| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Danish Agricultural Advisory Service, Danish Cattle Federation, Skejby, Denmark
US Dairy Forage Research Center, Madison, WI
University of Aarhus, Faculty of Agricultural Sciences, Foulum, Denmark
1 Corresponding author: dob{at}agrotech.dk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: degradation • neutral detergent fiber • kinetics
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Earlier studies indicate that methods (in situ, in vitro, in vivo) can differ considerably in estimated NDF degradation parameters (Huhtanen et al., 2006). However, in vivo measurements of digestion kinetics are few in number as this method is very resource demanding. Therefore, the in vitro (IV) method has been widely used in the United States (Mertens, 1993), whereas the in situ (IS) method has traditionally been used as basis for generating NDF degradation parameter estimates in Denmark (Weisbjerg and Hvelplund, 2005). The IS method is the basis for estimating NDF degradation parameters for the new common Nordic feed evaluation system (Eriksson et al., 2007). The IS and IV methods have strengths or weaknesses, dependent on whether the estimated feed degradation characteristics should express the actual or potential feed degradation, respectively. For example, the IS method may not estimate potential NDF degradation because results can be negatively effected by the rumen environment of the host animals as influenced by numerous factors as discussed in detail in review articles by e.g., Vanzant et al. (1998), Huntington and Givens (1995) and Nocek (1988). On the other hand, the IS method may predict actual degradation because it benefits from the feed being degraded in the rumen environment under typical feeding conditions. To be useful for feed evaluation, it is therefore essential to examine the behavior of the different methods for estimating the potential rate and extent of degradation and their alteration by factors like grinding.
The objective of this experiment was to examine the impact of screen size during grinding of samples and fermentation methods on estimates of aNDF degradation parameters, using samples of 5 common Danish feeds. For this examination the traditional IV method with media pH near 6.8 and the IS method using cows fed at production level were evaluated, in addition to an IV method where pH was lowered to about 6.0. The pH during IV incubation is known to have a significant effect on the aNDF degradation profile (Grant and Mertens, 1992b), and the latter method was included to evaluate effect of pH explaining some of the differences between the traditional IV and the IS methods used in this experiment.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In Situ Incubation
Four rumen-cannulated lactating Holstein cows ranging from 255 to 606 DIM, yielding from 12.7 to 20.0 kg of milk, were used for the IS incubation and as IV inoculum donors. Four cows were fed ad libitum 1 of 2 different mixed rations once daily at 0630 h during the total experimental period of 6 wk. On DM basis, the mixed rations (A, B) were composed of corn silage (30, 27%), alfalfa silage (25, 28%), a mix of concentrates (41, 45%), and alfalfa hay (4, 0%) and contained approximately (30, 28%) amylase-treated NDF (aNDF) and (18, 17%) CP. Mix of concentrates for ration A consisted of 52% medium ground corn, 15% roasted soybeans, 15% whole cottonseed, 8% soybean meal (44% CP), 6% corn gluten feed, 1% blood meal, 1% grease, and 2% vitamins and minerals (as-is). Mix of concentrates for ration B consisted of 74% coarse ground corn, 25% soybean meal (44% CP), and 1% vitamins and minerals (as-is). The 2 cows fed ration A and B, respectively, had an average daily DMI of 21.9 and 27.8 kg per cow.
Samples for IS incubation were prepared by transferring 0.5 g of ground feed material into 5 x 5 cm Dacron bags with a pore size of 50 (±15) µm. Sets of the 16 different feed samples and one empty bag (blank sample) were incubated for 6, 12, 24, 48, or 96 h, respectively, in each of the 4 cows. Incubation was replicated after 2 wk. Within replicate, incubation of all samples was initiated at the same time at 1000 h. Rumen pH was measured in all 4 cows before incubation, and every 3 h during the first 24 h of each IS incubation by collecting ruminal liquid in a small beaker from the cranio-ventral rumen and immediately measuring the pH. After incubation, bags were rinsed in cold tap water and frozen. Immediately before analysis, samples were thawed at room temperature, and together with 4 sets of zero-hour samples (not presoaked), were washed twice in a washing machine for 13 min in approximately 20 L of water.
