Development of Buffer Systems for pH Control and Evaluation of pH Effects on Fiber Digestion In Vitro

Development of Buffer Systems for pH Control and Evaluation of pH Effects on Fiber Digestion In Vitro R. J. GRANT1 and D. R. MERTENS Agricultural Research Service, USDA US Dairy Forage Research Center Maclson, WI 53719 ABSTRACT

in media pH from 6.8 to 5.8 resulted in a marlced reduction in NDF digestion; the largest apparent difference was extended digestion lag time. (Key words: fiber, artificial rumen, citrate, phosphate-bicarbonate)

An in vitro buffering system capable of pH control between pH 5.8 and 6.8 was developed to examine the effect of media pH on disappearance of NDF at various times of fennentation and to assess initially the effect of media pH on kinetics of NDF digestion. The pH conditions selected for evaluation of these buffer systems were 5.8, 6.2, and 6.8. Use of Mcllvaine's solution with sodium bicarbonate was not successful because of rapid drifting of pH downward during fennentation. To evaluate the effectiveness of citric or phosphoric acids as components of phosphate-bicarbonate buffer systems, alfalfa silage and a mixture of alfalfa silage and corn grain (1:1 mixture, dry basis) were fennented for 0, 12, 24, 48, and 72 h. The pH of each flask was measured at 0, 4, 12, 24, 48, and 72 h postinoculation, and pH was readjusted with bicarbonate solution when necessary. Drifting of media pH downward was more noticeable when phosphoric acid was used to adjust the buffer pH than with citric acid. Citric acid had no adverse effects on NDF digestion compared with phosphoric acid when used to adjust a phosphatebicarbonate buffer system. Alfalfa hay, bromegrass hay, and corn silage were incubated for 0, 12, 24, 48, 72, or 96 h at pH 5.8 or 6.8 using the phosphatebicarbonate buffer system adjusted with citric acid Estimation of kinetics of NDF digestion indicated that a decrease

Abbreviation key: pK = dissociation constant. INTRODUCTION

Received September 12. 1991. Accepted January 3. 1992. IDepartment of Animal Science, University of Nebraska, Lincoln 68583-0908. 1992 ] Dairy Sci 75:1581-1587

A reduction in fiber digestion may occur when readily available carbohydrates constitute a significant portion of the ruminant diet (1). This reduction could be caused by a direct effect of readily available carbohydrates on fiber digestion or by an indirect effect associated with reduction in ruminal pH when large proportions of starch are fed in ruminant diets. Mertens and Loften (8) examined the effect of starch on kinetics of forage fiber digestion in vitro using a model of fiber disappearance from the rumen developed by Mertens (7). These researchers found that additions of starch resulted in linear increases in lag time of fiber digestion, whereas digestion rate was unaffected when the pH of the in vitro system was maintained at 6.8. Furthennore, the potential extent of digestion was decreased with starch addition. They noted that these results could not explain depressions in fiber digestibility of the magnitude of in vivo studies in which starch was fed (9). Terry et al. (15) demonstrated minimal reduction in cellulose digestion with 40% glucose at pH 6.8. An alternative mechanism by which these carbohydrates, such as starch, depress fiber digestion might be a reduction of ruminal pH (14). The potential for negative effects of pH on fiber digestion exists in the high producing dairy cow, for which energy requirements for milk production typically are met with a diet containing 50 to 60% concentrates. Robinson et al. (10) fed dairy cows a diet containing

