Effects of partial replacement of barley with rapeseed oil or birch wood in comparison to barley and oats on the performance and blood metabolites of lactating cows

Barley was partially replaced with either rapeseed oil (RSO), unextracted steamexploded birch wood (Wl) or extracted steam-exploded birch wood (W2), and fed to 10 lactating cows (5 multiparous and 5 primiparous cows) in comparison to rolled barley and oats. The experiment was designed in two balanced 5x5 Latin squares. The experimental periods lasted for4 weeks and the cows were fed with a fixed amount of 8 kg (air dry basis) concentrate in two equal portions. Timothy dominant unwilted grass silage was provided ad libitum. The nutrient digestibility, blood metabolites, milk yield and composition were studied. Grass silage intake decreased by replacing barley with RSO and increased with W 1 and W2. Oat feeding also increased silage intake compared to barley feeding. Oat feeding decreased the digestibility of the organic matter(OM) compared to barley. No significant difference was observed between diets in plasma concentrations of insulin, free fatty acids, acetoacetate, 6-hydroxybutyrate and glucose. Replacing part of the barley with RSO increased milk yield by 1.3 kg and 0.87 kg compared to oats and barley feeding, respectively. Replacement of barley by low energy birch wood decreased milk yield. Both RSO and oat feeding decreased milk protein content. Milk urea concentration was elevated by oats and depressed by RSO and birch wood. RSO and oat feeding decreased the short chain fatty acid content of milk, with a concomitant increase in stearic and oleic acids.


Introduction
Grain has been partly replaced or supplemented by fats and oil for dairy cows with the aim of increas- ing the energy density of the diet.Recently, fats have been included in an attempt to alter the fatty acid composition of milk, thereby improving its nutritional and physical properties.Direct feeding of oil/fat to a cow on the farm creates day-to-day handling problems due to their organoleptic cha- racter which limits their acceptance.An alternative approach to dietary manipulation of the fatty acid composition of milk fat is to feed lactating dairy cows with oats (Lampila and Heikkilä 1986; Martin  and Thomas 1988), because oats contain more lipid than barley or wheat, and its fatty acids are highly unsaturated (Morrison 1977).
The digestibility of lignocellulosic materials like wood and wood products has been improved by chemical or steam treatments (Tarkow and Feist 1969;Bender et al. 1970; Reddy et al. 1972).Steam explosion is a method, where wood chips are steamed using saturated steam of over 200°C and a high speed discharge of steamed wood into the air, utilizing the concept of making popcorn.After the process, the hemicellulose of hardwood is hydrolysed into a low molecular substance and it becomes water soluble.Cellulose is easily decomposed by enzymes, lignin is degenerated, and part of it be- comes soluble (Takigawa 1987).A 49-60 % digestibility has been obtained for steam-exploded aspen wood and white birch (Terada et al. 1986, ref.Takigawa 1987, Tuori 1990).Replacement of the dry matter of com silage by 20 % of steam-exploded aspen wood had no influence on milk yield or milk composition (Fisher 1980).In Japan, when steamed white birch comprised 15 % of dairy cow ration, it had no influence on yield, but 30 % of wood tended to decrease production (Miyamoto et  al. 1986,ref Takigawa 1987).
In Finland, birch wood is used as raw material by xylitol manufacturers.Using the steam explosion method for birch wood it would be possible to separate hydrolysed xylans for further processing.In this experiment, the effect of substituting barley with rapeseed oil or steam-exploded birch wood and water-extracted steam-exploded wood (xylans removed) on milk production was studied.The effect of oats and rapeseed oil feeding on the fatty acid composition of milk fat was compared to that of barley feeding.

Experimental design and animals
Five primiparous (PP) and five multiparous (MP) Ayrshire cows that had calved on average 46 days before the start of the experiment, were used in a two balanced 5x5 Latin square design (MP and PP cows in separate squares).Each experimental period lasted for four weeks.Data from weeks 3 and 4 were used for the statistical analysis.One cow ceased eating before the completion of the fifth period, and the data on the oat diet was calcu- lated only on four observations.

