“IIIIIIII I I
I I IIIIIIIIIII
mIla

 

L81 BRAHIES

W Illllll'lllllllllIlllllllllllllllll| I

3 1293 01019843

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

Voluntary Intake and Feeding Behavior of
Dairy Cows in Response to Rumen Fill
from Forage Fiber

presented by

Richard Gary Dado

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Animal Science

 

 

M a' pfgfesvsor

Date Iii/H/Qia

MS U is an Affirmative Action/Er; ual Opportunity Institution 0- 12771

 

LIBRARY
Mlchlgen State
Unlverslty

 

 

 

PLACE II RETURN BOX to remove this checkout from your record.

TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

MR 3 1 199/»

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU In An Affirmative Action/Equal Opportunity lnetitttton

 

VOLUNTARY INTAKE AND FEEDING BEHAVIOR
OF DAIRY COWS IN RESPONSE
TO RUMEN FILL FROM F ORAGE FIBER

By

Richard Gary Dado

Major Professor: Michael S. Allen

AN ABSTRACT OF A DISSERTATION

Subnfifledto
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Animal Science
1993

ABSTRACT

VOLUNTARY INTAKE AND FEEDING BEHAVIOR OF DAIRY COWS IN
RESPONSE TO RUMEN FILL FROM FORAGE FIBER

By
Richard Gary Dado

Understanding factors that affect voluntary feed intake and mechanisms associated with
intake control is vital to improve efficiency of milk production by dairy cows. Three
experiments were conducted to address the hypothesis that forage fiber, as defined by
neutral detergent fiber (NDF), and indigestible forage NDF limit intake by cows in early
stages of lactation because of rumen fill. Because cows consume multiple meals each
day, within-day measurement of feeding behavior is required to adequately study
intake. A computer data acquisition system was developed for continuous monitoring
of feed and water intakes, chewing activity, reticular motility, and ruminal pH. In
experiment one, measuring 12 cows for 5 d in crossover designs provided adequate
statistical power to detect meaningful treatment effects during a preliminary feeding
behavior study. In experiment two, 12 cows (17 days in milk) were challenged with
rumen fill as dietary NDF or inert bulk to determine if filling characteristics of forage
diets varied with fiber content. Inert bulk decreased intake for 35% but not for 25%
NDF diets. Rumen volume was similar for 35% NDF diets whether or not inert bulk
was present. These data support the hypothesis that high NDF diets limit intake
because of rumen fill. Changes in feeding behavior or rumen function were insufficient
to maintain intake during high rumen fill. In experiment three, 12 cows (13 days in
milk) were offered one of two 35% NDF diets with alfalfa silages that initially
contained similar NDF concentrations (40%) but different NDF digestibilities (40 vs.
45% after 24 h in vitro fermentation). Silage with higher NDF digestibility increased
intake by 1.0 kg/d and milk production by 1.9 kg/d. During the study, diets differed in

Richard Gary Dado

both NDF content (1.8% units) and NDF digestibility (3% units), perhaps from loss of
digestible NDF in the higher digestible forage during feed-out. Equal NDF intakes
suggest that intake was limited by NDF content and not by indigestible NDF. If
increases in NDF digestibility prior to ensiling promote decreases in NDF content after
ensiling, benefits from high NDF digestibility may be substantial. Future research is
needed to determine if other dietary factors influence rumen fill and if cows vary in

their ability to accommodate fill.

This dissertation is dedicated to the memory of my friend, Mark Peters.
Your time here was short, but our friendship remains.
May we be able to fulfill the dreams you left behind.

iv

ACKNOWLEDGMENTS

The Beatles once wrote that we "get by with a little help from (our) friends."
Certainly this work would have been impossible without the significant contributions of
many individuals whom I have met at Michigan State University. Special appreciation
is extended to Dr. Michael Allen, for serving as my program mentor and for providing
me freedom to develop an enjoyable research project. I also thank Drs. Herbert
Bucholtz, Ted Ferris, Paul Kindel, and Duane Ullrey for serving as members of my
graduate committee, and Dr. John Gill for statistical advice.

Several farm personnel were vital to the success of dairy trials conducted for this
project, including Bob Kreft, Bob Story, Bruce Kurzhals, Gordon Galloway, Randy
Bontrager, Irv Sikema, students Dave Princer, Jeff Hulbert, and Ron Rice, and student
milkers Gary, Dean, Tim, and Ed, among others. I also thank Bary Darling and his staff
for assistance with forage production, and Bernie Fehr for technical electronic expertise.

Assistance in the laboratory and friendship provided by David Main, Dewey
Longusky, and Vicki Pulling is also appreciated. Fellow graduate students Frank Amdt,
Joe Domecq, Al Jordan, Katherine Knowlton, Rick Kohn, Steve Mooney, Kitty O'Neil,
and MaryBeth Roe were invaluable for their assistance with projects, advice, listening,
and friendship. Graditude is also expressed to the more than 30 individuals who
willingly assisted with rumen explorations, including Jim and Becky Connors, Dave
Krueger, and the students from ANS 059D. Thanks also go to student workers Curt,
Chris, Dan, and Dan for your patience and perseverance.

Finally, I wish to thank my patient and loving wife, Gwen, for sharing with
feces, rumen evacuations, and hot-weighing, and for providing financial support,

cookies, and our new daughter, Bethany.

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... ix
LIST OF FIGURES ....................................................................................................... xiv
LIST OF ABBREVIATIONS ...................................................................................... xvii
PROLOGUE .................................................................................................................... 1
CHAPTER 1

Literature Review ....................................................................................................... 4
IMPORTANCE OF FEED INTAKE ................................................................... 4
LONG-TERM INTAKE CONTROL IN RUMINANTS .................................... 7

Physical Control of Intake ............................................................................. 8
Physiological Control of Intake ................................................................... 17
Integrated Theories of Intake Control .......................................................... 19
SHORT-TERM INTAKE CONTROL IN RUMINANTS ................................ 22
Control of Individual Meals ......................................................................... 23
Dietary Modulation of Feeding Behavior .................................................... 25
Methods of Measuring Short-term Intake and Behavior ............................. 29
A Model of Feeding Behavior ..................................................................... 31
FORAGES, FIBER, AND INTAKE .................................................................. 34
Variation in Forage Consumption Due to NDF Concentration ................... 35
Variation in Forage Consumption Due to NDF Digestibility ...................... 49
Plant Factors Affecting Fiber Digestibility .................................................. 58
Future Directions ................................................................................... . ...... 62
CHAPTER 2

Continuous Computer Acquisition of Feed and Water Intakes, Chewing,

Reticular Motility, and Ruminal pH of Cattle .......................................................... 63
ABSTRACT ....................................................................................................... 63
INTRODUCTION ............................................................................................. 64
MATERIALS AND METHODS ....................................................................... 66

Acquisition System ...................................................................................... 66

vi

 

Feed intake ............................................................................................. 66

Water intake ........................................................................................... 67
Chewing activity .................................................................................... 67
Reticular motility ................................................................................... 7O
Ruminal pH ............................................................................................ 70
Data collection ....................................................................................... 71
Data interpretation .................................................................................. 73
Validation ..................................................................................................... 78
RESULTS AND DISCUSSION ........................................................................ 82
CONCLUSIONS ................................................................................................ 89
CHAPTER 3
Variation In and Relationships Among Feeding, Chewing, and Drinking
Variables for Lactating Dairy Cows ........................................................................ 90
ABSTRACT ....................................................................................................... 90
INTRODUCTION ............................................................................................. 91
MATERIALS AND METHODS ....................................................................... 92
Cow Trial ..................................................................................................... 92
Statistical Analysis ....................................................................................... 93
RESULTS .......................................................................................................... 95
DISCUSSION .................................................................................................. 104
CONCLUSIONS .............................................................................................. l 13
CHAPTER 4
Intake Limitations, Feeding Behavior, and Rumen Function of Cows
Challenged with Rumen F111 from Dietary Fiber or Inert Bulk ............................. 114
ABSTRACT ..................................................................................................... 1 14
INTRODUCTION ........................................................................................... 115
MATERIALS AND METHODS ..................................................................... 117
Experimental Design and Data Collection ................................................. 117
Sample and Statistical Analysis ................................................................. 120
RESULTS ........................................................................................................ 123
Production and Nutrient Digestibility ........................................................ 123
Rumen Characteristics ............................................................................... 127
Feeding Behavior ....................................................................................... 132
DISCUSSION .................................................................................................. 135
CONCLUSIONS .............................................................................................. 147

vii

CHAPTER 5
Harvesting Alfalfa Forages with Similar Fiber Contents but Different

Fiber Digestibilities from Production-Scale Fields ................................................ 148
ABSTRACT ..................................................................................................... 148
INTRODUCTION ........................................................................................... 148
MATERIALS AND METHODS ..................................................................... 150
RESULTS AND DISCUSSION ...................................................................... 153
CONCLUSIONS .............................................................................................. 162

CHAPTER 6

Enhanced Intake and Production of Cows Offered Ensiled Alfalfa with

Higher Neutral Detergent Fiber Digestibility ......................................................... 163
ABSTRACT ..................................................................................................... 163
INTRODUCTION ............................................................................................ 164
MATERIALS AND METHODS ..................................................................... 166

Forage Harvest and Selection ..................................................................... 166
Experimental Design and Data Collection ................................................. 167

Sample and Statistical Analysis ................................................................. 171
RESULTS ......................................................................................................... 174
Forage and Diet Composition ..................................................................... 174

Cow Response ............................................................................................ 178
DISCUSSION .................................................................................................. 182
CONCLUSIONS .............................................................................................. 191
EPILOGUE .................................................................................................................. 192
LIST OF REFERENCES .............................................................................................. 195
APPENDDC .................................................................................................................. 212

viii

LIST OF TABLES

CHAPTER 1

Table 1. Summary of studies where various forms and quantities of inert
bulk have been added to the reticulo-rumen to examine the effect
of decreased rumen capacity on voluntary intake of ruminants .............. 10

Table 2. Empirical equations developed with stepwise multiple regression to
predict DMI for lactating cows ............................................................... 21

Table 3. Summary of studies that examined the effect of dietary NDF
concentration or source as part of complete rations on intake and
production of relatively high producing lactating dairy cows.

Different NDF concentrations were obtained by changes in fiber
source, forage quality, or foragezconcentrate ratio .................................. 37

Table 4. Summary of studies that examined the effect of dietary N DF
digestibility as part of complete rations with constant NDF
concentrations on intake and production of lactating dairy cows.
Different NDF digestibilities were obtained by changes in fiber

source or forage quality ........................................................................... 55

CHAPTER 2
Table 1. Monitoring equipment. ............................................................................ 66
Table 2. Computer, software, and data acquisition boards .................................... 72

Table 3. Output from the feeding behavior data interpretation program
generated for eating, ruminating, and drinking bouts for one cow
measured for l d ...................................................................................... 79

Table 4. Intake and chewing activity of 10 cows measured for 3 (1 using
computer acquisition and manual observation ........................................ 82

Table 5. Relationship of intake and chewing activity of 10 cows measured
with computer acquisition and interpretation (independent data)
and manual observation (dependent data) ............................................... 84

CHAPTER 3
Table 1. Summary statistics for milk production and feeding behavior of six

primi- and six multiparous lactating Holstein cows measured for 96
10 d ..........................................................................................................

ix

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

CHAPTER 4
Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Pearson correlation coefficients among feeding behavior variables
for individual eating, ruminating and drinking bouts .............................. 99

Pearson correlation coefficients among feeding behavior variables
for 12 cow means averaged across 10 d of measurement. ..................... 101

Pearson correlation coefficients among feeding behavior variables
for six primiparous and six multiparous cow means averaged
across 10 d of measurement ................................................................... 102

Variance component estimates for feeding variables using three
linear models .......................................................................................... 103

Estimated cow numbers required for 80% probability of detecting
significant contrast differences between two treatment means for
feeding variables for random and complete block designs .................... 105

Estimated cow numbers required for 80% probability of detecting
significant contrast differences between two treatment means for
feeding variables for covariate and Latin square designs ...................... 106

Ingredient and nuuient composition of 25% NDF, low fiber diet
(LF) and 35% NDF, high fiber diet (HF) offered to 12 cows during
the first 14 wk of lactation ..................................................................... 118

Nutrient composition of forages and concentrates used to formulate
low fiber (LF) and high fiber (HF) diets ................................................ 119

Milk production and feed intake for 12 cows receiving low fiber
(LF) or high fiber (HF) diets without (+0) or with (+B) added
rumen inert bulk .................................................................................... 125

Apparent total tract digestibility of nutrients, fecal output, fecal
composition, and fecal fiber particle size from 12 cows receiving

low fiber (LF) or high fiber (HF) diets without (+0) or with (+B)

added rumen inert bulk .......................................................................... 126

Rumen digesta characteristics for 12 cows receiving low fiber (LF)
or high fiber (HF) diets without (+0) or with (+B) added rumen
inert bulk ................................................................................................ 128

Concentration and molar proportion of VFA in rumen fluid from
12 cows receiving low fiber (LF) or high fiber (HF) diets without

(+0) or with (+B) added rumen inert bulk .............................................. 129
Rumen digestion kinetics of NDF and rumen apparent digestibility

of NDF from 12 cows receiving low fiber (LF) or high fiber (HF)

diets without (+0) or with (+B) added rumen inert bulk ........................ 131

Table 8.

Table 9.

CHAPTER 5
Table 1.

Table 2.
Table 3.

Table 4.

CHAPTER 6
Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Eating, ruminating, and drinking activities for 12 cows receiving
low fiber (LF) or high fiber (HF) diets without (+0) or with (+B)
added rumen inert bulk ........................................................................ 133

Reticular contractions and ruminal pH for 12 cows receiving low
fiber (LF) or high fiber (HF) diets without (+0) or with (+B) added

rumen inert bulk ..................................................................................... 134
Characteristics of alfalfa forage harvested during the 1992 growing

season from two adjacent 8 ha fields ..................................................... 155
Composition of alfalfa forages before and after ensiling ....................... 158

Correlation matrix among descriptors of forage quality for all five
forage silages .......................................................................................... 159

Potential improvement in intake and milk production in early
lactation cows because of an increase in NDF digestibility between
forages D and E ...................................................................................... 161

Characteristics of low digestible fiber (LDF) and high digestible
fiber (HDF) alfalfa silages prior to commencement of animal
experiment. ............................................................................................. 168

Ingredient and nutrient composition of low digestible fiber (LDF)
and high digestible fiber (HDF) diets offered to 12 cows during the
first 10 wk of lactation ........................................................................... 169

Nutrient composition of forages and concentrate sampled during
animal experiment and used to formulate low digestible fiber
(LDF) and high digestible fiber (HDF) diets ......................................... 175

In vitro NDF digestion characteristics of low digestible fiber (LDF)
and high digestible fiber (HDF) alfalfa silages and diets sampled
during animal experiment. ..................................................................... 176

Milk production and body characteristics of 12 early lactation cows
receiving low digestible fiber (LDF) or high digestible fiber (HDF)
diets ........................................................................................................ 178

Nutrient intake, fecal composition, fecal output, and apparent total
tract digestibility for 12 early lactation cows receiving low
digestible fiber (LDF) or high digestible fiber (HDF) diets ................... 180

Eating, ruminating, and drinking activities for 12 early lactation

cows receiving low digestible fiber (LDF) or high digestible fiber
(HDF) diets ............................................................................................ 181

xi

Table 8. Rumen digesta characteristics for 4 early lactation cows receiving
low digestible fiber (LDF) or high digestible fiber (HDF) diets ........... 183

Table 9. Rumen fluid pH and VFA concentration of digesta from 4 early
lactation cows receiving low digestible fiber (LDF) or high
digestible fiber (HDF) diets ................................................................... 184

 

Table 10. Rumen digestion kinetics of NDF and rumen apparent digestibility
of NDF from 4 early lactation cows receiving low digestible fiber
(LDF) or high digestible fiber (HDF) diets ............................................ 185

APPENDIX ,-

Table 1. Continuous computer acquisition of feed and water intakes,
chewing, reticular motility, and ruminal pH of cattle: computer
board assignments in Labtech Notebook software setup ....................... 212

Table 2. Continuous computer acquisition of feed and water intakes,
chewing, reticular motility, and ruminal pH of cattle: channel
characteristics of algorithm functions for data acquisition run under
Labtech Notebook .................................................................................. 213 ‘

Table 3. Continuous computer acquisition of feed and water intakes,
chewing, reticular motility, and ruminal pH of cattle: channel
assignments for a complete acquisition run under Labtech
Notebook. Setup is designed for monitoring all five activities for
12 stalls ................................................................................................ 217

Table 4. Continuous computer acquisition of feed and water intakes,
chewing, reticular motility, and ruminal pH of cattle: feeding
behavior data interpretation computer program for summarizing
individual cow data into bouts of eating, ruminating, and drinking ...... 221

Table 5. Intake limitations, feeding behavior, and rumen function of cows
challenged with rumen fill from dietary fiber or inert bulk:
probability values for effects used to describe variation in
experimental variables ........................................................................... 228

Table 6. Intake limitations, feeding behavior, and rumen function of cows
challenged with rumen fill from dietary fiber or inert bulk:
apparent total tract digestibility of nutrients using acid detergent
lignin as an internal marker from 12 cows receiving low fiber (LF)
or high fiber (HF) diets without (+0) or with (+B) added rumen
inert bulk ................................................................................................ 232

Table 7. Intake limitations, feeding behavior, and rumen function of cows
challenged with mmen fill from dietary fiber or inert bulk: rumen
digesta parameters 2 h prefeeding for 12 cows receiving low fiber
(LF) or high fiber (HF) diets without (+0) or with (+B) added
rumen inert bulk ................................................................................... 233

xii

 

Table 8. Intake limitations, feeding behavior, and rumen function of cows
challenged with rumen fill from dietary fiber or inert bulk: rumen
digesta parameters 2 h postfeeding for 12 cows receiving low fiber
(LF) or high fiber (HF) diets without (+0) or with (+B) added
rumen inert bulk .................................................................................... 234

Table 9. Intake limitations, feeding behavior, and rumen function of cows
challenged with rumen fill from dietary fiber or inert bulk:
concentration and molar proportion of VFA in rumen fluid 2 h
prefeeding from 12 cows receiving low fiber (LF) or high fiber
(HF) diets without (+0) or with (+B) added rumen inert bulk ............... 235

Table 10. Intake limitations, feeding behavior, and rumen function of cows
challenged with rumen fill from dietary fiber or inert bulk:
concentration and molar proportion of VFA in rumen fluid 2 h
postfeeding from 12 cows receiving low fiber (LF) or high fiber _
(HF) diets without (+0) or with (+B) added rumen inert bulk ............... 236 ‘ ..

Table 11. Enhanced intake and production of cows offered ensiled alfalfa
silage with higher neutral detergent fiber digestibility: probability
values for effects used to describe variation in experimental
variables ................................................................................................. 237

 

Table 12. Enhanced intake and production of cows offered ensiled alfalfa
silage with higher neutral detergent fiber digestibility: apparent
total tract digestibility of nutrients using acid detergent lignin as an
internal marker for 12 early lactation cows receiving low digestible
fiber (LDF) or high digestible fiber (HDF) diets ................................... 240

xiii

 

LIST OF FIGURES

CHAPTER 1

Figure 1. Intake model illustrating the physical limitation and physiological
regulation theories of intake control with constant energy
consumed along line a-b and constant rumen fill obtained along
line b—c in the general case (a) and for a 600 kg cow producing
various quantities of 4% FCM (b) ........................................................... 20

Figure 2. Feeding behavior model for sheep (Forbes, 1980). Objectives of
the model were to predict meal size and frequency during eating as
determined by physical and physiological satiety factors. MER =
metabolizable energy requirements ......................................................... 33

Figure 3. Modified model from Dado and Allen (1993) illustrating ruminal
volume and passage compensation as concentrations of dietary
NDF increase. Passage from the rumen increases when maximum
rumen volume is reached. Intake of NDF is constant and DMI
decreases only when both rumen volume and passage are
maximum. The level of dietary NDF concentration where
thresholds occur are both diet and animal dependent. ............................. 50

Figure 4. Ruminal digestion curve of NDF. Y values represent the
percentage of original NDF that remains after a given length of
microbial fermentation. Fiber with faster fractional rates of
digestion femrent more quickly. Fiber with higher maximum
extents of digestion have less fiber remaining after extended
periods of digestion. Fiber digestibility is time dependent. ................... 52

CHAPTER 2

Figure 1. Scaled output of feed manger, water meter, and pH monitors
throughout 1 d for a cow fed twice daily ................................................. 68

Figure 2. Analog output from chewing and reticulum motility monitors

during episodes of eating and ruminating. Values presented are
scaled pressure transducer output. ........................................................... 69

Figure 3. File output of feeding behavior data from Labtech® Notebook in
ASCII format. ........................................................................................... 74

xiv

Figure 4. Frequency of inter-period intervals for eating and ruminating
activity of twelve cows measured for twelve days. Data are
expressed as survivorship curves with percent frequency
determined cumulatively and backwards (Metz, 1975). The chosen
minimum inter-bout interval of 7.5 minutes is indicated ........................ 76

Figure 5. Frequency of inter-drinking intervals of twelve cows measured for
twelve days. Data are expressed as a survivorship curve with
percent frequency determined cumulatively and backwards (Metz,
1975). The chosen minimum inter-bout interval of 4 minutes is
indicated ................................................................................................... 77

Figure 6. Relationship between reticular conu'actions determined by manual
observation of signal waveform and computer algorithm
interpretation for 88 observations. Data are expressed as total
number of contractions within a 30-min period. The line represents
perfect correlation .................................................................................... 87

Figure 7. Relationship between ruminal pH determined in vitro by manual
sampling and in vivo by computer acquisition for 40 observations.
The line represents perfect correlation .................................................... 88

CHAPTER 3

Figure 1. Distribution of chewing time within a day expressed as the mean of
12 cow means. Feed was offered when cows were away from stalls
for milking ................................................................................................ 98

Figure 2. Distribution of DMI within a day expressed as the mean of 12 cow
means. Bars represent the standard error of each mean. Feed was
offered when cows were away from stalls for milking .......................... 108

CHAPTER 4

Figure 1. Mean milk production and DMI by week of experimental period
for 12 cows receiving low fiber (LF) or high fiber (HF) diets
without (+0) or with (+B) added rumen inert bulk ................................ 136

Figure 2. Rumen digesta plus inert bulk volume 2 h prefeeding and 2 h
postfeeding for 12 cows receiving low fiber (LF) or high fiber (HF)
diets without (+0) or with (+B) added rumen inert bulk ........................ 140

Figure 3. Rumen digesta volume as a function of rumen NDF pool size for
12 cows receiving low or high fiber diets without or with added
rumen inert bulk ..................................................................................... 141

Figure 4. Model of NDF and DM intake response to increases in dietary
NDF content. Intake of NDF increases until maximum rumen
volume, at which point DMI decreases. Based on a maximum
rumen volume of 100 L, maximum NDF intake of 6.5 kg/d, and
baseline DMI of 22.8 kg/d, the threshold for intake limitation is
28.5% NDF ............................................................................................ 142

XV

Figure 5. Disuibution of individual meal sizes expressed as a percentage of
the total number of meals within each treatment for 12 cows
receiving low fiber (LF) or high fiber (HF) diets without (+0) or
with (+B) added rumen inert bulk .......................................................... 144

Figure 6. Distribution of DMI within a day expressed as a percentage of total
daily DMI for 12 cows receiving low fiber (LF) or high fiber (HF)
diets without (+0) or with (+B) added rumen inert bulk. Bars
within a time period with different letters differ (P < .05) ................... 145

Figure 7. Distribution of rumination within a day expressed as a percentage
of total time spent ruminating for 12 cows receiving low fiber (LF)
or high fiber (HF) diets without (+0) or with (+B) added rumen
inert bulk ................................................................................................ 146

CHAPTER 5

Figure 1. Dimensions of alfalfa plots from which forage was sampled and
harvested ................................................................................................ 151

Figure 2. Concentration of NDF in alfalfa forages sampled at various days
prior to harvest. Objectives were to obtain forage with 40% NDF
at day 0. ................................................................................................. 154

Figure 3. Neutral detergent fiber concentrations and NDF digestibility for
loads of chopped alfalfa obtained pre- and post—ensiling from
harvests D and E. Digestibility of NDF was determined from 24 h
in vitro fermentation with buffered rumen fluid .................................... 157

CHAPTER 6

Figure 1. Digestion of low digestible fiber (LDF) and high digestible fiber
(HDF) alfalfa silage NDF determined from in vitro fermentation in
buffered rumen fluid for time indicated ................................................. 177

Figure 2. Concentration of digestible NDF (DNDF) and indigestible NDF
(INDF) for low digestible fiber (LDF) and high digestible fiber
(HDF) alfalfa silages following 120 h in vitro fermentation for
samples taken prior to commencement of the lactation study (Pre)
and samples taken during the lactation study (Post) .............................. 187

Figure 3. Contribution to digestible DMI from digestible NDF (DNDF)
intake and digestible neutral detergent solubles (DNDS) intake for
low digestible fiber (LDF) and high digestible fiber (HDF) diets
offered to 12 early lactation cows. ........................................................ 190

APPENDIX
Figure 1. Schematic of complete data acquisition system for continuous
monitoring of feed intake, water intake, chewing activity, reticular

motility, and ruminal pH of cannulated cows at Michigan State
University Dairy Teaching and Research Farm ..................................... 241

xvi

LIST OF ABBREVIATIONS

+0 = No added rumen inert bulk.

