| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Animal Sciences, The Ohio State University, Columbus 43210
2 Corresponding author: firkins.1{at}osu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: liquid feed • animal fat • monensin • milk fat production
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
In addition to serving as a highly digestible energy course, liquid feeds (LF) are used in the dairy industry for other, nonnutritive reasons. Although DMI was not increased, cows preferred sweetened diets (Murphy et al., 1997), so LF might stimulate DMI in certain dietary conditions in which palatability is decreased. Field observations generally support the potential benefit of LF to decrease cow sorting behavior against long hay particles. Water decreased sorting against dry forage particles (Leonardi et al., 2005), but data are limiting for LF and results for individually fed animals might not represent field expectations (Leonardi and Armentano, 2007). Consequently, although LF are a convenient carrier for various feed ingredients such as NPN, which would require premixing, or supplemental fat, which could require melting before mixing, our aims were restricted to evaluating the nutritional qualities for individually fed cows.
Because LF could decrease ruminal ammonia concentration (Broderick and Radloff, 2004; DeFrain et al., 2006), adding urea to LF might improve its overall efficacy (Sannes et al., 2002). A slow-release source was generally comparable with soybean meal in diets containing sugars (Golombeski et al., 2006). Dietary fat improved milk production, but molasses or whey did not further enhance productivity and, in fact, tended (P 
0.10) to decrease milk fat and protein percentages (Maiga et al., 1995). However, in that study, the measured NSC was >35%, which might be higher than optimal for adding sugars in dairy rations. The calculated NFC was 41 to 42% and likely would be 2 to 3 percentage units higher if neutral detergent insoluble CP (NDICP) was accounted for in the NFC calculation. Therefore, our hypothesis was that LF in combination with different fat and NPN sources can substitute for typical sources of energy and protein in dairy rations while maintaining milk production or even increasing it if DMI is stimulated and dietary NFC is kept at moderate concentrations. Our first objective (trial 1) was to evaluate LF as a source of energy, without or with different sources of fat and NPN, with the hypothesis that milk production would be maintained or slightly improved compared with more conventional diets. The fat-containing LF was added in addition to conventional fat sources, including a control with unsaturated fat, to demonstrate that this form of feeding should not be detrimental to NDF digestibility or depress milk fat production appreciably. To differentiate the benefits of added energy from fat that is suspended in LF products, the objective for trial 2 was to compare lactation performance for LF without or with added fat and to test the hypothesis that increasing LF with added fat would stimulate milk production without depressing milk fat percentage if NFC was decreased. A diet containing a soybean lipid byproduct was added in both studies to serve as a control with increased potential for milk fat depression (Jenkins and McGuire, 2006).
Rumensin (R; monensin sodium; Elanco Animal Health, Greenfield, IN) is an antibiotic ionophore that is commonly added to dairy rations to improve feed efficiency. Methanogens and protozoa are not sensitive or can adapt to monensin (Nagaraja et al., 1997). However, even with some adaptation by some bacteria, persistence of bacterial response to monensin via increased propionate concentration, decreased CH4 production, and decreased ammonia concentration probably accounts for its improvement in feed efficiency (Russell and Houlihan, 2003). Because hyperammonia-producing deaminative bacteria are inhibited (Eschenlauer et al., 2002; Russell and Houlihan, 2003), supplementation of sugars in diets containing R could further exacerbate an ammonia limitation.
