Professional Documents
Culture Documents
L. C. GRIEL, JR.
Department of Veterinary Science
The Pennsylvania State University
University Park 16802
for biosynthesis of cell material (22). To were fitted with rumen and duodenal cannulas.
achieve optimal growth, the rate of ATP The latter was a Y-type plastisol cannula with
production (carbohydrate fermentation) must 2.5 cm i.d. and 135°C internal flange angle
equal its utilization (protein synthesis), or fer- (provided by B. P. Glenn, USDA-ARS, Belts-
mentation becomes uncoupled (12). This need ville, MD). The cannula was placed approxi-
for balance implies that rumen availabilities of mately 15 cm anterior to the bile ampulla.
carbohydrates and protein should be matched Cows were housed in individual tie stalls and
to maximize microbial protein synthesis. This milked and fed three times daily at 0600, 1400,
hypothesis is supported by the work of and 2200 h.
Herrera-Saldana et al. ( l l ) , who showed that The experimental design was a 4 x 4 Latin
microbial protein synthesis was maximized square with 20-d periods. The four treatments
when starch and protein sources (barley and consisted of diets formulated to contain high
cottonseed meal), which have relatively rapid and low RANSC and high and low RAP. The
rates of digestion, were fed. Exact amounts of treatments were 1) high RANSC and high
rumen-available carbohydrate and protein that RAP, 2) high RANSC and low RAP, 3) low
maximize microbial yields are not known. RANSC and high RAP, and 4) low RANSC
Hoover and Stokes (14) suggested that, when and low RAP. Ingredient composition of the
high percentages of grain are fed to high four diets is listed in Table 1. For high and low
producing dairy cows, DIP may be more limit- RANSC, high moisture shelled corn (HMSC)
ing than carbohydrate for microbial growth. and dry ear corn (EC) were used, respectively;
The objective of this experiment was to for high and low RAP, canola meal and blood
evaluate the effects of diets that vary in rumen- meal were used, respectively. The EC was
available nonstructural carbohydrate (RANSC) ground using a 2.2-cm screen. The HMSC was
and protein (RAP)on bacterial protein synthe-
stored in a sealed silo, obtained every 4 to 5 d
sis, rumen fermentation, flow of nutrients to
from a nearby commercial dairy farm, rolled,
the small intestine, milk production, and milk
component yield. and stored in plastic containers at 4°C. Soy-
bean meal (48% CP) was added to all diets as
a source of RAP so that low RAP diets were
MATERIALS AND METHODS
not grossly deficient in an available N source
Four multiparous Holstein cows (50 rt 8 for rumen microbes (26). High RANSC diets
DIM at the onset of the trial) were used. Cows contained more corn silage and more total
TABLE 1. Ingredient composition of experimental diets formulated for high and low xumen-available nonstructural
carbohydrate (RANSC) and high and low rumen-available protein (RAP) as a percentage of the DM.
High RANSC Low RANSC
High LOW High Low
Ingredient RAP RAP RAP RAP
Corn silage 23.5 37.1 14.0 24.3
Alfalfa-grass silage 25.2 16.0 26.0 20.7
High moisture shelled corn 31.7 30.7 . . . ...
Ear corn ... ... 39.5 38.8
Canola meal 7.0 ... 7.8 ...
Blood meal ... 4.8 . . . 4.3
soybean meal 7.3 5.7 7.4 6.2
Molasses 3.0 3.0 3.0 3.0
Urea .1 .1 .1 .1
Mineral-vitamin premix' 2.2 2.6 2.2 2.6
'High RAP contained dicalcium phosphate, 27.2%; sodium bicarbonate, 21.9%; trace-mineralized salt, 21.9%;
limestone, 11.4%; CaSO4, 4.4%; Mg03, 4.4%; Se premix (200 ppm), 5.7%; vitamin A, 650,400 IUkg; vitamin D,
453.600 IUkg; vitamin E, 4760 I U k g , 3.6%. Low RAP contained dicalcium phosphate, 34.6%; sodium bicarbonate,
19.3%; trace-mineralized salt, 19.3%; limestone, 11.6%; CaS04, 3.9%; MgOg, 3.9%; Se premix (200 ppm), 4.6%;
vitamin A, 650,400 IUlkg; vitamin D. 453,600 IUkg; vitamin E, 4760 IUkg, 3.0%.
