Use of rapeseed and pea grain protein supplements for organic milk production

Grass-red clover silage was fed ad libitum. In experiment 1 a duplicated 4 x 4 Latin square design was used. A mixture of oats and barley was given at 8 kg (C). Three isonitrogenous protein supplements were a commercial rapeseed meal (218 g kg dry matter (DM); RSM), crushed organic field pea (Pisum sativum L.) (452 g kg DM; P) and a mixture of pea (321 g kg DM) and organic rapeseed (Spring turnip rape, Brassica rapa L. oleifera subv. annua) (155 g kg DM; PRS). Cows on P and PRS diets produced as much milk as cows on the RSM diet. Milk yield was higher but protein content lower with PRS diet than with diet P. In experiment 2 a triplicated 3 x 3 Latin square design was used. A mixture of oats (395 g kg ), barley (395 g kg ) and a commercial heat-moisture treated rapeseed cake (210 g kg 1 ) was given at 8 kg (RSC). The second diet (ORSC) consisted (g kg) of oats (375), barley (375) and cold-pressed organic rapeseed cake (250). The third diet (RSCO) consisted (g kg) of oats (395), barley (395) and commercial heat-moisture treated rapeseed cake (250) and additional rapeseed oil (0.38 kg) to balance fat content between ORSC and RSCO diets. There was no dietary effect on the yield of energy corrected milk. Milk yield was higher with RSCO diet compared with other diets.


Introduction
Organic dairy farming is increasing in Finland, creating a demand for feed evaluation and for-mulation of diets which can be applied to organic milk production.More information is needed concerning the utilization and relative feeding values of organically cultivated crops in order Khalili, H. et al.Protein supplements for organic milk production to improve organic dairy cow feeding and feed budgeting.In organic farming systems mineral fertilizers are compensated by recycling of nutrients and nitrogen is supplied through biological nitrogen fixation.Organic farms can attain the same level of productivity as conventional farms, provided that legume-dominated swards account for 3-40% of the cultivated area and are correctly positioned in the rotation (Granstedt 1992).In organic milk production, cows are fed mainly with home-grown feeds with a high proportion of forage in the diet and a mixture of grass-clover silage is often used as a basal forage in Finland.Feeding red clover-grass silage has increased silage intake (Heikkilä et al. 1992, 1996, Vanhatalo et al. 1995) and duodenal microbial N flow (Vanhatalo et al. 1995) and consequently improved milk production compared with grass silage of similar digestibility.Although red clover-grass silage is a suitable forage for dairy cows, duodenal infusion of casein increased milk yield when cows were given redclover grass silage diet as the sole feed (Khalili et al. 1995a).Protein supplementation has also increased milk production of high yielding dairy cows given grass silage-concentrate diets (Castle 1982, Thomas and Rae 1988, Tuori 1992, Huhtanen 1998).
Field pea is a familiar grain legume in Finnish conventional and a most important grain legume in organic farming systems.In recent decades it has been neglected as a feed for dairy cows.There are only a few results available concerning the effects of pea in milk production.Rapeseed is a valuable crop.Rapeseed meal and cake are the most common protein supplements in Finnish dairy cow diets and have produced considerable production responses (Tuori 1992, Huhtanen 1998).Rapeseed is, however, a newcomer in organic farming systems and it has proved difficult to provide adequate crop protection within the constraints of organic farming.Cold-pressed rapeseed cake, a by-product from rapeseed oil production, is a new protein feed, which has not been evaluated for dairy cow feeding.Native Finnish commercial feeds, such as rapeseed products, are permitted in organic animal production, but their contribution should not exceed 15%.
Two experiments were conducted in order to investigate the effects of conventional or organic protein supplements on milk production.The results here are discussed in relation to the Finnish amino acids absorbed (AAT) -protein balance in the rumen (PBV) system which Tuori et al. (1998) concluded to be the most accurate protein evaluation system for predicting differences in dietary protein value for dairy cows.