1581

1582

GRANT AND MERTENS

32% starch in the concentrate at 24 kg/d and found the average ruminal pH to be 6.1. The diurnal pattern for ruminal pH indicated that pH was below 6.2 for approximately 70 to 80% of each day. Cellulolysis was shown to be inhibited by pH below 6.0 to 6.2 both in vitro and in vivo (9, 15). Indeed. ruminal pH can be a significant factor in determining competition among bacteria (11). Ruminal microbes are quite sensitive to changes in pH; most prefer a pH range of 6.5 to 6.8 (11, 14). Cellulolytic bacteria, in particular, generally are more sensitive to low pH than are amylolytic species (16). Thus, fiber digestion may be depressed at low pH because of negative effects of pH on cellulolytic bacteria. The disparity between in vitro conditions, in which pH is maintained near 6.8, and the ruminal environment, where diurnal fluctuations in pH occur, was noted by Mertens and Loften (8). This disparity suggests that lower pH under conditions of high grain feeding may be responsible for altered fiber digestion kinetics. Hoover et al. (5), using continuous culture fermentation, demonstrated that either a high (7.5) or low (5.5) pH severely compromised fiber digestion of a 60% concentrate diet. Kinetic analysis of fiber digestion may delineate those factors most affected by low pH, which accompanies the feeding of readily digested carbohydrates. With these observations in mind, the study of fiber digestion kinetics at various controlled pH conditions below 6.8 becomes important. The objectives of this series of experiments were 1) to develop an in vitro buffering system capable of pH control between 5.8 and 6.8, 2) to examine the effect of media pH on NDF remaining at various times of incubation. and 3) to make some preliminary estimates of the effects of pH on digestion kinetics of NDF in vitro. MATERIALS AND METHODS

Description of Buffers Evaluated

The initial approach in development of a buffering system capable of maintaining pH between 5.8 and 6.8 was use of a broad-range buffer such as Mcllvaine's solution (2). This solution contains .1 M citric acid and .2 M NazHP04 and can be prepared for different pH Journal of Dairy Science Vol. 75, No.6,

1m

at constant ionic strength (2). An in vitro buffer system operates under continuous C~; thus, bicarbonate solution described in Goering and Van Soest (3) was added to obtain a pH of 5.8 or 6.8 under these conditions. Previous research with Good buffers (4) and Mcllvaine's solution in our laboratory indicated that bicarbonate ions are a crucial component of an in vitro ruminal buffering system kept under continuous C~ pressure. The Mcllvaine-bicarbonate solutions contained 43.9 ml of .1 M citric acid (prepared from citric acid monohydrate, C@lg0THZO), 49.5 ml of .2 M NaZHP04, and 6.6 ml of bicarbonate solution (3) for 100 ml of pH 5.8 buffer. A pH 6.8 buffer contained 21.5 ml of .1 M citric acid, 72.8 ml of.2 M NaZHP04, and 5.7 ml of bicarbonate solution (3) for 100 ml of total volume. A second, more direct approach involved adjustment of the buffer described by Goering and Van Soest (3) to desired pH with citric or phosphoric acids. The phosphate-bicarbonate buffer described by Goering and Van Soest (3) maintains pH at approximately 6.8 under a C~ atmosphere at 39"C. Use of phosphoric acid to adjust pH of buffer systems was described by Terry et al. (IS). The approximate solutions used to obtain pH 5.8 and 6.2 buffers using citric acid are as follows. For pH 5.8, to 960 ml of buffer solution described by Goering and Van Soest (3) were added 40 ml of 1 M citric acid. For pH 6.2, to 982 ml of buffer (3) were added 18 ml of 1 M citric acid. Buffer solutions were prepared just prior to use. For the pH 6.8 buffer treatment, both the Goering and Van Soest (3) buffer (which maintains pH at 6.8) and that system adjusted to pH 5.8 (with citric or phosphoric acids) and readjusted to pH 6.8 (with bicarbonate solutions) (3) were evaluated with respect to fiber digestion. In summary, four pH conditions (5.8, 6.2, 6.8, and 6.8 readjusted) and two acids (phosphoric or citric) were evaluated using Goering and Van Soest buffer as the basic solution. As a check against any error in mixing the buffer solutions, two 4O-ml volumes of each buffer solution were mixed, warJlk:d to 39"C under C~, and checked to verify pH. This procedure should be performed prior to any in vitro experiment to verify buffer pH and to allow calculation of any additional acid or bicarbonate necessary to obtain desired pH of the buffer solution.