+B = Added rumen inert bulk.

ADF = Acid detergent fiber.

BW = Body weight.

CP = Crude protein (N x 6.25).

CV = Coefficient of variation.

DHIA = Dairy Herd Improvement Association.
DIM = Days in milk.

DM = Dry matter.

DMI = Dry matter intake.

FCM = Fat-corrected milk (usually 4%).

HDF = High digestible fiber.

HF = 35% NDF diet.

HPLC = High performance liquid chromatography.
kd = Fractional rate of digestible NDF digestion in the rumen.
kp = Fractional rate of NDF passage from the rumen.
LDF = Low digestible fiber.

LF = 25% NDF diet.

MIBI = Minimum inter-bout interval.

NDF = Neutral detergent fiber.

NEL = Net energy for lactation.

NRC = National Research Council.

OM = Organic matter.

P = Probability that difference was due to chance alone.
r = Correlation coefficient.

RIB = Rumen inert bulk.

RMSE = Root mean square error.

SAS = Statistical Analysis System.

SCM = Solids-corrected milk.

SD = Standard deviation.

TDN = Total digestible nutrients.

TMR = Total mixed ration.

VFA = Volatile fatty acids.

xvii

 

 

PROLOGUE

Dairy cows efficiently convert large quantities of nutrients to milk components
that supply valuable food products for human consumption. A sustainable dairy
industry is possible because benefits from milk production outweigh input costs.
Among domesticated animal species, dairy bovines supply the majority of milk for
human food because of large milk output in relation to their maintenance requirements.
High levels of feed intake are a major component of such efficiency.

Feed intake sets limits to production of dairy cows in early stages of lactation.
Increasing feed consumption during this period not only improves milk production but
also minimizes the negative health consequences associated with severe negative energy
balance. Greater feed consumption in early lactation also increases milk output
throughout the entire lactation period of the cow, since the shape of milk production
curves is relatively constant across different levels of production. Higher producing
cows convert nutrients to milk more efficiently than lower producers, because fixed
nutrient requirements for body maintenance are diluted across more units of milk
output. Such improvements in gross efficiency reduce production costs and
environmentally challenging waste excretions associated with each unit of milk
produced. Finally, increased intake capacity allows more flexible ration formulation
and greater potential utilization of more fibrous forages and by-product feeds that are
not suitable for human consumption. Undoubtedly, higher levels of feed intake by cows
is beneficial to the entire dairy industry and to human society in general.

Understanding limits of feed intake and mechanisms associated with its control

remains an intense area of dairy cattle nutrition research. Several mechanisms work in

l

 

2
concert with one another to produce a number of signals that are summarized by intake

centers of the central nervous system that ultimately determines commencement and
cessation of eating. Potential signals previously investigated include distention of the
gastrointestinal tract, accumulation of digestion endproducts, such as VFA, hydrogen
ions, peptides, or glucose, and concentration of circulating hormones or
neurotransmitters. Lactating cows consume multiple meals per day, thus daily intake is
the summation of many individual meals. To accurately assess the influence of diet on
feed intake control mechanisms, quantitative examination of within-day feeding
behavior is essential. Knowledge of other within-day animal responses to feeding also
may contribute to a more complete description of diet/intake relationships.

Forages serve as major components of dairy cattle diets. Not all forages are
consumed to the same degree, perhaps due to an intrinsic property that varies among
forages. Concentration of cell wall (fiber) in forage plants has been implicated as a
major factor associated with intake. In early lactation, feed intake by dairy cows may
be limited by the rumen-filling properties of forage fiber. Under this scenario, forages
with high concenu'ations of fiber are consumed to a lesser degree than those with low
concentrations. If maintenance of feed intake is a high priority, cows may adjust their
feeding behavior to accommodate diets with high concentrations of fiber.

Just as all forages are not consumed equally, all types of forage fiber may not be
consumed equally. Forage fiber consists of different molecules, whose number and
interactions differ between forage species and maturity. These variations contribute to
differences in digestion of fiber in the rumen and may alter the rumen-filling effects of
fiber, and forage intake. Fiber digestibility varies widely in forages of the same species;
however, this characteristic is not currently measured in routine forage analyses. Before
fiber digestibility is adopted for use in ration formulation, it must be demonstrated that

differences in forage fiber quality influence animal performance.

 

 

 

3
Objectives of this research project relate to the influence of forage quality on

feed intake by cows in early stages of lactation and were to:

1) Develop, validate, and implement a computer data acquisition system for the
continuous monitoring of feed intake, water consumption, chewing activity,
reticular motility, and ruminal pH of stall-housed dairy cows.

2) Challenge early lactation cows with rumen fill in the form of dietary NDF and
rumen inert bulk to determine if limitations to intake from runrinal fill vary with
NDF content.

3) Measure feeding behavior and rumen activity of early lactation cows to
determine if within-day behavior changes allow cows to adapt to higher fill
conditions and maintain normal daily feed intakes.

4) Harvest production-sized fields of alfalfa with similar NDF concentrations but
different NDF digestibilities in sufficient quantities to conduct a lactation study.

5) Determine the effect of forage fiber digestibility on intake and milk production
of early lactation dairy cows.

An overview of past research and current theories on intake as they relate to
rumen fill in ruminants is presented as a literature review in Chapter 1. Description of a
computer data acquisition system to quantitatively measure feeding behavior and rumen
function is reported in Chapter 2. Chapter 3 describes an experiment with lactating
cows receiving a common diet, that examined relationships among feeding variables.
Such information was useful in designing and conducting subsequent experiments
where feeding behavior was measured. Objectives 2 and 3 were addressed in a lactation
study with 12 cows that is reported in Chapter 4. The harvesting of alfalfa forages with
differences in NDF digestibility is described in Chapter 5, and the results of a 12 cow
lactation study using these forages is presented in Chapter 6.

The overall hypothesis of this research is that forage fiber and rumen
indigestible fiber limit voluntary feed intake in early lactation dairy cows because of
rumen fill. To test this hypothesis, intake must be shown to be limited independently by
both high fiber content of the diet and low fiber digestibility. Secondly, adequate
measurements must be made to quantitatively evaluate whether these intake limitations

were caused by physical filling of limited reticulorumen capacity.

CHAPTER 1

Literature Review

IMPORTANCE OF FEED INTAKE

The objective of domestic dairy cattle production is to optimize milk production,
usually defined as maximum milk output per unit cost. Because cost of digestible
nutrients is high, maximum production per unit of digestible nutrient is desired. Initial
increments of digestible nutrients are utilized to maintain body tissue and basal
metabolism. Such use is considered constant for a given animal regardless of
production (Bauman et al., 1985). Any additional increments are available for
productive processes, including growth, fetal development, and milk production.
Optimal milk production, therefore, results from maximizing total milk output to
decrease the percentage of total digestible nutrients used for maintenance.

High milk production requires large quantities of digestible nutrients. The
genetic correlation between milk production and estimated net energy consumption was
.82 for 548 cow lactations over a nineteen-year period (Miller et al., 1972). Cows with
different genetic abilities to produce milk were found to have similar metabolic
efficiencies (Custodio et al., 1983; Bauman et al., 1985), suggesting that a constant ratio
exists between digested/absorbed nutrients and milk production. In early lactation,
conversion of body tissue to milk allows cows to temporarily achieve production levels
in excess of digestible nutrient consumption. However, there are limits to tissue
mobilization. In addition, cows must return body stores to levels present before

lactation commenced by increasing digestible nutrient consumption in later lactation.

5
Thus, higher total lactation production must be accompanied by greater digestible

nutrient consumption across the lactation.

Consumption of digestible nutrients is a function of the rate of nutrient intake
and its corresponding digestibility. Of these two, nutrient intake appears to hold greater
importance. From 50 to 70% of the variation in digestible nutrient consumption
between animals is due to differences in nutrient intake (Mertens, 1985). Intake has a
coefficient of variation of 50% among different feedstuffs while digestibility has
variation of 30% (Van Soest, 1982). Reid (1961) suggested that improvements in
accurate determination of forage digestibility and nutrient content could increase
efficiency of ration balancing by only 1 to 5%, but accurate determination of forage
intake would increase efficiency of balancing by 40 to 50%. Finally, there is a
mathematical limit to the extent that digestibility can be improved (e.g., cannot exceed
100%), but there is no theoretical limit to the extent that nutrient intake can increase.

Nutrient consumption may be increased one of three ways: increased daily
consumption of DM (DMI), increased nutrient density of the DM, or increased DMI and
nutrient density. Often an increase in nutrient density of the diet will increase DMI as
well, especially with poor quality diets (Waldo, 1986). Increased nutrient density is
commonly achieved by replacing less digestible forage with highly digestible
concentrates. Such substitution not only increases energy density but also improves
energy digestibility and possibly DMI. There are limits, however, to the extent to which
forages may be replaced by concentrates (Kesler and Spahr, 1964). Rumen fermentation
rates that are too fast result in digestive and metabolic problems for the cow, an animal
not designed by nature to survive on extremely low forage/fiber diets. As with
digestibility, there is also a mathematical limit to the degree that concentrate can
substitute for forage: there is no known limit to DMI.

Intake and milk production are positively correlated, both phenotypically and
genetically (Freeman, 1975). Brown et al. (1977) developed multiple regression

 

6
equations to describe intake and milk yield for 492 cows and found that DMI was a

significant determinant of milk yield. Maximizing feed intake is especially critical for
cows in early lactation. Of interest among researchers is the phenomenon where
increases in intake lag behind increases in milk production immediately following
parturition. Because cows are in negative energy balance in early lactation, increased
DMI could prevent large losses of body tissue reserves or allow for increases in milk
production. Performance may be improved not only for this period but also for the
entire lactation (Broster, 1972).

Improving voluntary DMI requires a thorough understanding of factors involved
in the control and influence of intake for ruminants in general, and for early lactation
cows specifically. Several mechanisms that may control intake have been previously
investigated and reviewed (Baile and Forbes, 1974). Physical fill and physiological
control are two major mechanisms commonly cited. Because energy requirements are
not met by nutrient consumption in early lactation, it is believed that physical limitations
to intake predominate during this period. Daily intake of lactating cows usually
involves consumption during discreet meals. A thorough study of those factors limiting
intake must examine factors responsible for initiation and termination of individual
meals. Forages provide the majority of nutrients to ruminants world-wide. Even for
high producing dairy cows, forage quality may be crucial during limitations in intake
and digestible nutrient consumption, especially since forage cannot be completely
removed from their diets. Examination of forage quality in relation to intake is
warranted.

Objectives of this literature review are to discuss past research and current
theories on intake limitations as they relate to rumen fill in ruminants. Part I of this
review will discuss the long-term control mechanisms of intake in ruminants, including
physical, physiological, and integrated control mechanisms. Part II will address short-
term intake control, including control of individual meals, modulators of feeding

7
behavior, and methods of measuring short-term intake. Finally, part III will focus on

forages and fiber as they relate to intake and rumen fill. Variation in forage
consumption, factors affecting forage composition, and the role of fiber digestibility as
an influence on intake will be examined.

LONG-TERM INTAKE CONTROL IN RUMINANTS

Regulation of feed intake under ad libitum feeding has been differentiated into
those factors affecting long-term control versus those affecting short-term control (Baile
and Forbes, 1974). Long-term regulation involves maintenance of energy inputs and
outputs to achieve balance in BW and productive function. Short-term regulation
involves within-day activity that defines meal initiation, termination, and distribution.
Undoubtedly, short-term controls contribute to long-term balance and may more
accurately describe daily fluctuations in intake. However, long-term regulation has been
utilized more in models to predict intake because of more sufficient data to support such
theories (Illius and Allen, 1993). As additional within-day data are accumulated,
perhaps short-term regulation will replace traditional theories used to describe intake.
Regardless of relative factors, the ultimate coordinator of feeding behavior and intake
control is the central nervous system, and more specifically, the lateral and ventromedial
portions of the hypothalamus (NRC, 1987). Several reviews have discussed how such
coordination may take place (W yrwicka and Dobrzecka, 1960: Baile and McLaughlin,
1987; Miner, 1992).

Long-term regulators of intake have been categorized into three major
mechanisms: factors associated with physical limitation, factors associated with
physiological control, and factors associated with psychogenic adjustments (Mertens,
1985). Physical regulation implies that cows will consume a diet until the physical
capacity of the digestive tract is exhausted, or until the maximum rate of a physical
process associated with digestion (e.g., rumination) is reached. Physiological regulation

commences when a cow has consumed sufficient nutrients to meet energy requirements.

8
Such a process implies that products of digestion (e.g., metabolites, hormones) send

feedback to the hypothalamus to indicate energy balance. Psychogenic adjustment
considers other intake control factors associated with environmental conditions (e.g.,
heat or cold stress), behavior mechanisms (e.g., social interaction), or disease states that
are difficult to assign to general categories of intake control. Among these, palatability
may be the most important factor, especially under conditions of imprinted feeding
behavior. Nevertheless, palatability may exert an influence only for short periods of
time and may be more appropriately considered a short-term regulator of intake. In
general, one might consider psychogenic adjustments as "fudge factors" to describe
intake observations that do not fit model expectations. The major mechanisms of
physical and physiological control will be briefly examined as they relate to intake of
early lactation dairy cows.
Physical Control of Intake

The concept that feed intake might be limited by its bulkiness was initially
proposed by Lehmann (1941) more than fifty years ago. He suggested that ruminants
can consume only a certain amount of indigestible matter. Consequently, poorly
digested feeds would be consumed to a lesser extent than more digestible feeds. Since
then a number of reviews have been published supporting this basic concept (Balch and
Campling, 1962; Bines, 1971; Grovum, 1987). This mechanism is based on the premise
that the gastrointestinal tract has a finite volume and that feed material must be digested
and absorbed or passed out of the tract before additional intake can occur. Although the
abomasum and intestines probably have finite volumes, there is general consensus that
filling of these organs does not limit voluntary intake (Van Soest, 1982; Grovum, 1987),
primarily because fecal output does not remain constant over various levels of intake.
Ruminant digestion requires fractionation and fermentation of dietary components

before they can pass through the omasal orifice and leave the rumen. Accumulation of

 

9
digesta within the reticulorumen is thought to be where fill limits intake. The presence

of baroreceptors within these organs supports this possibility (Leek, 1969).

Several studies have investigated this mechanism by adding various quantities of
inert bulk to the rumen via permanent cannulae. The general hypothesis of these studies
is that inert bulk in the reticulorumen decreases utilizable space in the organ. Lower
rumen volume should result in decreased intake if intake is limited by rumen capacity.
A summary of results obtained from these experiments is presented in Table 1. A total
of 38 treatment comparisons were identified from 11 published studies. Variables of
interest included species, animal number, physiological state, BW, form of bulk, diet,
amount of bulk addition, DMI, and DMI depression due to bulk addition. Experiments
were conducted with sheep and cattle, with more recent studies conducted with lactating
dairy cows (Johnson and Combs, 1991; 1992). Water-filled bladders, air-filled bladders,
and polystyrene cubes have served as sources of inert bulk. Total added bulk varied
from a low of .2 L for a sheep experiment to 48 L for a study with Holstein steers.

Intake achieved during each bulk treatment was compared to intake for controls.
Depression in DMI due to bulk addition was calculated as the decline in intake,
expressed in grams, per liter of added bulk (Table 1). Not all treatments resulted in
depressions in DMI, although 92% of all treatments did. Average intake depression was
94 :1: 78 g of DM/L of bulk, with a minimum of -39 and amaximum of 300 g/L
(excluding studies involving reticular and abomasal filling; Grovum, 197 9).

If the existence of a finite rumen capacity limits intake, then intake should
decrease for every treatment in proportion to the amount of inert bulk added.

Obviously, this was not the case. Measurement of rumen volume may help explain such
discrepancies. Unfortunately, very few studies examined changes in digesta and total
rumen volumes as a result of bulk addition to see if bulk actually lowered utilizable
space in the rumen. Seven treatment groups included measurements for rumen volume

(Table l). Rumen digesta volume was lower upon bulk addition in every study,

10

 

 

 

 

3 mm. ..3.. 8. .. m 8......
3 .3 33.. e... .. a 5...... 85...... as.
.3 an 53.. .... .. . 3.58... ................5 3.. m2... 2.85 4 82.. 53m
..3.. an. .. o 8.8
.3 .. 3.. ma .. S. 5...... .85 .. m 8. .38 5. 8“ 358...... $3. .888...
3.. 3. a 8.. 85...... 8:38.... 5...... 55...... e .58 .880
3 e3 3.. as .. 3.. .88 .85 .. m 8. 38 3.. 8m 3582.. B... .888...
3.. v... .. 3... 85...... 8.5.883 .88 55...... 3 .58 5580
.3 e: 3.. a... .. S .88 .85 .. m 8. 8.8 3.. 8m 3582.. $3. .888...
3.. m... c .5. 85...... 88.8.38 5...... 55...... a .58 .5880
.3 2. ..3.. e m .. .3 no a...“ 58...... 88.. 8....
3 a. ..3.. m m .. a .5 .38 85...... 3.. 83 52...... .. .58 83......
.3. .8 .. c .5. 85...... ..52....
M... .3 3...... 3.... .353 .2...
3.. N. 3.. a. .3 E 53.. e e .. R .6 88 58...... 88.. 8....
3.. 8 3.. c. .3 en ..3.. e .. .. e. .6 8.3 8.8.883 3.. 33 52...... v .58 582
3.. 8 3.. a ..3.. m w .. o .5. 85...... 35.83
.3 R .853. a... .. .3 ".u 88 .5 8 a .5 .88 .8.
.3 8 .853. a... .. mm .5 e... 88...... 38858.. m .58 8.8. .5.
.853. q... .. .85 . 8. 3... 8.5.8.5. .8385... 35.35....
3513.35; 42m. #3... .33 my... 3.. 3.8 535831.83
.88. Snowfl— omeeoon Ho .5552. we .Eom ammo—3&5 .3532

 

 

8:35.55. .8 9.8... 3.8.8.? .8 3.898 :08... 3838.... we Euro

2.. 055...... o. 55.8-3339. 05 o. 32... noon 2:... £3 to... we 82.5.... 98 «8.8.. 95......» 20:3 8:53 we Em A 03...?

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

.383... 3.3.83 53.
£E>°EE o a:

 

ll

 

 

 

 

 

.853. R 3 .. 3 5. .3.... 22358 8.5.8.... 3.... .58...
3 33. 3.. .3 .. . 5385...... 5 .53
3 33. 3. 33 .. 33 ".9. 3333 .33. 3.3. 58...... 23.8.. 333.
3 33 3. 33. .. 33. .85 853-853 3. 33 2.5.5 3 38... 8.85
3333 3.3a .33. 55338.3...
3. 33,. 3 28.... 8.2.2.
3 .3 3. 33. 2 3 85... 5 .53 58...... 2283 33.3.
m.— 33. o 3.532.. 3.. 3:83 8.3.8.33 m.— 33 233:. n .38.... 5.580
3 33. 3. 33. 3 3 5355.2 5 .53
3 :3 3. 3.. .. 3 39.. 3.33 .30 .33. 58...... 2283 3.....
3 333 3. 3N .. 3. .85 83.3.8.3; 3. 33 2.5.5 3 38... 582..
3 333 3. 33. .. 3. 55 33 5. 3...
3. 33. 3 282.. 8.2.3..
3 3.. .3. 3.3 .5 3.3. .. 33 3...
.3 33 .3. 3... .5 3.-.. .. 33 .35.... 88 .. .. 333.
3 33 .3. 3.3 .3 3.. .. 33 5585 2338... 53.. . 238 .5 .3
.3. 3.... 3 833.853 .3... 853
3 333 .3. .3. .. 3 3......
.3 333 .3. 33. .. 3 .835 233...... 333.
3 333 .3. 33. .. . 838.... 833.253 3. 3333 35...... 3 38... 833
.3. 33. .. 3
3535.353: .53 .33.. 3.5.. 333 35...: 3:33.33
.38... 3.835 833.com .3 .5552. 3.. .53.. 33322335 .3552

3.5.... .. 2......

12

 

 

 

 

 

.. 3. .. m. .3 8. .3.. EN .. v... .92 9... .8 9m. 8.... .85 .38 522... .8. .858

.. 9. .. m. .3.. 3... c .88 8.. 2.2.. 8.8.8.... 9. .8 2... .N w 2.8 .8... =85...

.. 2. .. 8 3 8 .3.. ...~ .. a... .52 98 ..o 9.. .8... .85 .38 522... .8. .858

.. 8. .. 8. .3.. a... o .88 .8. 2...... 8......2.5 9. m8 2... a. e 2.8 .5. =85...

.. 3... .. .3 .3 .N .3.. an. .. no .52 9.. .8 9.. .9... .8.

.. 3.. .. .2. .3 c. .3. s... .. 3. 2...... 8.... .. a. 88...; 88> .8.

.. .3 .. .2 .3 8 .3.. m... .. ..N 8...... a... ................p 9. 8 2.85 e .8... .3383

.. 3.. .3.. .2 .. o .88 92.

. . .3 o... .3.. 3.... .. a .8... 9.. .6 8... 5. 2.... 83... a...

.3 8. .3.. 8.. .. . .5. 82.8.... 9. cm 3858.-.... c .8... .5 8

.— 3. A 0... Eu.— c: o 2:8... @359.» .558952 EOE—am

.3 m... .3.. R. 2 N .8... 8... ..u 9.. 5. 3.3 8.8 a...

.3 on. .3.. 9.. .. . 8.. 828...... 9. on «828.-.... o .8... .... 8

.. n... .. m... .3.. o... a 838.2... 8.8.8 .8...2....8 58......

.3 m. .853. 8. .. N 88.... 288 .8.

.3 8 .853. 8. .. . 55....55 v 888... 25.5 m .8... 8......

.8532 2. c .. m 8. .2... 22.58 .............55 8.. .58...

.3 8. .853. 3 .. a .88.... 258.. a...

.3 8. 8.3.. v. .2 55.855. 288... 255.. m .8... 8......

.853. .8 a .. m .o. .2... 22.58 8......853 55 8.8...