In Vitro Incubation
Sets of the 16 feed samples and a blank sample were incubated IV in duplicate for 6, 12, 24, 48, or 96 h. This in vitro incubation (IVn) followed the procedure described by Goering and Van Soest (1970) with some modifications of both preparation of incubation medium and inoculum. The constituent solutions (micromineral solution, buffer solution, and macromineral solution) and the reducing solution used for incubation medium were prepared as follows:
Micromineral Solution.
A total of 13.2 g of CaCl2·2H2O, 10.0 g of MnCl2·4H2O, 1.0 g of CoCl2·6H2O, and 8.0 g of FeCl3·6H2O were dissolved in distilled water by agitation, and volume was brought to 100 mL.
Buffer Solution.
A total of 4.0 g of NH4HCO3 and 35.0 g of NaHCO3 were dissolved in 1.0 L of distilled water.
Macromineral Solution.
A total of 5.7 g of Na2HPO4 – anhydrous, 6.2 g of KH2PO4 – anhydrous, and 0.6 g of MgSO4·7H2O were dissolved in 1 L of distilled water.
Incubation Medium.
A total of 2 g of trypticase, 400 mL of water, 0.1 mL of micromineral solution, 200 mL of buffer solution, 200 mL of macromineral solution, and 1 mL of resazurin were added in order and mixed.
Reducing Solution.
A quantity of 625 mg of cysteine hydrochloric acid was dissolved in 95 mL of water. Four milliliters of 1 N sodium hydrozide and 625 mg of sodium sulfide nonahydrate were added and dissolved.
Immediately before initiating the IS incubation, ruminal fluid and contents from the 4 cows were collected in 1 L warmed and insulated beakers for preparation of IVn inoculum. After discarding the top layer of the collected ingesta from each cow, it was strained through 2 layers of cheesecloth to separate solids from liquids. From each cow, 250 g of rumen solids was mixed with 250 mL of incubation medium and 12.5 mL of reducing solution, blended for 1 min, and then strained through 4 layers of cheesecloth together with 250 mL of rumen liquid. Twenty milliliters of the resulting 1:1 inoculum mixture was transferred to 125-mL Erlenmeyer flasks containing 0.5 g of feed sample, 30 mL of incubation medium, and 1.5 mL of reducing solution that was warmed in a 39°C water bath and had been purged with CO2 for a minimum of 30 min. All IV fermentations were maintained under continuous CO2 pressure via a tube connected to each flask. The IVn incubation was replicated after 2 wk, initiating the first incubation (IVn-rep1) and the replicate (IVn-rep2) the same days, respectively, as the IS incubations. The pH of the 1:1 inoculum mixture of strained ruminal fluid and media from blended solids from all 4 cows was 6.89 and 6.38 for IVn-rep1 and IVn-rep2, respectively.
The same number of samples and replicates were used for a pH-adjusted in vitro incubation (IVa) of feed samples. The IVa incubation was carried out the same way as described for the IVn incubation, only the pH of the incubation medium was reduced from 6.8 to 6.0 by adding 1 M citric acid (Grant and Mertens, 1992a). In this experiment, 60 mL of citric acid was added to 940 mL of original incubation medium to obtain 1,000 mL of adjusted incubation medium with a pH value of approximately 6.0. The IVa incubation (IVa-rep1) was initiated about 1 wk after terminating IS and IVn incubations, and the replication (IVa-rep2) was initiated 1 wk after IVa-rep1. The pH of the 1:1 inoculum mixture was 5.89 and 6.05 for the IVa-rep1 and IVa-rep2, respectively. For all IV incubated samples the aNDF contents of the feeds were used as zero-hour aNDF-residue (aNDF0).