Feeds and feeding
The concentrate feed was formulated as shown in Table 1.The first two grain feeds were rolled bar- ley (barley) and rolled oats (oats).An equal amount (w/w) ofrolled barley was replaced either by 0.4 kg rapeseed oil (RSO), 2 kg unextracted steam-explod- ed birch wood (Wl), or 2 kg water-extracted steam-exploded birch wood (W2).Wl and W 2 were lignocellulosic materials which were steam- exploded (Stake-process) under a pressure of 14 kg per cm 2 at 210°C for 5 min to break the cell wall material and to facilitate digestion (Bender et al.  1970).Degraded xylans were removed from W 2 by water extraction.Both Wl and W 2 were in a form of finely ground wad and were easy to mix with barley.RSO was from a double zero (00) variety of Brasica campestris.All cows received 1.3 kg (16 %ofconcentrate) soya bean meal as pro- tein supplement.Each ingredient replacing rolled barley was mixed manually with the barley before each meal.
The cows were fed individually twice a day at 0600 h and 1400 h.Concentrate was offered 8 kg (air dry) d' 1 and first cut unwilted grass silage was given ad libitum throughout the trial.All diets were supplemented with a mineral mixture fed at a rate of 250 g d l .The apparent digestibility of feeds was determined on five multiparous cows using chro- mium oxide (Cr 2 0 3 ) as an external marker.Cr 2 0 3 impregnated paper (20 g Cr per day in two por-    tions) was introduced to the rumen using a stomach tube for eleven days on each period.The efficiency of utilization ofmetabolizable energy (ME) in milk production was calculated on the basis of digestible organic matter (5 MP cows).

Recording and sampling
The cows were weighed on two consecutive days at the beginning of the experiment, as well as in the middle and at the end ofeach period.The daily feed offered and refusals were recorded.The weekly feed samples were collected and pooled into one sample per diet per period.The pH and dry matter (DM) content were determined from the weekly sample of grass silage for quality control purposes.
An extra weekly sample was frozen at -18°C for the determination offermentation characteristics.
The milk yield was recorded at each milking and sampled at four milkings during the two last weeks of each period for the determination of fat, protein, urea, lactose and fatty acids.Blood samples were collected from the jugular vein (Vena jugularis) into vacutainers on day 24 at 0,2, 4 and 6 h post- morning feeding.Plasma was separated by centri- fuging at 2000 rpm for 10 min and deep frozen at -18°C until analysed for insulin and free fatty acids (FFA) concentration.Samples for 13-hydroxybutyrate (BHBA) and acetoacetate (AA) analyses were immediately precipitated with 2 ml of 0.6 M perchloric acid to avoid loss of AA and kept at -18°C (Työppönen and Kauppinen 1980).Faecal grab samples were collected rectally twice a day at 0800 and 1500 h for 6 days, starting on the 23 day, and stored at -18°C.At the end of the collection period, samples were thawed and pooled into one sample per treatment.

Chemical analysis
Feeds and milk samples were analyzed as described previously (Tesfa et al. 1991).Faecal samples were analyzed for DM by drying at 105°C to con- stant weight, and for organic matter (OM) by ashing in a muffle furnace at 550°C for 24 h.Fresh faecal and dried feed samples were determined for crude protein by the Kjeldahl method.Dried faecal and feed samples were used for the determination of neutral detergent fibre (NDF) and acid detergent fibre (ADF) (Goering and Van Soest 1970).Chromium in faeces was determined by atomic absorption spectrophotometry (Williams et al.  1962).
The blood glucose concentration was analyzed by the enzymatic glucoseoxidase and FFA by an enzymatic colorimetric method (Wako chemicals).Insulin was analyzed by radioimmunoassay using a double antibody solid phase technique (Phadeseph Insulin RIA).AA and BHBA were analyzed accord- ing to Hansen and Freier (1978) with a Gilford 3500 analyzer at a temperature of37°C and a wave- length of 340 nm.Milk urea was determined with an enzymatic colorimetric system according to Rajamäki and Rauramaa (1984).