The N:purine ratio was used to calculate the The four diets were evaluated in situ to
flow of bacterial N to the small intestine. determine RANSC and RAP. Fresh samples
Fecal grab samples were obtained at every were ground with dry ice using a Wiley mill
second duodenal sampling time (9 samples per (4-mm screen). An amount to yield 5 g of DM
period), frozen (-20°C) immediately, thawed at was weighed into duplicate 10- x 20-cm poly-
the end of the period, composited by cow, ester bags (Ankom, Fairport, NY).Bags had a
dried at 55°C and ground through a Wiley mean pore size of 53 f 10 p , and sample size
mill (1-mm screen). (DM) to surface area ratio was 12.5 mg/cm2.
Blood was drawn (20 ml) from the coc- Bags were incubated in the rumen of 2 cows
cygeal vein at 0900 and 1700 h on d 19 and 20 per period (in situ diet was the same as the diet
of each period into heparinized tubes and cen- fed) beginning on d 19 for 0, 1, 2, 3, 6, 12, 18,
trifuged (3000 x g for 15 min at 4'C), and the 24, 48, and 72 h after soaking at 39°C in
plasma was frozen (-2O'C) for analysis of urea distilled water for 15 min. Two blank bags
N (Technicon Autoanalyzer@; Technicon were included in each incubation and used to
Corp., Tarrytown, NY). correct for material entering the bags. Bags
Duodenal digesta flow and digestibilities were inserted in reverse order and retrieved as
were determined using Yb-marked insoluble a group, rinsed in cold tap water, and washed
DM. Ytterbium was bound to the insoluble in a washing machine under the gentle cycle
DM of each of the four TMR. Each diet was
soaked in tap water for 24 h with occasional for 10 min. Bags then were dried at 55°C for
stirring. Soaked material was rinsed with tap 48 h. The residues at 0 to 24 h were analyzed
water, filtered through three layers of cheese- for NSC and N as described previously. A
cloth, and hand squeezed. Insoluble material linear regression model was fitted to the log-
then was soaked in tap water in a ratio of 1:3 transformed residuals to determine disappear-
(wt/wt), and .6 g of YbCL3.6H20 were added/ ance rates. Rumen availability was calculated
100 g of wet material. After 48 h with occa- as
sional stirring, the Yb-marked feed was
washed and filtered as previously described RANSC or RAP = a + {b X [kd/(k,j + kp)l)
and dried at 60'C for 48 h. After air equilibra-
tion, 120 g of Yb-marked insoluble DM were where
weighed into 1-L plastic bags for dosing. a = 100 - antilog intercept (amount
Marker was dosed into the rumen at 0600 and
soluble at 0 h),
1630 h on d 11 to 20 of each period. Measured
Yb concentration averaged 14.1 mg/g of in- b = 100-a,
soluble DM across the four diets. k,j = digestion rate of b, and
Diets, duodenal digesta, and feces were ana- k, = passage rate (assumed to be .07/h).
lyzed for DM, CP, and ash (l), ADF, NDF, For the 16 regressions, r2 2 .93 except for one
ADF CP, and NDF CP (25), and Yb (10). replicate for CP disappearance (high RANSC
Organic matter truly digested in the rumen was and low RAP; r2 = .86).
calculated as OM intake minus the difference Data (except in situ data) were analyzed as
between OM and bacterial OM flow to the a 4 x 4 Latin square using the general linear
duodenum. An equivalent calculation was models procedure of SAS (27); cow, period,
made for N. and treatment were the effects in the model.
Protein solubility of diets was determined
using a borate-phosphate buffer (15). For anal- Single degree of freedom orthogonal contrasts
ysis of nonstructural carbohydrate (NSC),sam- were used to determine the effect of RANSC,
ples were ground using a UDY@ cyclone mill RAP, and their interaction. Significance was
(1-mm screen; UDY Corp., Ft Collins, CO); declared at P < .lo.
the procedure of Smith (30), modified to use
ferricyanide as the colorimetric indicator, was RESULTS AND DISCUSSION
used. Amino acid analysis was by the hydroly-
sis procedure described by Lynch et al. (18), In Situ Evaluation and Diet Composition
except that the hydrolysate was passed through
a Whatman filter (Number 541; Whatman In- The in situ estimates of RANSC (as a per-
ternational, Ltd., Maidstone, England). centage of NSC) was 17% higher for the high
TABLE 3. Chemical composition of experimental diets formulated for high and low rumen-available nonstructud
carbohydrate (RANSC) and high and low rumen-available protein (RAP).
High RANSC Low RANSC
High Low High Low
Item RAP RAP RAP RAP SEM'
DM, % 60.6 53.7 68.2 61 .O 1.4
NEL.' McaVkg of DM 1.64 1.66 1.61 1.61 . .