853...; .2.— ..a. .2... a... 3.. a... a;
.88. «.835 8.208 .8 3.5.5.. .o .Eom ammo—3&5 82.52

3:8. .. 2a....

l3

 

 

.0823 2:0: 583:
._ E .— 3 .3 am- a} 8.8 .— 8 "32 82." do 8: 88. .8... 38 588.: «a. .388
4 8 a 8 28. 3“ a .88 Ba 8% 8:888: 8. n2. :5 a: a a w 3.8 as 8853.
a 5 4 Ma .3 on- a} «.8 ._ mm "52 $8.? .6 s: 8.83 8.8. $8 52%: 82 .358
a x: a x: 28. 3a a .88 Ea 08% 8:888; 9. m2. :5 a: a mm m 8.8 as 8888
4 m2 .. 3 d» m a} Eu 4 we "52 8:8 .8 $2 «33 .85 $8 528.: «8. .388
,_ 8 a 8 29. 3a a .88 as 088 8.8.883 E E 25 88 a 2:8 88 885.8
a; 38 a3 85 as E. a?
.58. Sauna 0888a ac .555 .3 8.8”— 3032335 59:52

:83 ._ 85

14
although not every decrease was statistically significant. Likewise, every study

indicated that total rumen volume (digesta + bulk) increased upon bulk addition.
Therefore, rumen space did appear to decline after bulk addition, however, expansion of
the rumen compensated to some degree in order to decrease the extent of this volume
decline (Mowat, 1963; Waybright and Varga, 1991; Johnson and Combs, 1991, 1992).
Finite rumen capacity probably exists; however, no rumens were at capacity prior to
bulk addition for any of these studies.

Besides failing to be at maximum rumen volume prior to bulk addition, other
conditions may help explain why bulk does not elicit proportional declines in intake.
Many authors have been quick to point out that physical fill is not the cause of limits in
intake (Carr and Jacobson, 1967; Ketelaars and Tolkamp, 1992). However, other
reasons for failing to achieve intake declines after bulk addition must first be discarded
before such a broad statement can be made. Variation in rumen capacity has been
observed frequently. Capacity varies with body size across species (Demment and Van
Soest, 1983), stage of lactation (Remond, 1988), extent of fattening and pregnancy
(Forbes, 1969), and time after feeding (Wright and Grainger, 1970; Aitchison et al.,
1986). Because of such variation, it is plausible that different thresholds for maximum
capacity exist for ruminants in different physiological states (Egan, 197 2). Therefore, it
must be shown that maximum capacity has been reached before one can expect added
bulk to decrease intake. In addition to variation in rumen digesta volume, there may be
variation in the volume of the gaseous head space above the digesta in the rumen. Such
volumes have not been reported but need to be considered when defining maximum
capacity available to hold digesta.

Multiple regression analysis was performed with data from Table l to determine
what factors might contribute (P < .01) to variation in intake depression due to bulk.
Intake depressions for sheep were 114 g/L greater than for cattle. In addition, animals
of relatively smaller BW within species had greater depressions in intake. Finally,

 

15
animals that tended to be in physiological states with higher nutrient requirements

(growing or lactating) had intake depressions 79 g/L greater than animals in
maintenance states. Factors not significant included the relative amount of fiber in the
diet, the amount of added bulk (as a % of BW), and the level of intake (as a % of BW).
These estimates support the concept that animals differ in their susceptibility to intake
declines because of added bulk and in the extent that maximum capacity has been filled,
even when all were under ad libitum feeding conditions. Smaller animals under higher
states of production appear more likely to have rumen fills near maximum capacity.
Future studies with rumen bulk can improve their odds of reaching maximum capacity
by utilizing smaller animals under high levels of production. Relatively small early
lactation dairy cows provide such a model.

Two other reasons may explain why bulk does not elicit equivalent declines in
intake. First, the degree of rumen fill depends not only on animal differences in rumen
capacity but also on characteristics of digestion and passage in the rumen. Addition of
rumen inert bulk increased the fractional dilution rate of liquid in the rumen of sheep
(Waybright and Varga, 1991) and the fractional passage rate of fiber from the rumen of
lactating cows (Johnson and Combs, 1991; 1992). Increased removal rates may allow
these animals to maintain intake even though rumen volume has diminished. Second,
animals challenged with rumen bulk may adjust their feeding behavior or rumen activity
to compensate for the added fill. Baumont et al. (1990) reported that eating and
ruminating bouts increased when bulk was added to the rumen of sheep and time spent
eating and ruminating remained unchanged even though intake declined. Okine et al.
(1989) placed 24 kg of mass in the ventral rumen of steers and observed 15% increases
in the duration of reticular contractions and 46% increases in fractional passage rates of
digesta from the rumen. Changes in ruminating and reticular activity may increase

digesta flow, providing additional space in the rumen. Increases in eating and

16
ruminating frequencies may serve to keep the rumen full for as much of the day as

possible and allow for normal intakes in spite of limiting rumen capacity.

Not all forages or diets produce the same level of rumen fill. Blaxter et al.
(1956) calculated "ballast", or indigestible DM in the rumen using meal patterns,
digestibility, and passage data and found more digestible feeds resulted in less ballast
and greater appetites. Feeds that ferment and pass more quickly through the rumen are
believed to be less filling because they occupy space within the rumen for a shorter
period of time (Aitchison et al., 1986). Consequently, these feeds must differ in some
component that produces varying degrees of rumen fill. Various components of
ruminant diets have been implicated. Several recent models that predict intake consider
rumen DM as the component of fill (Forbes, 1980; Fisher et al., 1987; Illius and Gordon,
1991; Hyer et al., 1991). Others use rumen OM (Pienaar et al., 1980), rumen fiber
(Mertens and Ely, 1979; Bywater, 1984) or dietary fiber content (Mertens, 1987). In
some instances, it is the indigestible portion of the rumen fiber pool that is considered as
fill (Mertens and Ely, 1979; Lippke, 1986).

Bulk studies comparing air- and water-filled balloons indicate that digesta
volume, not digesta mass, is the characteristic of digesta which results in fill (Mowat,
1963). Van Soest (1982) has elaborated on this concept by suggesting that cell wall, or
NDF, is the fill agent because of its relationship with diet bulk density and volume. The
relationship between diet NDF content and intake has been well investigated (Mertens,
1973) and will be discussed in greater detail in part 111. However, few NDF studies have
measured rumen volume to assure that it is rumen fill that links NDF and intake. It is
likely that no one dietary component can be used to predict rumen fill because fill is a
concept based on dynamic properties of digestion and passage. As improvements are
made to existing rumen kinetic models, perhaps estimates of "rumen fill content" and
. "intake potential" can be appropriately defined from routine analyses and assigned to
feeds for dairy cattle.

 

17
Physiological Control of Intake

Physiological control of intake is based on the premise that animals will
consume nutrients until energy balance is achieved, and has been the subject of many
reviews (Jones, 1972; Journet and Remond, 1976; De Jong, 1986). Once balance is
attained, intake will cease. This mechanism requires three basic assumptions. First, it
must be assumed that the basic urge to eat remains until the maximum genetic capacity
for productive purposes has been fulfilled. Recently this premise has been scrutinized
by Ketelaars and Tolkamp (1992), who believe that animals seek to maximize the
efficiency of oxygen utilization, not energy. Second, if energy is the indicator of
balance, then energy intakes should be constant for a given animal regardless of the
energy density of the diet. Data supporting this argument have been reported by Conrad
et al. (1964) and Montgomery and Baumgardt (1965). However, in a review on intake
control, Grovum (1987) cited several studies conducted with sheep, beef cattle, and
dairy cattle where digestible energy intakes were relatively constant at moderate diet
energy densities but declined at high dietary energy concentrations. He suggests that
maintenance of rumen function receives priority over maintaining energy intake under
high concentrate feeding. Third, if energy balance is a threshold, it is implied that it can
be measured in the animal. Because "energy receptors" do not exist per se, some other
molecules associated with energy status must be monitored by receptors in the ruminant.
Two classes of energy status mediators have been identified (De Jong, 1986):
metabolites associated with energy metabolism, and hormones associated with
homeoretic and digestive control.

Rumen fermentation produces large quantities of VFA that are related directly to
total energy intake. Consequently, acetate, propionate, and butyrate have been
thoroughly studied as possible mediators of energy intake. The most common method
of studying VFA has been by ruminal or portal vein infusion. Early reviews suggested

that acetate receptors exist in the rumen and propionate receptors in the rumen wall

18
(Baile and Forbes, 1974). A more recent review suggests that this is not the case

(Grovum, 1987) and argues that depressions in intake fi'om VFA infusion in earlier
studies were confounded because of changes in osmolarity. Changes in rumen pH
associated with VFA infusions must also be considered, since pH receptors are known
to exist in the rumen. Concentrations of VFA may be more important during conditions
of slug feeding compared to ad libitum-fed ruminants (De Jong, 1986). Lactate may be
produced during fermentation of high concentrate diets and has also been suggested as a
modulator of intake (Baile and Forbes, 1974). Other potential metabolites include
glucose, free fatty acids, and amino acids. ‘

Two types of hormones have been studied as regulators of energy status. They
are hormones associated with peripheral energy metabolism (insulin, glucagon,
glucocorticoids) and gut hormones associated with feeding (gastrin, secretin,
cholecystokinin). Interactions among these hormones probably play the most important
role in monitoring energy status and intake. Any perturbation in metabolism will likely
initiate a hormonal response that, in conjunction with neural responses, ultimately
signals the hypothalamus and controls intake. However, the numerous levels of
hormone action and complexity of interaction will keep even the best future researchers
challenged indeed. Continued examination of metabolism at this level is required if an
understanding of intake regulation is to be obtained. Failure to understand relationships
at this level has forced intake control to be modeled at more aggregated levels.

Other possible mediators of energy status may play a role in certain situations.
They include rumen fluid pH and osmolarity (Carter and Grovum, 1990), placental
lactogen (Byatt et al., 1992), melatonin, somatotropin, and insulin-like growth factor I.
These, like any additional mediators that may someday be discovered, are associated
with digestion or metabolism of nutrients in some manner. As with physical control
mechanisms, threshold values for any of these mediators is likely to vary with animal

and physiological state of production. Integrating these concepts into mechanistic

19
models of intake control may be the most efficient means of understanding these

relationships.
Integrated Theories of Intake Control

Cessation of eating implies that satiety has been invoked by the hypothalamus
and the brain. Lack of any one mechanism that induces satiety in all feeding situations
suggests that satiety may result from the summation of many individual inputs. Grovum
(1987) listed several possible contributors, many of which have already been discussed.
They include distention of the reticulum, rumen, or other part of the digestive tract;
effects of VFA in the stomach, ruminal veins, and liver; effects of ionic strength in the
blood or digesta; effects of digesta pH; effect of gut contents and peptides on smooth
muscle tone in the digestive tract; effects of circulating hormones; and the effects of
various neurotransmitters. In many studies, supraphysiological levels of these factors
were required to achieve depressions in intake. Few studies have examined possible
cooperation or synergism among these many possible factors. Acetate infusion with or
without rumen distention in sheep indicated suppression in intake with either two factors
alone and an additive effect when combined at various levels (Adams and Forbes,
1981). With Friesian cattle, 7 .5 L of rumen distention, 250 g/h of acetate infusion into
the rumen, and 167 g/h of propionate infusion resulted in no significant intake
suppression individually, but did depress intakes of hay under various combinations of
all three factors (Anil et al., 1987). More investigation into interactions among multiple
satiety signals is required before a comprehensive understanding will be achieved.

Attempts have been made to combine the main mechanisms of physical fill and
physiological regulation to determine which mechanism dominates and to predict intake.
Conrad et al. (1964) were one of the first groups to combine these ideas and suggested
that intake was limited by undigested material in the digestive tract when diet DM
digestibility was below 66% and limited by physiological controls at higher energy

densities. Since then, other workers have integrated these concepts into static,

20

III/ll u ’1 III! III \
6.0 till!!!I”$$$$I$$$$$I:IIIIl/II’$I$:I$II$ kg
$33$6,355,n”m.o$$$$$$$$$$$$$$’$" “‘
”tug”Ia4I”tan/rxxxxg’x’”
’
22:4:22232333225”
[$555M m m ”It”

 

 

 

2 4.0
5
0
i
5 2.0
Theoretical Intake
1 I l 1 l
100 so so 40 20 0
Energy Units
1 I l i l I
0 2o 40 so so 100
Fill Units
(a)
5.0 r \
40 kg FCM

 

Dry Matter Intake (% Body Wild)

 

 

25 30 35 40 45

Neutral Detergent Fiber (% Ration DM)
(13)

Figure 1. Intake model illustrating the physical limitation and physiological regulation
theories of intake control with constant energy consumed along line a—b and constant
rumen fill obtained along line b—c in the general case (a) and for a 600 kg cow producing
various quantities of 4% FCM (b).

21

Table 2. Empirical equations developed with stepwise multiple regression to predict

 

 

DMI for lactating cows.

E E Pr ed' . .

Chandler DMI (kg/d) = [3.62 + .076(FT) - .17(BW) + .01017(MY x MF) -

and Walker .026(CF)] x BW

1972

Brown LnDMI (kg/d) = .52 - .00083(DIM)+.148(LnDIM)+

etal., 1977 .00068(BW) + .339(LnMY) + .0993(MFd) +
.018(CF) - .00056(Cl:")2

Vadiveloo DMI (kg/d) = -4.14 -.095(WL) + 4.0401(log WL) + .015 (BW) +

and Holmes .208(MY) + .43(C)

1979

Yungblut DMI (kg/d) = 3.37 + .34(LN) + .0096(BW) + .3362(MY) +

1981 .528(MF) - .106(ADF)

NRC, 1989 DMI (kg/d) = -.293 + .372(FCM) + .0968(BW)-75

Kertz et al. DMI (kg/d) = .2286(DIM) - .00218(DIM)2 + .00000705(DIM)3 +

1991 .00804(BW) + .3134(FCM)

Rayburn and DMI (DIM < 84; kg/d) = .0117(BW) + .0749(DIM) + .281(FCM)

Fox, 1993 DMI (DIM > 70; kg/d) = .023(BW) + .0201(DIM) + .286(FCM) -

where:

.097 9(NDF)
ADF = acid detergent fiber (% of diet DM)
BW = body weight (kg)
C = concentrate intake (kg/d)
CF = crude fiber (% of diet DM)
DIM = days in milk
FCM = 4% FCM (kg/d)
FT =- type of feeding (+1 = winter, -1 = summer)
LN = lactation number
MF = milk fat (%)
MFd = milk fat (% as a decimal)
MY = milk yield (kg/d)
NDF = neutral detergent fiber (% of diet DM)
WL = week of lactation

 

22
mechanistic models to predict intake (Forbes, 1980; Bywater, 1984; Mertens, 1987;

Fisher et al., 1987); these models have been summarized by Illius and Allen (1993).
Most predictions are based on the premise that intake will vary according to energy
demands (Forbes, 1977) but will be constrained to a maximum based on rumen capacity
and dietary fill. Components associated with predictive equations vary with the model.
One such model proposed by Mertens (1985, 1987) uses NDF as a common currency for
both the "fill" content and energy content of the diet. Intake capacity of the animal is
based on maximum NDF intake and energy requirements are based on maintenance and
production requirements. An illustration of intake responses to various diet qualities
based on this model is presented in Figure 1. Other systems have been developed that
consider the filling effect of forages in different ways (J arrige et al., 1986).

Many intake prediction equations for lactating cows have been reported that are
based on empirical relationships rather than mechanistic relationships (Table 2). Under
most practical applications of ration formulation, these equations tend to be used more
often than their more complicated mechanistic model counterparts (NRC, 1989).
Routine use requires that inputs to the equations be known; during ration balancing
seldom does one have access to all the information required by most mechanistic
models. Intake has been shown to be most related to daily milk production, BW, DIM,
and dietary fiber content in these multiple regression equations (Table 2). Their
accuracy of prediction depends on the extent that conditions during predictive use
deviate fi'om conditions used during equation generation.

SHORT-TERM INTAKE CONTROL IN RUMINANTS

As previously stated, long-term intake regulation seeks to obtain balance
between maintenance and production requirements and energy inputs. Overall balance
is essential if BW is to be maintained and survival assured. Long-term intake
measurements are frequently defined in terms of the amount of feed consumed during a

one day period (Forbes, 1986). Lactating cows under ad libitum feeding conditions,

23
however, do not consume feed continuously throughout the day, nor do they consume

their entire daily intake in one meal. Intake within a day consists of a varying number
of individual eating bouts (meals) whose frequency, duration, and intensity of
consumption ultimately determine average daily intake (Bines, 1976). A more complete
understanding of intake control may be obtained if factors affecting short-term meals
can be defined. Short-term measures also are essential to understand interactions
between particular diets and animal response.
Control of Individual Meals

Any study of individual meals requires an adequate definition of a meal. Meals
are generally defined by the minimum size of eating required to be considered a meal,
the maximum amount of time during which the minimum meal size must be consumed,
and the minimum amount of inactivity required between two periods of eating
(Heinrichs and Conrad, 1987). Not all meals are easily defined because of variation in
eating intensities and variation in frequency and duration of breaks taken by animals
during eating. Minimum intermeal intervals are often defined, with eating separated by
intervals of time less than these values considered as part of the previous meal (Forbes,
1986). Wangsness et al. (1976) studied feeding of sheep and defined a meal as eating
activity of at least 1 min after an intermeal interval of 20 min. A study with only two
lactating cows found the appropriate meal definition for their animals was a minimum
eating length of 2.5 minutes and an 8 min intermeal interval (Heinrichs and Conrad,
1987). Meal definitions should not be arbitrarily assigned, but selected only after
analysis of a fiequency plot of intermeal intervals (Savory, 197 9). Extrapolation of
meal definitions across species or across studies within the same species may lead to
inappropriate meal descriptions.

Ruminants begin to consume food when in a state of hunger and stop when
satiated. Unfortunately, these two states have no precise physiological meaning

(Forbes, 1986) and cannot be verbally described by cows to curious researchers. Most

24
research has focused attention on regulation of intake via satiety mechanisms since these

are more easily observed (Gallouin and Focant, 1980). In the central nervous system,
satiety appears to be associated with the ventromedial hypothalamus and hunger with
the lateral hypothalamus (Baile and McLaughlin, 1987). Lesions of the ventromedial
hypothalamus usually produce excessive eating (hyperphagia) and lesions of the lateral
hypothalamus produce aphagia (NRC, 1987). Whether cows commence eating because
of hunger or because of lack of satiety has not been determined. If a high degree of
correlation exists between meal size and length of preceding intermeal interval, a satiety
mechanism controlling intake is implied; if the correlation is between meal size and
length of postrneal interval, a hunger mechanism initiates intake (Savory, 1979). Metz
(1975) conducted feeding behavior experiments with seven non-lactating cows and
found meal size was correlated more positively with its preintermeal interval than with
its postintermeal interval, suggesting that meals stop once a fixed level of repletion is
reached. This supports the concept that satiety, and possibly rumen fill, is controlling
meal size. Relationships between intermeal intervals and meal size vary diurnally
(Metz, 1975), indicating that different mechanisms may be influencing intake at
different times of the day.

Short-term regulators of meal size and meal distribution are not unlike those
described for long-term intake control, although short-term regulators must respond
more rapidly. Many of the same factors affecting long-term intake have been implicated
as playing a role in meal regulation (Baile and Forbes, 1974). They include factors such
as palatability, ruminal distention, rumen fluid and body osmolarity, ruminal pH, VFA,
other blood metabolites, and hormones. For example, rumen fill was measured before
and after feeding two qualities of forage to tlmee cows (Freer and Campling, 1963).
Rumen pool sizes of DM differed between the two diets prior to feeding (9.5 vs 5.2 kg)
but were similar after feeding (16.2 vs 14.7 kg). They suggested that each feed was

consumed until satiation fiom physical distention occurred and that rumen capacity was

 

25
limiting meal size for both diets. This conclusion is also supported by a summary of

data compiled by Baile and Forbes (1974). Other measures of rumen volume before and
after feeding have not resulted in such consistency of fill (Campling et al., 1961; Bines
and Davey, 1970), suggesting that other measures of satiety may have occurred,
especially with more digestible diets.

Regardless of the actual mechanisms that control meal sizes and distributions, it
is often useful to measure within day eating activity (i.e., feeding behavior) to examine
how dietary manipulations alter behavior and ultimately affect intake. Effects of diet on
ruminating activity may also lead to differences in meal patterns and are useful
measurements. An understanding of how diet affects intake is valuable because diet
manipulation provides the primary tool in which to improve animal performance. With
adequate data, perhaps a complete model of feeding activity can be developed to
describe the occurrence and size of individual meals.

Dietary Modulation of Feeding Behavior

Before discussing how diet affects feeding behavior, a review of general aspects
of eating, rumination, and digestion is warranted. These comments are specific for stall-
housed dairy cows and are based on reviews by Camng and Morgan (1981),
Beauchemin (1991b), and Albright (1993). Cows consume between 6 and 20 meals per
day, with larger meals occurring immediately following feed offering. Rate of
consumption is most rapid at the beginning of the meal and slows until meal end. Meal
size is highly variable, but total eating time is less variable, with 4 to 7 h spent each day
eating. Comminution during eating is adequate to permit swallowing of food boluses
and to add small amounts of saliva. Whether there is an upper limit to meal size based
on physical abilities to process feed during eating is unknown; nevertheless,
oropharyngeal metering is considered to be improbable (Baile and Forbes, 1974).

Rumination occurs during 10 to 20 individual ruminating bouts throughout the

day, with each bout lasting from 1 min to 2 h or more. Total time spent ruminating

 

26
ranges from 5 to 10 h/d. Each rumination bolus consists of longer fiber portions of

digesta near the reticular cardia and is chewed for 45 to 60 s at a relatively constant rate
(1 chew/s). The primary purpose of rumination is to decrease digesta particle size,
increase particle surface area, and increase saliva in the digesta pool. Total time spent
chewing ranges from 9 to 16 h/d during which 30,000 to 50,000 chews occur.
Requirements for rest from chewing may limit the maximum amount of time spent
chewing each day. If such is the case, intake may be restricted to that which can be

physically processed during eating and rumination. Once ingested and hydrated, feed

fun;

particles are fermented by growing rumen microorganisms which produce VFA that
must be buffered by saliva. The rate of fermentation and absorption in the rumen and
the rate of passage out of the rumen determine the residence time of feed particles.
Clearance from the rumen enables consumption of more feed. Motility of the
reticulorumen enhances fermentation and plays a role in nutrient removal from the
rumen.

Feeding behavior may be affected by two types of dietary manipulation; they are
feeding management and diet formulation. Feeding management involves those aspects
of feed consumption associated with physical aspects of supplying feed, and includes
issues like animal housing, feed availability, and feeding frequency. Behavior of
confinement-fed cows differs from pasture-fed cows (O'Connell et al., 1989). Pasture
feeding occurred predominantly during daylight hours and rumination during darkness.
In confinement, these same mid-lactation cows consumed feed and ruminated
throughout all hours of the day, with no indication of diurnal variation. Harb et a1.

( 1985) compared eating rates for cows housed in individual stalls compared to group
housing and feeding. Eating rates increased when cows competed during group feeding.
Other studies examining group feeding demonstrated that socially dominant cows ate
first compared to more submissive animals. As feeding space was reduced, dominance
tended to be more positively related to total daily intake (Friend and Polan, 1974; Friend

27
et al., 1977). Time spent ruminating tends to increase as time spent eating decreases

during competitive feeding.

Shortening the time of access to feed from 24 h to 5 h/d has been shown to
increase eating rates and rumination time and decrease DMI (Campling and Morgan,
1981). However, when feed access times were lowered to 8 h/d, DMI, time spent
eating, and time spent ruminating did not change (Erdman et al., 1989). In a review on
feeding management, Robinson (1989) stated that the sequence in which feedstuffs are
offered and the frequency of feeding are potentially important contributors to intake and
cow performance. His comments were based on the concept that more regular
consumption of balanced nutrients would increase rumen microbial fermentation
efficiency due to the more steady supply of nutrients. Such improvements in cow
performance have been difficult to show experimentally (Gibson, 1984; Nocek, 1992),
and few studies have included behavior measurements.