Although monensin should increase ruminal concentrations of propionate, the main glucose precursor (McGuffey et al., 2001), it is not clear if this response would be maintained with diets containing sugars, which commonly increase ruminal butyrate concentrations (Maiga et al., 1995; Broderick and Radloff, 2004; DeFrain et al., 2006), sometimes at the expense of propionate (Golombeski et al., 2006). Finally, some dairy herds have milk fat depression when high amounts of ruminally available carbohydrate are fed with R, particularly when there is supplemental unsaturated fat in the diet; these responses might be transient (Bell et al., 2006), requiring more long-term feeding trials to assess the efficacy of R when combined with supplemental fat and sugar. Also, increasing the grain percentage of the diet promotes accumulation of isomers from bio-hydrogenation that are responsible for milk fat depression (Jenkins and McGuire, 2006), although the negative aspects associated with increasing starch availability might depend on dietary situations (Bradford and Allen, 2004). Sucrose has been suggested to promote milk fat depression, whereas the flow of trans-10 isomers from fermenters (Ribeiro et al., 2005) or rates of biohydrogenation (Ribeiro et al., 2007) were not affected when incubations were buffered to prevent pH decline. The objectives for trial 3 were to investigate our hypothesis that increasing supplementation of LF plus urea concomitant with decreasing NFC concentration to maintain a similar concentration of rumen-degraded NSC would maintain or improve digestibility, stimulate DMI, and increase milk production and increase or maintain milk fat yield even when R was added to the diet.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The diets had an equal mix of corn silage and alfalfa silage that comprised 50% of the diet DM (Table 1
). Roasted soybeans (RSB) were added to replace soybean meal and blood meal to provide a source of fat while maintaining approximately similar RUP concentrations. Tallow (T) was added to diets to provide more fat. Finally, an extra increment of fat was provided in the form of LF to be near maximum recommended limits. All diets were balanced to be comparable in minerals and vitamins. The LF contained cane molasses, corn steep liquor, and suspension agents. All LF products for trials 1 and 2 were provided by Cargill Molasses Liquid Products Division (Elk River, MN), which was subsequently sold to Westway Trading Corporation (New Orleans, LA). The animal fat was choice white grease (WG), and this LF product containing WG was marketed as Synergy 200 (Westway). Urea (U) and a small amount of ammonium salts provided NPN (Synergy 20/20). In the last diet, the WG was replaced with neutralized soap stock (soybean oil refining lipid, SL). On a DM basis, the LF + WG, LF + WG + U, and LF + SL products had approximately 30.8, 30.3, and 33.3% crude fat; 33.5, 31.2, and 16.7% sugars; and 6.2, 30.3, and 6.7% CP, respectively. Using a procedure described by Sukhija and Palmquist (1988), the 16:0, 16:1, 18:0, 18:1 trans, 18:1 cis, 18:2n-6, and 18:3n-3 concentrations were 17.4, 2.2, 11.0, 6.2, 37.5, 13.1, and 1.0 g/100 g of fatty acids (FA) from the WG; 13.6, 0.2, 4.1, 0.3, 14.5, 56.2, and 7.9 g/100 g from SL; and 33.0, 3.8, 12.8, 3.5, 40.7, 7.9, and 0.6 g/100 g from the tallow, respectively. The concentrations of FA in tallow were comparable to our previous source (Pantoja et al., 1996). The RSB were not analyzed directly, but based on FA concentrations of the RSB diet (data not shown), the fatty acid profile was typical for RSB. All LF were blended into a separate concentrate mix for each diet in trial 1.
|
Monthly TMR were subsampled, dried at 55°C, and ground. Samples were analyzed for NDF with preextraction with hot ethanol and amylase addition (Van Soest et al., 1991), ADF, acid detergent lignin, and Kjeldahl N by AOAC (1990). We did not correct NDF or ADF for ash. Fatty acids were determined by GLC (Sukhija and Palmquist, 1988), and macrominerals were determined by inductively coupled plasma emission spectroscopy from the Ohio State University STAR lab (Wooster, OH). The NDF and ADF residues were further analyzed for Kjeldahl N concentration. All nutrient analyses were weighted for the number of cows on those respective diets per month and, using monthly composites, for the entire experiment. The NEL concentration was estimated using chemical analyses according to the equation of Weiss (1993) and scaled to 3 x maintenance (NRC, 1989).
Milk samples were weighted for individual milk volumes for a weekly average. Energy-corrected milk was calculated based on a 3.5% fat and 3.2% protein basis (Tyrrell and Reid, 1965) and NEL output in milk was calculated (NRC, 2001) using standard equations. Statistical analyses were performed using Proc Mixed of SAS (SAS Institute, Cary, NC) with a randomized complete block design, with block as a random variable and repeated measures within cow using models described by Reveneau et al. (2005). In all cases, the treatment x time interaction was judged not to be significant (P> 0.20), so main effect means were compared using an F-test-protected LSD.
Six cows were used in a 6 x 6 Latin square design with 6 periods of 3 wk, with the last week for sampling. Cows were bolus-dosed with a gelatin capsule containing 5 g of Cr2O3 twice daily (10 g/d) 5 d before collection and during the 4-d collection period. Fecal samples were taken before Cr2O3 dosage (for background Cr correction) and at 8 times staggered over the collection period. Fecal samples were composited for analysis of Cr concentration (Williams et al., 1962), N by the Kjeldahl method, and NDF as described previously. The TMR and refusal samples were taken every day during the 4-d sample collection period for the same chemical analyses. The data were analyzed using Proc Mixed of SAS, with period and treatment as fixed effects and cow as a random effect. In all cases, treatment differences were based on a P 0.10 to protect against Type II error (Berndtson, 1991), and least-significant difference was used to separate treatment means at the same treatment P-value reported.