(70of DM)
CP 17.5 17.3 17.6 17.8 .I
Soluble CP 5.9 5.4 5.4 5.1 .2
UIP~ 6.0 8.2 6.1 8.7 . . .
ADF 21.9 20.1 21.2 21.1 .3
NDF 34.2 34.4 35.6 35.6 .4
Forage NDF 25.1 26.5 21.0 23.0 . . .
NSC4 35.8 37.7 35.0 34.6 .5
Ash 6.2 5.3 6.0 5.6 .1
ADF-Insoluble CP 1.1 .9 1.2 1 .o <.1
NDF-Insoluble CP 2.9 3.6 2.9 3.7 .1
IStandard error of the pooled treatment means (n = 16).
'Estimated from NRC (21).
3Calculated as ( 1 0 0 - rumen-available protein)/100 x CP. Rumen-available protein measured in situ.
4Nonstructural carbohydrate
The NSC content was 2 percentage units of OM digested in the rumen was similar
higher for the high RANSC (36.8%) than for across diets; however, the higher flow of OM
the low RANSC (34.8%) diets (Table 3). Acid to the small intestine for cows fed the low
detergent-insoluble CP was low for all diets, RANSC diets did not result in compensatory
and NDF-insoluble CP was higher for the low digestion of OM in the lower gastrointestinal
RAP diets than for the high RAP diets, most tract because the quantity of OM apparently
likely because of the blood meal in the low digested in the total tract was higher for cows
RAP diets. Blood meal contained 37% NDF- fed the high RANSC diets (15.9 vs. 15.1 kg/d;
insoluble CP. P < .01). Postruminal OM digestion tended ( P
Processing corn resulted in relatively small e .15) to be higher (3.9 vs. 3.3 kg/d) for cows
particle sizes; 67 and 88% of the particles were fed the low RAP diets, apparently because of
less than or equal to 3.35 mm for HMSC and
EC, respectively (Table 4). Because processing
can affect rumen digestion rate and availability
of corn (9).the objective was to minimize the TABLE 4. Particle size distribution of rolled high moisture
differences in particle size of the two forms of shelled corn (HMSC) and ground ear corn (EC).
corn; however, the rolled HMSC had a higher
proportion of larger particles (66 vs. 48% Screen
size HMSC EC SEI
21.17 mm) than did the ground EC.
- (46 retained on screen)' -
~
the higher amount of feed protein escaping (49.1 vs. 55.4%) in the rumen were affected
rumen fermentation in these diets (Table 8). adversely by the high RAP diets compared
The percentages of NSC digested in the with the low RAP diets (P e .08). The high
rumen (P < .02) and the percentages dqested RAP diets appeared to promote a more vigor-
in the total tract (P < .01) were higher for cows ous fermentation, as evidenced by the higher
fed the high RANSC diets than for those fed total VFA concentration (Table 7). Degrada-
the low RANSC diets (Table 5). The amount tion of NSC proceeded at a faster rate (Figure
of NSC digested postruminally was higher for 1A) for the high RAP dets than for the low
cows fed the low RANSC diets than for those RAP diets (12.3 vs. 8.0%/h). Rumen pH also
fed the high RANSC diets (1.9 vs. 1.3 kg/d; P was lower for the high RAP diets at 6 and 8 h
e .07). The higher total tract apparent digesti- postfeeding (Figure 2), which would be ex-
bility of NSC for the high RANSC diets com- pected to correspond to the time of maximal
pared with the low RANSC diets (92.6 vs. fiber digestion (16). The higher rate of NSC
digestion with the high RAP diets could have
89.7%, respectively) was due to the higher
reduced fiber digestion because of its effect on
rumen digestibility of the high RANSC diets. rumen pH or because of competition for sub-
A RANSC by RAP interaction existed on the strate between fiber- and NSC-digesting organ-
quantity of NSC digested in the total tract. isms (13). Differences in fiber digestibility be-
This interaction may have been caused by tween the high and low RAP diets also may
slight but nonsignificant (P > .lo) differences have been the result of differences in the rela-
in NSC intake among diets. tive proportions of corn and alfalfa-grass
The intake of NDF was higher (P < .01) for silages in these diets (Table 1). Alfalfa con-
the low RANSC diets than for high RANSC tains a higher amount of lignin and has a faster
diets (9.2 vs. 8.5 kg/d; Table 6), which rate, but lower extent, of digestion than do
reflected higher NDF content of the low grasses (31).