Experiment 1
The experiment was conducted on Siikasalmi research farm of University of Joensuu in Liperi with 8 mid-lactating (80 ± 38 days in milk) Finnish Ayrshire cows, of which four were in their first lactation and the others in their third lactation.Cows were individually fed at 0730 and 1530 and milked at 0600 and 1600.The experiment was conducted as two balanced 4 x 4 Latin squares, each having 21 day periods, comprised of 14 days for dietary adjustment and 7 days for data collection.Cows were divided into two blocks according to parity and within blocks cows were randomly allocated treatments.The four concentrates including 250 g of a mineral mixture were given at 8 kg (as fed) and consisted of a mixture (1:1) of oats and barley (control diet) and three isonitrogenous protein supplements of a commercial rapeseed meal (RSM diet), crushed field pea (Pisum sativum L.) (P diet) and a mixture of field pea and crushed rapeseed (Spring turnip rape, Brassica rapa L. oleifera subv.annua) (PRS diet).A commercial mineral mixture contained (g kg -1 ) Ca (160), P (64), Mg (80) and Na (90) and 20 g d -1 of NaCl.Oats, barley, field pea and rapeseed were organically cultivated on Siikasalmi research farm.Organic crop production is based on Council Regulations Vol. 8 (1999): 239-252.(EEC no 2092/91).Pea grain and rapeseed were crushed with a hammer mill to pass through 4 and 2 mm sieves, respectively.Rapeseed meal was solvent extracted (Raisio Feed Ltd).Composition and calculated feeding values of experimental concentrates are shown in Table 1.Second cut grass-red clover silage was fed ad libitum, and the daily portions were adjusted to provide refusals of 50 to 100 g kg -1 of the amount offered.On Siikasalmi research farm, herbage dry matter from a second cut contained on the average 60% grass (timothy (Phleum pratense) and meadow fescue (Festuca pratensis)) and 40% clover, mainly as red clover (Trifolium pratense).Herbage was ensiled after wilting (6 h) with a formic acid based additive (6 l t -1 ) into a tower silo.

Experiment 2
The experiment was conducted on Siikasalmi research farm as three 3 x 3 Latin squares, each having 21 day periods comprised of 14 days for dietary adjustment and 7 days for data collection.The experiment was conducted with 9 midlactating (79 ± 40 days in milk) Finnish Ayrshire cows, of which three were in their first lactation and the others were in their fourth lactation.Cows were divided into blocks according to parity (one block for first lactating) and the remaining 6 cows were divided into two blocks according to milk yield.Within the blocks cows were randomly allocated treatments.Three isonitrogenous concentrates including 250 g of a mineral mixture were given at 8 kg (as fed).Experimental concentrates consisted of a mixture of oats and barley supplemented either with commercial rapeseed cake (control RSC diet), coldpressed rapeseed cake (ORSC diet) or a mixture of commercial rapeseed cake and rapeseed oil (RSCO diet) (Table 2).A commercial mineral mixture contained (g kg -1 ) Ca (160), P (64), Mg (80) and Na (90) and 20 g d -1 of NaCl.Oats, barley and rapeseed were organically produced as described earlier.On-farm home made rapeseed cake (ORSC) was made from whole rapeseed by pressing oil from the seed with a screw press (Tabypressen 40, Sweden).The fat content of the remaining rapeseed cake was about half that of whole rapeseed.Commercial rapeseed cake (RSC) was manufactured by removal oil by pressing and a subsequent heat-moisture treatment (Öpex R , Mildola Ltd).Rapeseed oil was produced by Raisio Feed Ltd.Grass-red clover silage was fed ad libitum as described earlier.

Measurements and analytical procedures
Milk yields and silage intakes were recorded daily.Silage dry matter (DM) content was determined by oven drying at 105 o C for 24 h, and DM content was corrected for volatile losses according to Huida et al. (1986).Feed samples were collected on days 15-19 of each period and pooled within the period.During the five days of each collection period, clean faecal grab samples were taken from four (experiment 1) and The values derived from feed tables (Tuori et al. 1995).
from six (experiment 2) cows at 0600 and 1600, and pooled within each cow and frozen.Pooled samples were subsequently thawed and dried at 60 o C and stored at room temperature for chemical analysis.Samples of feeds and faeces were analysed for organic matter by ashing at 600 o C for 12 h, nitrogen (Kjeldahl-N) and neutral detergent fibre (NDF) (Robertson and Van Soest 1981).Fat was extracted with diethyl ether (AOAC 1990) after boiling for 1 h in 3 N HCl.
Silage fermentation quality was determined as described by Huhtanen and Heikkilä (1996).Diet digestibility was measured using acid insoluble ash (AIA) as an internal marker (Van Keulen and Young 1977).Milk samples were taken during four consecutive milking on days 17 and 18 of each period and analysed for fat, protein and lactose by an infra-red milk analyser and milk urea content as ammonia (McCullough 1967) after hydrolysis by urease.