Trial 2
Sixty cows (30 primiparous and 30 multiparous) were fed a common diet formulated for fresh cows for wk 2 and 3 of lactation and then blocked and assigned to 1 of the treatment diets from wk 4 to 19 of lactation as described for trial 1. The treatments included a control diet that contained no supplemental fat or LF (Table 2). A second diet contained LF without any added fat. The same product containing LF plus WG as in trial 1 was fed at a low, medium, or high inclusion rate, with the medium amount being at a comparable amount of LF as the diets containing LF without fat or LF + SL. Diets were mixed and fed daily as in trial 1 with the exception that the LF were weighed daily for each individual batch of feed. To accomplish this, a bucket was tared for residual LF that remained from the previous weighing (the tare was adjusted for cold weather). The LF was slowly poured over the TMR while mixing and with care to avoid pouring on the ribbon auger or internal side of the mixer wagon. All lactation measurements and feed analyses were taken as described for trial 1.
|
Trial 3
Diets were formulated to contain 21.0% forage NDF based on chopped alfalfa hay and corn silage (not kernel processed; Table 3). Whole cottonseed was added as a source of physically effective NDF and NEL. However, it also would provide linoleic acid, which would be bio-hydrogenated in the rumen. In contrast with trials 1 and 2, the LF used in trial 3 was provided by Quality Liquid Feeds (Dodgeville, WI). The LF was composed of cane molasses, corn steep liquor, and other ingredients but in different combinations than the LF used in trials 1 and 2. This product was marketed as TMR 20 (Quality Liquid Feeds) and was blended separately with the rest of the ingredients, as in trial 2. Dry ground corn was substituted by LF and soybean hulls to manipulate NFC concentration (calculated using software from NRC, 2001). The LF feeding rates were based on the assumption of optimum feeding rates of about 2.4% added sugar, with maximum values of 3.4% added sugar (Broderick and Radloff, 2004). Based on measured analyses of 53% sugars in the LF, this optimum and maximum would be 4.5 and 6.4% LF. Therefore, we formulated LF rates to be 3.25 and 6.50%.
|
A mix of protein supplements was added to meet predicted NRC (2001) requirements for RUP for dairy cows producing 40.9 kg/d of milk with a predicted DMI of 25.6 kg/d. To prevent a potential limitation of ruminal ammonia, urea was added to maintain at least an 8.0% safety factor for RDP supply relative to predicted RDP requirements. Rumensin (11.5 g of monensin/909 kg of DM) was added to be slightly greater than the manufacturer’s (Elanco Animal Health) lowest recommended inclusion rate (11 g/909 kg of DM).
Starting on wk 9 of lactation, 40 multiparous and 20 primiparous Holstein cows were fed the control diet for 2 wk for standardization and covariate adjustment. Then, cows were blocked by parity, date of calving, and milk production (average from the 2-wk standardization period) and randomly assigned to either stay on that diet or be switched to one of the other 4 diets for 12 wk. The cows were all administered Posilac (Monsanto, St. Louis, MO) commencing on wk 9 of lactation and administered every 2 wk thereafter. All diets were mixed daily and fed for 10% orts. Monthly composite samples of TMR, forages, cottonseed, and grain mixes were analyzed for nutrients using standard methods at Dairy One (Ithaca, NY). All analyses are comparable with previous descriptions except that NDF analysis used pretreatment with acetone and incubation with amylase plus sulfite; fat was analyzed by ether extract (AOAC, 1990); and water-soluble monosaccharides (Hall et al., 1999) were preextracted before analysis for starch using gelatinization with an autoclave, hydrolysis with glucoamylase, and analysis by glucose oxidase with standardization against dextrose (x 0.9). All animal handling, data collection, and data compilation were done as described previously, with the exception that BCS was monitored every 3 wk during the last 12 wk of the experiment.