RANSC diets and higher DMI (Table 5 ) . Intake sometimes is reduced when a more
Digestion of NDF (50.6 vs. 55.2%) and ADF readily fermentable NSC source is substituted
Journal of Dairy Science Vol. 76, No. 4, 1993
1098 ALDRICH ET AL
TABLE 6. Effect of amount of rumen-available nonstructural carbohydrate (RANSC) and rumen-available protein (RAP)
on intake, duodenal passage, and digestibilities of ADF and NDF.
High RANSC Low RANSC
Effect ( P <)
High Low High Low
Item RAP RAP RAP RAP SEM C1 P2 c x P3
ADF
Intake, kg/d 5.5 5.0 5.7 5.3 .2 NS4 .07 NS
Passage to duodenum, kg/d 2.8 2.2 2.8 2.4 .2 NS .02 NS
Digested in rumen apparently, 96 48.5 56.2 49.7 54.6 3.0 NS .08 NS
Digested postruminally, kg/d -.5 -.5 -.9 -.8 .I .05 NS NS
Total tract digestion, kg/d 2.1 2.3 2.0 2.1 .2 NS NS NS
% of ADF intake 39.4 45.5 35.6 39.4 2.1 .06 .06 NS
NDF
Intake, kg/d 8.5 8.6 9.5 9.0 .2 .01 NS NS
Passage to duodenum. kg/d 4.3 3.7 4.5 4.1 .2 NS .08 NS
Digested in rumen apparently, t 48.8 55.8 52.4 54.6 1.7 NS .04 NS
Digested postruminally. kg/d -.7 -.7 -1.2 -1.0 .2 .ll NS NS
Total tract digestion, kg/d 3.5 4.1 3.8 3.9 <.1 NS .01 .05
% of NDF intake 41.5 47.3 40.7 44.4 .8 .06 .01 NS
'C = The RANSC effect.
zP = The RAP effect.
3C x P = The interaction between RANSC and RAP.
4P .2 .lo.
for a less fermentable source (4, 20). Whether RANSC diets. This fraction is assumed to be
this reduction is because of a lower rumen pH degraded rapidly in the rumen. At 6 and 8 h
(20) or because of the palatability of the NSC postfeeding, rumen pH was affected more by
source (barley) is uncertain, although the protein availability when the high RAP diets
reduction in intake on barley diets observed by produced lower rumen pH.
Casper et al. (4) was not associated with a The high RAP diets also stimulated higher
reduction in rumen pH. In contrast, Herrera- total VFA (P < .08) concentrations in the
Saldana et al. (11) observed no differences in rumen (145.8 vs. 137.1 pmoVml; Table 7). In
intake when barley replaced milo in diets for continuous culture fermenters, total VFA
lactating cows. production was increased linearly as DIP
(peanut meal) increased (32). Although
Rumen pH, NH3 N, VFA, RANSC had no effect on acetate concentra-
and Plasma Urea N tions, propionate concentrations tended to be
Rumen pH was significantly lower (6.28 vs. higher (30.7 vs. 27.0 pmoVml; P = .16), and
6.39) when cows were fed the high RAP diets acetate to propionate ratios were lower (3.08
( P < .01, Table 7). Cows fed the low RANSC vs. 3.49), for cows fed the high RANSC diets
diets also had lower (P < .07) rumen pH, but (P < .03). Cows fed the high RAP diets also
the magnitude of the difference was small. The had lower (3.12 vs. 3.45; P < .07) ratios
change in pH over time postfeeding (Figure 2) because of higher concentrations of propionate
indicates that the low RANSC diets had (31.3 vs. 26.4 pmoVml; P < .07). These ratios
smaller fluctuations than the high RANSC may have resulted from the considerable quan-
diets, which is particularly evident between 1.5 tities of propionate derived from the degraded
and 6 h postfeeding. The difference in pH at protein (35). Butyrate concentrations were
these hours may have been the result of the lower (16.3 vs. 18.8 pmoVml; P < .06), and
higher fraction of soluble NSC (Figure 1A) for isovalerate concentrations were higher (2.2 vs.
the high RANSC diets than for the low 1.9 pmoVm1; P < .04),when cows were fed
Journal of Dairy Science Vol. 76, No. 4, 1993
NONSTRUCTURAL CARBOHYDRATE AND PROTEIN 1099
Production of C4 and C5 VFA would be
expected to be higher when more DIP sources
are fed; however, measured concentrations re-
flect both production and utilization by rumen
6.5 - microbes; therefore, isobutyrate and valerate,
but not isovalerate, increased when the high
6.4 -
I RAP diets were fed. Others, also (4, 20) noted
Q
6.3 - variable effects of carbohydrate and protein
6.2 - sources on concentrations of the minor VFA.