Chemical Analysis
Feed content of ash, CP, crude fat, starch, aNDF, and ADF are presented in Table 1
. Feeds were analyzed according to conventional procedures. Ash was analyzed by heating to 525°C (AOAC, 2000), CP was calculated based on analysis of total N (Hansen, 1989), crude fat was analyzed by Soxhlet extraction with petroleum ether after hydrolyzing with HCl (Stoldt, 1952), starch was analyzed by enzymatic calorimetric technique (Knudsen et al., 1987), aNDF was analyzed by neutral detergent extraction (Mertens, 2002), and ADF was analyzed by acid detergent extraction (AOAC, 2000). The aNDF residue of IS incubated feed samples was analyzed using an Ankom200 Fiber Analyzer (Ankom Technology Corporation, Macedon, New York) following the user’s instructions (Ankom Technology, 2005). Dacron bags (5 x 5 cm) were used directly in the Ankom200 Fiber Analyzer. The aNDF content in feeds and the aNDF residues from IVn and IVa incubated feed samples were analyzed following the official AOAC method described by Mertens (2002), using Berzelius beakers with reflux. Using different methods for analyzing aNDF content between IS and IVn or IVa methods means that incubation method is confounded with analytical method.
|
 |
In model 1, Y is the pH; 
i is fixed effect of method (i = IVn, IVa); βij is fixed effect of replicates-within-method (j = 1, 2); 
k is fixed effect of feed (k = CS, GS, B, BP, RC); 
l is fixed effect of screen size used for grinding (l = 1, 2, 4, 8 mm); 
is fixed effect of incubation time (m = 6, 12, 24, 48, 96 h); ik, im, km, βijm, and ikm are the respective interactions; and 
ijklmn is the residual.
Degradation parameters for aNDF were estimated based on aNDF residues expressed in percent of DM, using the Marquardt method in Proc NLIN in SAS (SAS Institute, 2000) applying model 2:
 |
In model 2, I is the nonlinear parameter estimate of the indegradable fraction of aNDF (g/g of DM); aNDF0 is the amount of aNDF in the feed at time = 0; D0 (= aNDF0 – I) is the potential degradable fraction of aNDF (g/g of DM); kd is the fractional rate of aNDF degradation (h–1); t is time (h); and L is the discrete lag time (h). Following the recommendations of Weisbjerg and Lund (2004), duplicate measurements of aNDF residues at the same incubation time within IV method and replicate were averaged before parameter estimation. For the IS method, the aNDF residue of the 4 cows at each fermentation time within replicate was averaged before parameter estimation. This approach resulted in one set of degradation parameters per replicates-within-method. For each set of degradation parameters, the effective degradation (ED) of aNDF was calculated to provide the combined influence of all aNDF degradation parameters on aNDF degradation. The ED was calculated using the revised formula of McDonald (1981) that includes the effect of lag time as described by Mertens (1977) and Stensig et al. (1994), based on a 1-compartment model for rumen fiber turnover assuming the rumen fractional rate of passage (k) to be 0.02 h–1 (formula 1). More complicated models for description of rumen fiber turnover might better express the actual aNDF degradation (Lund et al., 2007).
 |
Initial statistical data analysis using a full model including feed, method, grinding, and all 2-factor interactions showed that feed interacted with all other variables investigated. Therefore, to investigate the influence of sample grinding and fermentation method, statistical analyses were carried out within feed. Estimates of aNDF degradation parameters were analyzed in SAS (SAS Institute, 2000), using Proc GLM applying model 3:
 |
In model 3, Y represents I, D0, kd, L, or ED, respectively; is the fixed effect of incubation method (i = IVn, IVa, IS); is the fixed effect of screen size used for grinding (j = 1, 2, 4, 8 mm); ij is the interaction between method and grinding; and ijk is the residual. The P-values presented refers to the F-test of model 3. The Waller-Duncan k-ratio t-test was used for mean separation when F-test had a P < 0.05. The mean separation test used the main factor interaction (ij) as the error term. This conservative approach was chosen because grinding might be perceived as a split-plot factor and because of the occurrence of interactions between the main factors in model 3 for a few parameters within some feeds.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
. Some concentrate feeds (BP and B) ground at a 4-mm screen had a higher geometric particle size than silages (GS, CS) ground at a 8-mm screen. Similar results are reported by Michalet-Doreau and Cerneau (1991) who found significantly different mean particle sizes among concentrates in response to grinding through same screen size. As discussed in the review of Huntington and Givens (1995), feeds disintegrate differently to grinding due to differences in physical structure. The distributions of particle size resulting from grinding are dependent on the proportion of the constituent plant parts (leaf:sheath:stem ratio), but different plant species also may behave differently (Huntington and Givens, 1995). In cereal and corn whole-crop forages, the seed fraction might disintegrate differently compared with vegetative plant parts.