Statistical Analysis
The data were analyzed by the least square analysis of variance using: Model I.For apparent digestibility of nutrients: Y ijkl =p+C+D.+ P k + e ijk| , where pis the mean, and C, D, and P are cow, diet and period effects, resprctively, and e jk| is the residual effect.Model 11.For production parameters: Y i Jklm = M + S,+ C(S,) + P k + D j+ (SP)]k+ (SD), j + where C, P, and D are as above.S is the square (parity) effect and SP is the square*period effect, SD is the square*diet effect and e jjklm is the error.Model 111.Blood metabolites were analyzed using a split-plot analysis of variance using the following model: where T m is the sampling time effect, e ikj is the main plot error, T m (C j (S 1 )) is the interaction between cow and sam- pling time; (PT)km, is the interaction between period and sampling time, (DT).m , is the interaction between diet and sampling time, and e ikj|m is the sub-plot error and p is the constant.The differences between treatments were tested with Tukey's test (Steele and Torrie 1960).Values given in Tables 3-8 are least square means oftreatments.

Results
The mean chemical composition and feed values of feeds used are given in Table 2.As can be seen from Table 3, the total DM intake in barley and oat diets was equal.However, the cows fed with oats had more leftovers of concentrate.Silage intake decreased by 0.5 kg when RSO replaced part of the barley and increased by 0.6 kg with oat feeding.
The main difference in feed intake was the increased (1.1 kg DM) silage intake by cows fed on W 2 increasing their total DM intake by 1 kg com- pared to barley or oats fed cows.Multiparous cows consumed 2.1 kg DM more than the PP cows.

Feed apparent digestibility
The digestion coefficients of dietary nutrients are given in Table 4. Oat feeding decreased (P<0.05) the digestibility of OM compared to barley.Crude protein (CP) digestibility was higher (P<0.05) for the oat group than for the RSO group.(MAFF 1975); FFU, fattening feed units = 0.7 kg starch (Salo et al. 1982).Table 5.Effects of dietary treatment and cow status on blood acetoacetate (AA), 13-hydroxybutyrate (BHBA), glucose and free fatty acids (FFA) (mmol I 1) and insulin (mU T 1).

Blood metabolites
Although there was some variation in blood meta- bolites between cows fed on the different diets, none of these differences were statistically significant, except for BHBA on Wl and W 2 (Table 5).
There was a slight increase in AA, FFA, BHBA and insulin concentration on RSO.Insulin concentrations slightly decreased on W 1 and W 2 diets.The blood metabolites of both multiparous and first lactation heifers were generally similar.However, blood insulin (23.1 vs 37.8), blood FFA (0.151 vs 0.176) and BHBA (0.786 vs 0.878) were lower for PP cows than for MP cows.No significant change in body weight (0.22 vs 0.10 kg d 1) was observed between cow parity.

Milk yield and composition
The overall milk yield of cows fed with oats, Wl and W 2 is lower as compared to barley fed cows.A more significant (P<0.01)difference was observed between cows fed on oats and RSO (Table 6).Slightly more milk fat was produced by animals fed on Wl as compared to those fed on W 2. Replace- ment of barley with RSO decreased milk fat con- tent and increased milk yield by 5.4 % (0.87 kg d 1) compared to barley, but this increment was not sta- tistically significant (Table 6).A decrease (P<0.05) of milk protein content was observed in cows fed with oats and RSO compared to those fed with bar- ley.The milk urea content of cows fed oats was increased (PO.01); those fed with Wl and W 2 showed a decrease (P<0.01)compared to those fed with barley.There was no significant difference between those fed with RSO and barley.Similar to their total feed intake, milk yield and FCM yield was significantly (P<0.001) higher for MP cows compared to PP.The efficiency of utilization of ME in milk pro- duction is given in Table 7.There was no signific- ant (P>0.10)difference in efficiency of utilization Table 6.Effects of diet and cow status on daily milk yield (kg) milk composition (g kg '), and live weight changes (LWC). of ME in milk production between cows fed with oats and RSO, but cows fed with oats were more efficient than those fed with barley, when cor- rections were made for changes in live weight and for maintenance.Differences in body weight change were not significant (P>0.10) between treatments.However, the trend suggests that cows fed on barley were mobilizing more tissue than cows on other diets.