Five cows (n = 2 rumen cannulated) were used in a 5 x 5 Latin square for a digestibility experiment as described for trial 1 with the exception that periods were 4 wk, Posilac was administered every 2 wk during the trial, and pellets containing Cr2O3 (5%) and soybean hulls (95%) were dosed into the rumen for the rumen-cannulated cows. For those 2 cows, ruminal samples were taken by cannulas on 0, 3, 6, and 9 h after feeding. For the other 3 cows, gelatin capsules containing Cr2O3 were dosed twice daily as described in trial 1. One rumen sample was taken per period by stomach tube at 3 h after the morning feeding. The concentrations of VFA were determined using GLC (Harvatine et al., 2002). There was no treatment x time interaction for VFA for the 2 cows (albeit power was insufficient to test this interaction), and a mean was used along with the single value for the noncannulated cows.
All statistics were performed using Proc Mixed of SAS. For the digestibility trial, cow was modeled as a random effect in the Latin square design. For the production trial, the data were weighted for parity (Reveneau et al., 2005). Evaluation of treatment x time interactions were modeled with the ar(1) covariance structures (which consistently had the lowest Bayesian information criterion score). In all cases, treatment differences were based on P 0.10 to protect against Type II error (Berndtson, 1991). There were no treatment x time interactions (P > 0.10), and least-significant difference was used to separate treatment means at the same treatment P-value that is reported in the tables.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
). As planned, adding each increment of fat from RSB, tallow, or WG increased the concentration of FA and NEL proportionately.
|
). Compared with the control diet, all diets with the greatest amount of fat (RSB + T and all 3 LF diets) decreased (P 0.08) milk protein percentage but had no effect on milk protein yield. Compared with the control (P 0.07), adding RSB decreased milk fat percentage, but RSB + T increased it. Cows fed both of the diets with LF + WG (without or with U) had similar fat percentages, but those fed LF + SL tended to have a lower fat percentage compared with LF + WG + U (P = 0.07) and LF + WG (P = 0.12) such that LF + SL was comparable with that from the RSB treatment. Milk fat yield was lower (P < 0.05) for cows fed RSB but greater for those fed RSB + T than any of the LF diets, which were comparable to control.
|
Trial 2
As in trial 1, all nutrient analyses were comparable among treatments relative to formulation values, and changes in FA and NEL reflected expected rates of fat intake (Table 6). Trial 2 had a control without LF and a diet containing LF without fat (the second diet). The LF diet without fat was comparable on an LF basis to the LF diets containing WG (fourth diet) or SL (sixth diet), which are labeled in Table 7 as medium rates. Adding fat to LF diets increased (P < 0.02) DMI (LF vs. LF + WG contrast), but there was a quadratic response (P < 0.02) for the feeding amount of LF + WG. Milk production increased linearly (P < 0.01) with increasing LF + WG, and the quadratic contrast was P = 0.10.
|
|
0.09) milk protein yield. The average milk protein yield from the LF + WG or LF + SL treatments was greater than the LF treatment. Increasing amount of LF + WG in the diet increased milk protein yield linearly (P < 0.01), but the quadratic contrast (P = 0.02) indicates a plateau in the response. Milk fat percentage was greater (P < 0.04) for cows fed LF vs. the average of the medium amount of LF + WG or LF + SL, but milk fat yield was similar between those treatments. Both milk fat percentage and yield were markedly lower (P < 0.01) for cows fed LF + SL than the medium amount of LF + WG. Increasing the amount of LF + WG increased milk fat yield the most for the medium feeding rate (the linear and quadratic contrasts were P < 0.01 and P < 0.10, respectively).
The yields of both ECM and NEL output in milk were greater (P < 0.05) for the average of the medium amount of LF + WG plus LF + SL compared with LF, but, as with protein and fat yields, there were both linear and quadratic responses (P 0.07) with increasing amount of LF + WG. Because of a lower DMI, the ECM/DMI was increased for the high feeding amount of LF + WG (P < 0.05 linear increase). The NEL output in milk and its efficiency ratio followed similar trends as for ECM or ECM/DMI. Because of numerically greater DMI and lower ECM, the ECM/DMI ratio was lower (P = 0.08) for LF + SL than for the medium amount of LF + WG. The LF diet had a greater BW and BW change than the control (P < 0.06) and was greater (P < 0.01) for BW or tended to be greater (P = 0.13) for BW change than for the average of the LF + WG diets.