|
Zero-Hour aNDF Content
The aNDF0 content of both IV and IS methods increased with increasing particle size, as illustrated in Figure 1b
, especially for B and CS. Using samples of dried winter wheat straw ground to pass 4.5-, 1.0-, and 0.25-mm screens, respectively, Udén (1992) found no difference in NDF0. In contrast, Mertens (2002) mentions that the finer ground samples in general are extracted more completely. The AOAC method (Mertens, 2002) for aNDF determination is validated for samples ground at 1-mm screen size and for feeds with starch content up to 50% of DM. The starch content of B used in this experiment was close to this upper starch limit. We observed increasing difference between aNDF0 and the aNDF residue after 6 h of incubation as particle size increased in samples of B and CS. The higher aNDF0 with increasing particle size (Figure 1b) and rapid loss of aNDF residue during the first 6 h of fermentation suggest that starch was extracted incompletely from aNDF0 in B when grind size exceeded 1 mm.
Figure 1b illustrates that aNDF0 was higher in IV samples analyzed using the Berzelius beakers with reflux (gray lines), compared with IS samples analyzed in the Dacron bags (black lines), probably due to particle loss from the bags during the washing procedure when using the IS method. For CS, GS, BP, and RC, the difference between the IS and the IV methods tended to decrease as particle size increased, indicating a reduced loss of particles during the washing procedure as particle size increased. In B, the influence of starch contamination might exceed the influence of particle loss on measurements of aNDF0, which may explain the increasing difference in aNDF0 found between IS and IV as particle size increased.
pH During Incubation
The LS means for pH of the media after fermentation were 6.7 and 6.0, for IVn and IVa respectively, which were close to target values of 6.8 (IVn) and 6.0 (IVa). There was a significant interaction (P < 0.0001) between replicate-within-method and incubation time affecting ending incubation pH. The influence of replicates-within-method could be due to variation in microbiological activity of the inoculums (Rymer et al., 2005). However, the magnitude of this interaction was in general small, averaging < 0.1 pH between replicates. Significant interactions were also found between feed, method, and incubation time as illustrated in Figure 2. It appears that the environment in the flasks developed differently, dependent on the feed and method pH (IVn vs. IVa). The decrease in pH during incubation mainly depends on the amount of fermentation acids produced and thereby the fermentability of the incubated feeds. For normal (IVn) as well as pH-adjusted (IVa) methods, GS and RC generated higher pH during the entire incubation interval compared with the other feeds, and RC often had pH above the initial incubation pH. Perhaps the high IV pH of these feeds was due to ammonia release during protein degradation. The difference in pH development during incubation among feeds could arise from differences in buffering capacity of the feeds (Jasaitis et al., 1987). Grant and Mertens (1992b) also found different changes in incubation medium pH between feeds as fermentation progressed, when comparing alfalfa hay, bromegrass hay, and corn silage. They observed the largest changes for corn silage, where initial pH of 6.8, 6.2, and 5.8 after 96 h of incubation declined to 6.65, 6.15, and 5.66, respectively. Their results do not indicate that changes in pH during incubation were dependent on initial pH level, as found in our experiment where differences in pH among feeds and among incubation times within feed were greater for IVa compared with IVn (Figure 2; e.g., for CS the initial pH was 6.8 and 6.0 and after 96 h of incubation was reduced to 6.58 and 5.68, respectively). In the experiment of Grant and Mertens (1992b), pH was reduced from 6.8 to 6.2 or 5.8 by replacing 18 and 40 mL of a 1,000-mL batch of incubation medium, respectively, with 1 M citric acid. In our experiment, 60 mL out of 1,000 mL of incubation medium was replaced by 1 M citric acid to obtain pH 6.0, which might have reduced the medium buffer capacity compared with that used by Grant and Mertens (1992b), and resulted in larger pH changes during incubation of some feeds.