Gill and Castle (1983) observed that the number of eating bouts increased from
16 to 24 when feeding frequency increased from 2 to 22 times per day, however time
spent eating and ruminating did not change. Ullyatt et al. (1984) fed sheep once or 24
times per day and noted increased DMI and eating times with multiple feedings, but no
change in time spent ruminating. However, they forced eating patterns to be different
between diets since sheep fed once daily only had access to feed for 3 to 4 h. This study
emphasizes the difl‘iculty of conducting behavioral work. If various feeding
management factors are to be accurately studied as they relate to feeding behavior and
intake, it is imperative that treatments and measurements be designed so experimental
animals are free to alter their behavior for the correct reason. In other words, let the
animals decide if treatments really work! Under ad libitum feeding, increased frequency
of feeding does not automatically imply more numerous eating bouts.

The second type of diet manipulation is diet formulation, which consists of

changes in the type and form of feedstuffs selected for inclusion in daily offerings to

28
cows. Variation in dietary concentrations of energy and fiber have been noted most

often as feed components that alter feeding behavior. Baumgardt et al. (197 3) found
rate of eating was highly correlated with digestible energy content (.99) and bulk density
(.93) of the diet with sheep. Rumination time has been associated with the roughage
value of feeds (Balch, 1971) that is commonly defined in terms of fiber content. Cell
wall intake and daily ruminating time was highly correlated in both sheep (.99; Welch .-
and Smith, 1969) and cattle (.94; Welch and Smith, 1970). Beauchemin (1991b) ‘
summarized several studies where different concentrations of dietary NDF were offered
to dairy cows. In general, eating and ruminating times increase as NDF content I ‘
increases but chewing per unit of fiber remains relatively constant. Meal frequency has
been shown to increase with increases in NDF content (Beauchemin and Buchanan-
Smith, 1989), however, such results are not consistent (Colenbrander et al., 1991). Not
all NDF elicits the same amount of chewing activity (Mooney and Allen, 1993).
Chewing variation for fiber is likely due to differences in size of the fiber particles.
As forage particles increased in size, time spent ruminating increased (W oodford
and Murphy, 1988; Colenbrander et al., 1991), and occasionally time spent eating
increased (Jaster and Murphy, 1983; Grant et al., 1990). There is an upper limit to the
ability of particle size to increase chewing activity. Therefore, chewing behavior is
similar between coarsely chopped silages and long hay (Beauchemin, 1991b). Adequate
particle size reduction is necessary to increase fermentation sites for rumen digestion
and to allow for particle passage. Smaller particles were found to digest faster during in
vitro fermentation (Cherney et al., 1988), and more dense particles were found to pass
more quickly from the rumen (Ehle, 1984). Small particles tend to be more dense than
larger particles (H00per and Welch, 1985); sufficient specific gravity also is required for
particle passage. Reduction in particle size may also decrease the effective volume of
fiber, decreasing its rumen filling effect and stimulate greater intake. Campling and

Freer (1966) noted increased intake when low digestible straw was ground but not when

29
a more digestible grass was ground. Eating and passage rates also increased with

grinding. Whether particle size affects frequency or distribution of eating is generally
not known.
Methods of Measuring Short-Term Intake and Behavior

Daily intake is typically measured by recording the mass of feed offered to cows
and subtracting the mass of feed remaining the following day. Because cows consume 4‘
multiple meals each day, more frequent weighing of feed is necessary to determine
masses of individual meals. To measure eating and ruminating activity, chewing is
typically monitored. Both feed and chewing measurements require continuous
monitoring if complete data are to be obtained. Several methods of measurement have
been developed in the past to obtain feeding behavior data for grazing, group-fed, or
stall-fed ruminants. Some rely on manual observation, while others utilize automation.

Manual observation has been difficult. Although they may be great learning
experiences, manual approaches are labor intensive, limited to short time periods, and
somewhat subjective. Kertz et al. (1981) weighed feed every 4 min for 28 min
following feeding of 10 cows to estimate eating rates of different forms of a diet. Nocek
and Braund (1985) weighed feed hourly for 4 d with four cows to estimate eating rate
and pattern of intake throughout the day. Unfortunately, timed weighing may not
describe meal size accurately since cows may consume more than one meal an hour. In
grazing experiments, meal sizes have been measured by weighing animals before and
after grazing (Erlinger et al., 1990). Manual measurements of chewing activity have
been made by recording cow activity periodically (every 5 min, Woodford and Murphy,
1988; every 15 min, Erdman et al., 1989). Such methods must assume that the same
activity was present during the entire period represented by the sample observation. In
all manual measurements it must also be assumed that the measurement process itself
does not alter behavior. This may not be true, especially if feed mangers are being
weighed periodically.

30
Automatic systems of measurement attempt to alleviate a part or all of the labor

required for manual observations. They require a monitor on the feed manger or head of
the cows and a data logging system to record signals from the monitors. Determination
of meal size automatically during grazing is obviously extremely difficult. Chacon et al.
(197 6) attempted this by automatically monitoring bite number with a microswitch to
detect jaw movement and a mercury head switch to differentiate grazing and ruminating
activity. Meal size was calculated as the product of bite size and bite number, where
bite size was estimated via weighing of food boluses from an esophageal fistula.
Methods to measure chewing activity during grazing are c0nsiderably easier and have
been summarized (Penning, 1983). Many different monitors have been developed to
detect jaw movement. They include electrodes in the masseter muscle of the jaw,
carbon-filled stretch electrodes, pneumatic bellows, and electronic microswitches,
straingauges, and pressure transducers. Data collection devices have included strip-
chart recorders, oscillographs, and analog/digital convertors for computer acquisition.
Signal transfer between monitors and collection devices range from cassette tapes on the
animals, long cords to a central location in the paddock, radiotelemetry, to static ram
chips mounted on the chewing collars. Once collected, the signals have been interpreted
either manually or via computer algorithm (Penning et al., 1984).

Intake and behavior of group-fed cows have been monitored with systems
developed by Forbes et al. (1986) and Mason et al. (1991). Group-fed monitoring faces
similar challenges to data collection as does grazing, in that there is not a one-to-one
correspondence between cow and manger as their is in stall feeding, and that free-
roaming animals cannot be hard-wired into a collection system. These problems have
been overcome by using electronic identification systems and, as mentioned, use of
radiotelemetry or animal-mounted recorders.

Automatic determination of meal size in stall-fed animals has been conducted

using two basic methods. Multiple allocation of pre-weighed feed was used in systems

31
described by Minson (1966) and Wangsness et al. (1976). Continuous electronic

weighing of feed mangers with load cells has been utilized by several groups (Suzuki et
al., 1969; Metz and Borel, 1975; Heinrichs and Conrad, 1987). This second method has
advantages in that meal weights are determined electronically and that it is not limited in
the number of individual meals an animal can consume. To detect chewing activity,
many of the same methods have been used in stall-feeding as in grazing. These are
summarized by Beauchemin et al. (1989), who have developed one of the most
automated systems to date. Pioneers in this field include Balch (1971), Sudweeks (Law
and Sudweeks, 1975), and Welch (1982). Other unique ways of determining feeding
behavior include use of time-lapsed photography (V asilatos and Wangsness, 1980),
insertion of pressure transducers in the esophagus (McLeod and Smith, 1989), and
estimation of rumination via detection of the three phase contraction of the reticulum
with internal balloons and pressure transducers (N orgaard, 1989).

The optimum system for measuring feeding behavior should be accurate,
continuous, accommodating to a sufficient number of cows, automatic, and capable of
recording several cow activities simultaneously for an extended period of time. Current
developments in computer technology and electronic data acquisition make this
possible.

A Model of Feeding Behavior

Qualitative description of factors involved in short-term eating patterns have
limited usefulness if they cannot be quantitatively related to specific behavior variables.
Integration of these factors into a simulation model to predict meal patterns and daily
intake has been developed by Forbes (1980), and is the only model of its type known to
this author. Forbes assumed that feeding commences in response to an insufficient
quantity of absorbed substrates in relation to energy requirements and terminates when a
relative excess of energy has been consumed or maximum gut fill has been reached.

Consequently, the model incorporates both physical and physiological aspects of intake

32 .
control. Objectives of the model were to simulate minute-by-minute changes in

digestion and metabolism of a lean mature sheep to drive satiety factors that explain the
size and frequency of individual meals. An overview of the model is presented in
Figure 2. Feed components consisted of a rapidly fermentable and absorbable soluble
fraction, a more slowly fermentable insoluble fiaction, and an indigestible fraction. At
each iteration (1.0 min interval) energy available from previous bites are summated and
compared to requirements. If absorbable energy is below a defined threshold, eating
begins or continues, if above a slightly higher threshold, eating ceases or does not begin.
Such thresholds are similar to a household thermostat, alloWing an interval between on
and off thresholds so eating is not rapidly turning on and off. Total gut capacity
available to hold feed started at 12 L and decreased as abdominal fat or conceptus
increased.

During simulation, both gut capacity and energy adequacy limited meal size at
different times during the day (Forbes, 1980). The number of individual meals and total
daily intake was similar to data from a living animal of similar size and state. Meal
patterns were more difficult to predict; however predicted meal sizes were realistic.
Eating patterns of actual sheep are not consistent day to day, that is, eating does not
occur at exactly the same day after day for ad libitum fed animals. The model, however,
did behave in this fashion, suggesting that other factors need to be considered when
predicting meal distributions, or that a form of stochasticity should be incorporated.
Other pertinent factors suggested by Forbes was rumen and blood concentrations of
products of digestion, distention of different parts of the digestive tract, feed palatability,
competing drives, and social interaction. The model also showed promise in predicting
changes attributable to differences in growth, fattening, pregnancy, lactation, and feed
quality. Although not perfect, this modeling attempt illustrated that the basic concepts
associated with the control of meal size may be realistic. One might suspect that similar

modeling of lactating dairy cow meal patterns would be considerably more complex,

33

 

l
Advance time by1 min
1.
Yes <—— lsitnight? ——> No

 

Lowei'MEH J
Jr

 

No 6— Any food left? ——> Yes

 

ulv
No <—- ls animal already eating? __) Yes
- Yes (— Absorption > MERithreshoId?
No -—-—> Start eating
' Higher MER

Yes €—-— Absorption > MER + threshold?

Stop eating [1°

Yes <—— Gut contents 1 Gut capacity?
No

 

 

 

l

Animal not eating Animal eating

Store time
1

 

‘1' Increase gut contents

Absorption of energy from previous meals
Exit at undigefted particles
Decrease gut contents

st (——— Absorption>MER? -—-> [if

Increase reserves 4' Decrease reserves
Total tat

Abdominal tat
Gut capacity

 

 

 

Figure 2. Feeding behavior model for sheep (Forbes, 1980). Objectives of the
model were to predict meal size and frequency during eating as determined by physical
and physiological satiety factors. MER = metabolizable energy requirements.

9"

34
however, such attempts have not been made. Models will never simulate reality

completely, but they certainly will continue to provide a mechanism to ask pertinent
questions and design future experiments.

In conclusion, feeding behavior is a function of the interactions between hunger
and satiety. To consider that any one factor is responsible for meal control under all
situations is fallacious. Perhaps this concept was best stated by Forbes (1988):

"We can no longer consider the various theories of intake control as

alternatives but rather as complementary and contributing to a

multifactorial control system. No single factor is essential for normal

feed intake and many manipulations which stay within the physiological

range have effects which can only be picked up by close attention to

details of feeding behaviour."

Measurement of feeding behavior in dairy cattle is in its infancy. Continued efforts to
measure short-term intake will enhance our understanding of and ability to improve
animal performance.

FORAGES, FIBER, AND INTAKE

Dairy cattle, because they are ruminants, harbor a population of anaerobic
microorganisms in their reticulorumen that live in synergy with their host animal. The
ruminal environment provides a continuous supply of nutrients, warmth, moisture,
anaerobiosis, and end-product removal for the microorganisms. In return, the microbes
ferment cellulosic plant material that results in production of end-products which
provide a substantial portion of the animal's energy needs. To sustain this symbiotic
relationship, the animal must maintain a ruminal environment where variables such as
pH, osmolarity, and oxidation-reduction potential are appropriate for microbial growth.
The slow, continuous fermentation of roughage provides this type of environment; it is
for this reason that dairy cattle have forage requirements.

Diets for dairy cattle typically contain 40 to 60% forage for lactating cows, and
60-100% forage for other types of dairy cattle. Forage consists of the entire vegetative

portion of legume or grass plant species and is characterized by the composition and

G

35
structure of individual plant cells. Not all forages are utilized by dairy cattle to the same

degree. Forages differ in rate of digestion, extent of digestion, and intake potential by
cattle. Variation in forage quality, and how to measure it, has been the subject of
investigation for many decades (Wolff, 1856; Goering and Van Soest, 1970). It was,
and still remains, unfeasible to conduct a biological assay (i.e., animal trial) on every
forage wishing to be fed to cattle, consequently, it was necessary to develop laboratory
methods to measure characteristics of forages that were highly related to animal
performance. The detergent fiber analysis system was originally developed to describe
forage quality in terms of nutritionally available and unavailable plant components (Van
Soest, 1967). Measurement of the plant cell wall, as estimated by the NDF assay, was
considered effective in dividing forages into completely available (neutral detergent
solubles) and partially available (NDF). Although not chemically homogenous, the
NDF fi'action has been useful in describing forage quality because of its negative
relationship with digestibility and intake (Van Soest, 1965).
Variation in Forage Consumption Due to NDF Concentration

Dietary cell wall content has been suggested to be the most important single
factor affecting ad libitum consumption of forages (Van Soest, 1982). Mertens (197 3)
collected data from trials where 187 forages were fed without supplementation to sheep
and found the correlation between cell wall content of the forage and DMI was -.76.
Cell wall intake, as a percent of metabolic BW, was also relatively constant, indicating
that cell wall content reflects the ruminal fill characteristics of the diet, although rumen
fill was never measured. Another study where standardized intake of sheep was
determined also demonstrated a strong negative relationship between cell wall content
and intake (Osbourn et al., 1974). If NDF content is directly related to rumen fill, a
strong relationship between NDF content and intake indicates that rumen capacity was

the predominant limitation to intake for these sheep.

36
A more vigorous test of the NDF/intake relationship would include not only

studies solely with forage, but also those using mixed diets of forage and concentrate.
Addition of more digestible, less fibrous concentrates will dilute the fill effect of forage
fiber and allow limitations other than capacity to influence intake. One might expect
intake to be correlated negatively with NDF content at high dietary NDF but correlated
positively with NDF content at low dietary NDF. Such results would support the
physical/physiological intake control mechanism proposed by Mertens (1987) as
discussed in part I of this review. Mixed diets are much more common for lactating
cows than are all forage diets. However, early lactation coWs experience a period of
negative energy balance where physical limits to intake are more likely. Examination of
recent studies where lactating cows have received diets varying in NDF content may be
useful.

A summary of 26 studies utilizing 673 animals conducted within the last ten
years with lactating cows was compiled from the literature (Table 3). Cows were
relatively high producing and averaged 29.6 kg/d of milk; cows ranged from 7 to 200
DIM at the start of their trial. Each experiment quantified dietary NDF concentrations
from which forage NDF in the diet DM was calculated. Objectives of the experiments
involved the study of forage quality (e.g., Beauchemin, 1991a), non-forage NDF sources
(e. g., Clark and Armentano, 1993), or the interaction between a dietary component, such
as fat (Grant and Weidner, 1992), and fiber level on animal performance. Different
NDF concentrations were obtained by changes in fiber source, forage quality, or
foragezconcentrate ratio. The minimum dietary NDF content used was 19.6% and the
maximum was 44.8%. Minimum DMI obtained was 12.9 kg/d and the maximum was
26.2 kg/d. Regression of DMI on dietary NDF concentration across all 103 treatment
observations resulted in a negative relationship (b = -. 18; P < .002; r = .31). Regression
of DMI on forage NDF concentration resulted in a similar relationship (b = -.15) that

was slightly stronger (r = .37), suggesting that non-forage NDF may not be as related to

37

 

:2 In earn £< n
a: a: sauna ”8.... 8+5. 4 .8.
a: a: saun— §E§+=< m 2... 8 a m. a.»
a: 8.: emu"— 888 8888+? N 5888.3
a: M8 82...". ”5t _
z: 8.8 earn. ”589:... m 82
88 En saun— uemzaiaaoia. a own R n 4 as
4.8 Sm sci". "8+5. _ 838m
.8 saun— umo+m< v 82
38 earn 888.2889? n 8883
new semi". ”8+? N am 2 a m 8255
8.8 8.8"”. 88818888 829.. _
«.8 8m sane 888 .833 a. 82
4.8 we“ exam ea 8:55 $8 883 m .8 a
EN EN sumac 8988 .885 N am 8 «a u 28..
we“ «.8 sane 82.8 322382838 _
a ”.2 sang ”888:8 283+? a. 82
«.2 8.8 earn ”888 8:88 85+? m 85852
we: v.3 ear". ”9‘ a 2. 2 _ 8.. 85
qe «.8 8.816888 ”$888 888 _
"Hz , 388.8% .8 ”833%

 

 

~85 "5.2 .55: 85 38 23.

 

.oufi 3838888988 .8 5:23 owns.“ .858
“RE 5 mowed—6 .3 355% 0.83 meounbeooeoo mDZ Bogota 838 .33 $682: win—Goa .32 32628 we actuate; .88
8.85 no £858 829:8 .«o SE 3 0038 8 coup—E350 "52 50:. we fiche 05 85.55 35 8:53 we gm .m 0389

38

 

«E. n: «.8 «.8 «Sun. um< Case as m 82
«E. «.8 «.2 92 gel ”5. u: « one 8. c : 358 2a
«5.. «.3 «.8 o.«« «Scum a... om _ 8.5-354
v.«.« «.2 «Sum fi< m 32
«.8 in «Eng 2.2 « Se 3 e 2 358 c5
_._« «.8 gown". 2.2 _ 3535..—
vé. a? «3 833.3 «82mg? .. 3%
3m «.8 95 833-33 §$m9m< « .2»
mm «.3 vs 32.3..“ §«+mu+m< « 8m «2 w a 8.8.6
ma 3% was fiéfium 38:8 8+? _
3m «in; 358:: 3238 v as
3... $1". imam m Bozo: 2a
«.2 «Sum ”Be 8.2» .828 « e: S m .558
can «Scum €28 8 _
3: «.2 «Mann. cumufiaa v 32
«.2 3m «Smum a ”«niaa m a a
_.o« «.8 «.«vnm ”«guzfima « 3o «« v« « 22:
«.8 v.8 89mm ”mania? _
S.« «.5 gang a? o: e
§« «.3 «Eng ”58 m
«.8 3M eaum 42 u: e 28—
«._« «.2 «$3. ”5.8 m cm. a e c 55383
n.«_ _.«m «33. "5. 62:8 2... «
n: n: 83...". ”5. 68.8 «=8 _
"Hz 5%

 

N85 "52 ESP— 85 :30 ~60”—

Gcoov .m 035.

39

 

I'll“
2m 3m gnu". a? u: v 82
EN 93. gang ”5. u: m 228 fig 2a
3m 2m *2"... "5. om N 2...: h 3 : 3&8
:8 can gown". 32 um _
3... 93 32mm ”5 o: v 82
:8 3a gnu". 32 02 m 2.23 5:5. 2a
3a a? gem". a? B N :5... N. a S 3293
c8 92 $83. $2 um _
3..” 9% $5. u_"_u8+m< m 82
«.8 2m $5. ”8 N 8m 5 «a 2 .20
“.8 2m §mum a< _ 3555.8
2. 92 88"". “8+? a 52 .3 3
man Sm gang ”8+3 _ m2 8 3 .0555
5:3 w? ”.8 gm $8"; ”8 m
€an Se 3... ”.3. $3 nausea e «32
55".. «.9 Rm «.2 gown". a? o.— m an an n 2 a a
.523 .3 can can *3 £< us a 325
55¢ Em ma ”.8 eoeum 32 um _
2N Aion". 068+? a $2 a.»
3:. $1". $253+? _ «on g N. 2 £03
mmz all? .
«35 "52 SE. 35 38 <3

 

3.88 .m oBaF

 

 

[IL
EN a: $91... ”5. 3 e 8“ age
”.8 Eu swoon". ”5. om m an 38 2a
«.8 2m gen". um< 3 a a: 2 a. mm .302
EN SN $9} 2.2 B _ 8m
3m 3m 88"". ”5. 3 e 5 ~82
v.8 2m .39! a? om m can .25 2a
«.8 Zn swoon”. nm< o.— N as 2 a. a .302
EN 3." $8"; a... om _ 8n
n: new $21.". ”5. 2
«.2 2a a?! ”3. a.
«.2 :3 $3. ”5. m
as can gaun— u=< 5 N8.
2: 2: fimum £4. a 3:55
fine 9: 5a «.3 gnu". 42 n no 8 9. 8 Egg
fine Sm mom 92 “Sun". i... ..
Que 2o 3: n: 83"". a? m
fine «.8 3: cam $8"; ”5. a
2."... $6 3: 3m gonna ”5. _
”.8 NS §mum as a a £2 .a 3
EN «.2 gang 68% a _ 8n 3 a a 3%:
3m $81.". a? on v 32
v.3 $93. m2 um m .2»
SN “Snug um... um u a... 8 v a 28°83
N: gun". e? um _
32 33%; .
«.05 "32 ES: 35 So 13.

 

3.83 .m 03a...

41

 

w.«« :2 $an ”5. 3 e

«5 «.2 «.29:— ”:< u: m

«.m« 2m *2"; ”5. om e 82
2a «.2 «an»; a< 3 m 8... 8:3 3 8 3% 9a
«.8 «.2 swan". "2 o: « :83:
a.«« «.8 82"". 0? um _

«.2 «an «annu— um< o.— m 82
Ba Ba gmmum ”a... o: « 33m Ea
_.«« Sm «hum a... om _ «c 82:. mm n« .5202
«.2 m._« gnu". a... .2... 32 « 32 .3....
o.«« a: «Snug 2.2 “Ba .3.. _ 3 M: «n v« 3555.8

 

 

 

353 .m 035...

42

 

 

 

8.. o..« e 8 o E «2 8 m 8. _ _« 2 « « 8 m
2.. c«« V o« « 8 8. 8.« «a _ _« «a « a 8 c
8« 8« _ _« _ 8 3.. «« m S c _« 8 « « 8 « «
««« ««« u«« e «« an. R « S « _« :8 o 8 «
§ 8« c «« _ a «: ««.« S e 8 8 « a 8 _
«.8 « 8 8. S n «« «.«« «.2 G « « 8 «
w «« _ S «: «m n «m 8.0 w a R « « 8 « v
«8 in «: a: «v «3 v2 8« «8 _
«a 8 «3 c.«« _«.« «.«« v
«fl «.8 ««.c c.«« 8 v a «« «
«.«« o«« «to «.«« a « c «« « «
«.«« «.«« «3 «.«« a « « o« _
«.«« «8 :5 «3 a 8 «.««. «e c.«« z e c 8 ««.« e _« v
«8 8« :5 «a a 8 33 R c «« «S. « 8 8.« w «« «
««« «8 «S «8 « 8 «.3 «e m «« «_ o e 8 «c.« «.3 « «
8« e8 5. «8 « : «.««. K « «« £6 _ _« «m « e 8 _
3.. e8 _ «« «.«o a e «.8 «.«« «._« v
«8 «8 «.«« 3o ««.« «.8 ««.« «.«« «
«a. «c« 3. «.8 «E «.«« o«.« c.¢« « _
«3 8« _.«« 3e «3 «.«« 2.« «.«« _
“an. as... fl 3 $2 $2 2:. an Ad ._d> in. flu 32> d: d:
«5.3 «3. «Bow «3...: «a «2 .55. «285 £52 029 «"52 «2a “a: «a: 35 10m

 

 

583m

383 .« 23.