Calculations and statistical methods
Intake of metabolizable energy (ME) and supply of amino acids absorbed from the small intestine (AAT) were calculated according to Finnish feed tables (Tuori et al. 1995).ME intake was also calculated based on intake of digestible organic matter (DOM) (AIA method) assuming ME content of 16 MJ kg -1 DOM (MAFF 1984), and corrections were made for the difference in the fat intake.Data from both experiments were analysed by SAS Systems for Linear Models for a Latin square design (Littell et al. 1992).The model was: where µ is the overall mean, B, C and P are the random effects of block, cow and period, respectively, D is the fixed effect of treatments and f ijlkm is the random error term.The model used for calculating digestibility data from experiment 1 did not include block because there was only one square.In experiment 1, treatment effects were further separated into single degree comparisons using contrasts.The three contrasts were: effect of common commercial protein supplement (C diet vs. RSM), comparison of RSM with organic protein supplements (RSM vs. P and PRS) and effect of inclusion of rapeseed with pea (P vs. PRS).In experiment 2, pair wise comparisons were performed based on the least significant difference (LSD T-tests) procedure.For experiments 1 and 2, two different but very common commercial rapeseed feeds in Finland were used to minimize the effects of different amounts of fat between experimental feeds.In experiment 1, a low fat RSM was used because pea had a low fat content.In experiment 2, a higher fat RSC was used since cold-pressed rapeseed cake had high fat content.RSCO diet was included in order to study the effects of higher fat content 12.9 13.5 12.9 RSC=concentrate (C)+rapeseed cake; ORSC=C+organic rapeseed cake; RSCO=C+rapeseed cake+rapeseed oil. 1 oil removed by pressing and a subsequent heat-moisture treatment (Öpex-treatment).
2 ecologically produced and cold-pressed.
3 oil supplement to balance differences in fat content between diets ORSC and RSCO 4 Neutral detergent fibre.
5 Amino acids absorbed from the small intestine.
of organic cold-pressed rapeseed cake compared with commercial rapeseed cake.

Experiment 1
The fat content of rapeseed was higher than other protein feeds (Table 3).Generally chemical composition of experimental feeds were typical.
There was no difference (P>0.05) in silage intake (Table 4).Among protein supplements supply of AAT was highest (P<0.05) and digestibility of OM and NDF were lowest (P<0.05) with RSM diet.Partial replacement of pea with rapeseed (PRS diet) decreased AAT supply (P<0.01),but improved fat digestion (P<0.001) compared with pea without rapeseed (P diet).
RSM diet increased (P<0.05)yields of energy corrected milk (ECM), milk and milk protein and milk urea content, but decreased (P<0.001)N utilization (N in milk N -1 intake) compared with C diet (Table 5).Milk protein content was higher (P<0.05) with RSM than with organic protein supplements.Replacing part of field pea with rapeseed (PRS diet) resulted in increased yields of ECM and fat and improved feed efficiency (P<0.05),but decreased milk protein and urea content compared with P diet.
Experiment 2 Chemical composition of rapeseed cake was typical (Table 6).The nitrogen content of organic cold-pressed rapeseed cake was lower, but the fat content was higher than that of commercial rapeseed cake.In both experiments, grass-red clover silages had low proportions of soluble N (Kjeldahl-N) in total N which is typical due to reduced proteolysis during ensiling of red clover-grass silage (Jones et al. 1995).
Silage intake was similar between commercial RSC and organic ORSC diets, but fat intake and digestibility were higher (at least P<0.05) with ORSC diet (Table 7).Inclusion of oil with RSC decreased (P<0.05)intakes of silage and consequently that of nitrogen and AAT, but improved nitrogen and fat digestibility.
Increased (at least P<0.05) nitrogen utilization due to a lower nitrogen intake with organic ORSC diet compared with RSC diet was the only 2 Calculated from the intake of digestible organic matter determined in cows using acid insoluble ash as an internal marker. 3,4Calculated using values from feed tables (Tuori et al. 1995). 5Protein balance value. 6NDF=neutral detergent fibre.
significant difference between these diets (Table 8).Adding oil with RSC increased yields of milk and lactose (P<0.05),but decreased protein content compared with RSC diet without oil.Decreased silage intake increased feed efficiency and N utilization with RSCO diet compared to RSC diet.