BUFfERS AND pH EFFECf ON FIBER DIGESTION

In Vitro Experiments

To evaluate Mcllvaine-bicarbonate solution, timothy and alfalfa hay were incubated at pH of 5.8 or 6.8 through 72 h of fennentation. Each forage substrate was ground through a Wiley mill (l-mm screen; Arthur H. Thomas, Philadelphia, PA), and .5 g of sample was weighed into 125-ml Erlenmeyer flasks. Mcllvaine-bicarbonate buffer solution (40 ml) of pH 5.8 or 6.8, 2 ml of reducing solution (3), and 10 ml of inoculum were injected into flasks. The inoculum source was a nonlactating, fistulated Holstein cow consuming 4 kg/d of com grain with alfalfa and grass hay for ad libitum intake. The pH of the inoculum at the time of inoculation averaged 6.2. The pH of each flask was measured at 0, 24, 48, and 72 h. To evaluate the Goering and Van Soest buffer adjusted with either citric or phosphoric acids, forage and com samples were ground through a I-mm screen with a Wiley mill, and a .5-g sample of alfalfa silage and a 1:1 mixture (dry weight basis) of alfalfa silage and com grain were weighed into 125-ml Erlenmeyer flasks. Goering and Van Soest buffer solutions (40 ml) of pH 5.8 (adjusted with citric acid), pH 5.8 (adjusted with phosphoric acid), pH 6.2 (adjusted with citric acid), pH 6.2 (adjusted with phosphoric acid), pH 6.8 (reduced to pH 5.8 with citric or phosphoric acids and readjusted to 6.8 with bicarbonate solution), and pH 6.8 (Goering and Van Soest buffer alone) and 2 ml of reducing solution (3) were injected into each flask. All flasks were incubated at 39·C under continuous C~. Incubation times were 0, 12, 24, 48, and 72 h. During this evaluation experiment, when pH drifted downward by more than .2 pH unit, bicarbonate solution (3) was used to correct the pH to the desired level. The inoculum amount and source were the same as in the first experiment. Average inoculum pH was 6.3. The experiment to evaluate the effect of buffer pH on NDF digestion was replicated three times using citric acid to obtain desired pH conditions. The 3 x 2 x 6 factorial arrangement of treatments consisted of three forage substrates incubated at a pH of 5.8 or 6.8 for 0, 12,24,48,72, or 96 h of fennentation. Alfalfa was chosen to represent a legume hay, bromegrass to represent a cool season grass,

1583

and corn silage to represent a forage with significant amounts of starch that might have an impact on fiber digestion. The in vitro method used was basically that described in Goering and Van Soest (3) except for modifications with respect to buffer solutions. The inoculum source was the same as for the first two experiments; average inoculum pH was 6.3. For all experiments, NDF was determined by a modification of the methods of Goering and Van Soest (3). Modifications included addition of 2 ml of amylase solution at the filtering step only and use of .5 g of sodium sulfite. Dry matter was determined by drying a 2.O-g sample for 24 h at 105·C. All NDF values are reported on a blank-corrected, ashfree, DM basis. Analysis of variance of the mean NDF remaining at each time was by the general linear models procedure of SAS (12). Mean separation for significant factors was by StudentNewman-Keuls test (12) using P < .05. The experimental model included terms for replicate, sample, pH, and sample by pH effects. For kinetic analysis, the data were transformed logarithmically, and linear regression was used to estimate the parameters. Also, the nonlinear regression procedure of SAS (12) was used to obtain estimates for comparison with the loglinear method. The model for the kinetics of fiber digestion was that described by Mertens (6, 7). RESULTS AND DISCUSSION Development of Buffer System