When a less fermentable NSC source (corn)
6.1 - was fed, butyrate concentrations were higher
6.0 - than when a more fermentable NSC source
TABLE 7. Effect of amount of rumen-available nonstxuctural carbohydrate (RANSC) and rumen-available protein (RAP)
on rumen pH, NH3 N, VFA, and plasma urea N.
High RANSC Low RANSC
Effect (P <)I
High Low High Low
Item RAP RAP RAP RAP SEM C* Pi
PH 6.29 6.40 6.26 6.37 .01 .07 .01
NH3 N, mgldl 12.4 8.9 14.8 10.4 .9 .07 .o1
Plasma urea N, mddl 15.9 17.1 16.8 15.7 .7 NS4 NS
(pmoUml)
Total VFA 147.9 135.6 143.7 138.7 4.1 NS .08
Acetate (A) 90.7 86.5 90.7 90.7 1.9 NS NS
Propionate (P) 33.9 27.4 28.7 25.3 2.2 NS .07
A:P 2.9 3.3 3.3 3.6 .1 .03 .07
Butyrate 17.1 16.4 18.9 17.7 .7 .06 NS
Isobutyrate 1.2 1.1 1.2 1.1 <.1 NS .06
Valerate 2.7 2.0 2.3 2.0 .2 NS .03
Isovalerate 2.3 2.1 1.9 1.9 .1 .04 NS
'The interaction between RANSC and RAP was not significant (P > .lo).
2C = The RASNC effect.
3P = The RAP effect.
4P > .lo.
TABLE 8. Effect of amount of rumen-available nonstructural carbohydrate (RANSC) and rumen-available protein (RAP)
on intake, duodenal passage, and digestibility of N.'
High RANSC Low RANSC
Effect (P <)
High Low High Low
Item RAP RAP RAP RAP SEM C2 P3 C X
~~ ~ ~~ ~
N
Intake, g/d 700 691 755 722 17 .04 NSS NS
Passage to duodenum, g/d 555 580 566 648 32 NS .15 NS
Digested in rumen apparently, % 20.6 15.9 24.7 10.6 3.4 NS .03 NS
Digested in rumen truly, 7% 58.3 46.9 57.9 43.5 5.4 NS .04 NS
Total tract digestion, g/d 477 462 465 444 18 NS NS NS
% of N intake 68.3 66.9 61.9 62.1 1.3 .01 NS NS
Fecal NDF N, % of N 32.7 31.1 27.9 28.9 .9 .01 NS NS
Digestibility: % 89.6 89.7 89.5 89.2 .4 NS NS NS
NAN
Passage to duodenum, g/d 536 565 549 633 31 NS .12 NS
Bacterial N
Passage to duodenum, g/d 262 214 234 237 10 NS .04 .03
OMTD? g/kg 17.9 14.5 15.2 17.8 1.0 NS NS .01
NSCS digested, g/kg 36.7 29.3 35.5 42.7 2.7 .05 NS .03
'Least squares means.
2C = The RASNC effect.
3P = The RAP effect.
% x P = The interaction between RANSC and RAP.
5P > .IO.
6Total tract N digestibility corrected for NDF N (assuming that NDF N represents undigested feed N).
'The OM truly digested in the rumen.
8Nonstructural carbohydrate.
Greater quantities of NSC may have been fer- essential AA were higher when cows were fed
mented in the cecum, which would stimulate the low RAP diets. Amount of RANSC had no
microbial protein synthesis and, hence, fecal effect on AA flow to the small intestine.
output of N (23). When NDF N in feces was Generally, AA flows were lower than those
assumed to represent undigested feed N [not reported by McCarthy et al. (20) except for the
corrected for keratin (36)l and used to estimate flow of Lys, which was slightly higher (252
true digestibility, total tract digestibility of N and 260 vs. 246 and 228 gld) when cows were
averaged 89.5% across treatments and was un- fed the low RAP diets than when they were
affected by amount of RANSC or RAP (P > fed diets supplemented with fish meal (20),
.15). respectively. Blood meal supplied 29% of the
total CP in the low RAP diets. Lysine content
Passage of AA t o the Small Intestine of blood meal is high relative to other protein
supplements (6) and, because of its low rumen
The low RAP diets increased essential (P e degradability (34), would be expected to de-
.Ol), nonessential (P < .06), and total AA (P e liver a substantial quantity of Lys to the small
.02) flows to the small intestine (Table 9). The intestine.
