.
4 Intake as a percentage of requirements for maintenance and milk production.
In the study at hand, milk protein content was unaffected by treatment, which agrees with results of Huhtanen et al. Reference Huhtanen, Jaakkola and Saarisalo6, but is contrary to those of Moran and CrokeReference Moran and Croke7. The present study showed a significantly higher milk fat content in treatment M (P=0.01). However, this difference was only a trend for milk fat yield (P=0.09). It is commonly agreed that the amount of ruminally fermentable fiber and consequently the production of acetic acids have a major impact on milk fat content. Nevertheless, several studies comparing fibrous versus starchy supplements did not report any effect on milk fat percentageReference Huhtanen, Jaakkola and Saarisalo6, Reference Moran and Croke7, Reference Castillo, Kebreab, Beever, Barbi, Sutton, Kirby and France17, although NDF intake was significantly higher with fibrous carbohydrates.
In the present study, milk urea concentration was within the recommended range of 15–30 mg 100 ml−1 in both treatmentsReference Steinwidder and Wurm26. However, in treatment M, urea concentration was slightly but significantly lower (19.7 versus 20.8 g, P=0.04). Hristov and RoppReference Hristov and Ropp27 fed two diets supplemented either with ruminally fermentable non-structural carbohydrates (barley and molasses) or ruminally fermentable fiber (beet pulp, brewer's grain and corn) and found a significantly lower milk urea concentration for the ruminally fermentable fiber diet. However, whether the reduced urea content observed in the present study resulted from reduced protein intake, increased ruminally fermentable energy or improved ruminal N capture is unclear.
In general, the efficiency with which dietary N is utilized by dairy cows (N in milk in % of N intake) is low, ranging from 22 to 39%Reference Givens and Rulphin2, Reference Huhtanen, Nousiainen, Khalili, Niemeläinen and Topi-Hulmi28, Reference Castillo, Kebreab, Beever and France29. It is generally agreed that low dietary N efficiency is partly related to high dietary CP content and an unbalanced ruminal energy and protein supply, and that it can be improved through the supplementation of feedstuffs rich in carbohydratesReference Cohen, Stockdale and Doyle30, Reference Tamminga31. The present study showed a slight tendency (26.1 versus 25.5%, P=0.08) for an improved dietary N efficiency in treatment M. Unfortunately, in the present study, herd size was too small to allow for a third, non-supplemented treatment group; hence no statement for a diet without supplementation can be made.
With the data at hand, it is difficult to make a reliable statement on changes in body weight and BCS, because cows were in different stages of lactation and cows were not kept separated from water troughs and grass–clover silage before weighing. Generally, treatment did not show a significant effect on body weight, which averaged 649 kg (P=0.56), and BCS, which averaged 3.2 points in both treatments (P=0.83). The lack of differences in body weight and BCS between treatments may be partly because of the rather low amount of supplements fed and partly because of the lack of difference in DMI and NEL intake.
Conclusions
The present study gives an indication that with even small amounts of supplemental energy nutrient intake, protein supply and milk composition can be influenced. Although only 11% of total DMI was fed as maize silage, CP intake was significantly reduced, RNB and milk urea concentration significantly decreased, and dietary N efficiency slightly improved as compared with the grain-supplemented treatment. However, with the present trial, no reliable statement can be made whether these improvements are due to reduced CP intake, improved ruminal energy supply or greater ruminal N capture. Milk yield was not negatively influenced by feeding maize silage instead of grain; however, the dietary proportion of supplements was probably too small to observe responses in milk production. It is concluded that inclusion of small amounts of maize silage instead of grain did not seem to have any negative effects. However, higher supplementation rates and lower dietary CP contents might lead to greater differences.
Acknowledgements
The authors would like to thank the Agricultural High School Ursprung, A-5161 Elixhausen, for providing its stable and dairy herd, the farm staff for their contribution to the experimental work, Dr Andreas Steinwidder of the HBLFA Raumberg Gumpenstein for his contribution to experimental planning and ration formulation and Dr Roswitha Baumung for her statistical assistance. For funding the project, the authors would like to express their thanks to the Austrian Ministry of Agriculture and Forestry, Environment and Water Management, the Salzburger Landesregierung, the Raiffeisenverband Salzburg and Bio Austria.