|
Average pH measured in the 4 cows during the first 24-h period of IS incubation is presented in Figure 3. Within the total 24-h period the average pH was 5.8, but there was a large diurnal variation in pH during IS incubation. Neither the replicate effect nor the replicate by time interaction were significant for IS ruminal pH. However, cow, cow x time, and cow x replicate interactions were different (P < 0.01), indicating the value of replicating results and averaging observations across cows within replicate to accurately detect differences among methods, feeds, or grinds.
|
, 3, and 4.
|
|
|
Interactions between grinding screen size and method were observed for the ED and aNDF degradation parameters of some feeds (Table 2). The interactions between method and grinding for ED of GS and B were created by the significant interactions for the kinetic parameter D0 for these feeds. The highly significant interaction (P = 0.0001) between method and grinding for D0 of B occurred because D0 increased markedly as screen size used for grinding increased, and this tendency was more pronounced in IV compared with IS methods. This interaction could be caused by starch contamination of the aNDF0 because D0 is the difference between aNDF0 and I, which should not be affected by starch contamination after 96 h of incubation. Less significant interactions (P < 0.05) between method and grinding for D0 were found in GS and for L in BP. In GS, D0 decreased with decreasing particle size for IS incubated samples, whereas D0 was almost independent of particle size in IV incubated samples. Mean particle size of GS was small, resulting in loss of D0 and to a lesser extent I for IS compared with IV when grinding screens had apertures <2 mm.
Influence of Diet, Cow, and Time of Inoculum Collection.
Multiple cows and diets were used to provide diversity in the microbial population for IS and IV fermentations. To make comparisons consistent with the composite inoculum used for IV fermentations, parameter estimates for IS were based on the average residues measured in the 4 cows fed 2 different diets. Our experimental design replicated the results of IS, IVn, and IVa methods, and the variation of replicate-within-method is a part of the error term used to detect differences. In general, replicate-within-method variation was greater than variation within-replicates, providing more conservative estimates of probabilities of differences among methods.
Because IVn-Rep1, IVn-Rep2, IVa-Rep1, and IVa-Rep2 were run on wk 1, 3, 5, and 6 of this experiment, the effects of method and time of inocula collection are confounded. However, the 4 cows used to generate a composite inoculum were fed the same diets throughout the experiment, which should minimize any systematic differences in inocula over time. The average pH of the ruminal contents used to create the inocula was not different among cows or runs (5.8, 5.6, 5.9, and 5.9 for IVn-Rep1, IVn-Rep2, IVa-Rep1, and IVa-Rep2, respectively) suggesting the inocula were similar. To evaluate systematic differences in inocula, the differences in results between IVn-Rep2 and IVa-Rep1 or IVa-Rep2 and between IVn-Rep1 and IVa-Rep1 or IVa-Rep2 provided estimates of IVn versus IVa comparisons that were 2, 3, 4, or 5 wk apart in time of inoculation. There was no significant trend in these method differences, which also indicates no systematic change in inoculum during the time of this experiment. Although there is confounding in method and time of inocula collection in our experiment, we postulate that the difference in results for IVn and IVa method is due primarily to differences in the in vitro fermentation conditions (media pH of 6.7 or 6.0 for IVn and IVa, respectively). This conclusion agrees with that of Grant and Mertens (1992b) who compared in vitro fermentations at different pH using the same inoculum.
Influence of Incubation Method.
Incubation method affected all degradation parameters for all feeds except for RC where significant difference between methods was only found for kd. In Table 3, influences of method on individual degradation parameters and the ED are summarized. Within the IVa and IVn methods the sum of the D0 and I fractions were approximately equal, but the sum was lower (P < 0.01) for all feeds except RC using the IS method due to losses during washing of zero-hour bags. There was a tendency for D0 to be lower and I to be higher for IVa compared with IVn with all feeds except RC, but only the differences for CS were significant (P < 0.05). Grant and Mertens (1992b) observed that IV pH did not influence the D0 and I of forages incubated without starch at different buffer pH (6.8, 6.2, and 5.8) in contrast to our CS results. However, they observed that adding starch increased I at all IV pH. The inoculum donor for their study was a single nonlactating cow fed forage only, compared with our study which used a composite inoculum from 4 lactating cows fed a mixed forage-concentrate ration. Differences in the microbial organisms between inoculum sources may explain the discrepancy between the results of the 2 experiments.