43

nll. .
wee who N 3 _ we 3 ac e 8 v 9. c 2. a 0 min A: m x mm M
mi. 95 cda a we mm no 6 an e wc m we a w ««.«« mm m — mm N 2
«no vwo fiw. v we me mad 6: c mm m K m w _.ea 3 m 0 mm m

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

«no «8 m a. m we $6 93 E? m be _ h m a a. m N. mm m
So mg m a. c «o N— 0 Q3 m.w¢ a S o c v —N a N w mm N o.
«.3 mm: 0 cu m No S c ado adv h an m w c MN and u an _

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

 

 

 

 

 

 

« «. « «e .«. 2 c «.«« «.«« a. ..« « « «« ««« a 8 a
m 2 a a. 8. v. o . 2. «.3. 8 «« .. o .« «.. « a 8
« «« o «« 8... «..« s. « a R v
. «o . 8 ««« «.«« S « « «« «
. «« a a. 8 .. «.«« cm « e «« « «
. 8 « 8 «o .. «.8 a « h «« .
««« ««« «v «.« «««. ««« ««« «8 v
. 8 « «m «o .. 8 «2 a «« v8 .. «« «
. 8 « .«« «« v «v ««« « «« 8 « « .«« « «
«.« .3 3 ««« .««. ««« ««« .«« .
««« ««« . o. o
«3. 5. « v. «
««« .8 a «. .. o
««« .«« . s «
.«« .«« ««« «
«9. «8 «a. .
1...... «a... a. 9. as . . . . . .91 .8
«as. «.3 «2.8. «2...... «a «a... 2%. 5.8.5 v.52 £6 «"57. «2n. 5.: «a: .05 .«3.

 

 

3:8. .« 29¢

 

N— o m.m_ m m m cm 2
an e 99 m m v mm

8 n 93 _.v e on v

we ¢ 5.2 fin a mu n

ma v 5.3 fin n mu N 3
3 v afi— v.m 5 mm _

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

 

 

 

 

 

a: 8. Sm «.«« mm «.«. ««« «.«« a
««« .8 «.«« «.«« 3. «a. ««« «.8 . v.

«8 3. a «m «E «.«« n on m

«3 «8 a.“ 8.« 3.. h a v
«3 m: «.«« 3; «.8 .«« m m.

«3 a: « a. 8.5 «.8 e am a

«3 «a. «.8 «Z «.«« a «m .

«8 .8 «««~ 3.. a... «a n: ««« a
«.«« a «n m 8 «.2. $6 .. 8 .v m a a . a.

1.35 as: «I a. $2 $2 23 :49 45 . . . .
«as. «.«m «2.8 «3...... «a «3. .35. 58.5 252 as... «"32 «2a «:2 as: .05 .«3.
.55....”

 

 

3.83 .m 03¢

45

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

 

 

 

 

2. e QNN 9. m m on N
mm m h.—N 3 m N am

5% m w— 11m a 8 v

mm m c w— mm.m N._n m
cu m c E aN.m v Nm N .N

_n m NdN Dam m cm .

2N mm Nb «.3. EN QN c an 2

NNN S. 2. 2..” N mN _ m m aN a

mm. wv Nb Sn... _ MN hN c an m

mm. 3. Q. 36 m «N m m N «N N
S: Nv we and m.NN QM v.Nm 0 ON

IN on E. 5.x v.mN 9m .:.N m

NmN wv me Ed m N m m c “N v

SN em Nb mNd N 9N o N wdm m

3. mm 2. ans KEN N.m c an N

G— E 2. 9:. ANN an :m _

w :. 3. n QN— a... m N 8 N
:K mwv Qm— can N .N _ 2

Se mhm on: NN. c m. N. 5 a6 «.3 5:. 2 N QmN c6 ..Nm v

0% EM moo w: c 2 N we ”.0 adv v.2. Nmé «6N an v mm m
chm m .m coo w: m .N m m... a... 2». N.:. N 6 EN ed fimm N m—

nmm 3N c3. _ _ _. oéN ado e6 v.9. c.Nh wN.m aiN Wm cém _
.:. .:. 3 .. .... 2 ... D ) : .. . . ... .... .. . _. . . .. z). . ... ... ...
«as. «.«m «2.8 «3...... «a “on“... 2.3. ««««5 «192 «Sc «”52 «En ...: ma: 35 ...3.

a

3.88 .m 032.

223.. «2.2.8.... .o 2...... «2.3 .5... 6.2.5 22....
.:-... 20.2.... 0... 2.0.. 09.8.... .0 0.... 2.20208”...
3220.20... n ..n. .8300... u 0.0...
0.2.5.8030 o>.> 2. .2039... u 300
.««. .o 0......
8.2.2.2.. 2..« 2. .3 00.20.00... «0 222220220. .0 «.25.. A... .8... 020.0. .0 22.3.8030 ”.97...
.3... 82o. 25.. 22 .0... .o .828 ".22.
.3... 2n. .0... .0 .20.2o0 mn—ZN
.«.20.200 .02.. 020.8. .20.0..... 2... «0...... 023200200038. 02.. 002.8 020.0. 02.28 0:. ....3 «.0...
022.02. 0N o. 0N «00220.02 «0...... 822200200033. m2..¢..> ..o "52 .0 «00.28 nib? ......» «.0... 002.02. a. o. . «00220.02.

0N... vwm N60 N00 ..m. an...” m.nN 0
wmv 00m 0.0 c. .0 ...w. Rum m.nN 0
0m.. 8m ...m0 3.0 ....N on...“ 0.0N v
...m NNN NR 50... m.NN 02d EhN m ON
00.. NmN m..0 Nm... m.NN Na... cSN N
E... 0.N N.N0 0m.0 c.NN a... NNN .

can 0... ma m Wm. 00 m c N N
Nmm NmN mm 0 «.2. mm m h 0N . ..N
90.. v... w: 00m 00... mmN man man v
cam wNv v: N..0 no.0 w MN 00 n 0 an m
Nam on... w... 0.0m 5.0 a..N E. n n ..N N .N
NaN cam ma 1N0 36 ......N cm m . on .

 

 

 

 

 

I02... ...... .... .... a... 2:3...» «.2 . . . .
«25.. «.2. «2.8 «2...... h... .3. 22>. 2.022 2.22 022 «222 «:2 2.2 «2.2 .22 ......
202......

 

 

6.200. .m 030...

47
DMI as forage NDF. Multiple regression that included the study number in both of

these models greatly increased model fit, however, parameter estimates for the effect of
NDF concentration on intake did not change appreciably. To examine effects in early
lactation only, regression of DMI on NDF concentration was performed for studies
where cows were less than 100 DIM at the start of their trial. The linear fit of DMI on
dietary NDF content was more strongly negative (b = -.30; r = .40), as was that for DMI
on forage NDF content (b = -.21; r = .45). Fifty treatment observations had cows
starting their experiment at less then 80 DIM. Regression of DMI on dietary NDF
content was once again more strongly negative (b = -.42; r = .59).

Results from this analysis support the concept that DMI decreases in early
lactation as NDF concentration of the diet increases. Approximawa 35% of the
variation in DMI for cows starting treatments less than 80 DIM can be explained by
differences in dietary NDF content. Conducting forage studies to examine NDF effects
is extremely challenging. Analysis for NDF content of mixed forage/concentrate diets is
diffith because of starch contamination of the residual fiber (Van Soest et al., 1991).
Often researchers have little ability to control the quality of forages harvested on
experimental farms and have to use what is available. Even if they know forage
characteristics prior to cutting, the potential for these parameters to change during
harvest, storage, and feeding is immense (McGechan, 1990). Animal selectivity when
offered high or low fiber diets is also very common. Therefore, what one thinks they are
feeding may not be what is being consumed. Just measuring DMI as affected by dietary
factors can be difficult. Abrams et al. (1987) used the intake of a standard forage as a
covariant to reduce the error of estimating DMI and suggested that differences in
physiological state, production, body condition, environment, genetics, and
experimental error can all contribute to difficulties in measuring DMI precisely.
Consequently, it may be surprising that any linear relationships can be observed
between DMI and NDF content at all!

48
Several mechanisms may allow cows to compensate for increases in NDF

content of the diet without decreases in DMI. Some of these thoughts have previously
been suggested by Van Soest (1982). As mentioned in part I of this review, the rumens
of cows consuming control diets may not be at maximal fill, so when NDF content of
the diet increases, additional ruminal fill can be tolerated. This was found when forage
NDF concentrations increased across maturities of alfalfa (Nelson and Satter, 1992b).
Additional distention may be tolerated by cows if they have a greater drive to eat or are
consuming poor quality diets, such that their intake of digestible energy is farther from
their physiological set point. Such ideas have been included in intake models. Williams
et a1. (1989) modified the intake model of Mertens (1987) by allowing rumen capacity
for NDF to vary between .8 and 1.2% of BW as energy demands increased. They also
adjusted intake upward by 10% of the difference between the physical and physiological
intake estimates (Mertens, 1987) if the latter was greater. Rate of passage of digesta
from the rumen may increase as NDF intake and distention increase (W oodford et al.,
1986; Llamas and Combs, 1991). Cows may also increase their eating and ruminating
activity (W oodford et al., 1986; Grummer et al., 1987) to allow for faster diet digestion
and passage from the rumen. These changes would allow faster clearance of digesta
from the rumen and enable cows to consume diets with higher concentrations of NDF.

Limits to DMI from ruminal fill of NDF might imply that NDF intake is
constant. Based on the observations from Table 3, this may not be the case. For these
cows, NDF intake ranged from 3.7 to 9.6 kg/d. Rayburn and Fox (1993) examined the
variation in NDF intake from production studies and found that NDF intake as a percent
of BW was related positively to DIM, FCM, and dietary NDF concentration. The
positive relationship between NDF intake and NDF concentration was also observed
from regression of data in Table 3. This supports the concept that additional distention
may be tolerated as diet quality declines.

49
Perhaps these concepts of rumen capacity, digestion, passage, and NDF intake

can be summarized best by a modified model (Figure 3) originally proposed by Dado
and Allen (1993). As the concentration of NDF in the diet increases, NDF intake and
rumen volume increase until a maximum rumen volume has been reached at the
capacity threshold. Beyond this point, rumen volume remains constant. Diet NDF
content can continue to increase without a decrease in DMI because of passage
compensation. During this period, eating, ruminating, and rumen motility activity
increases to promote passage from the rumen. Eventually, the cow cannot increase
these activities any further and reaches a point where no additional NDF can be
consumed (NDF intake threshold). Any additional increases in diet NDF concentration
results in a decrease in DMI. It is proposed that the level of diet NDF at which these
thresholds occur are both diet and animal dependent. Such a mechanism would allow
the cow to maintain DMI at the expense of digestibility in the rumen. However,
maintaining DMI is more critical to total digestible energy consumption than
digestibility because of potential post-ruminal digestion.
Variation in Forage Consumption Due to NDF Digestibility

Not all sources of NDF may elicit the same response from dairy cattle as their
concentrations are increased in the diet. Briceno et al. (1987) performed similar
statistical comparisons between DMI and dietary NDF content using 20 experiments
from the University of Florida that included 1688 cow-period observations from early to
mid-lactation and various roughage sources. Dietary NDF content was related
curvilinearly to DMI, however, there was a significant interaction between source of
roughage and dietary NDF concentration. They suggested that NDF could only be used
to formulate dairy rations within roughage source and postulated that differences in
intake response were due, in part, to variation in NDF digestion characteristics among

fiber sources.

50

    
    

NDF intake

Vol —>
max

Rumen volume

DMI

 

Capacity NDF intake
threshold threshold
| l

 

Passage compensation

0% Dietary NDF% 100%

Figure 3. Modified model from Dado and Allen (1993) illustrating ruminal volume
and passage compensation as concentrations of dietary NDF increase. Passage from the
rumen increases when maximum rumen volume is reached. Intake of NDF is constant
and DMI decreases only when both rumen volume and passage are maximum. The level
of dietary NDF concentration where thresholds occur are both diet and animal dependent.

51
Not all NDF has the same composition. Availability of the NDF fraction to

rumen microorganisms is variable and depends not only on the chemical composition of
the NDF but also on the structural interrelationships between the components.
Tremendous effort has been put forth by botanists and plant biochemists to describe
plant cell wall structure and by nutritionists and physiologists to describe cell wall
degradation and utilization by animals. A recent international symposium highlighted
progress in these areas (USDA-ARS, 1993). Nutritionists have measured fiber
digestibility either directly with in vivo digestion studies (Quicke et al., 1959), with in
vitro fermentation methods that utilize rumen fluid (Goering and Van Soest, 1970) or
enzyme preparations (Clarke et al., 1982), or with nylon bags placed in situ within the
rumen of a fistulated animal (Marinucci et al., 1992).

Ruminal fiber digestion is a dynamic process. Extent of fiber digestion (i.e.,
digestibility) depends on the length of time spent in the rumen (or in vitro flask, or in
situ bag). A typical digestion curve of two types of fiber is presented in Figure 4.
Longer times of fermentation result in greater extents of digestion, except at very long
times when a relatively indigestible residue of fiber remains and extent of digestion is
complete. One of the first models to kinetically describe fiber digestion was developed
by Waldo et al. (1972). Since then numerous other models have been developed (Illius
and Allen, 1993), however, most utilize the same principles outlined by Waldo et al.
(1972). Fiber digestion can be described in terms of the fractional rate of its digestion
and the maximum extent of its digestion. Maximum extent of digestion is measured
with very long times of fermentation; rate of digestion is generally determined as the
negative slope of the natural log of digestible fiber remaining versus time. Such a
calculation assumes fiber digestion is a first order chemical process with respect to
digestible fiber pool size. Large variation in rate and maximum extent of fiber digestion
in different forage species have been reported (Smith et al., 197 2; Varga and Hoover,

1983). When fiber digestibilities are reported from in vitro or in situ measurements, the

52

 

 

100
if so-
- —u-— Faster rate, lower extent
3 . —o— Slower rate, higher extent
2
2 60 -
i-
.2 .
I:
3
u. 40 ‘
a
Z .

20 r

O . I ' r ' r ' T ‘

 

0 20 40 60 80 1 00
Time of fermentatlon (h)

Figure 4. Ruminal digestion curve of NDF. Y values represent the percentage of
original NDF that remains after a given length of microbial fermentation. Fiber with
faster fractional rates of digestion ferment more quickly. Fiber with higher maximum
extents of digestion have less fiber remaining after extended periods of digestion. Fiber
digestibility is time dependent.

53
length of fermentation must also be reported. Typically, fermentation times are chosen

which are hoped to represent the average retention time of fiber in the rumen. In vivo,
fiber digestibility is a function of the proportion of fiber that is potentially digestible, the
rate of fiber digestion in the rumen, and the rate of fiber passage from the rumen (Allen
and Mertens, 1988b).

Digestibility of fiber has been shown to vary both across and within forage
species (U SDA-ARS, 1993). Because lactating dairy cows derive from 10 to 20% of
their digestible energy from fiber (Galyean and Goetsch, 1993), an increase in fiber
digestibility should result in additional energy for productive purposes. What is less
well known is whether differences in fiber digestibility affect the voluntary intake of
forages and their diets. A more rapid or greater extent of fiber digestion may enhance
clearance of fiber from the rumen, stimulating an increase in DMI. Glenn and Waldo
(1993) suggested that such an understanding is critical in improving the efficiency and
performance of ruminants. Crampton and co-workers (Crampton et al., 1960; Donefer
et al., 1960) developed a nutritive value index for forages based on the intake potential
and digestibility of individual forages. They believed that previous forage quality
indexes failed to predict animal performance because they did not include estimates of
intake. They found both relative intake and the nutritive value index for sheep to be
highly correlated (r > .83) with cellulose digestibility after 12 h of in vitro incubation,
which suggests that fiber digestibility influences DMI. Muller et al. (1972) fed corn
silage stover from normal or a brown mid—rib mutant variety to sheep. Both silages
contained equal NDF concentrations (60%) but differed in NDF digestibility by 12%
units (39 vs. 51%) after 36 h in vitro. Both in vivo DM digestibility (56 vs. 62%) and
DMI (1.9 vs. 2.4% of BW) increased with high fiber digestibility. Steers were fed
straws with varying maximum extents and rates of DM digestion; their intakes varied in

relation to these digestion parameters (Orskov et al., 1988). Because the straws

54
contained similar and high concentrations of NDF (mean = 83% of DM), such intake

responses were likely due to differences in fiber digestion.

Few attempts have been made to directly measure the affect of fiber digestibility
on intake of lactating cows. A summary of six experiments where such measurements
were made is presented in Table 4. To assess only the effect of fiber digestibility, it was
imperative that these studies used dietary treatments with equivalent fiber
concentrations. Studies 1 to 4 altered forage:concentrate ratios to achieve equivalent
NDF contents, while studies 5 and 6 used forages that contained approximately equal
fiber contents so ratios could remain constant. In only one 'study (Llamas-lamas and
Combs, 1990) was DMI significantly greater (P < .01) with higher fiber digestibility.
Cows on this experiment had substantially higher DMI than the other studies, which
may have resulted in their ability to detect intake differences. Studies 1, 2, and 4 are
difficult to interpret because of the wide variety of fiber sources utilized. They
formulated rations based on the rate of fiber digestion of feedstuffs measured in situ
(Varga and Hoover, 1983) to achieve diets with different rumen fill characteristics.
Studies 5 and 6 were ideal to examine digestibility effects because they were not
confounded by differences in fiber source and forage:concentrate ratio. Both of these
studies showed improved milk production with greater fiber digestibility; however, their
cows were not in early lactation so intake differences were not detected. Any of the
studies investigating the effect of NDF content (Table 3) had the potential of having
differences in fiber digestibility, unless shown to the contrary. Most studies never
measure fiber digestibility so they do not know if their NDF content differences are
confounded by differences in NDF digestibility. To isolate fiber digestibility, the ideal
approach is that of Robinson and McQueen's (1992); however, improvements might
include use of higher fiber diets (40% NDF is not high for a study using grass forage)

and cows in negative energy balance limited by rumen capacity. Comparison of

55

 

 

0.00 0. 00a 0000"". 00.0... E8 .28.. 0
.28 0020.. 0.... 0. .0... 0.00"... 00...... .28 00-0.... .58.. 0 .00.. 08. .... 0
0.". 0.00 0.00 0....» ......u. 00.0... .:8 .28.. 0 0.0 .00.. 0. 0 .00..

01.0 0.00 0.00 0.8.. 000"". 0.0... .28 0.0.0.... .50... .
«an... 00 ..0 0.0.. 0.0"". 000.... .59.... 080 08. 0 08.
..."... 00 .0 0.... 0.00". 0.0... 28...... 8.. 0.00 0 800.8. 88002 ......
00a... 00 .0 0.0.. 0.0"". 00.0... .58... .8. ....00 0 000 ...... 0. 0. 80.003.

an... 00 00 0.0.. 000". 00.0... .52.... .8. .08. . ,
«an. 0.. .00 .... .0... ”0.00". ”02> 0 00... .... 0
00".. 00 0.00 .... .8. 0.00". 02> . 000 0 00 0 000>
0...... 50 0.00 0.00 .80". 0.0.. 0.0.... .8 ...... 0 08. 00.80 ......
flue 5.0 0.00 0.00 0.00am 20.. 0.0.... .8 2.8 . 000 00. 0 0 0022.85...
an. 0.00 ..00 .... ..0... 0000"". 02> 0 08. .... 0
an... 0.0 0.00 .... 30. .00"... .«z> . 000 0 ..0 0 0......
0.0.0.2328 1,220.50- 202u0mz 000.0. 200.02".— 000.200 "52 00.20.20: «20..0>umz>

8080... 30.0 0. 00 0. .0 002 000-..... 00.... .0000". .«2> 0
8000... .0... 0. .. .00 002 0.80.... 30. .000". 02> 0 08. .... 0
8080... 30.0 0.0 ....0 002 080-..... ..0... ”00?. 02> 0 ...0 .0. 0 . 0.0.

8000... .0... 0 ....0 002 080-..... 30. 0.0». .«2> .
.0... 00.3%: ...: 80.8....

00.0 .02 .50.... 0.0 300 . ......

 

 

 

 

0.2.0.... 0w0.0. .0 00.200
.00.. 2. «0w20..0 .3 002.030 0.03 «0.2.3000»... ".97. .20.0...Q .0300 .30.. 0220.00. .0 20203.0... ..20 2.0.2. 20 «20202200200
....QZ .250200 5.3 «20.0. 0.0.0800 .0 .22. «0 22.320030 ”.97. go... .0 .00..0 0... 00222020 .0... «0.0.20 .0 2.0.2.2.... ... 0.00...

56

 

 

 

 

 

 

mm w v a 86 m g 0
am a Qua an m Qua
2 be me mad can on an 8 m @5— emd Qua v
S be we wad m2... 3 E. 3. m mdn N0...” 0 0N m
S we we and «2... he 2. a. a flea mad m cm N m
S 8 me mad an? X. ch 82 2N and m cm _
:3 5o WMN v.8 .3 mad em 24. o w— Na...” mdn N
03. m: wfim flan we. 86 8. 85 a 5 3d m Om _ 0.
mg. 35 cdm 0.8 mm mod 3 0.9. m we 9 w véa N2.” _ mm N
«3. ewe. fix. v.8 8— mm... a: can m E. m w flow 3. m 0 mm _ m
an m a 2 mfim _ an N
an e e w— 2...“ N mm _ N
95 cs. 0.3 0% NO— 8.0 9% m we a m v.5 fin m «N v
:5. 2b. Qmm 0.0m cm: mud adv N 3. c n «.3 hm m «N m
58. as. m.—N mac «0: wwfi —.Nm — mv e n We. 0m 0 5 N _
08. 03. v.8 flaw c: Sun m.mv m cm m m v.2 NM 5 cm _
0.22M 0.0m «2.0m @323..— hi “‘00... 2.5. 80030 £92 035 mun—Z mzn :52 mag .05 20m
2.3—

 

 

4308 v 030%

57

3:5 33582 a $8 ”.58 28% 835 08E
.:-5 :25: 05 Soc 0323 uc 82 133883
.3333:— n E ”8808 u u<h
589593“. o>m> E 323% H mm?
.23 Mo £5;
dam—mam Ba 5 3 @0532: an Evacuate.“ mo Enos 6 “can 30% mo banana»? mn—Zv
.33 ”was Boa 29 3% mo 22:8 "Ezm
.33 ED 3:. no 38:8 mflzu
.moufinumowmc “one amp—8 820% :5 8%.. Sgcoocoouomfiom 28 858 owauom 08am 2: 53 30%
3205 c 2 m mooceouom .89: Babcoocoonowfiom wig; 8 "52 mo 88:8 «5.9; .23 80% 09:05 v 9 ~ 80:98.85

:83 v 03¢

58
by-product feeds that differ substantially in NDF digestibility (Stern and Ziemer, 1993)

is another approach to study fiber digestibility.
Plant Factors Affecting Fiber Digestibility

To understand why plant cell wall is not uniformly available to ruminal
fermentation requires an appreciation for the diversity of molecular composition and
structure within different forages. Cell wall, as isolated by the NDF procedure, consists
of the general molecule types of cellulose, hemicelluloses, and lignin (Van Soest, 1982).
Other chemical methods are available to isolate each of these components in a fashion
more amenable to traditional plant cell wall chemists (Theander and Westerlund, 1993),
however, the gravimetric NDF method has been preferred by nutritionists because of its
relative speed and nutritional uniformity of neutral detergent solubles. Cellulose and
hemicelluloses are part of the polysaccharide fraction of plant cell walls. Cell wall
polysaccharides are generally classified according to their structure into one of five
groups: glucans (which includes cellulose), rhamnogalacturonans and associated
arabinans, mannans, xylans, and glucuronomannans (Aman, 1993). Polysaccharide
names are based on the structure of their monosaccharide residues.