Milk fat fatty acids
There was no significant (P>0.10)difference between diets on certain fatty acids such as C 4 C 6:o' C 10:l' C 12;0 C .4:PC 18:3 alld C 20:0 ( Table 8)' Concentrations of C l 2 C ]4 .o and C l 6 0 decreased on oats and RSO feeding, but no apparent difference was observed between barley, W 1 and W 2 feeding on these fatty acids.Substitution ofRSO for barley significantly decreased (PO.OOl) C l6O and increased C ]g o and C |B ] fatty acids compared to barley.

Discussion and conclusion
Two hypotheses were proposed in this experiment.
The first hypothesis was that if part of the grain in the ration is replaced by fat/oil, the daily milk yield will not be higher than with the control concentrate feed, or it will be lower than with the control concentrate feed due to the negative effect of fatty acids on the rumen environment.The second hypothesis was that low energy lignocellulose feeds can sustain milk production when grass silage is provided ad libitum.The results confirmed however, that the daily milk yield can be increased by replacing part of barley with RSO but not with birch wood.

Feed intake and digestion
We observed increased silage DM intake and decreased concentrate intake by cows fed with oats compared to barley fed cows.This is contrary to the findings ofMartin and Thomas (1987) with lacta- ting milking cows.The reduced silage intake by cows fed on RSO agrees with the work of Steele (1985) where soya bean oil or crushed soya beans were added to silage and fed with supplementary protein.W 1 and W 2 fed cows consumed more silage to compensate the low energy content of the concentrate diet.Although there was a substantial difference in energy content of the test feeds, it seems that the difference in energy intake was com- pensated by higher silage intake by cows fed on W 1 and W 2. The decrease in apparent digestibility of OM by cows fed with RSO was expected due to its toxic effect on cellulolytic organisms.However, the reduction in OM digestibility with RSO feeding observed in this trial was not significant.The absence of a negative effect on nutrient digestibility with RSO feeding to cows is in agreement with the reports of Palmquist and Conrad (1978).
Under conventional feeding systems, the feeding of wood industry by-products to milking cows is not customary.A basic argument against its use is its low digestibility.Contrary to this, the apparent digestibility coefficient of OM was higher in cows fed W 1 than in those fed with oats.In fact, it was observed that RSO and oats depressed the digestibility of NDF and ADF more than W 1 did .The lower digestibility of OM, NDF and ADF observed with oats compared to barley and RSO feeding was not expected.We assume that this depression can be attributed to the exceptionally high fibre content of the oats used in this trial.