Trial 3
The dietary CP ranged from 17.0 to 17.5% (Table 8). Forage and cottonseed NDF averaged 19.7 and 4.6%, respectively. The NFC was uniformly about 1 percentage unit higher than formulated values, but this is likely because the NDF procedure uses sulfite, whereas NDICP does not. Consequently, underestimation of protein-free NDF would subsequently overestimate NFC, verifying our diets to be on target with formulations. The NSC was about 10 percentage units lower than NFC because of soluble fiber from legumes (alfalfa and soybean hulls) and from fermentation acids in corn silage. Of the NSC component, LF substitution for corn grain increased sugar and decreased starch concentrations comparable to treatment design. Other nutrient concentrations were comparable with formulations. We verified by commercial laboratory testing through Elanco Animal Health that our actual measured concentration of R was within specifications of expected R concentration.
|
). Cows fed the 37% NFC diet with 3.25% LF tended (P < 0.12) to have greater DMI than both the control and the 40% NFC diet with 3.25% LF. When R was added to the 6.5% LF diet, DMI decreased (P < 0.08). Despite the changes in DMI, milk production was not affected by treatment.
|
0.08) for both of the 6.5% LF diets (without or with R). Cows fed the 3.25% LF diet with 40% NFC had a lower MUN than the 3.25% LF diet with 37% NFC.
Milk fat percentage was not affected by treatment, but the greatest (P 0.08) milk fat yield was for cows fed the 3.25% LF diet containing 37% NFC. The ECM was greater (P 0.08) when 3.25% LF was added to 37% NFC diets compared with control or 3.25% LF in 40% NFC diets, and ECM yields for cows fed 6.25% LF without or with monensin in 37% NFC diets were intermediate. The NEL output in milk followed similar trends (P 0.09) as did ECM except that the difference between the 3.25% LF treatment with 37% NFC and the 6.5% LF + R treatment exceeded the significance threshold. There were no differences in either measure of efficiency, BW, BW change, or BCS.
Nutrient digestibility and VFA concentrations were not affected by treatment (Table 10). Cows in this experiment averaged 25.7 kg/d of DMI (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Trial 1: Different Fat Sources and NPN in LF Added to Concentrates
Adding more dietary fat generally reduced milk protein percentage (Table 5
). The greater milk protein percentage for the RSB + T treatment might be related to a slightly higher measured CP percentage for that diet (Table 4). Jenkins and McGuire (2006) recently discussed a cumulative understanding of how supplemental fat usually depresses milk protein percentage without changing protein yield. There were no effects from adding urea to the LF + WG diet. Sannes et al. (2002) reported that sucrose decreased the outflow of urinary purine derivatives (a surrogate for estimating microbial protein synthesis) but decreased ruminal NH3-N and MUN, so cows fed LF + WG (even without U) appears not to have been limited in NH3-N under our circumstances.
Supplemental sources of unsaturated fat can depress fiber digestibility and potentially decrease milk fat percentage, although RSB should have a much more limited response because of the slower release of oil from the seed (Jenkins and McGuire, 2006). However, total tract NDF digestibility was not affected by treatment (Table 5), so the increase in milk fat percentage and yield by providing tallow to RSB or WG in the LF treatments compared with the RSB treatment should be a result of increased intake of supplemental saturated fatty acids. Fatty acid digestibility was not affected by treatment, so increasing saturated fatty acid supply should increase transfer to the milk and dilute the negative effects of absorbed intermediates from biohydrogenation of unsaturated fatty acids on de novo milk fatty acid synthesis (Shingfield et al., 2006). Compared with the RSB treatment, ECM/DMI and NEL output in milk/DMI were greater for cows fed RSB + T, although DMI, ECM, and NEL output in milk were not affected. The RSB diet tended to have slightly greater predicted NEL concentration than RSB + T, which was greater than control (Table 4), but measures of gross feed efficiency did not seem to be closely related to NEL concentration.
Trial 2: Source of Fat and Amount of LF Blended in TMR
Compared with the control, cows fed the LF diet without fat increased milk protein yield, BW, and BW change (Table 7). Even though the NFC concentration was slightly greater (Table 6), the LF might have provided more ruminally available carbohydrate to stimulate propionate production and microbial protein synthesis, which could enhance milk protein production (Jenkins and McGuire, 2006). Increasing insulin concentration could also limit adipose tissue mobilization or enhance lipid uptake, explaining the increased BW change for cows fed LF (Table 7). Increasing insulin concentration associated with increased ruminal carbohydrate digestibility might decrease DMI (Bradford and Allen, 2007), so lack of decrease in DMI might be associated with improved palatability of sugar-containing diets (Murphy et al., 1997).