Pectic polysaccharides are common in primary cell walls but are not included in
the NDF fraction because of their rapid availability to ruminal fermentation (Van Soest,
1967). Cellulose is the most abundant organic polymer on earth and consists of beta 14
linked glucopyranosyl units (Aman, 1993). The extended polymer forms a flat ribbon
supported by intra- and intermolecular hydrogen bonds. Multiple cellulose polymers
aggregate into highly ordered crystalline and noncrystalline regions. Hemicellulose is a
general name for polysaccharides soluble in dilute acid or alkali and consist mainly of
substituted xylans. Lignin is a general name for phenylpropanoid molecules deposited
in cell wall in highly diverse, almost amorphous, polymers (core lignin) or linked to
polysaccharide (Jung and Deetz, 1993). Lignin monomers include coniferyl, sinapyl,
and p-coumaryl alcohols. During forage plant growth, lignin is deposited predominantly

59
into the secondary cell wall with some also found in the primary wall and middle

lamella. For the plant, lignin serves to decrease water permeability and add structural
strength to the maturing cell wall.

Fiber digestibility is variable across forages because of differences in the relative
proportion of cell wall components and because of variation in structural linkages
between these components. Lignin is the cell wall component most closely associated
with indigestibility. Not only is lignin considered to be unavailable to rumen
fermentation, it also complexes with cell wall polysaccharides making them less
susceptible to microbial attachment and enzymatic degradation (Chesson, 1988). Lignin
does not depress fiber digestibility the same across all forages. Smith et al. (197 2)
found a greater depression in NDF digestibility after 72 h of in vitro fermentation for
grasses compared to legumes as lignin concentration in the NDF increased, but no
relationship between lignin concentration of NDF and rate of NDF digestion. The
1ignin:NDF ratio was not highly correlated to NDF digestibility after 30 h in vitro
digestion for some forage cuttings, but was for others (Allen et al., 1991). It appears
that not all lignin has the same composition nor does it link to carbohydrate to the same
degree in all cell walls. The diversity of lignin structure continues to make this a
complex area of research (Ralph and Helm, 1993). Diversity in polysaccharide
composition also creates variability in cell wall degradation. Cell wall polysaccharides
interact through structural relationships associated with ionic, hydrogen, and covalent
bonds (Hatfield, 1993). Linkages among different polymer types creates additional
variation. Differences in structure within a given polysaccharide, such as cellulose
crystallinity, also add diversity. Lack of uniformity in structural linkages and sugar
composition generates a complex molecular puzzle for microbial enzymes. Not only do
these enzymes need to be able to recognize the specific linkage and conformation of the
fiber polymers, they also must have sufficient space to access their binding locations. A

60
greater understanding of the entire cell wall matrix is necessary before specific

predictions can be made about cell wall digestibility.

Even though forage cell wall chemistry is complex, there are a few general
trends among different types of forages and forage maturity that allow for some
precision in estimating fiber digestibility. Grasses contain higher concentrations of
NDF, but the maximum extent of NDF digestion in grasses is greater because of lower
lignin concentrations. Because of lower cell wall content, legume DM is generally more
digestible even though legume fiber is less digestible. In general, the ratio of lignin to
carbohydrate in indigestible residue is l: 1.4 (Van Soest, 1982) so lignin concentrations
tend to define the limit of cell wall digestion. Rate of cell wall degradation, however, is
faster in legumes, perhaps because a larger portion of their lignin is present in core
lignin compared to grasses (Jung, 1989). Noncore lignin has the potential to bind with
noncellulosic polysaccharides; as its concentration increases, it depresses digestibility
more than equivalent increases in core lignin. These factors may help explain why fiber
digestibility of grasses declines faster with maturity compared to legumes (Sharma et
al., 1988). Legumes demonstrate greater intake potential compared to grasses, perhaps
due to lower ruminal digesta volumes, greater fragility, or faster digestion rates (Thorton
and Minson, 1973). Legumes generally contain higher concentrations of pectic
polysaccharides while grasses contain more hemicelluloses, although it is unknown
what effect these have on fiber digestibility. An extensive study with legumes,
temperate grasses, and tropical grasses has been summarized that illustrates fiber
differences between forage types and their effect on ruminant performance (Reid et al.,
1988). Differences between legumes and grasses have been incorporated into intake
models for dairy cattle (Williams et al., 1989).

As forage plants grow, concentrations of NDF generally increase and protein
decrease, suggesting a significant decline in forage quality with maturity (References in

Table 3). Early cut forages had faster rates and maximum extents of DM digestion

61
compared to later cut mixed forages (Cleale and Bull, 1986). With perennial grass

species, NDF content (40 to 63%), lignin content (2 to 5%), and in vitro indigestible
fiber (6 to 26%) increased as maturity advanced by 50 d, while maximum extent of fiber
digestibility decreased from 80 to 44% (Cherney et al., 1993). They suggested that
maturity differences were more significant than species differences. Similar declines in
fiber digestibility due to maturity were found in a more diverse set of forages (Smith et
al., 1972). The significance of decreases in forage quality because of maturity were
illustrated by Kawas et al. (1991), who observed that not only did intake and production
of dairy cows decrease as maturity of alfalfa at cutting increased, but also that decreases
in forage quality could not be compensated for by increases in concentrate feeding.
Several mechanisms exist for possible improvements in fiber digestibility.
Researchers are beginning to measure fiber digestibility in forage breeding programs
and are observing sufficient genetic variation to warrant selection on this trait (Buxton
and Casler, 1993). Genetic improvements in certain compositional characteristics of
forage quality, such as crude protein content and detergent fiber concentrations, have
already been obtained. Improvements in fiber digestibility will probably be centered
around changes in the 1ignin:NDF ratio. As previously noted, certain mutant strains of
corn, as well as sorghum, millet, and sudangrass, contain lower concentrations of lignin
that has been shown to improve fiber digestibility (Cherney and Cherney, 1991).
Isolation of the gene(s) responsible for this mutation and its incorporation into other
forage species remains a possibility. Other techniques may be available to increase fiber
digestibility after the forage has already been harvested (Fahey et al., 1993). Physical
alterations, including pelleting and grinding, irradiation, steam treatment, and
mechanical separation of plant parts, have been investigated with limited success.
Chemical treatments have included hydrolytic cleavage with alkali, such as sodium
hydroxide, ammonia, or urea, and oxidation with ozone, sulfur dioxide, and other

oxidizing agents. Grasses have been more easily improved with chemical treatments

62
because of their higher proportion of ester linkages compared to other plant lignins.

Finally, biological treatments have been implemented and include the use of white-rot
fungi (Agosin et al., 1985), cellulases, and bacterial inoculants in fermented silages (Pitt,
1990). Most of these treatments work best on poor quality forages with high fiber and
lignin concentrations. Additional efforts are being made to improve their utility in
higher quality forages more commonly fed to lactating cows.

Future Directions

If fiber digestibility is found to have an important influence on feed intake in
dairy cows, it will be necessary to understand what aspect of fiber digestibility is most
important. Fiber digestion has been defined in terms of rate of digestion, lag until
fermentation, or extent of digestion at various times post commencement. What
definition is most highly related to intake? Mertens (197 3) observed that in vitro
digestion after 6 h was the most highly correlated time point with intake for sheep. This
may, however, only reflect the effect of fiber content on intake. Miller and Muntifering
(1985) suggested that the amount of potentially digestible fiber is more important than
rate of digestion, rate of passage, or digestion lag in determining in vivo digestibility.
Rumen content mass in lactating cows was related more to fiber rate of digestion and
passage than to total tract digestibility or maximum extent of fiber digestion (Shaver et
al., 1988a). Finally, Von Keyserlingk and Mathison (1989) found that forage intake in
steers was related most to 36 h in situ degradation values.

Because fiber digestion depends on such a plethora of plant and animal .
characteristics, it may seem impossible to ever quantify its true affect on intake. Models
of fiber digestion offer researchers the opportunity to visualize how these different
factors interact with one another (Allen and Mertens, 1988b). Besides suggesting what
potential experiments might be crucial to more accurately describe fiber digestion,

models may also enable researchers to assess the importance of fiber digestibility under

numerous plant, animal, and environmental conditions.

 

 

CHAPTER 2

Continuous Computer Acquisition of Feed and Water Intakes, Chewing,

Reticular Motility, and Ruminal pH of Cattle

ABSTRACT

Monitoring of feeding, chewing, and rumen activity was integrated into one data
acquisition system for continuous measurement of 12 dairy cows. Feed mangers were
hung from single-point load cells for measurement of feed disappearance from individual
stalls. Water flow meters were inserted in supply lines for each stall and generated pulse
output for electronic summation of water intake. Jaw movements were detected with a
water-filled tube connected to a pressure transducer under the cow's jaw to determine
chewing activity. Similar tubes were used to detect contractions in the reticulum. Rumen
pH was monitored continuously with an electrode and pH transmitter. All signals were
processed and recorded on a microcomputer using commercially available computer
hardware and software. One file was written for each animal monitored. Data were
interpreted using algorithms developed with SAS®. Two studies were conducted with 10
lactating cows to evaluate the performance of acquisition hardware, protocols, and
interpretation algorithms. Interpreting behavior of many cows with one algorithm
sacrificed accuracy of bout time borders for some individual cows. Nonetheless, high
correlations (r 2 .85) between computer-interpreted and manually determined variables
indicawd acceptable performance of the acquisition system. With continuous measurement
of many animal feeding variables, a more complete understanding of dietary effects on
digestive function and animal performance is possible.

63

 

 

 

64
INTRODUCTION

Interactions among dietary characteristics and ingestive behavior of ruminants have
received considerable research attention (Balch, 1958; Freer and Campling, 1965; Jaster
and Murphy, 1983). Intake, feeding rate, salivation, rumination, and digesta passage are
animal traits that may respond to changes in diet characteristics such as nutrient
composition, physical properties, or palatability. Interpreting effects of dietary
manipulation on digestive efficiency and animal performance requires quantitative
measurement of responsive animal variables. More frequent measurement is necessary to
obtain reliable information from rapidly changing variables. Continuous measurement
ensures that even the most rapid fluctuations are detected. Numerous animal monitoring
systems have been developed to provide such information.

Feed consumption of individually housed ruminants has been recorded throughout
the day by frequently offering known quantities of feed in separate containers for each
meal (W angsness et al., 1976) or by electrical measurement of strain gauge devices placed
below (Brewer et al., 1977) or suspended above (Metz and Borel, 1975; Heinrichs and
Conrad, 1987) feed mangers. Loose housed animals have been monitored using
computer-controlled feeders with animal identification capabilities (Mason et al., 1991).
Water intake has been recorded continuously by Metz and Borel (1975). Automatic
methods to detect chewing activity have been reviewed by Penning (1983) for grazing
ruminants and by Beauchemin et a1. (1989) for ruminants fed in stalls. All methods use
jaw movement as an indication of chewing, which is detected with electronic switches
(Metz and Borel, 1975; Luginbuhl et al., 1987), pneumatic transducers (Law and
Sudweeks, 1975), or hydraulic transducers (J aster and Murphy, 1983). Ruminant
forestomach motility is essential for fractionation, fermentation, and passage of rumen
digesta (Ruckebusch, 1988) and frequently is measured to assess factors affecting

digestive function (Stevens and Sellers, 1968). Monitors for detection of motility include

balloons pneumatically linked to chart pens (Balch, 1958), pressure-sensitive

 

 

65
radio-telemetry capsules (Riley and Cook, 1974), and catheter tubes connected to pressure

transducers (Okine et al., 1989). Digesta pH in the reticulorumen was measured in vivo
as early as 1941 (Smith, 1941). Since then, continuous pH measurements have been
made by various groups (Johnson and Sutton, 1968; McArthur and Miltrnore, 1968) for
several weeks using electrodes designed to withstand the rumen environment.
Advantages of all automatic monitoring systems include reduced labor required for
visual or manual determinations, improved accuracy, and measurement of some activities
that are manually impossible to monitor. Multiple animal responses may result from a
single dietary change due to ruminant digestion complexity. Monitoring two or more ' ‘
variables allows interactions among activities to be examined and may improve accuracy of
measurement and interpretation for each activity because of such interdependence. Some
studies have monitored more than one response variable continuously and simultaneously
(Freer and Campling, 1965; Metz and Borel, 1975); these data, however, are rare.
More recent studies incorporated monitoring devices into complete data acquisition
systems (Penning et al., 1984; Luginbuhl et al., 1987 ; Beauchemin et al., 1989). Striving
for complete automation, these systems consist of activity monitors, data collectors, and
interpretation programs. Unfortunately, these systems have been limiwd by inabilities to
record activity over long periods, to measure many experimental animals simultaneously,
or to monitor more than one activity. Recent advances in data acquisition software and
computer technology enable flexible acquisition control, rapid data collection, and large
data storage. It is now possible to expand the number and length of experimental
observations and the number of activities that may be monitored automatically and
continuously.
Objectives of this experiment were to develop and construct a data acquisition

system that continuously and simultaneously monitors feed and water intake, chewing

activity, reticular motility, and ruminal pH of cows housed in stalls. A ftn'ther objective

 

 

66
was to evaluate monitoring equipment and interpretation algorithms against manual

observation for each activity.
MATERIALS AND METHODS
Acquisition System

The complete measurement system consists of monitoring equipment for 12 cows,
a computer, software, and data acquisition boards for data collection, and a computer
program for data interpretation (Appendix Figure 1). Tie stalls at the Michigan State
University Dairy Teaching and Research Center were modified to accommodate
monitoring equipment. Equipment wiring was designed to enable detachment so cows
could be milked in a parlor. All electronic equipment was purchased from commercial
sources to ensure compatibility among original equipment and replacement parts.

Feed Intake. A 425-L polyethylene feed manger (Model #99075; Bastian Material
Handling Corporation, Indianapolis, IN) was suspended with four cables from a single S
beam load cell (Table 1) for each stall. The 225-kg capacity load cell was mounted to
metal framework 2 m above the floor. The stall and frame were designed in a manner

Table 1. Monitoring equipment.

 

 

E . . D . l I I I U E 5
Feed intake Load cell LCCA-sool Analog 0-225 kg .25 kg
Water intake Flow meter F'I'BGIOS-F‘Rl Pulse .9-94 um .05 L
Chewing Pressure PX236-015l Analog 0-1.05 kg/cm2 .016 kg/cm2
activity transducer

Reticular Pressure PX236-015l Analog O-l.05 kg/cm2 .016 kg/cm2
motility transducer

Ruminal pH pH electrode 450CD2 Analog 0-14 pH units .02 pH

pH transmitter PH’I‘X-9ll Milliamp 0-14 pH units .02 pH

 

lOmega Engineering, Inc., PO Box 2284, Stamford, CT .
2Sensorex, Inc., 11661 Seaboard Circle, Stanton, CA.

67
similar to that developed by Metz and Borel (1975), that minimized cow interaction with

the manger except during eating. Loadcells were supplied with 10 V of power that could
be switched on and off independently and generated output of 30 mV over a 0- to 225—kg
range. Load signals required 200x amplification prior to input into the analog converter.
Supply current and return signal for all monitored activities were transmitted with 24-
gauge four-conductor shielded wire (Belden 9534; Belden Wire and Cable, Richmond,
IN). An example of feed weights measured with this design for one cow during one day
is presented in Figure 1.

Water Intake. Individual drinking cups and in-line water flow meters (Table l)
were provided for each cow. Cumulative water flow could be read directly from the meter
register or electronically totaled. Meters calibrations indicated that 26.6 :l: .35 electronic
pulses/L were registered using an input of 5 V. Additional signal conditioning was
required to amplify pulse signals prior to processing. Counter-timer circuits accumulated
pulses for each cow over time (Figure 1).

Chewing Activity. Jaw movements were monitored and counted to estimate
chewing activity. A water-filled rubber bicycle tube (31.8 cm id x 30.5 cm) was fixed to
a nylon halter under the cow's jaw and connected to a pressm'e transducer (Table 1) via a
compression fitting on the valve stem (Jaster and Murphy, 1983). The tube was
supported by a piece of latex to ensure contact of the tube with the jaw while maintaining
elasticity for free jaw movement. A retractile cable (Belden 8415) allowed head movement
and connected the halter and transducer to the power source (10 V) and analog converter.
Opening of the jaw decreased the volume of the tube causing increased pressure that was
read as an analog signal by the computer. Typical analog signals for eating and ruminating
activity are presented in Figure 2. Analog signals were sampled at 4 Hz, enabling
detection of chewing rates up to 2 Hz for cows receiving mixed forage and concentrate
diets. A potential chew was identified when the derivative of the sampled analog signal
exceeded a negative threshold value that was constant for all cows monitored. This

68

 

 

 

 

 

 

 

 

307
’6:
£5 25.
E.
'5 20..
3
'8 15..
a:
LI.
10 I I I I T I
so -
=0} 40‘
f‘g 30-
.5
t. 20-
2
“3 1o-
3
o I I I I I I
7.25-
I 7.00..
O.
a“: 675
E . ..
3
II 6.50-
625 T f | T I I
o 4 s 12 16 20 24
Hour of day

Figure l. Scaled output of feed manger, water meter, and pH monitors throughout 1 d
for a cow fed twice daily.

69

Reticular contractions

 

 
   

Reticular contractions

Analog signal (mV)

Chewing

O 20 40 60 80 100 120

fime$)

Figure 2. Analog output from chewing and reticulum motility monitors during
episodes of eating and ruminating. Values presented are scaled pressure transducer
output.

70
threshold value, determined by trial and error, depended on characteristics of the monitors,

such as degree of water filling in the tube and proximity to the cow's jaw. This criterion
allowed accurate peak counting, regardless of baseline location. Detection of higher
chewing frequencies are possible with this system, however a rate of 4 Hz allowed more
accurate peak detection when using the derivative threshold method.

Reticular Motility. A pressure transducer and water-filled rubber tube similar to
that used for monitoring chewing was used to detect pressure changes from reticular
contractions. Ends of the tube were tied to form a "balloon" 20 x 7.5 x 3 cm. The tube
contained 1.0 kg of lead mass to keep the device within the reticulum. Electrical wires
were protected from the rumen environment with plastic tubing (Tygon R-3603; 1.3 cm in
diameter, American Scientific Products, McGaw Park, IL), and retractable four-conductor
wire was used exterior to the rumen to allow cow movement. Monitoring devices were
placed in the reticulum through a permanent rumen fistula and secured to its plug. Figure
2 illustrates typical analog signals received from a motility monitor supplied with 10 V
during eating and ruminating activity. Motility signals were sampled at 4 Hz and averaged
over a period of 1 3. Potential reticular contractions also were identified by the derivative
of a smoothed signal, similar to that performed for chewing activity. Two or more
potential contractions occurring within 15 s of each other were required for verification of
a reticular contraction. This criterion helped keep reticular pressure changes due to
coughing, licking, and other activities from being recorded as contractions.

Ruminal pH. Continuous pH measurement required use of an electrode and
transmitter. An epoxy body, liquid-filled combination pH electrode (Table l) with
recessed bulb and sealed reference electrode provided durability necessary to withstand the
rumen environment for extended periods. The wire lead and upper quarter of the electrode
were protected from digesta with a 1.6-cm diameter plastic hose. A metal shroud attached
to the end of the electrode allowed digesta to pass freely around the pH bulb while keeping

it from resting against the ventral rumen wall. Against the wall, electrode pH was

7 1
approximately .7 pH units greater than within the ventral sac, possibly because of rapid

absorption of VFA or elevated ammonia concentrations near the wall. Values within the
ventral sac digesta probably are more important for characterizing ruminal fermentation
than are readings near the rumen wall. The electrode was held about 8 cm off the floor of
the ventral sac by attaching 1.0 kg of lead sealed in plastic to the metal shroud and
securing the plastic tubing to the cannula plug through which it passed.

Outside the cow the electrode lead was plugged into a four-wire pH transmitter
(Table l). A separate transmitter was used for each cow monitored. Transmitters
provided 4 to 20 mA direct current output over any two units of pH that was converted to
a millivolt signal at the computer by use of a shunt resistor. In addition, pH was displayed
visually on the transmitter panel where two point calibration and manual temperature
compensation could be performed. Transmitters required a 110-V standard power supply.
Electrode leads were suspended from elastic cord to keep wires away from cows' hooves.

Dming periods of pH measurement, each electrode-transmitter combination was
calibrawd at pH units of 4 and 7 before insertion of the electrode into the ventral sac.
Rumen pH for one cow for l d are presented in Figure 1. Electrode calibration was
diffith to maintain over several days because of static build-up, faulty solid-core
electrode leads, and rumen fluid leakage. Electrodes were recalibrated once every 2 d to
help to maintain accuracy. Recent changes in tubing, connectors, and electrode leads have
improved electrode performance.

Data Collection. All electronic signals were received by an IBM PC-AT
compatible computer (Table 2) located in a room adjacent to the cow stalls. One, eight-
channel analog to digital converter and two, five-channel counter-timer boards (Table 2) fit
into three half-slots of the computer. Each of four channels of the analog to digital
converter was expanded to 16 channels by using an analog multiplexer (Table 2) for a total
of 68 available analog inputs. Each multiplexer provided analog signal filtering and

independent attenuation. Pulse output from water meters was measured via the two

72

Table 2. Computer, software, and data acquisition boards.

 

Specifications

 

 

Device Modem;
IBM PC-AT compatible Premium/2861
computer
8 channel analog-digital DAS-83
converter
Analog multiplexer EXP-163
Counter-timer board CTM-OS3
Uninterruptable power P024004
supply
Data acquisition Labtech Notebook5
software Version 5.0

80286 microprocessor at 10 MHz,
0 wait state, 80287 co-processor,
640 kbyte base memory,

40 Mbyte fixed hard disk,

88 Mbyte removable hard disk2

Expandable to 128 channels; 12 bit,
2.4 mV resolution at :l: 5 V firllscale;
2 counter-timer channels

16 channel with user-selectable gains

5 independent up/down counters at
16 bit resolution

120 VAC, 800 watt; on-line output,
inverter powers load continuously

Menu-driven program for data acqui-
sition and process control with real-
time graphic data display; fully com-

patible with data acquisition boards

lAST Research, Inc., 2121 Alton Avenue, Irvine, CA.

2SyQuest Technology, 47071 Bayside Parkway, Fremont, CA.

3Omega Engineering, Inc., PO Box 2284, Stamford, CT.

4C1ary Corporation, 320 West Clary Avenue, San Gabriel, CA.
SLaboratory Technologies Corp., 400 Research Drive, Wilmington, MA.

73
counter-timer boards and two available counting channels on the analog to digital

converter. Separate screw terminal boards were required for each counter-timer board and
the converter. This hardware setup enabled each of the five activities to be measured on
12 cows simultaneously with one computer.

Collection protocols were carried out with a commercially available data acquisition
software program (Table 2; Appendix Table 2). The menu-driven program provided
flexibility, with easy adjustment of channels monitored, sampling rate, channel triggering,
and data processing. Real-time graphical displays permitted efficient acquisition and
algorithm debugging. All acquisition boards were compatible with the software and were
controlled automatically once defined in a collection protocol.

To collect information for five activities from 12 cows, 145 channels were
required: 1 time channel, 12 counter channels, 48 analog input channels, and 84
calculation channels. Channels were sampled at various rates depending on the activity
and algorithm; however, channels written to files all were sampled at .2 Hz. One file was
written for each cow monitored with all five activities contained in each file. An example
of a file produced from an acquisition run for 1 cow is presented in Figure 3. Water
values are total pulses since commencement of the run, load and pH values are their
respective measurements at the listed time, and chews and contractions are their total
within the previous S-s interval. During collection, files were written directly to the
computer's internal hard disk. The 88-Mbyte removable hard disk has adequate capacity
for 14 d of continuous collection for 12 cows.