Blood metabolites and milk yield
It is expected that silage based diets that contain less fibre than hay based diets would interact with RSO or fat from oats to increase propionate production in the rumen (Steele and Moore 1968) and to increase blood glucose.This increased out- put of glucose then concomitantly increases plasma insulin which redirects fat synthesis from mammary tissue towards adipose tissues (Vernon 1980).However, in the present study neither glucose nor insulin concentration was significantly increased on RSO or oat feeding.Regarding plasma glucose concentration on RSO feeding, the result is contradictory to the previous observation with increasing level of rapeseed oil feeding to cows (Tesfa et al. 1991).There were no significant changes in the BHBA or AA of blood plasma either.Therefore, the slight depression in the milk fat content seems not to have been caused by an alteration in concentration of blood metabolites on RSO diet.The observed slight increase in BHBA on RSO diet probably resulted from the oxidation of dietary fatty acids (Steele 1984).The increase in BHBA, therefore, cannot be attributed to any pathological condition as the cows were in a positive energy balance during the observation.In the case of W 1 fed cows, the elevated BHBA concen- tration could be due to high butyrate production in the rumen from the xylose contained in the birch wood.
An increased milk yield with oats compared to barley at a similar level of DM intake by the cows has been reported (Heikkilä et al. 1988; Lampila and Heikkilä 1986; Martin and Thomas 1987;   Moran 1986).In our study, despite similar DM intakes, cows fed with oats produced less milk than barley and RSO fed cows.Moss and Frier (1981)   also reported increased FCM yield on oat diet, al- though oats is inferior to other grain on the basis of its ME content.In the present study, oats depressed milk, FCM and protein yields compared to RSO.
RSO diet slightly increased milk and FCM yield and sustained milk protein yield but decreased milk fat and protein contents compared to barley.The reported depression in milk fat content with oats (Martin and Thomas 1987) and with canola oil (Beaulieu et al. 1990) is in agreement with our observation with oat and RSO feeding.
The feed values of non-conventional feeds such as W 1 and W 2 are low compared to grain for high yielding milking cows.Owing to this fact, reduced milk yield was inevitable with W 1 and W 2 compar- ed to both grain sources.The slightly higher or sus- tained milk fat content of cows fed with W 1 and W 2 can be viewed as a result of high intake of silage and higher fibre content of the birch wood- barley concentrate.This would have provided high output of acetate which in turn would be used to enhance de novo synthesis of short chain fatty acids of milk fat (Moore and Christie 1979).However, the short chain fatty acid concentration of milk fat from cows fed with W 1 and W 2 was lower compar- ed to that of barley fed cows.The milk protein content has been found to res- pond to diets containing high fat negatively (Depeters et al. 1987, Steele 1985).Compared to con- trol diets, additional fat containing diets may cause a decrease in milk protein content but an increase in milk yield (Banks et al. 1980), the protein yield remaining unchanged.Contrary to these findings, Steele et al. ( 1971) observed an increase in milk protein yield on soya oil feeding.
When cows consumed complete diets containing 600 g kg' 1 DM as oats or barley, the oat diet reduced the milk protein content (Moran 1986).Martin  and Thomas (1988) also observed decreased milk protein yield and content when cows were fed with oats.Although the effect of oats on the milk protein content varies slightly, our observation is in agree- ment both with the work of Martin and Thomas (1988) and that of Moran (1986).However, the higher negative effect of oats than RSO on the milk protein yield and content was not expected.The increase in milk urea content from cows fed with oat diet can be explained as a result ofhigher intake of crude protein.

Milk fatty acid composition
The effects of oats on the fatty acid composition of milk fat resembled those observed in response to dietary fat containing diets (Storry 1981; Clap- perton and Banks 1985; Tesfa et al. 1991).De- pression in short-and medium chain fatty acids on oats and RSO was expected as the dietary unsatura- ted fatty acids are thought to reduce de novo synthesis of C 6 0 -C l 6 0 fatty acids within the mammary gland through a reduced supply of acetate, which apparently inhibits mammary acetyl-CoA decarbo- xylase (Clapperton and Banks 1985; Mattos and   Palmquist 1974).The increase in the proportions of C |B o and C in milk fat on RSO and oat diets is due to the hydrogenation of unsaturated fat in the rumen and subsequent intramammary desaturation of the C lBO formed (Banks et al. 1980; Christie  1981).RSO raised the proportion of Cl8:1 trans- fatty acid in milk fat, similar to the reports ofKen- nelly (1987) with diets containing full-fat canola seed.The elevated trans-fatty acid is a result of the hydrogenation of oil in the rumen as has been observed in our previous study (Tesfa et al. 1991).
Based on the results obtained in this study, both oats and RSO diets showed a substantial reduction in C, nn -€,<" fatty acid content of milk fat with a iu:u io:u J concomitant increase in C lBo and C lg| .However, oat feeding and replacing barley with RSO was observed to result in decreased milk protein content.Although there is a need, both nutritionally as well as technologically, for changing the fatty acid composition of milk fat, the emphasis should not be put on replacing grain with fat.Because replacing grain with fat reduces the availability of fermentable carbohydrate which could be used as energy sources by rumen microbes.Therefore, oil should be used as a supplement to the basal diet rather than as a replacement.

Table 2 .
Chemical composition of experimental feeds.
See Table2.for NDF and ME.CP, crude protein.

Table 4 .
Digestibility coefficients of dietary nutrients.

Table 8 .
Fatty acid composition of milk fat, g 100 g' 1 of total fatty acids.