In the contrast for LF + fat (the average of the medium amount of LF + WG and the LF + SL treatments), DMI increased compared with LF alone (Table 7), which was unexpected. Because the dose response for increasing amounts of LF + WG (control vs. low, medium, and high LF + WG diets) was quadratic, the greater amount of LF + WG probably decreased DMI through chemo-static regulation mechanisms (Allen, 2000). We calculated the iodine value to be about 135 for the LF containing SL, whereas the iodine value was about 69 for the LF containing WG. Although unsaturated fat in the SL would be expected to potentially depress DMI (Allen, 2000) through decreasing meal size (Harvatine and Allen, 2006b), again no such effect on DMI was detected until this high feeding rate. The DMI was not stimulated by the LF diet without fat compared with control, so palatability issues associated with LF without fat appear not to explain the responses. A mechanism for stimulation of DMI by adding moderate amounts of fat to the LF is not clear but deserves further study.
Milk production was increased when fat was added to LF diets compared with LF alone (Table 7). There was also a linear (P < 0.01) and quadratic (P = 0.10) effect of amount of LF + WG. There was a larger relative increase in milk production from the control compared with the lower feeding amount of LF + WG, followed by a gradual peaking with increasing increments of LF + WG. There was less symmetry in the quadratic response, which would increase its P-value. Consequently, the combination of a significant linear contrast with a strong trend for a quadratic contrast supports a plateau response. Similar responses were detected in yield of milk protein, fat, ECM, and NEL output in milk. Thus, in support of generally quadratic results using molasses (without fat) reported by Broderick and Radloff (2004), the medium feeding rate of LF + WG would be considered optimal for milk production under our circumstances.
As explained previously, Maiga et al. (1995) did not see an apparent benefit from sugar-containing supplements with fat at 7.4 to 8.3% of DM, which approximated our high amount of LF + WG. As we increased the amount of LF + WG, the NFC decreased from 41.9% in the control to 38.8% in the high feeding amount of LF + WG (Table 6). Moreover, increasing amount of LF + WG increased milk protein yield while cows increased their consumption of supplemental dietary fat. Perhaps milk protein depression associated with fat feeding was less severe because we only added moderate amounts of fat and provided high-quality RUP sources or because the LF + WG helped to prevent suppressions in DMI that are correlated with milk protein depression from supplemental fat (Wu and Huber, 1994).
Although milk production was maintained when cows were fed the LF + SL diet, milk fat percentage and yield were decreased compared with the medium amount of LF + WG. As explained by Harvatine and Allen (2006a), lipid-induced milk fat depression without depressed DMI should increase BW gain. The BW gain was numerically but not statistically increased with the LF + SL diet compared with the medium amount of LF + WG. We noted similar trends and responses for decreases in other variables, particularly for efficiency of ECM or NEL output in milk per unit of DMI, when SL was added to the LF. Although soybean soapstock decreased the rate of NDF degradation in situ (Abel-Caines et al., 1998), we did not detect any differences in kinetics of degradation of NDF from either ground alfalfa or or-chardgrass hays (data not shown). Therefore, the results seem to be more likely associated with increased intermediates of biohydrogenation of linoleic acid (which was the major fatty acid in SL) and resultant depression of milk fat percentage that persisted long-term (Shingfield et al., 2006). From our responses in both trials 1 and 2, the usage of SL at these feeding amounts should be considered with respect to potential for milk fat depression.
Trial 3: Response to Amount of LF Blended into TMR Balanced for Rumen-Degraded Carbohydrate and Containing Rumensin
Compared with the control, the concentration of ruminally degraded NSC was formulated to be the greatest in the 40% NFC diet with 3.25% LF, the lowest in the 37% NFC diet with 3.25% LF, and to be intermediate and equivalent to the 6.5% LF diets (see earlier discussion). Therefore, we expected that a benefit in enhanced palatability might increase DMI if ruminal carbohydrate availability did not suppress DMI (Allen, 2000). The stimulation of DMI in the 3.25% LF diets by diluting NFC from 40 to 37% supports our hypothesis. As in many but not all of the ionophore studies summarized by Ipharraguerre and Clark (2003), adding R to the 6.5% LF diet depressed DMI. Our inclusion rate (which met label requirements for the lowest concentration to improve feed efficiency) was lower than that of some studies that reported depressed DMI (Bell et al., 2006; Odongo et al., 2007), but the decrease in DMI was not positively related to R feeding rate (Phipps et al., 2000).