Method IVa decreased kd (P < 0.05) for BP and RC compared with IVn, whereas no significant effect was found for CS, GS, and B. Although we observed no difference in kd of CS and GS between in vitro pH 6.0 versus pH of 6.7, Grant and Mertens (1992b) found that kd of NDF was reduced in forages when samples were incubated at pH 5.8, compared with incubation at pH 6.2 and 6.8, especially when starch was included in the fermentations. During IVa incubation of GS, average pH was above 6.0 (Figure 2), which might explain the lack of differences in kd between IVa and IVn for this feed. We observed that kd was less for IS than for IVn in all samples. The kd of the IS method was also less than that observed for the IVa method for B, CS, and GS. Lower kd from IS incubation compared with IV incubation was also found by Dewhurst et al. (1995). The pH magnitude (averaging 5.8 and 6.0 for IS and IVa, respectively) and pattern of IVa and IS fermentations were different (Figures 2 and 3), which may explain the lower kd of the IS method. Huhtanen et al. (1995) observed that rates of digestion measured IS were less than those calculated from ruminal evacuation. They suggested that slower rates of digestion in situ might be related to hindered flow of microorganisms, nutrients, and buffers through the small pores and open area of IS bags.
The L of the IVa method were higher than those of IS or IVn methods for CS, GS, and BP (numerically higher for RC). The L of the IS method was similar to the IVn method for all feeds except CS. Reducing IV buffer pH from 6.7 to 6.0 in present experiment increased L in agreement with Grant and Mertens (1992b), but the effect was larger. Increasing L due to decreasing pH could be related to reduced/delayed microbial adherence. Mourino et al. (2001) observed increased L and decreased adherence of bacteria to cellulose within 6 h of inoculation as the pH of the IV medium decreased from 6.5 to 5.5.
The ED of aNDF was calculated to determine the combined effects of D0, kd, and L on degradation (Mertens, 1977; Stensig, et al., 1994) for each method. The ED of the IS method was lower than that of the IVn method for all feeds except RC, which also was numerically lower. The ED of CS and GS was similar for IS and IVa methods, but IS was lower than IVa for barley and higher for BP. The ED of aNDF in RC was not different among methods.
Differences in aNDF residue amounts after IS and IV incubation in this experiment can be divided in 2 groups: 1) differences mainly due to technical conditions and 2) differences mainly due to biological conditions.
Compared with IV incubation in flasks, the amount of aNDF residues after IS incubation in Dacron bags may be reduced due to particle losses from the bags during incubation and washing after incubation (Hvelplund and Weisbjerg, 2000). Particle loss is probably the main reason for the lower sum of I and D0 within feed, obtained for the IS compared with the IV methods. When calculating degradability, residues of the feed fraction of interest (e.g., DM, OM, aNDF) after incubation are related to the initial content of the fraction. Using the IS incubation method, particle losses can lead to underestimation of the NDF0 content, which could result in an overestimation of the degradabilities. However, there is no indication that the ED obtained from the IS incubation were overestimated compared with the IV results obtained in this experiment. When using bags with a pore size of 42 µm in combination with feed samples that were either not ground or ground to pass a screen with 6-mm openings for IS incubation and compared with samples ground to pass a 1-mm screen for the IV incubation, Dewhurst et al. (1995) found higher OM degradability for most feeds when estimated by IS incubation compared with IV incubation. Investigating DM degradability of untreated and ammonia-treated barley straw, Graham and Åmen (1983) found higher degradability of DM after 168 h of IS incubation compared with IV incubation. In the experiment of Graham and Åmen (1983), straw was ground to pass a 1-mm screen, and nylon bags with a pore size of 20 µm was used for IS incubation. The difference in degradability was explained by inhibited IV digestion due to build-up of fermentation end products, in combination with higher particle losses from the IS incubation.