Mcllvaine's solution contains citric acid and potentially could buffer the fermentation media more strongly within the pH range desired (5.8 to 6.8). The three dissociation constants (pK) for citric acid (25·C) are 3.13, 4.76, and 6.40 (2). However, Mcllvaine's solution with added bicarbonate did not maintain the desired pH beyond 24 h of fermentation (Figure 1). The loss in buffering capacity was especially apparent for pH 6.8. Because we were unable to maintain pH, no NDF digestion experiments were perfonned using this system. Because of the tricarboxylic nature of citric acid, it was tested as an acid to lower the pH of the buffer system described by Goering and Van Soest Journal of Dairy Science Vol. 75, No.6, 1992

1584

GRANT AND MER1ENS

TABLE 1. In vitro NDF residues as influenced by pH, substrate, and acid used to adjust buffer system to desired pH. Fermentation media pH 6.8 1

pH 6.2

pH 5.8 Time

Pbosphoric acid

(h)

- - - - - - - - - Alfalfa silage NDP

o 12 24 48 72

o 12 24 48 72

Citric acid

39.0

39.0

36.8

36.5

Phosphoric acid

38.5 28.2 25.0 23.3 21.6

Citric

Pbosphoric

Citric

~~

~~

~id

(~

pH 6.8 2

of initial DM)3 - - - - - - - - -

39.8 29.6 25.0 23.6 21.8

39.3 29.1 24.6 20.6 20.4

39.2 30.5 24.4 20.8 19.5

38.5 31.0 29.1 28.5 26.0 23.2 23.6 20.2 22.4 22.4 19.3 - - - - - - Alfalfa silage and com grain NDF (% of initial DM)4 - - - - - 25.9 26.2 25.0 26.1 27.1 28.6 28.0 22.3 21.7 20.6 20.2 18.9 18.4 18.0 16.0 16.9 13.8 14.3 13.5 14.2 14.5 14.7 14.6 12.1 12.9 11.4 12.0 11.1 14.9 13.2 10.7 11.6 10.8 10.9 9.5 IMedia adjusted to pH 5.8 and then readjusted to pH 6.8 with bicarbonate solution. 2Buffer described by Goering and Van Soest (3). 3mlF remainiDg after each fermentation period for either phosphoric or citric acids. 4substrate was a 1:1 mixture (DMbasis) of com grain and the silage used to generate the data in the lOp of the table.

(3). Terry et al. (15) previously had used phosphoric acid for the same purpose, but drifting of media pH downward was noted as a problem. The pK of phosphoric acid are 2.15 and 7.20 at 2S·C. The pK of citric acid are closer to the desired pH range of 5,8 to 6.8. Thus, although McDvaine's solution proved unsuitable as an in vitro buffer, the citric acid appeared to be ideally suited as a component of the phosphate-bicarbonate buffer of Goering and Van Soest (3), provided that it exerted no negative effect on NDF digestion. To evaluate the effect of citric acid on in vitro fermentation, citric acid was compared with phosphoric acid, which was evaluated earlier by Terry et al. (15), and shown not to affect fiber digestion. The percentage of NDF remaining at each time was not affected by the acid used to obtain the desired pH for fermentation, whether citric or phosphoric acid (fable 1). The NDF values for citrate- and phosphate-adjusted buffers were nearly identical at all times of fermentation for all pH conditions for both substrates. However, pH effect on NDF digestion was pronounced; pH 5.8 flasks showed the least active fiber digestion. Generally, fiber residues were similar at pH 6.2 and 6.8, although for alfalfa silage the residues were larger for pH 6.2 at 48 and 72 h. Journal of Dairy Science Vol. 75, No.6, 1992

The buffer with pH 6.8 described in Table 1 was obtained by first lowering the pH from 6.8 to 5.8 and then readjusting to 6.8 with bicarbonate solution. The NDF residues for the Goering and Van Soest buffer (pH 6.8) were similar to the concentrations obtained after pH readjustment Throughout 72 h of fermentation, no additional bicarbonate solution needed to be added

7.0

... :I:

.

. Q.

6.2

'ij

:IE

,.