Discussion
Feed intake and digestibility

Experiment 1
Increased proportion of protein in the concentrate has been shown to increase silage intake (Thomas and Rae 1988).In recent studies a re-placement of energy supplement with rapeseed meal has increased silage dry matter intake 0.247 kg per kg RSM (Huhtanen 1998).Here the corresponding increase of 0.225 kg in silage intake per kg RSM was similar to that observed with grass silage-based diets in conventional milk production.However, in the present experiment this increment was not significant.With organic protein supplements silage intake was consistent with commercial RSM.According to Oldham and Smith (1982) and Thomas and Rae (1988), increases in dietary crude protein content generally increase dry matter digestibility.
Here RSM did not improve organic matter digestibility compared with C diet, which is in agreement with data of Tuori (1992) and Huhtanen (1998).Furthermore, RSM had slightly lower organic matter digestibility compared with the other two protein supplements.RSM has a In silage: pH 4.00; in dry matter (g kg -1 ): water-soluble carbohydrates 12.5; lactic acid 68.0; acetic acid 18.3; in total N (g kg -1 ): ammonia N 36.0;soluble N 487.In vitro cellulase digestibility of silage OM=0.699.
2 C=concentrate mixture of oats and barley. 3OM=organic matter; 4 NDF=neutral detergent fibre. 5Amino acids absorbed from the small intestine. 6Protein balance value, calculated using effective protein degradability (EPD) value 82 for silage and EPD values from feed tables (Tuori et al. 1995) for other feeds.RSC=concentrate (C)+rapeseed cake; ORSC=C+organic rapeseed cake; RSCO=C+rapeseed cake+rapeseed oil. 1 Calculated from the intake of digestible organic matter determined in cows using acid insoluble ash as an internal marker. 2,3Calculated using values from feed tables (Tuori et al. 1995). 3Protein balance value. 4NDF=neutral detergent fibre.

Khalili, H. et al. Protein supplements for organic milk production
lower D-value than other components of concentrates (Tuori et al. 1995).In harmony with earlier studies (Huhtanen 1998), RSM did not improve NDF digestibility.Higher NDF digestibility with field pea containing diets are probably due to changes in NDF composition rather than as a result of changes in rumen cellulolytic activity.Cell walls of barley and oats are poorly digested, whereas cell walls of pea are highly digestible.Although AAT supply (corrected for negative PBV) was more than AAT requirements for maintenance and milk production with the control diet, RSM diet slightly increased silage intake.One reason for increased silage intake with RSM could be the increased supply of AAT and an improved balance of AAT.Post-ruminal casein infusion has increased the intake of red clover silage more than ruminal infusion (Khalili et al. 1995b).Post-ruminal metabolism and/ or increased rumen NDF fill affected intake with duodenal infusion.According to Heikkilä et al. (1998) there was a close relationship between AAT content of the diet and silage intake.In spite of high fat content of crushed rapeseed, silage intake was not decreased with PRS diet compared to the control or P diet (P>0.05).Here inclusion of 1.07 kg DM (fat 418 g kg -1 DM) of rapeseed did not impair NDF digestibility.In an earlier study (Steele 1985) addition of 1.9 kg (oil content 194 g kg -1 ) crushed soybeans decreased silage intake compared with a silage and soybean meal diet.

Vol. 8 (1999): 239-252.
Experiment 2 There was no difference in silage intake between commercial heat-moisture treated RSC and organic non-treated ORSC supplements.Rinne et al. (1999) reported similar silage intakes between RSM and heat-moisture treated RSC.These experiments show no benefits in silage intake when rapeseed cake is heat-moisture treated to reduce protein degradability in the rumen.High concentrations of supplementary fat have often decreased feed DM intake though this is not always the case (Coppock and Wilks 1991).Here higher fat content of organic ORSC did not significantly decrease silage intake compared with RSC.However, when oil was added with RSC, silage intake was reduced in agreement with Steele (1985) and Mohamed et al. (1988).Fat inclusion has reduced the digestibility of OM and fibre (Palmquist and Jenkins 1980, Tesfa 1992, Jenkins 1993, Khalili et al. 1997).Here the rea-son for decreased silage intake was not decreased digestibility of OM or NDF.Fat digestibility was increased with the ORSC and RSCO diets compared with the RSC diet in agreement with Khalili et al. (1997).