There were no significant differences in milk production or milk protein yield, but milk protein percentage was decreased with all diets containing LF compared with the control, but most of this response would be expected to be a result of dilution because protein yield was not affected (Table 9). Milk protein percentage was lower or tended to be lower when LF diets were fed in trials 1 and 2 and could not be differentiated into effects of LF vs. added fat in trial 2. Molasses did not depress milk protein percentage in the study of Broderick and Radloff (2004) but tended to decrease it in the study of Maiga et al. (1995). Generally, milk protein percentage is depressed by several factors (Jenkins and McGuire, 2006), but as discussed subsequently, none of these factors seem relevant for LF feeding. In trial 3, the LF diets had lower concentrations of ground corn (Table 3). Although the greater concentration of starch vs. sugar in the control could be supporting more microbial protein synthesis because of more continuous carbohydrate supply and synchronization with RDP than with LF diets, in vivo support for this synchronization response for dairy cattle is equivocal (Firkins et al., 2006), and LF concentrations were balanced to not exceed recommended sugar concentrations (Broderick and Radloff, 2004). Using the procedure of Griswold et al. (2003), the concentration of ruminal peptides in the 2 rumen-cannulated cows was actually numerically greater when the cows were fed the control treatment than the others, and all NH3-N concentrations were >10 mg/dL (data not shown). In cows fed 20% of the diet as various carbohydrate sources, NH3-N irreversible loss from the rumen decreased compared with a control; this response was attributed to decreased deamination when the cows were fed glucose and to increased incorporation of NH3-N into microbial protein when they were fed starch (Hristov et al., 2005). Molasses and sugars often decrease rumen NH3-N and MUN (Sannes et al., 2002), but this response was quadratic in 2 studies (Broderick and Radloff, 2004), and our MUN data do not support any major change in N metabolism in the rumen. The MUN was greater for the 6.25% LF diets, but this response likely was a result of higher CP in the diet formulations with increasing LF, even though the analyses (Table 8) did not detect a difference. An increased propionate concentration and resultant insulin response to increase milk protein percentage (Jenkins and McGuire, 2006) also are not corroborated by our VFA data (Table 10), which were not affected by treatment, or by milk lactose, which ranged only from 4.74 to 4.77% among treatments (data not shown). Therefore, the decrease in milk protein concentration by LF treatments is likely just a simple dilution response, which seems consistent with typical responses in which milk protein yield was unchanged or increased with increasing milk yield when sugars were supplemented.
Milk fat percentage was not affected by LF in trial 3 (Table 9). Cows in trial 3 were later in lactation and were injected with bST, but long-term administration should not affect milk component concentrations (NRC, 2001). Milk fat percentage in trial 1 decreased with LF + WG compared with the RSB + T diet but not when compared with the control (Table 5), and increasing amount of LF + WG did not affect milk fat percentage in trial 2 (Table 7). These results could be explained, in part, from those diets in trial 1 having greater NFC concentrations than optimal, whereas trial 2 and especially trial 3 evaluated LF when ruminal carbohydrate availability was controlled. However, adding molasses or sucrose to diets comparable or higher in NFC than in our trials (Cherney et al., 2003; Broderick and Radloff, 2004) did not affect milk fat percentage or increase it, and NFC concentration within or among studies had little apparent effect. Also, sucrose addition did not direct the flux of biohydrogenation toward the trans-10 pathway (Ribeiro et al., 2007), so we would not expect sugars to depress milk fat unless the pH declined significantly. Perhaps the stimulation of DMI by LF helps protect the rumen from conditions that lead to accumulation of FA intermediates from biohydrogenation that promote milk fat depression (Bradford and Allen, 2004).
Although we could not measure pH in our trial for the noncannulated cows (and there was no difference among treatments for the 2 cannulated cows; data not shown), Broderick and Radloff (2004) noted that pH or VFA proportions did not respond to increasing dietary concentrations of molasses. Our VFA data (Table 10) also were not affected by diet. As discussed previously, many studies with sucrose or other sugars often increase the molar proportion of butyrate (Maiga et al., 1995; Golombeski et al., 2006). In contrast, butyrate molar percentage did not increase when sucrose was added to diets containing high-moisture corn (Cherney et al., 2003; Broderick and Radloff, 2004). Because high-moisture corn has a very high ruminal NSC digestibility but does not depress ruminal pH compared with ground dried corn (Firkins et al., 2001), butyrate concentration might only increase when LF promotes lactate production and subsequent conversion to butyrate (Firkins et al., 2006).