,.. 5.0

l.--_~_..L-_""""_~_~_--"-_--'-_

•

10

20

30

40

50

60

70

Time of Fermentation (h)

Figure 1. Change in pH of fermentation media after various limes of fermentation when McIlvaine's solution and bicarbonate mixture (3) were used as in vitro buffer system. Substrates included timothy at pH 5.8 (-0-), timothy at pH 6.8 (---), alfalfa at pH 5.8 (-ll.-), and alfalfa at pH 6.8 (-A-). Note ineffectiveness of buffer system for maintaining pH after 24 h of fermentation.

1585

BUFFERS AND pH EFFECT ON FIBER DIGESTION

TABLE 2. Percentage of NDF remaining at selected times of in vitro fermentation as influenced by substtate and pH of buffer media. Bromegrass hay

Alfalfa hay Time

pH 5.S

pH 6.S

pH 5.S

43.5a 39.8a 29.3 25.2 23.8 24.4

42.5b 36.3b 28.1 25.1 23.6 24.2

68.9 66.051.~ 39.4a 34.~ 31.7a

12 24 48 12

96

Com silage pH 5.8

pH 6.8

SEM

NDF (% of initial DM) - - - - - - - - -

(h)

o

pH 6.8

6S.1 5S.2b 45.3 b 34.Sb 29.9b 21.1b

34.9b 28.8b t9.t b l3.8b 12.5b lUb

36.035.029.4a 21.5 a 11.3a 15.Sa

.226 .624 .S87 .142 .319 .362

a,~eans within a substrate column at each time with unlilce superscripts differ (P < .05).

to flasks containing citric acid for the forage substrate (Figure 2). In contrast, 1 to 1.5 ml of additional bicarbonate solution were added at 24 h to the flasks containing phosphoric acid to adjust pH to treatment levels (Figure 2). For the substrate containing starch (alfalfa silage and com grain), approximately 3 ml of additional bicarbonate solution were added for the citric acid flasks, whereas 5 ml were needed for the phosphoric acid flasks at 24 h. Terry et al. (15) demonstrated that romen liquor, adjusted to pH 5.5, stored at that pH for 1 h, and then readjusted to pH 6.9 with sodium carbonate, did not lose its original digestive capacity. They concluded that additional cations, anions, and excess acidity resulting from preparation of low pH media had no effect on the course of fiber digestion when this media was used for in vitro fennentation. However, as occurred in our study, these researchers had difficulty in maintaining constant pH at the more acidic pH conditions when phosphoric acid was used for pH adjustment. Citric acid maintained pH at the desired level with less additional bicarbonate solution than phosphoric acid. Given that citric acid exerted no apparent negative effects on NDF digestion, we used this phosphate-bicarbonatecitrate buffer system for our experiment to evaluate buffer pH effects on NDF digestion. Effect of pH on NDF Digestion The pH of the fennentation media dramatically affected NDF disappearance, as in the buffer evaluation study (fable 2). With alfalfa hay, pH effects were significant (P < .05) at 0

and 12 h. Beyond 24 h, the NDF remaining was essentially identical for both pH levels. Alfalfa hay is typically a rapidly digesting forage (8), so this result was not surprising. The NDF of bromegrass hay remaining at each time other than 0 h differed significantly (P < .05) for each pH. As with alfalfa and bromegrass, pH of 5.8 increased the amount of NDF remaining for com silage, although in this case the differences were significant (P < .05) at all times of fennentation. With this variation in forage response to pH, it is understandable that a significant (P < .0001) pH by substrate interaction occurred for all times other than 0 h. The main effects of substrate and pH were significant (P < .0001), as was

7.•

s.• s.•

..

L-~_~_~-,-_--,_~-,-_--,_

"

"

72

Time 01 Fermentalion (h)

F~ 2. Effectiveness of the phosphate-bicarbonate buffer system descn'bed by GoeriDg and Van Soest (3) at maintaining pH of 6.8 and the ability of this buffer system to maintain pH of either 5.8 or 6.2 when adjusted with citric or phosphoric acid for alfalfa silage substrate. Buffee and pH combinatioos iDcluded citrate at pH 5.8 (---), citrate at pH 6.2 (--0-), phosphate at pH 5.8 (-.-), phosphate at pH 6.2 (-~-), and buffer at pH 6.8 (-11-). Numbers 1 and 2 denote addition of bicarbonate solution (3) to readjust pH to desired value.