Data Interpretation. Completed data files were interpreted using data management
procedures of SAS (1985) on a microcomputer (Appendix Table 4). Objectives of
interpretation were to designate time as eating, ruminating, drinking, or idle and determine
chews, feed consumption, water consumption, pH changes, and contractions during each

period. Activities were determined with a precision of 30 s. Time definitions shorter than

74
"Cow l, Expt. 91RD001"
"Feeding Behavior Validation"
"The time is 12:00:01.50"
"The date is 1-28-1991"
“Time" "Water" "Load" "Chews" "Contr." "pH"

"sec" ".01L" "kg" "chews" "counts" "units"
5 0 24.8 6 0 6.73
10 0 24.8 4 0 6.73
15 0 24.8 5 0 6.73
20 0 25.0 1 1 6.73
25 0 24.9 4 l 6.73
30 0 24.8 5 1 6.7 1
35 0 24.8 7 0 6.72
40 0 24.7 6 0 6.73
45 0 24.7 7 0 6.73
50 0 24.8 5 0 6.73
55 0 24.8 1 0 6.75
60 0 25.0 2 1 6.74
65 0 24.8 5 2 6.74

Figure 3. File output of feeding behavior data from Labtech® Notebook in ASCII
format.

75
30 3 led to inconclusive evidence for the presence or absence of certain chewing activities,

especially when cows changed from one activity to another slowly or erratically.

Jaw movements met three criteria to be considered as chews. No drinking could
have occurred during periods of jaw movement, chewing rate must have equaled or
exceeded .45 chews/s over a 180-5 period, and chewing rate also must have exceeded .3
chews/s over a 30-s interval. The first chewing rate threshold permitted accurate
differentiation between active chewing and idle jaw movement; the second threshold
defined more accurate time borders between events. Any jaw movement not meeting
requirements was considered idle activity. Frequent disturbances of the feed manger by
the cow were used to differentiate ruminating and eating chews. If the standard deviation
of manger weights over a 1 10-s range exceeded .45 kg, chews for that particular 5-s were
denoted as eating chews, otherwise they were considered ruminating chews. Additional
robustness was gained by comparing eating and ruminating chew totals over longer
periods of time and minimizing inconsistent chewing patterns. Stable feed manger
weights, averaged over 110 s that occurred in the absence of chewing, were used to
determine meal size. Beginning and ending pH for each bout of activity were the average
of 24 measurements recorded over 2 min. Drinking periods occurred when water flow
was indicated.

Determination of the minimum interbout interval for eating, mminating, and
drinking activity was performed using procedures of Men (1975). A minimum interbout
interval is the minimum time interval between two of the same activities required to
consider these two periods as separate events. For each type of activity, the frequency of
all inactive intervals was determined. The smallest interval was .5 min, and the largest
exceeded 300 min. Frequency was plotted as percentages of all intervals cumulatively and
backwards on a log scale to produce survivorship curves (Men, 1975) (Figures 4 arxl 5).
Deviations from linearity indicated that probabilities of beginning new periods of activity

were not random. Plateaus of the line occurred when a shortage of intervals exiswd and

..L
O
O

 

' \

7.5 minutes

 

Percent cumulative frequency (log scale)
8

1 . . . . l

0 50

76

Eating

/

Ruminating

j I j A

1 00 1 50 200

Inter-period interval (min)

Figure 4. Frequency of inter-period intervals for eating and ruminating activity of
twelve cows measured for twelve days. Data are expressed as survivorship curves with
percent frequency determined cumulatively and backwards (Men, 1975). The chosen
minimum inter-bout interval of 7.5 minutes is indicated.

77

_L
O
O

    

4 minutes

 

I I I A L J

50 1 00 1 50 200
Inter-drinking interval (min)

 

Percent cumulative frequency (log scale)

..l.
O
C

Figure 5. Frequency of inter-drinking intervals of twelve cows measured for twelve
days. Data are expressed as a survivorship curve with percent frequency determined
cumulatively and backwards (Men, 1975). The chosen minimum inter-bout interval of 4
minutes is indicated.

78
provided a point of demarcation between those intervals that were likely to occur within

the same bout of activity and those that were breaks between different bouts. Based on
curves generated from data gathered on 12 lactating cows measured over 12 d, minimum
interbout intervals of 7.5 min were chosen for both eating and ruminating activity (Figure
4) and 4.0 min for drinking activity (Figure 5). These criteria were included in the
interpretation algorithms to aid in differentiating bouts.

Output generated after data interpretation are presented in Table 3 for one animal
monitored for 1 d. Bout start and stop times are expressed as fractions of a day to aid in
comparison of activities across days. Actual activity times for each bout are the sum of all
30-s intervals of the same activity within the bout. Because of presence of small breaks
within activity bouts, actual time spent performing a particular activity could be equal to or
less than the overall time difference between the start and stop time of a given bout.
Overall and actual activity durations have been used in another study (Wangsness et al.,
1976). All calculated rates involving time were determined using actual activity times.
Rumination bouts less than 5 min were deleted as well as eating bouts when less than .05
kg of feed was consumed. Idle bouts were periods of time where eating, ruminating, or
drinking were not occurring. Output was saved as permanent SAS data sets for
subsequent statistical analysis. Run-time on the microcomputer (IBM PS/‘2 Model P70
386-20 MHz; International Business Machines Corporation, Armonk, NY) for data
interpretation was 15 min for each cow and day combination.

Validation

Two validation studies were performed to determine the operational performance
of monitoring equipment and interpretation algorithms. Feed intake, water intake, and
chewing activity of 10 lactating dairy cows (4 primiparous) were measured both manually
and automatically for 3 d in the first study. Cows averaged 79 DIM (SD = 11) and
produced 34.4 i 5.9 kg/d of milk. Cows were monitored electronically and adjusted to
equipment and diet for 16 d prior to manual data collection. Diets were mixed once daily

79

Table 3. Output from the feeding behavior data interpretation program generated for
eating, ruminating, and drinking bouts for one cow measured for 1 d.

 

 

 

 

 

 

Time
Bout since Activity Feed Water Chew 9L 1
S S l . . l . l Cl 5 S 2

(00- (no.

(no) —-(d)2--- —-<min)—— (kg) (L) (no) 7min) Imin)
Eatingbouts
1 .017 .030 13.0 .36 768 59.1 6.72 6.76 1.77
2 .097 .135 97.5 46.0 3.94 3294 71.6 6.91 6.61 1.84
3 .173 .211 54.5 48.0 5.17 3400 70.8 6.67 6.54 1.93
4 .330 .364 171.5 44.0 3.45 3266 74.2 6.88 6.69 1.79
5 .472 .485 156.5 14.5 1.22 972 f 67.0 6.97 6.84 1.73
6 .520 .541 50.0 30.5 2.04 2209 72.4 6.75 6.70 1.71
7 .691 .712 215.5 29.5 1.90 2198 74.5 7.22 6.92 1.77
8 .788 .824 109.5 41.5 3.72 3059 73.7 7.01 6.74 1.75
9 .832 .839 11.0 9.5 .82 669 70.4 6.81 6.72 1.89
10 .907 .919 98.5 15.5 1.41 1127 72.7 6.96 6.78 1.83
11 .932 .956 17.5 31.0 2.81 2207 71.2 6.90 6.61 1.77
12 .970 .983 19.5 15.0 1.13 1011 67.4 6.58 6.58 1.78
Rumimtingbouts

1 .013 .017 7.0 414 59.1 6.75 6.72 1.57
2 .035 .050 26.0 20.0 923 46.2 6.73 6.87 1.67
3 .059 .070 13.5 15.0 674 44.9 6.89 6.89 1.73
4 .085 .097 22.5 17.0 819 48.2 6.89 6.91 1.71
5 .147 .165 72.0 25.0 1176 47.0 6.76 6.89 1.49
6 .222 .237 81.5 22.0 1137 51.7 6.63 6.82 1.55
7 .247 .269 14.0 31.5 1532 48.6 6.78 6.83 1.69
8 .316 .330 68.0 20.0 1355 67.8 6.77 6.88 1.50
9 .378 .407 69.5 41.0 2567 62.6 6.80 6.78 1.42
10 .424 .446 25.0 31.0 2069 66.7 6.80 6.77 1.32
11 .466 .472 29.0 9.0 496 55.1 6.77 6.97 1.56
12 .551 .579 113.5 38.5 2108 54.8 6.71 6.54 1.58
13 .590 .605 15.0 21.0 1028 49.0 6.71 7.00 1.51
14 .636 .652 45.0 16.5 857 51.9 7.03 7.09 1.70
15 .671 .691 27.0 24.5 1314 53.6 7.17 7.22 1.58
16 .724 .744 47.0 28.5 1691 59.3 6.83 6.85 1.73
17 .758 .767 20.0 7.5 323 43.1 6.88 6.99 1.52
18 .852 .880 122.0 35.0 1676 47.9 6.71 6.81 1.63
19 .893 .906 19.5 18.5 1109 59.9 6.79 6.76 1.46

80
Table 3. (cont)

 

 

 

Time

Bout since Activity Feed Water Chew pH 1

WM time inlakLinlakLJhms—raLm—ML

(no. (no.

(no) —+d)2--— ~—(min)—-— (kg) (L) (no) /min) Imin)
Drirkingbouts
1 .026 .026 .50 2.16
2 .123 .125 139.5 1.33 6.09
3 .132 .133 10.0 .42 2.19
4 .192 .194 86.0 .83 3.33
5 .208 .208 19.0 .67 3.29
6 .346 .347 198.0 .75 3.33
7 .364 .365 24.5 .50 2.19
8 .508 .509 206.5 .25 1.14
9 .541 .543 47.0 1.17 5.33
10 .712 .714 244.0 1.42 5.52
11 .811 .817 140.5 1.17 5.11
12 .840 .841 33.5 .17 .83
13 .914 .916 106.0 1.08 4.35
14 .951 .952 51.0 .50 2.12
15 .983 .983 45.0 .17 .46

 

lCont = Contractions.
2Time expressed as fraction of a day.

 

81
and offered for ad libitum intake as a TMR twice daily when cows were away from stalls

for parlor milking. Ration ingredients included alfalfa silage (20% CP, 40% NDF; DM
basis), corn silage (6% CP, 40% NDF), high moisture shelled corn (9% CP, 9% NDF),
soybean meal (55% CP, 15% NDF), a commercial protein pellet (65% CP, 4% NDF),
and necessary minerals and vitamins to provide a ration with 17% CP, 31% NDF, .8%
Ca, and .5% P. Manual observations of chewing activity were recorded once every 5 min
for the 3-d period. Hourly and daily summaries of chewing times and bouts determined
from manual observation were compared with acquisition system values for each cow.
Feed and water intakes were determined manually twice daily and compared to system
totals.

A second study was conducted with 10 ruminally cannulated cows to evaluate
performance of automatic reticular motility and rumen pH measurements. Cows were
monitored during early lactation and received a mixed ration (25 or 35% NDF; 18% CP)
with the same ingredients as those of the first study. Motility and pH monitors were
placed in the reticulorumen at various times during 1 wk, and computer measurements
were made at .2 Hz as previously described. Reticular contractions were determined
manually by visual observation of analog signals produced by each motility monitor on the
graphic display of the acquisition program. Rumen liquor samples were obtained from the
same location in the ventral sac as the in vivo electrode using a covered beaker and
immediately measured for pH with another electrode exterior to the cow. Computer
estimates and manual estimates were compared for both 30-min contraction totals and pH
readings for each cow.

For both validation studies, computer and manually derived variable means were
tested for equality using paired t tests (Snedecor and Cochran, 1980). Statistical
procedures similar to those used by Beauchemin et al. (1989) were employed to examine
further the relationships between the two methods of observing cow activity and to

partition sources of variation among these methods.

 

82

RESULTS AND DISCUSSION

Cow adjustment to the monitoring equipment and collection stalls was minimal

because of little change in variable means across time during the adaptation period. Feed

removed by cows from the suspended manger but not consumed was less than .10 kg/d

per cow. Equipment malfunctions were limited to short circuits in connectors for

chewing, motility, and pH monitors, ruptured chewing and motility tubes, and drifts in

pH calibration. Improvements have been implemented to alleviate most of these problems.

Currently, greater than 90% of all attempted measurements are obtained successfully.

Daily means of intake and chewing activity for cows on the first validation study

are presented in Table 4. Based on the definition of chewing bouts described earlier, mean

numbers of eating and ruminating bouts were 10.6 and 14.4 /d, respectively, which are

similar to other data reported for cattle (Men, 1975; Deswysen et al., 1987a).

Table 4. Intake and chewing activity of 10 cows measm'ed for 3 d using computer

acquisition and manual observation.1

 

 

 

 

__C0nlrl.llt¢r ManuaL____l21ffmce_
Mm SE Mean SE M.

DMI, kg 23.5 .7 23.7 .7 -.2** .1
Water intake, L 83.0 2.8 83.0 2.8 0 .1
Eating bouts, no. 10.6 .4 11.0 .4 -.4* .2
Eating time, min 306 12 315 12 -9* 4
Eating chews, no. 19456 834 . . . . . . ...
Ruminating bouts,2 no. 14.4 .4 14.3 .4 .1 .2
Ruminating time,2 min 495 9 491 9 4 4
Ruminating chews,2 no. 31874 786 .. . . . . . ..
Total chewing tirne,2 min 801 14 806 15 -5 6
Idle chews, no. 2729 108

 

lData expressed as daily means of 26 observations.
2Does not include rumination (mean = 33 min/d) observed dining milking.

*P < .05.
“P < .01.

83
The chosen minimum interbout interval for chewing activity of 7.5 min appears adequate

for differentiating bouts and is in agreement with the 7-min criterion described by Dulphy
(1971). A mean of 54,059 total chews/d were measured of which 36% were denoted as
eating chews, 59% as ruminating chews, and 5% as idle chews (nonchewing jaw
movements). Because cows were away from monitoring equipment about 2 h/d for
milking, an average of 33 min/d of rumination (6% of all rumination), as determined by
manual observation, was not recorded by the acquisition system. Stall milking has since
been implemented to allow collection of all data.

No differences (P > .25) in means between measurements obtained with the
acquisition system and manual observation occurred for daily water intake, ruminating
bouts, ruminating time, and total chewing time (Table 4). Dry matter intake measured
manually was .2 kg (.9% of total intake) greater (P < .01) than that recorded
automatically due to omission of meals less than .05 kg in size for computer data. Eating
bouts and eating time per day also were greater for manual measurement for similar
reasons. I11 addition, differentiation of eating bouts for manual observations was
performed with a minimum interbout interval of 5 min, compared with the 7.5-min interval
used for electronic observations. Consequently, more bouts were likely to be identified
with the manual method. Magnitudes of these differences were small, suggesting good
perfomrance of the acquisition equipment and interpretive algorithms.

Relative agreement between automatic and manual methods of measuring cow
activity was estimated as the root mean square error (RMSE) (Beauchemin et al., 1989).
This estimate of total error was further decomposed into error associated with mean bias
(Table 4), regression, and random error (Table 5). Slopes between the two methods of
collection indicated the direction of regression error. Hourly totals of chewing activity
were correlated highly between the two methods (r 2 .94) with most error associated with
random disturbance. Daily total RMSE as a percentage of the mean was lowest for DM

and water intake (< 2.5%) and highest for eating and ruminating bouts. Eating and

Table 5. Relationship of intake and chewing activity of 10 cows measured with computer
acquisition and interpretation (independent data) and manual observation (dependent data).

 

1

Regression Random

 

RMSE error error Slope r
Hourly totals
Eating time, min 3.78 .41 3.75 1.00 .97
Ruminating time, min 3.46 .02 3.35 .97** .97
Total chewing time, min 4.67 .36 4.50 .97 .94
Daily totals
DMI, kg .57 .09 .39 .99 1.00
Water intake, L .52 .16 .49 1.01 1.00
Eating bouts, no. .94 .26 .80 .87 .91
Eating time, min 20.3 2.45 18.2 .96 .95
Ruminating bouts, no. .98 .25 .94 .88 .88
Ruminating time, min 21.7 6.91 20.2 .85 .89
Total chewing time, min 28.5 2.04 28.0 .97 .93

 

lRMSE = Root mean square error.
2Random error = Discrepancy between two procedtu'es not associated with mean bias

or regression.

"Slope significantly different from unity (P < .01).

85
ruminating time RMSE were 20.3 and 21.7 min/d, respectively. Most error was due to

random error (> 68%) for all variables; however, the slope for hourly ruminating time
differed significantly (P < .01) from unity. Correlations for daily totals were high.
Lowest correlations were obtained for ruminating bouts (r = .88) and ruminating time (r
= .89). Manual observations indicated the presence of several sporadic rumination bouts
of short duration nested within eating bouts. From scanning of raw computer collection
data, it was noted that some activities that were determined manually were incorrect. For
example, eating was noted manually at times when no activity at the feed manger was
detected electronically, as well as ruminating recorded during periods of feed
disappearance. This suggests that trained observers may err in differentiating eating from
ruminating. Less agreement between manual and automatic observations for rumination
events may have occurred because of incorrect manual observations or the inability of the
algorithms to handle rapid fluctuations in cow behavior.

Feed and water intake measurements by the acquisition system were most accurate
among all activities measured in the first study. Chewing interpretation was more difficult
because of variable chewing and feeding patterns among cows and the presence of jaw
movements that were not chewing. Other workers that have developed chewing
interpretation algorithms have differentiated eating and ruminating chews based on rates of
chewing (slower and less variable for rumination), duration of chewing, and presence of
pauses during rumination due to bolus swallowing and regurgitation (Penning et al., 1984;
Luginbuhl et al., 1987; Beauchemin et al., 1989). In the present study, chewing rates for
cows in early lactation were nearly identical for both eating and ruminating (61.6 and 62.9
chews/s, respectively). In addition, 1 of the 10 cows continued to have jaw movement
during rumination bolus swallowing and regurgitation, consequently limiting the
usefulness of this characteristic for distinguishing chew types. Use of feed disappearance
to make this distinction was a viable alternative. Separating idle jaw movements has been

performed using various criteria with most based on inadequate event numbers per unit

86
time (Penning etal., 1984; Beauchemin et al., 1989). Similar criteria were useful in this

study. Regardless of algorithm specifics, the greatest challenge in algorithm development
for any cow activity is establishing a compromise between robustness necessary to
interpret varied behavior patterns and event accuracy. When activities change rapidly, it is
difficult to achieve both correct activity identification and exact event time borders.
Inevitably, accuracy is sacrificed as more diverse behavior is interpreted with the same
algorithm. Requirements of individual experiments should dictate the relative position of
this compromise.

Comparisons between computer- and manually derived measurements for reticular
motility were made using 88 observations from the second validation study (Figure 6).
Monitors successfully remained in the reticulum regardless of animal position or
movement. Number of contractions in each 30-min period ranged from 22 to 58. Total
error between measurement methods was 1.6 contractions/30 min with most (80%) due to
random error. Mean contraction counts for computer (38.8) and manual (39.3)
observations were significantly different (P < .01) due to missed computer counting of
small peaks from lying cows. This error, however, was small and occurred infrequently.
Correlation between methods was greater than .98. Previous studies of rumen motility
have relied typically on manual interpretation of strip-chart recordings (Okine et al., 1989).
Autonntion of collection and interpretation of these data could greatly expand the scope
and duration of motility studies.

Each pH electrode remained in the ventral rumen for the duration of the second
study. A total of 40 observations were compared between in vivo computer measured and
in vitro manual measured pH (Figure 7). Total error between methods was .17 units of
pH with 41% due to mean bias and 54% due to random disturbance. Correlation between
methods was .85. In vivo measurements averaged 6.17 and in vitro averaged 6.28 pH
units for a significant (P < .01) difference of .11 units. Other workers also have shown

lower values for in vivo pH compared to in vitro, with a mean difference of .17 units of

87

 

 

60 -
_>_. A
t .
g 50 ‘ ‘ ‘
3 A
> E A ‘ ‘
8 E J
3?, 40
° \.
35:1 1
g 30 i
E
o
o
20 . I u r v u v I
20 30 40 50 60
Contractions via computer acquisition
(no. I30 min)

Figure 6. Relationship between reticular contractions determined by manual
observation of signal waveform and computer algorithm interpretation for 88
observations. Data are expressed as total number of contractions within a 30-min period.
The line represents perfect correlation.

In vitro pH determined manually

 

 

 

5.9 6.2 6.5 6.8

In vivo pH via computer acquisition

5.6

Figure 7. Relationship between ruminal pH determined in vitro by manual sampling
and in vivo by computer acquisition for 40 observations. The line represents perfect
correlation.

89
pH (McArthur and Miltrnore, 1980; Smith, 1941). Smith (1941) suggested that rapid loss

of C02 from in vitro samples may have caused their elevated pH. Additional work is
nwded to determine the consequences of this disparity between in vivo and in vitro pH on
interpretation of past and future rumen pH studies.

C ONC L USIONS

A computer data acquisition system has been developed successfully to
automatically and continuously monitor feeding behavior and rumen function of 12 cows
in stalls. The system permits measurement of many events simultaneously for long
periods, is fast and very flexible, and does not require extensive training in electronics or
computer programming to construct or operate. Computer interpretation of collected data
provides additional automation and is essential for studies involving many cow
observations.

High correlations (r 2 .85) between computer interpreted and manually determined
variables suggest that the system and interpretive algorithms were successful in accurately
measuring animal behavior. Some accuracy in event interpretation was lost because of
attempts to achieve greater robustness. Such sacrifice is necessary, however, regardless
of the algorithm chosen in order to measure more diverse cows and to expand the
population of inference. With continuous measurement of many animal feeding variables,
a more complete understanding of dietary effects on digestive efficiency and cow

performance is possible.

CHAPTER 3

Variation In and Relationships Among Feeding, Chewing,
and Drinking Variables for Lactating Dairy Cows

ABSTRACT

Twelve Holstein cows (63 DIM; 6 primiparous) were offered a common diet and
monitored for 21 d (1 l d adaptation, 10 d collection) with a data acquisition system to
measure continuously feed and water intakes and chewing behavior. Objectives were to
examine relationships among feeding behavior variables for non-competing cows
producing various quantities of milk and to determine experimental designs with adequate
power to detect reasonable treatment differences in future experiments. Coefficients of
variation across cows ranged from 5 to 41% for the variables studied. Milk production
was correlated positively with DM and water intakes within and across parities. For
multiparous cows, production was related positively to meal size (r = .78) and length of
eating bouts (r = .7 5), and unrelated to meal number and eating rate. For primiparous
cows, production tended to be related positively to meal number (r = .55) and eating rate (r
= .87), and unrelated to meal size. Ruminating and total time spent chewing per unit of
DMI were correlawd negatively (r = -.58) with milk production both within and across
parities. These correlations suggest that differences exist between cows for chewing
efficiency. Reasons why high producing cows consume and chew more effectively
deserves further study. Contrast differences of 10% of means for variables examined had
an 80% probability of detection with a Latin square design utilizing 12 cows monitored for
5 d.

91
INTRODUCTION

Feed and water intakes by dairy cattle are related significantly to milk production
(Freeman, 1975; Holter and Urban, 1992). Increases in intake are associated positively
with both phenotypic and genotypic increases in milk production (Freeman, 1975). Intake
is defined as mass consumed per unit of time, with time commonly measured in days.
Shorter units of time, however, may be appropriate if they describe relationships that
cannot be detected with once daily measurements.