When 3.25% LF was added to the 37% NFC diet, milk fat yield was greater compared with any other treatment. This treatment also had greater yields of ECM and NEL output in milk compared with the control or 3.25% LF diet with 40% NFC, whereas both of the 6.25% LF diets were intermediate (Table 9). Thus, most of the response is a result of increases or tendencies for increases in DMI by cows fed LF when the NFC and ruminal carbohydrate availability are moderated by ration formulation (see earlier discussion). Yields of milk fat, ECM, or NEL output were relatively proportional to DMI such that there were no differences in ECM/DMI or NEL output in milk/DMI associated with feeding LF. Although VFA proportions were not changed, increasing the ruminal degradability of carbohydrate by adding 3.25% LF in the 40% NFC diet should increase the amount of VFA production, thus increasing the potential for DMI suppression with increasing ruminal fermentability of the diet (Allen, 2000; Bradford and Allen, 2007).
We expected a decrease in the acetate:propionate ratio associated with feeding R (McGuffey et al., 2001), but no responses were detected (Table 10). As discussed by Bell et al. (2006), there is some concern about long-term adaptation by ruminal microbes to monensin. However, Odongo et al. (2007) detected a long-term decrease in methane production from feeding monensin. In fact, physiological adaptation to R is more likely than a decrease in sensitivity through genetic mutation (Russell and Houlihan, 2003), so a concomitant ripple through the entire microbial ecosystem would be expected to interact with other dietary factors. For example, shifts in concentrations of FA isomers from feeding supplemental unsaturated fat can take longer for equilibration (Shingfield et al., 2006), so longer term changes in microbial population structure or fermentation pattern could be associated with various factors other than monensin sensitivity.
Adding R to the 6.25% LF diet decreased DMI without affecting ECM or NEL output in milk. Algebraically, the mean of 2 variables will not equal the mean of the ratio of those 2 variables; combined with covariate analysis, there might have been some more extreme values for individual ECM and NEL output in milk or DMI that normalized when the individual efficiency ratios were evaluated statistically. In summaries of ionophore research, monensin typically slightly decreased DMI and slightly increased milk production (Ipharraguerre and Clark, 2003), explaining the general response in improved feed efficiency (McGuffey et al., 2001). In contrast with other responses (Bell et al., 2006), adding R to the 6.25% LF diet was not expected to depress milk fat percentage because the ruminal carbohydrate availability was moderated by addition of nonforage NDF and because a major source of unsaturated fat was from whole cottonseed, which decreases trans isomer accumulation compared with processed cottonseeds (Reveneau et al., 2005). Milk fat depression by monensin is associated with increased accumulation of intermediates of biohydrogenation of unsaturated fat (Benchaar et al., 2006), but moderation of the diet fermentability also should help maintain ruminal protozoa populations that incorporate unsaturated fatty acids (Devillard et al., 2006) and bacteria responsible for reducing the trans intermediates of biohydrogenation to stearic acid (Firkins et al., 2006).
There were no differences in any measurements from the digestibility study (Table 10). The DMI ranged from 24.5 to 26.8 (SE = 1.5) kg/d (data not shown). The lack of treatment differences contrasted with changes in DMI in the longer term production trial, documenting the importance of continuous lactation trials to evaluate the true efficacy of treatments. Unfortunately, high variability in fecal output predictions, which are most likely explained by increased error associated with bolus-dosing of Cr2O3 in this study, decreased the accuracy of the mean values and the sensitivity for treatment comparisons. Modeling cow as a random effect considers the variance of cow in addition to the residual variance, often increasing the SE without changing power to detect differences (Harvatine et al., 2002). The heterogeneity of cows (cannulated vs. noncannulated) might further increase the among-cow response to treatments, as has been shown for cows fed different fat sources (Harvatine and Allen, 2006a). Despite these limitations, the lack of consistent differences in nutrient digestibility or VFA proportions among treatment means support the likelihood that major responses in lactation performance are associated with responses in DMI.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
3 Current address: Kent Feeds, Granger, IN 46530.
4 Current address: Universidad de Puerto Rico, Mayagüez PR 00681-9000.
5 Current address: Quality Liquid Feeds, PO Box 240, Dodgeville, WI 53533.
Received for publication November 16, 2007. Accepted for publication January 11, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. R. Oelker, C. Reveneau, and J. L. Firkins Interaction of molasses and monensin in alfalfa hay- or corn silage-based diets on rumen fermentation, total tract digestibility, and milk production by Holstein cows J Dairy Sci, January 1, 2009; 92(1): 270 - 285. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