Journal of Dairy Science Vol. 15, No.6, 1992

1586

GRANT AND MERTENS

effect of pH on lag time was significant only for bromegrass and corn silage. Rate of digestion was significantly reduced (P < .05) by a pH of 5.8 for the corn silage substrate only Bromegrass Com Media Alfalfa SEM bay silage pH bay (fables 3 and 4). With linear regression, pH significantly affected (P < .05) digestion rate Indigestible residue of corn silage but exerted no effect on rate for (% of initial DM) alfalfa or bromegrass. Media pH showed no 11.42 1.15 30.76 5.8 24.46 effect on indigestible residue, although residue 11.59 24.11 25.93 6.8 was greater (nonsignificantly) at pH 5.8. For - - Rate of digestion (/h) - slow rates of digestion, as seen in our low pH .041& .007 .101 .050 5.8 treatment, indigestible residue may not be esti.054 .067b .106 6.8 mated accurately with only 72 or 96 h of Digestion lag (h) incubation (13). The effect of low pH on end14.24& 9.68& 1l.80& 1.00 5.8 point of digestion was significant for 7.61 b 6.79b 6.48b 6.8 bromegrass hay and com silage using linear a,!>Yalues within a substrate column for each kinetic regression (Table 4). parameter differ because of pH (P < .05). The differences in lag time among forages may be due to differences in forage tissue characteristics that require physical and chemical alteration prior to bacterial attachment and the replicate effect (P < .(01). Examination of subsequent fiber degradation (8). Within each the standard error of the mean (fable 2) indi- forage and averaged over all forages, however, cates that the most variation among replicates low pH unquestionably increased the lag time occurred at 12, 24, and 48 h. prior to NDF digestion. This effect might be Although a minimal number of fermenta- attributable to adaptation of the microbial poption times was used in this experiment, useful ulation of the inoculum to a low pH environpreliminary information concerning fiber ki- ment. A microbial population adapted to pH netics could be obtained. Thus, in addition to 6.2 to 6.5, as in this study, which abruptly is analysis of results at each time of fermentation inoculated into a media of pH 5.8, might ex(fable 2), data also were analyzed to obtain hibit a lag time prior to digestion. Theoreticalkinetic parameters such as discrete lag time, ly, ruminal cellulolytic bacteria would be infractional digestion rate, and indigestible resi- hibited by low pH, perhaps via a chemiosmotic due. Although these results must be evaluated mechanism (11). cautiously because the number of fermentation times used to estimate kinetic parameters was small, they give additional insight into the possible mechanisms whereby pH alters diges- TABLE 4. Influence of substrate and pH of buffer on in vitro kinetics of NDF digestion estimated by logarithmic tion of NDF. transformation of data and linear regression techniques. Table 3 illustrates the effect of pH and substrate on digestion kinetics as estimated by Media Alfalfa Brornegrass Com SEM hay silage hay nonlinear regression. In contrast, Table 4 pH shows the estimates obtained when the same 96-h Endpoint of digestion data were analyzed by logarithmic transforma- - (% of initial DM) - tion with linear regression. Generally, non- 5.8 31.68& 15.78& .36 24.38 27.74b 11.29b 24.16 linear regression estimated digestion rates 19% 6.8 higher than linear regression of log- - Rate of digestion (/h) - .037& transformed data. .007 5.8 .088 .045 .082 .044 .053 b The largest influence that pH exerted on 6.8 NDF digestion appeared to be lengthening lag - - Digestion lag (h) - 10.51& 9.17& time (fables 3 and 4). For all substrates, lag 5.8 8.23 .45 5.35 b 6.08 3.27b time increased significantly (P < .05) at the 6.8 low pH when estimated by nonlinear regresa,!>Yalues within a substrate column for each kinetic sion. When linear regression was used, the parameter differ because of pH (P < .05).