flow of Ile was lower (100 vs. 117 g/d) for Schwab (29) determined that Lys was the
cows fed the low RAP diets (P e .Ol), and first-limiting AA for cows in peak and early
flows of Arg and Met were similar across lactation when the proportional passage of Lys
treatments. The flows of all other individual to the small intestine was 12.2 to 12.6% of
TABLE 9. Effect of amount of rumen-available nonstructural carbohydrate (RANSC) and rumen-available protein (RAP)
on duodenal AA flow.
High RANSC Low RANSC
Effect ( P e)'
High Low High Low
AA RAP RAP RAP RAP SEM C2 P3
Essential
A% 132 132 128 139 4 NS4 NS
His 53 74 50 82 3 NS .01
Ile 120 96 113 104 4 NS .01
Leu 259 310 263 331 11 NS .01
LYS 212 252 201 260 7 NS .01
Met 48 49 47 50 2 NS NS
Phe 131 158 131 168 5 NS .01
Thr 142 153 139 161 5 NS .02
Val 153 184 149 192 5 NS .01
Total essential 1248 1408 1221 1485 45 NS .01
Nonessential
Ala 181 210 179 222 6 NS .01
ASP 296 318 286 330 10 NS .02
Glu 37 1 375 370 401 15 NS NS
GlY 169 181 179 182 9 NS NS
Pro 135 143 139 154 7 NS NS
Ser 142 155 139 162 5 NS .02
TYr 101 100 98 103 3 NS NS
Total nonessential 1395 1482 1389 1553 53 NS .06
Total AA 2643 2890 2610 3038 98 NS .02
lThe interaction between RANSC and RAP was not significant ( P > .lo).
2C = The RASNC effect.
3P = The RAP effect.
4P > .IO.
essential AA flow. The low RAP diets and the diets (3.07 vs. 3.01%) and when they were fed
high RAP diets contained 17.5 to 17.9 and the high RAP diets (3.06 vs. 3.02%; P < .OS).
16.5 to 17.0% Lys as a percentage of essential An improvement in AA supply to the mam-
AA flow, respectively. Methionine was mary gland probably is not a plausible expla-
second-limiting or colimiting with Lys when it nation for the differences in milk protein per-
constituted 3.8 to 3.9% of essential AA flow centages observed in the current study because
(29). The low and high RAP diets contained neither the high RANSC diets nor the high
3.5 and 3.8% Met, respectively, as a percent- RAP diets increased the flow of AA to the
age of essential AA flow. Based on those small intestine. Increases in milk protein per-
criteria (29), Met may have been more limiting centages with increasing RANSC or RAP may
than Lys for all diets. have been related to slightly increased propi-
onate concentration in the rumen, which may
Milk Production end Composition have increased glucose supply to the mammary
gland (7) or stimulated insulin production,
Milk production averaged 39 kg/d across all which, in turn, may have enhanced AA uptake
treatments and was unaffected (P > .lo) by or protein synthesis by the mammary gland
amount of RANSC or RAP (Table lo). (24).
Production of 4% FCM was elevated by 1 kg/d The small but significant impacts of
(P < .01) for cows fed the low RANSC diets. RANSC and RAP on 4% FCM production and
Higher FCM for cows fed the low RANSC composition may have reflected the marginal
diets resulted from marginally higher milk but significant shifts in intake, rumen fermen-
production and fat percentages (P > .lo), tation end products, and nutrient flow to the
which resulted in higher fat yields (P < .Ol); small intestine. The direction of these changes
intakes of DM and NDF were higher when varied by diet and may have been compensa-
cows were fed these diets and may have stimu- tory in their effect on the milk production
lated higher FCM production. A RANSC by variables. For example, intake increased when
RAP interaction (P < .01) existed on the yield cows were fed the low RANSC diets, but total
of milk fat, because milk fat percentage was tract OM digestibility was lower. Improve-
somewhat lower when the high RANSC and ments in passage of bacterial N to the small
high RAP and the low RANSC and low RAP intestine when cows were fed the high RANSC
diets were fed. and high RAP diet were offset by decreased
Milk protein percentage was higher (P < passage of AA to the small intestine when the
.01) when cows were fed the high RANSC high RAP diets were fed. Passage of Met and
Journal of Dairy Science Vol. 76, No. 4, 1993
1104 ALDRICH ET AL