The experiments of Varel and Kreikemeier (1995), Robinson et al. (1999), and Spanghero et al. (2003) suggest that some differences between IV and IS incubations may be due to biological factors; however, their results are contradictory. Varel and Kreikemeier (1995) compared NDF degradation parameters obtained from IV and IS incubation, correcting the NDF residues from IS incubation for particle loss of 3%. They incubated samples of neutral detergent extracted bromegrass and alfalfa hay ground in a shear mill to pass a 1-mm screen and used Dacron bags with a pore size of 53 µm for the IS incubation. Their experiment revealed potential NDF degradabilities which were 6% lower and average kd that were 0.030 h–1 (approximately 48%) lower for IV compared with results obtained from the IS method. Varel and Kreikemeier (1995) also found lag time to be significantly higher for the IV method compared with the IS method. They explained the results with reduced activity of the rumen fluid inoculum because the rumen fluid is diluted 1:5 when inoculated into the tubes containing IV buffer solution. However, using a 1:4 ratio of ruminal fluid to medium, we observed shorter lags and faster kd for IVn than for IS when the same cows were used as IV donors and for IS incubations. After 48 h of incubation, Robinson et al. (1999) found higher NDF degradability when whole crop forages (ground to pass a 2-mm screen) were incubated IV using the a rotating jar IV system compared with IS. This agrees with our observations when we compare IS to IVn. In a similar experiment, Spanghero et al. (2003) also found higher NDF degradation after 48 h of incubation using a rotating jar IV system compared with IS, but their IS samples were ground more coarsely (4-mm screen) than those incubated IV (1-mm screen). These results and the variable responses in NDF degradation among feeds that we observed suggest that several factors and their interactions are responsible for differences between IS and IV methods.
It is interesting that when all degradation characteristics are combined to estimate ED, the IVa method agrees more closely with IS than did IVn. However, as discussed by Huhtanen et al. (2006), pH is one of many extrinsic factors influencing NDF degradation when measured using the IS. In addition to particle loss and pH, other differences in fermentation methods could contribute to differences in degradation characteristics between IS and IV measurements (e.g., differences in mixing of feed particles, concentration of individual VFA, and osmotic pressure in the fermentation bag or vessel).
In the present experiment, L appeared to be related to the magnitude and pattern of pH differences among methods (IVn < IS < IVa). Inoculum used for the IVn incubation was prepared from rumen contents originating from the cows used for the IS incubation and fermented with substrates in a buffer solution (Goering and Van Soest, 1970) designed to obtain stable pH of 6.8, which is thought to be optimal for fiber degradation (Grant and Mertens, 1992b). Lowest L was observed for IVn, which had the theoretical optimum pH for fermentation. Feed level and ration composition influence rumen microbial environment and degradation of NDF when pH declines to 6.0–6.2 (Huhtanen et al., 2006). In the present experiment, the cows used for the IS incubation and as IV inocula doners were fed TMR at production levels of intake. Average rumen pH of the 4 cows was 5.8, but there was a large diurnal variation in pH (Figure 3) during IS incubations compared with pH during IV incubations. For 67, 42, 42, and 0% of the 24-h period rumen pH was above 6.0 for the 4 cows, respectively. In comparison, pH after IVa incubation was below 6.3 during the whole incubation interval for all feeds and below 6.0 for CS, B, and BP. The L was significantly higher for the IVa method compared with the IS method, which could reflect a more pronounced delay in microbial adherence under incubation at continuously low pH (IVa), compared with incubation at low, but variable, pH (IS). Mourino et al. (2001) observed that the pH at which the inoculation occurred had a significant impact on lag time. Thus, L measured in vitro at low pH may be increased compared with IS at similar pH because the pH of the in vitro system is low at the time of inoculation and does not vary during fermentation as it does in the animal during bouts of rumination.
Degradation parameters for NDF are dependent on the incubation method, and therefore the method selected must be compatible with the way the ration evaluation system uses NDF degradation parameters. For feed evaluation, it is important that digestion characteristics representing the intrinsic potential of the feed are measured in a way in which fermentation system is not limiting. It appears that IV methods conducted at optimal pH obtain the largest potentially degradable fractions and rates that should be or approach the intrinsic limitation of the feed. However, if these degradation parameters are to be used in models of animal performance, they may need to be adjusted to reflect the conditions in the rumen. Conversely, degradation characteristics measured in suboptimal IS or IV systems represent a confounding of system and feed limitations. They will be useful in models only if the conditions under which they are measured are compatible with model assumptions and function. Because suboptimal conditions cannot be fully described with current state of knowledge, degradation characteristics measured under suboptimal conditions can only be compared within method or experiment and have limited utility in describing potential nutritive value of feeds.