TABLE 3. Influence of substrate and pH of buffer on in vitro kinetics of NDF digestion estimated by nonlinear regression techniques.

Journal of Dairy Science Vol. 75, No.6, 1992

BUFFERS AND pH EFFECT ON FIBER DIGESTION

Although it is uncertain whether low ruminal pH induced by the diet has an effect similar to that of low pH in vitro on NDF digestion, these data suggest that lag time rather than rate is the parameter most likely to be of practical importance in explaining in vivo results. Considering that the typical dairy cow may have ruminal pH below 6.2 for a large portion of the day, theoretical justification exists for further examination of the effect of pH on in vitro fiber digestion kinetics. REFERENCES I EI-Sbazly, K., B. A. Dehority, and R R Iohnson. 1961. Effect of starch on the digestion of cellulose in vitro and in vivo by rumen microorganisms. I. Anim. Sci. 20:268. 2 EIving, P. I., I. M. Markowitz, and I. Rosenthal. 1956. Preparation of buffer systems of constant ionic strength. Anal. Chern. 28:1179. 3 Goering, H. K., and P. I. Van Soest. 1970. Forage fiber analyses (apparatus, reagents, procedures, and some applications). Agric. Handbook No. 379. ARSUSDA, Washington, DC. 4 Good, N. E., G. D. Winget, W. Winter, T. N. Connolly, S. Izawa, and RM.M. Singh. 1966. Hydrogen ion buffers for biological research. Biochemistry 5:467. 5 Hoover, W. H., C. R Kincaid, G. A. Varga, W. H. Thayne, and L. L. Iunkins, Ir. 1984. Effects of solids and liquid flows on fermentation in continuous culture. IV. pH and dilution rate. I. Anim Sci. 58:692. 6 Mertens, D. R 1973. Application of theoretical mathematical models to cell wall and forage intake in

1587

ruminants. PhD. Diss., Cornell Univ., Ithaca, NY. 7 Mertens, D. R. 1977. Dietary fiber components: relationship to the rate and extent of ruminal digestion. Fed. Proc. 36(2):187. 8 Mertens, D. R, and 1. R. Loften. 1980. The effects of starch on forage fiber digestion kinetics in vitro. I. Dairy Sci. 63:1437. 9 Orskov, E. R, and C. Fraser. 1975. The effects of processing of barley-based supplemenls on rumen pH, rate of digestion, and voluntary intake of dried grass in sheep. Br. 1. Nutr. 34:493. 10 Robinson, P. H., S. Tamminga, and A. M. Van Vuuren. 1986. Influence of declining level of feed intake and varying the proportion of starch in the concentrate on rumen fermentation in dairy cows. Livest. Prod. Sci. 15:173. 11 Russell, 1. B., W. M. Sharp, and R L. Baldwin. 1979. The effect of pH on maximum bacterial growth rate and its possible role as a determinant of bacterial ~tition in the rumen. 1. Anim. Sci. 48:251. 12 SAS User's Guide: Statistics. 1984. SAS Inst., Inc., Cary, NC. 13 Sharma, B. K., and R A. Erdman. 1988. Rate and extent of in situ digestion of medium and high quality alfalfa and orchardgrass neutral detergent fiber as determined by extended periods of incubation time. I. Dairy Sci. 71:3509. 14 Stewart, C. S. 1977. Factors affecting the cellulolytic activity of rumen comenls. Appl. Environ. Microbiol. 33:497. 15 Terry, I. M, A. Tilley, and G. E. Outen. 1969. Effect of pH on cellulose digestion under in vitro conditions. 1. Sci. Food Agric. 20:317. 16 Therion, I. 1., A. Kistner, and 1. H. Komelius. 1982. Effect of pH on growth rates of rumen amylolytic and lactilytic bacteria. Appl. Environ. Microbiol. 44:428.

Journal of Dairy Science Vol. 75, No.6, 1992