Animal performance
There were differences in the protein contents of silages.Huhtanen (1998) concluded in his review that RSM in the diet has produced similar increases in milk protein yield over a wide range of silage CP contents.Depending on a given situation, protein supplements can cause different responses due to e.g.differences in supply of AA and/or AA profile and/or rumen degradability.Recently, Huhtanen (1998) discussed protein evaluation and concluded that in studies carried out after the AAT -PBV system was  (Sjaunja et al. 1990).Khalili, H. et al.Protein supplements for organic milk production adopted in Finland, the relationship between AAT supply calculated using feed table values (Tuori et al. 1995) and milk protein yield has been very close.Current results are discussed in relation to the Finnish AAT -PBV system.
Experiment 1 Rapeseed meal and cake are excellent protein supplements for grass silage-based diets in conventional milk production (Tuori 1992, Huhtanen and Heikkilä 1996, Huhtanen 1998, Rinne et al. 1999, Ahvenjärvi et al. 1999).In the present experiment cows improved milk yield 0.66 kg per kg RSM DM compared with the control diet.This was close to the mean value of 0.77 per kg RSM DM concluded by Tuori (1992) but much lower than the mean value of 1.05 per kg RSM as fed by Huhtanen (1998) with grass silage-based diets.Tuori (1992) and Huhtanen (1998) concluded that increased ME intake was one explanation for the increased milk production.Here milk yield increased 0.16 kg milk per each additional MJ ME (estimated using DOM) consumed with RSM diet indicating that additional ME intake was in line with increased milk yield.Estimated ME utilization for milk production was not affected by RSM (Huhtanen andHeikkilä 1996, Rinne et al. 1999).RSM diet did not affect feed efficiency here.
Post-ruminal casein infusion increased yields of milk and protein more than ruminal casein infusion when cows were given a red clovergrass silage diet (Khalili et al. 1995a).In agreement with recent reviews (Tuori 1992, Huhtanen 1998), RSM increased milk protein yield 36 g per kg RSM DM.Tuori (1992) reported a mean response of 29 g increase in protein yield per kg RSM DM and Huhtanen (1998) 39 g per kg RSM as fed.RSM here caused a positive response in milk yield although AAT supply was higher than requirements (for maintenance and milk production) and milk urea was in the normal range of the control diet.Increased milk production indicated a sub-optimal AAT balance with grass-red clover silage supplemented with oats and barley.Therefore, with RSM diet, improved balance of absorbed AAT was most probably one reason for the increased production of milk and protein.RSM is a good source of histidine and Vanhatalo et al. (1997) reported that histidine appears to be the first limiting amino-acid for milk protein production on grass silage based diets.
Organic protein supplements produced at least as much milk and milk protein as RSM.Milk yield increased 0.11 and 0.14 kg milk per additional MJ ME with P and PRS diets, respectively.These increments were clearly less than theoretical considerations of 0.19 (Tuori et al. 1995).Pea increased milk protein yield 18 g per kg pea DM due to a low protein content.Field pea has shown promising results in earlier experiments (Syrjälä-Qvist et al. 1981, Corbett et al. 1995, Khorasani and Kennelly 1997, Heikkilä and Toivonen 1997).In the study of Syrjälä-Qvist et al. (1981) replacing soybean meal with pea did not affect milk production or composition.Corbett et al. (1995) reported that pea can be substituted for soybean and canola meal as a protein source for high-producing cows fed lucerne and grain.Similarly in two recent experiments replacement of oats either with rapeseed meal or with pea caused no differences in milk yields but in one experiment protein yield was lower with the pea diet (Heikkilä and Toivonen 1997).In spite of the higher protein degradability of pea (soluble protein fraction was 58.6 % by Khorasani and Kennelly 1997) than that of rapeseed meal (soluble protein was 13.3% by Rinne et al. 1999) there was only a minor difference in AAT content between concentrate mixtures in P and RSM diets.High protein degradation of pea may have stimulated microbial protein synthesis and consequently contributed to observed milk production.Ruminal bacterial N yield was higher when replacement of soybean meal protein by pea protein increased (Khorasani and Kennelly 1997).In addition, pea contains histidine which is an important AA for grass silage based-diets as mentioned earlier but the methionine content of pea is low.On the other hand, according to Varvikko et al. (1996) abomasal infusion of methionine did not improve milk production on grass silage based diets.