Influence of Grinding.
Influences of grinding on degradation parameters found in the present experiment are summarized in Table 4 as mean values of the 3 methods. Only parameters D0 and kd in B, I in CS, and L in GS was influenced by sample grinding. The B had the largest particle size of all the feeds when grinded at screen size 2 and 4 mm. Increasing D0 and decreasing kd of B is probably an artifact of incomplete starch extraction from the aNDF0 of B ground at large screen size (4 mm) compared with small screen sizes (1 and 2 mm). The I found in CS was significantly higher when grinding at screen size 8 mm compared with grinding at screen size 4, 2, and 1 mm. It indicates incomplete extraction of neutral detergent soluble material from large particles. When no similar effects were seen for GS, it might be due to the fact that particle size of GS ground at 8 mm was lower than the corresponding particle size obtained in CS. Grinding significantly affected L in GS, but the effect was not consistent with development in particle size. Grinding had no influence on L in BP and RC, which could be due to the typically low L observed in concentrate feeds.
Udén (1992) suggested that grinding of feed samples and the resulting reduction in feed particle size can be expected to increase the kd and decrease L, as a result of increased surface area for microbial adhesion. Particle size interacts with method used for measurement of aNDF degradation because of potential losses from IS bags, but the biological impact of feed particle size on degradation characteristics is less clear. Robles et al. (1980) and Belyea et al. (1983) observed no effect of grinding particle size on rates of digestion. Mertens (1993) postulated that particle size may not be a major impediment for digestion because feed particles are porous rather than solid, which dramatically affects the particle size and surface area relationship. Huntington and Givens (1995) suggest that particle size might interact with incubation time because the effects of particle size are evident only at short incubation times.
Based on results from this experiment, there is little indication that the degradation characteristics of NDF are influenced by particle size when ground through screens with 1- to 8-mm openings other than effects related to the neutral detergent extraction. Although there must be situations in which the physical form of the feed limits or defines its degradation characteristics, it appears that increasing the particle size of ground samples had limited impact on the degradation of the feeds we evaluated.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The confounding between collection of inoculum for the IVa method and time was assumed to be of minor importance because donor cows were fed the same rations during the entire experimental period, and the variation of the difference between aNDF residues using the IVn and IVa method, respectively, did not seem to be systematically related to time interval between inoculum collections.
The IS method resulted in smaller fractions of potentially degradable and indegradable aNDF, possibly due to particle losses. In addition, the IS method had slower rates of digestion for all substrates, ascribed to variation in ruminal environment or fermentation conditions within in situ bags. The small differences in rates of degradation between IVn and IVa for forages suggests that the slower rate of digestion measured IS did not appear to be related to average pH during incubation.
Lag time increased markedly for all feeds when the pH of in vitro incubation was lowered from 6.8 to 6.0. The IVa method generated longer lag times for most feeds compared with the IS method, suggesting that the inhibitory effect of low pH on initiation of aNDF degradation is stronger when using the IV compared with the IS method. The diurnal variation in pH IS may be an important factor moderating the effects of IS pH on L.
Although the effect of fermentation methods on the pattern of degradation characteristics varied considerably among feeds, the extent of degradation of the IS method was related more consistently to IVa than to the IVn method. It appears that the IVn method, which obtained the largest extent of degradation for all feeds, estimates the degradation characteristics that represent the potential degradation of feeds (intrinsic feed value) that is important in feed evaluation for ruminants.
Received for publication June 4, 2007. Accepted for publication December 19, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. Huhtanen, A. Seppala, S. Ahvenjarvi, and M. Rinne Prediction of in vivo neutral detergent fiber digestibility and digestion rate of potentially digestible neutral detergent fiber: Comparison of models J Anim Sci, October 1, 2008; 86(10): 2657 - 2669. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