ECM yield was higher when a combination of pea and rapeseed was fed but protein content was clearly lower and consequently protein yield was similar as with P diet.There are ample results showing that feeding moderate amounts of fat increases milk production.In contrast, Steele (1985) reported that addition of crushed soybeans with silage based diets decreased milk production and protein content compared with soybean oil meal diet.ME intakes were similar between P and PRS diets indicating that ME intake did not explain the higher ECM production or lower milk protein content with PRS diet.The increase in ECM yield with PRS diet had some dilution effect (Wu and Huber 1994) for milk protein content.The fat content of rapeseed might also have reduced microbial protein synthesis as suggested by Mohamed et al. (1988).The supply of AAT was lower with the PRS diet.This affected the balance between energy and protein which might have contributed to the lower protein content with PRS diet.
Experiment 2 AAT supplies were much higher than requirements (for maintenance and milk production) and values of PBV and milk urea were high with all diets in the present experiment.Theoretically reduced ruminal protein degradability of RSC could have increased supply of histidine, the first limiting amino acid with grass silage based diets (Vanhatalo et al. 1997), and consequently increased milk production compared with coldpressed untreated rapeseed cake.There was, however, no differences in milk yield or composition between these two diets.This discrepancy between estimated protein value based on ruminal degradability, and milk production response was recently studied by Rinne et al. (1999).They compared RSM meal and heatmoisture treated RSC and did not notice any differences in milk production and milk constituent output.Also non-ammonia N flows from the rumen have been found to be similar for the two rapeseed feeds (Ahvenjärvi et al. 1999).The present results confirmed their earlier observa-tions that estimation of AAT content based on rumen degradation can be misleading.
Effects of fat supplementation have been variable and both positive and negative results have been reported in many experiments.Also here the effects of fat were different since higher fat supply with ORSC did not improve milk production compared with RSC.In contrast, although there was no differences in ME intake addition of oil with RSC increased milk yield compared with RSC without oil.Addition of oil/ fatty acids has decreased milk protein content (Mohamed et al. 1988, Tesfa 1992, Khalili et al. 1997) as was observed also with RCSO diet.
Here one reason for a decreased protein content was a dilution effect (Wu and Huber 1994) since protein yield was not affected.With the RSCO diet, the supply of AAT was decreased affecting balance between nutrients.This might be one reason for the decreased protein content.Mohamed et al. (1988) reported reduced plasma amino acid concentrations and that the ratio of serum glucose to unsaturated long-chain fatty acids was positively correlated with milk protein content.

Conclusions
Yields of milk and protein were similar between organic field pea and pea plus rapeseed and commercial RSM.Protein content was decreased when a combination of pea and rapeseed was given compared with pea.Equal amounts of milk and protein were produced when either commercial RSC or organic cold-pressed RSC were fed.However, when oil was added with commercial RSC milk yield increased.

Table 4 .
Effect of different diets on feed intake, estimated consumption of metabolizable energy (ME) and amino acids absorbed from the small intestine (AAT) consumption and digestibility (experiment 1, digestibility data for 4 cows used).

Table 5 .
Effects of different diets on milk yield, composition, yield of milk constituents, feed efficiency and utilization of N for milk production (experiment 1).

Table 7 .
Effect of different diets on feed intake, estimated consumption of metabolizable energy (ME) and amino acids absorbed from the small intestine (AAT) and digestibility (experiment 2, digestion data for 6 cows used).
a,b,cMeans in a row with different superscripts are significantly different (at least P<0.05).SEM=standard error of mean.

Table 8 .
Effects of different diets on milk yield, composition, yield of milk constituents, feed efficiency and utilization of N for milk production (experiment 2).