A complete understanding of daily intake requires that its individual components be
studied. Daily fwd intake can be described in terms of number of meals consumed per
day, the length of meals, and the rate of eating that occurs during meals. For daily intake
to increase, one or more of these three variables must increase. Feeding commences when
cows are in a state of hunger and stops when they are satiated. Therefore, daily intake is
the sum of several meals that are controlled individually by hunger and satiety mechanisms
(Savory, 1979). Within-day measures of feeding, ruminating, and drinking behavior
provide a mechanism to examine these components that contribute to daily intake.

Measurement of feeding and chewing activity has become common in nutrition
studies and has been reviewed by others (Beauchemin, 1991b; Campling and Morgan,
1981; DeBoever et al., 1990). Unfortunately, little is known about sources of variation in
feeding behavior studies and the likelihood of detecting true difl’erences between sample
means. Estimation of sample sizes required for effective studies would be useful. Only a
few studies have examined variation in feeding behavior among animals. Heifers
(Deswysen et al., 1987b), dry cows (Harb and Campling, 1985), and lactating cows
(V asilatos and Wangsness, 1980) have been studied; however, limited information is
available for high producing lactating cows. Identifying how high producing cows differ
in behavior compared with low producing cows is also of significant value. Such

information may help describe the mechanism involved with high intake and production.

92
Objectives of this experiment were to obtain feeding, chewing, and drinking

behavior information from lactating cows varying in milk production to examine
relationships among feeding variables across cows. A second objective was to estimate
the number of cows required to detect differences between potential treatments using
various experimental designs.
MATERIALS AND METHODS

Cow Trial

Six multiparous and 6 primiparous lactating Holstein cows averaging 63 DIM (SD
= 11) were housed in stalls designed for continuous collection of feeding activity. The
study was conducted in January under continuous lighting in a white-colored tie-stall barn
with no windows. Cows with similar DIM were selected to obtain a large range in milk
production and intake. All cows received a common diet of 29% alfalfa silage (DM basis),
29% corn silage, and concentrate for a total ration composition of 51% DM, 17% CP,
31% NDF, .8% Ca, and .5% P. The ration was mixed once daily and offered for ad
libitum intake as a TMR at 0300 and 1500 h when cows were parlor-milked. Cows
averaged 2 h/d away from feed and monitoring equipment during milking. Cow BW was
determined for 3 consecutive (1 immediately prior to and following behavior
measurements. Feed amounts offered and refused were recorded daily for use in
comparison to computer-collected data. Samples of feed and individual orts were collected
daily and composited weekly. Samples were dried in a forced-air oven at 55'C, ground
with a Wiley mill (l-mm screen; Arthur H. Thomas, Philadelphia, PA), and analyzed for
NDF (Goering and Van Soest, 1970) modified by addition of 4 ml of a 2% ct-amylase
solution (Sigma A-3051, Sigma Chemical Co., St. Louis, M0) to each sample,
substitution of triethylene glycol for 2-ethoxyethanol, and omission of
decahydronaphthalene and sodium sulfite. Values for NDF were expressed as a
percentage of DM determined from forced-air oven drying at 105‘C.

93
Feed intake, water intake, and chewing activity were measured continuously and

simultaneously for all cows with a data acquisition system (Chapter 2). Measurements
were made for 21 d with the last 10 (1 used for statistical comparisons. Behavior was
summarized for each cow-day combination according to eating, drinking, and ruminating
bouts. Minimum inter-bout intervals (MIBI) that define the occurrence of two separate
bouts of similar activity were used as described in Chapter 2. A minimum of 7.5 min
between events was required to define two eating periods or two ruminating periods as
separate bouts. For drinking, 4 min was required to define two periods as separate bouts.
A new meal was noted if feed disappearance 2 .05 kg, time elapsed since the previous
meal was 2 7.5 min, and chewing rate criteria were met. Variables for each bout included
start and stop times, duration, number of chews (for eating and ruminating), meal size (for
eating), drink size (for drinking), and ratios of these variables. Meal size was expressed
in units of DM and NDF. Calculations were required to account for changes in feed
moisture and NDF concentration during the day because of cow drooling and selectivity.
Because orts samples were not obtained after every meal, estimates of amounts of DM and
NDF in feed remaining after a meal were made by assuming linear changes in nutrient
mass between the beginning and end of the day for each unit of intake. Computer data
were screened and compared with manual data to eliminate errors that were due to
equipment malfunction, cow interference, and human error. If data were incomplete for
any time during a day, data for that cow-day were omitted from analysis. From the 10-d
collection period, 7.3 :l: 2.6 d/cow and 8.2 i 2.4 d/cow contained acceptable data for
chewing and drinking activities, respectively.
Statistical Analysis

Feeding activity was summarized for each cow-day combination to obtain either
daily means or totals of bout variables. Descriptive statistics were determined for all daily
variables; differences in means between parity groups were evaluated for significance

using analysis of variance. Pearson product-moment correlation coefficients were

94
calculated among individual bout variables using procedures of SAS (1985). Variables of

interest included bout duration, number of chews per bout, meal or drink size, and
interbout intervals. These intervals were the periods of time between the current bout and
the bout of similar activity immediately prior to or following the current bout, whether or
not the interval contained bouts of different activities. Correlation coefficients were
determined for variables summarized by cow to examine relationships across cows
differing in production.

To examine potential sources of variation among feeding variables, variance
components were estimated from mean square estimates for each behavior variable using

the general linear models procedure of SAS (1985). Three linear models were examined:

ij=ll+Cj+Dk® [1]
Yuk = 11 + Pi + Cj(i) + Dk(j) [2]
ij = it + leuk + Cj + 131:0) [3]

where Y was the variable of interest, assumed normally distributed with mean it and
variance (:2, Cj = the random effect of cow (j = 1 to 12), Dkg) = the random effect of day
nested within cow (k = 1 to 10), P; = the fixed effect of parity (i = 1, primiparous to 2,
multiparous), Cjfi) the random effect of cow nested within parity, and 01 = the slope of
regression on a defined covariate (x 1 jg). For each dependent variable, the covariate was
the mean of the same variable over the 5-d period immediately prior to the 10-d collection
period. Cows were denoted as experimental units, and effect of cow served as
experimental error and effect of day as sampling error.

The minimum number of experimental units required to detect significant contrast
differences between two potential treatments in future experiments was estimand for each
variable for four different experimental designs using procedures of Gill (1978). Eighty
percent probability (Type 11 error 5 20%) of dewcting contrast differences of 10% of
variable means with Type I error of 5% was desirable. Future experiments were assumed

to have four treatments, but results should be similar if studies have three or five

95
treatments. Experiment sizes were determined for completely random, randomized

complete block, covariate (adjustment for covariance), and Latin square crossover designs

(Gill, 1978). A variable 6 was iteratively calculated as

 

(p = (c-d / “(AZ/(68k + 36%)) and compared with sample size charts for the
power of an F test (Gill, 1978), where c = suggested number of cows per treatment, (1 =
suggested number of days in collection period, A = detectable contrast difference, (3%/c =
variance component estimate for day(cow), and 6% = estimated experimental error. For
example, the value of q) for milk production in a randomized design was 1.996 when c =
64, d = S, A = 3.3 kg/d, (3%/c = 2.58, and 63 = 43.2. This value was sufficiently large
to obtain Type II error less than 20%.

A different estimate for 6% was used for each experimental design. For the

random design, variance due to cow from model [1] was used for 6%. For the block

design, in which parity was the nuisance factor, variance due to cow(parity) from model

[2] was used. For the covariate design, variance due to cowlcovariate from model [3] was

the estimate for 6'2 . For the Latin square design, 6% was assumed to be equal to zero, so

only variance due to days was used. This represents the minimum error possible with
Latin square designs and assumes no interaction among cows, periods, treatments, or
squares.
RESULTS

A summary of daily milk production and feeding behavior variables is presented in
Table l for primiparous, multiparous, and all cows combined. Milk production averaged
33.1 kg/d per cow. Mean milk fat and protein concentrations were 3.3 and 3.1%,
respectively. Feed intake occurred during 11.0 eating bouts each day that lasted 28.8 min
and averaged 2.2 kg of DM in size. Average daily DMI was 22.8 kg. Mean daily DMI
does not equal the product of the mean number of eating bouts per day and meal size

because all variables were determined on a daily basis before summary statistics were

96

Table 1. Summary statistics for milk production and feeding behavior of six primi- and

six multiparous lactating Holstein cows measured for 10 c1.1

 

 

_Eliruirlar.0us___Mullirlarolls All cows
imam: Mean CV Mean CV Mean—Qt
(%) (%) (%)

Milk production,a kg/d 28.7 15.5 37.5 13.7 33.1 19.7
DMI,a kg/d 20.0 13.6 24.8 11.3 22.8 16.1
NDF intake,a kg/d 6.2 13.8 7.6 11.4 7.0 16.1
Meal size, kg

DM 1.8 17.0 2.5 29.8 2.2 30.6

NDF .56 17.3 .75 29.7 .67 30.5
Eating bouts, /d 11.3 17.3 10.8 25.4 11.0 22.1
Eating bout length, min 25.9 22.2 31.1 33.4 28.8 31.3
Eating time

min/d 284 16.5 314 16.8 301 17.3

min/kg DM 15.9 23.7 13.6 14.1 14.6 21.0

min/kg NDF 51.1 22.4 44.3 14.4 47.2 20.1
Eating chews, /d 18,276 22.8 19,256 16.9 18,832 19.6
Eating chew rate,

chews/min 62.7 7.4 60.8 8.7 61.6 8.3
Runrinating bouts, /d 15.4 17.5 12.9 13.3 14.0 17.8
Ruminating bout length,a min 29.7 15.9 36.0 19.9 33.3 20.9
Ruminating time

min/d 453 18.3 460 14.8 457 16.3

min/kg DM 22.9 21.1 18.7 18.9 20.5 22.5

ming NDF 74.1 20.8 60.9 19.0 66.6 22.3
Ruminating chews, /d 29,645 21.5 28,946 17.6 29,248 19.3
Ruminating chew rate,

chews/min 64.4 10.7 61.8 7.2 62.9 9.1
Total chewing time

min/d 738 13.9 774 11.4 758 12.6

ming DMa 37.2 15.9 31.4 13.6 33.9 17.1

min/kg NDFa 120.7 15.6 102.0 13.8 110.1 17.0
Total chews, [(1 47,921 19.1 48,201 14.3 48,080 16.4
Water intake,a Ud 63.2 19.5 89.5 15.0 77.6 23.8
Drinking bout size, L 5.4 33.2 7.2 43.8 6.4 43.1
Drinking bouts, /d 13.0 35.9 14.9 41.9 14.0 40.2
Drinking time, min/d 17.7 37.4 19.1 20.0 18.5 28.7
Drinking rate, Umin 3.9 31.0 4.6 11.0 4.3 22.4

 

 

 

lStatistics calculated from daily total or daily mean for each cow and day combination.

aMeans differ between parities (P < .05).

97
calculated. The 12 h of greatest intake, compared to any other consecutive 12 h period,

occurred between 1200 and 2400 h (59.2% of total DMI). Cows ruminated during 14.0
bouts/d lasting 33.3 min per bout. Rumination data do not include activity occurring when
cows were away from stalls during milking.

Cows spent 301 min/d eating and 457 min/d ruminating. Distribution of chewing
activity within a day is presented in Figure 1. Large meals were consumed after milking
and allocation of fresh feed. When not eating, cows spent considerable time ruminating.
Time spent chewing was distributed evenly throughout the day; 3 to 6% of chewing
occurred within each hour of the day when not milking. Chewing rates during eating and
ruminating were similar, averaging 61.6 and 62.9 chews/min, respectively (T able 1).
Total time spent chewing was 758 min/d during which 48,080 chews occurred. Daily
water consumption was 77.6 L/cow and required 18.5 min to drink. Cows averaged 14.0
drinking bouts/d of 6.4 L each. Parity means differed significantly (P < .05) for milk
production, DMI, NDF intake, ruminating bout length, total time spent chewing per unit
of DM and NDF, and water intake (Table 1). Coefficients of variation ranged from 8.3 to
43.1% (mean = 21.5%) across variables from combined parity data. Within parity CV
tended to be slightly lower than for all data combined, although average CV across all
variables were similar for both parities (prinriparous = 19.9%; multiparous = 18.6%).

Meal size, eating bout length, and number of eating chews were correlated (r 2
.90) among individual eating bouts (Table 2). Ruminating bout length and number of
chews (r = .97) and drinking bout length and water intake (r = .85) also were highly
correlated when summarized by their respective bouts. Eating and drinking activities were
correlated significantly with their respective pre- and post-interdrinking intervals,
however, ruminating activities were not highly related to pre- and post-rumination
intervals. Pre- and post-interbout intervals were not related for any activity.

Means averaged across the 10 d of measurement for each of the 12 cows were

used to determine relationships among feeding behavior variables summarized by cow

98

 

V

 

 

 

 

 

 

 

 

 

VIII/é.
7
VII/gig
rill/Ill;
r/l/l/l/lll/lfl/d
r.
11v
.m. r/l/A
m9 Wr 7
mm rill/Ill; -l
um 7
mm. Vill/l/l/l/l/Al
ma gill/lg
mm VI/ll/I/I/ll/é
r/l/l/Afl/l/A
Zl V/I/llllllallll/A
V/I/l/I/llle (
.111
m. gill/é
m VII/é;
A
al.,.o..r.n.4.mo.n.,_.....

 

eE: 9.35:0 >=uu .32 .0 mus—.88....

0

123456 789101112131415161718192021222324

Hour of Day

Figure 1. Distribution of chewing time within a day expressed as the mean of 12 cow

means. Feed was offered when cows were away from stalls for milking.

99

Table 2. Pearson correlation coefficients among feeding behavior variables for individual
eating, ruminating and drinking bouts.1

 

whet
Variable 12 34 67 8 101112

1 Meal size, kg of DM

2 Eating bout length, min .91

3 Eating chews, /bout .90 .98

4 Pre-intermeal interval,2 min .38 .38 .36

5 Post-intermeal interval,3 min .14 .15 .14 .03

6 Ruminating bout length, min

7 Ruminating chews, /bout .97

8 Pre-interrumination interval,2 min .12 .10

9 Post-interrumination interval,3 min .05 .06 .005

10 Drinking bout size, L

11 Drinking bout length, min .85

12 Pre-interdrinking interval?- min .48 .41
13 Post-interdrinking interval,3 min .34 .31 .03

 

 

1Total number of bouts: eating, 969; ruminating, 1232; and drinking, 1385.
Significant correlation (P < .01) for r > .07.

2The time interval between the beginning of a bout and the end of its preceding bout.

3The time interval between the end of a bout and the beginning of its succeeding bout.

100
(Table 3). Milk production, BW, daily DMI, and daily water intake were highly correlated

with one another (r 2 .89, 95% confidence interval: .65 S r S .97) and were correlated
to other variables in a similar manner. Production was correlated positively with meal size
and ruminating bout length; negatively with eating, ruminating, and total time spent
chewing per unit of DM consumed; and uncorrelated (r < .10) with daily number of eating
bouts. As length of eating bouts increased, meal sizes tended to increase (r = .72), and
number of bouts per day decreased (r = -.72). Similarly, as ruminating bout lengths
increased, the number of ruminating bouts per day decreased (r = -.80). Consequently,
daily eating and ruminating durations were not related significantly to meal size, daily
intake, or milk production, although their sum, total chewing duration, was related
positively to intake and production. Drinking bout size and number of drinking bouts
were not related to daily water consumption but were correlated negatively to each other (r
= -.80).

Correlations within parity were calculated to examine relationships among milk
production and eating variables unconfounded by the effect of parity (Table 4).
Relationships among milk production, BW, DMI, and water intake were similar to
combined data correlations. For primiparous cows, production was not related to meal
size but was related negatively (r = -.7 8) to length of eating bouts and time spent eating per
unit of DMI. For multiparous cows, production was related positively to meal size (r =
.78) and length of eating bouts (r = .75) but unrelated to time spent eating per unit of DMI.
Relationships within parity among other variables (e.g., rumination, data not shown) were
similar to relationships described for combined data (Table 3).

Sources of variation in feeding variables are presented as variance component
estimates in Table 5. Variation across cows without considering parity was greater than
variation across days for 19 of the 28 variables. Including effect of parity in the model
decreased variance due to cow for 19 variables; however, effect of cow(parity) was still

significantly greater than zero. The other 9 variables had increased variance due to

101

.2 as J a: .25 a: .2 2.: .29 no 5555 G: .255 a: .2 G: in .8 M258 as .255 a: .55 2 s
.2 a: .2 av .35 ..o 5255 as .255 E .55 Ge .2 6 .29 .6 we. 25 .255 6 as. S .281: ”ozone, 88 .2 2558

58.. N 28280 a: n.:.. N 26:89 my ".83 28:80

S ”cote—oboe “Smog—EG—

 

 

8.- 3. 8.- 8.- 8.- S. S. 8. 8.- S. a..- 8.- $.- 8: S.- E. :. 2. :. 25855558
2.. 8. 8.- 8. 8. a..- 2.- 8. 2...- oe. 3.- 8. 8. 8. 3. no. 8. 8. 85253958er
9.. 8.- we. 3.. 2..- e4. 8. 8.- 8. $.- 84. 8. t. K. 8. 8. 8. 8.55835
8.- S. we. a... 2. S. $.- 5. we. 8. 2. 3.. 8. 3. 8. 8. $2638:

2.- 8.- 8. 2.. N...- 8. 8.- 8. S.- 8. a..- 8.- 8.- 5.. 8: 958 88856382

8. 8.- 8. 8. 8.- S. 2.- 8. 8m. 8.- 8. 8. E. 8. 953263.82

2.- 3. E. 8.- :. 8.- 2. 8.- 8. 8. 3. we. 8. 2365688553.:

8. 2..- me. a... 8. 3..- 8. s..- 8.- 8.- 8.- 8.. 388832555 2

R. S. 8.- S.- eo: S. 8.- we. 8. 2.. 2.. 255525585

8.- 8.... 8.- me. 2.. S. we. 8. 8. 8. 8828888553.:

E.- :.- 8.- 2..- 2; S..- R: 8.- 8.- 2:8 m§e5§m 2

me. 8. 8. 8. 3. 8. 8. 2. aaoeomeeeme

8. 8. z.- :..- $.- 8: 8.- $885588 5

8. :. 8. me. 8. 8. Samoan:

8.- 2. 8. S.- 8. 55828358 e

8.- 8. 8. 3. answeeom m

8. on. 8. 85282..

8. 8. gen

8. 3mm

cones—coauaz _

284-522322525”Newemua 85858

 

 

383%

 

3:088:82: .«o e S 828 homage £808 38 Q 8.. 83.5.5? 83.2.3 wince“ mecca $58508 cone—oboe com-acn— .m 293.

102

Table 4. Pearson correlation coefficients among feeding behavior variables for six
primiparous and six multiparous cow means averaged across 10 d of measurement.l

 

 

 

 

31W

Variable 1 2 3 4 5 6 7 8 9
lMilkproduction,kg/d .97 .96 .16 .55 -.78 -.55 -.87 .97
2 BW,kg .81 .99 .13 .59 -.77 -.43 -.87 .95
3DMI,kg/d .68 .58 .03 .67 -.78 -.33 -.78 .92
4Mealsize,kgofDM .78 .33 .32 -.72 .40 -.25 -.37 .11
5 Eatingbouts,/d -.36 -.03 .34 -.78 -.86 -.08 -.28 .58
6 Eatingboutlength,min .75 .30 .27 .89 -.69 .56 .69 -.79
7 Eatingtime,min/d .56 .41 .63 .23 .20 .54 .76 -.52
8 Eatingtime,min/kgofDM .16 .17 -.03 -.07 .08 .35 .71 -.87
9Waterintake,L/d .81 .49 .89 .64 -.04 .65 .73 .09

 

1Primiparous cow data above diagonal; multiparous cow data below diagonal.
Significant correlation: (P < .10) for r 2 l.73l; (P < .05) for r 2 l.8ll; (P < .01) for r _>.
|.92|.

103
Table 5. Variance component estimates for feeding variables using three linear models.1

 

 

 

 

 

Medel
_LL1_ _[ZJ__ [31 41.2.3.1.

Mariam: Cow COWQWW
Milk production, kg/d 43.2M 24.3" 2.01“ 2.58
DMI, kg/d 10.2** 4.4“ .56 3.98
NDF intake, kg/d .94“ .40" .059* .41
Meal size, kg

DM .25** .16" -.023 .21

NDF .024“ .015“ -.002 .02
Eating bouts, /d 2.86** 3.16M .058 3.32
Eating bout length, min 48.5" 46.4“ 22.4” 36.7
Eating time

min/d 2133“ 2123" 372.5" 770.2

min/kg DM 5.53“ 4.69“ 2.17 ** 4.28

min/kg NDF 58.0“ 51.0“ 20.1M 36.8
Eating chews (x 105), /d 83.2“ 90.4“ 26.0" 59.4
Eating chew rate,

chews/min 12.1M 12.5" 3.50* 14.7
Ruminating bouts, /d 3.85" 2.51" 1.83“ 2.70
Ruminating bout length, min 24.2“ 15.2"”I 11.9** 26.1
Ruminating time

min/d 495 608 161 5079

ming DM 10.3M 6.49“ 1.28 11.8

min/kg NDF 101.3" 61.7“ 11.5 127.0
Ruminating chews (x 10's), /d -9.28 -7.2 ~10.1 327.4
Ruminating chew rate,

chews/min 18.8“ 19. 1** 3.03* 15.6
Total chewing time

min/d 2176“ 2129M 1198* 7198

min/kg DM 20.7 ** 12.9* 4.13“ 14.8

ming NDF 207.3“ 128" 37.7* 159
Total chews (x 10'5), Id 81.4* 98.0* 81.8* 549
Water intake, L/d 304.6" 129" 8.92* 60.0
Drinking bout size, L 7.02" 6.89” .44" 1.13
Drinking bouts, /d 27 .2" 29.11“ 3.26" 6.60
Drinking time, min/d 24.5" 26.7“ 5.07" 5.54
Drinking rate, um .82“ .77M .104M .16

 

1Model [1]: Y = cow + day(cow); Model [2]: Y = parity + cow(parity) + day(cow);
Model [3]: Y = covariate + cowlcovariate + day(cow).

* Variance component significantly > 0 (P < .05).

** Variance component significantly > 0 (P < .01).

104
cow(parity) because degrees of freedom were used for parity. Compared with model [1],

including the covariate effect decreased the residual effect of cow for all 28 variables; for 6
variables cowlcovariate was not significantly greater that zero (P > .05).

Contrast differences (10% of variable mean) between two potential treatments and
the total number of cows required to detect these differences under four experimental
designs are listed in Tables 6 and 7. Cow numbers declined as variables were measured
for additional days. Greater declines were achieved for variables with large day(cow)
variance compared with cow variance (e. g., ruminating duration). For random designs,
differences in eating chew rate (chews/min) required the fewest animal units (28 cows),
and drinking bout size required the most (1132 cows) when measured for 5 d (data not
shown). Cow number decreased for some variables when cows were blocked by parity.
Nonedteless, requirements still exceeded 32 cows for all variables (Table 6). A minimum
of 232 cows was required to detect differences for all variables when covariate effects
were included in the model (Table 7). When measured in a Latin square design, all
variable differences were detectable using 52 cows measured for l d or 12 cows measured
for 5 (1.

DISCUSSION

Feeding and drinking behavior means obtained in this experiment were similar to
those observed by others for lactating cows receiving similar diets (Andersson et al.,
1984; Beauchemin and Buchanan-Smith, 1989; Colenbrander et al., 1991). Some
variation in results among studies may be because of different production levels, variation
in definitions used to describe behavior, chemical composition and physical form of the
diet, and differences in management and environment. Beauchemin and Buchanan-Smith
(1989) offered three levels of NDF in alfalfa silage-based diets to multiparous, mid-
lactation cows. Cows receiving the 30% NDF diet spent 237 min/d eating and 413 min/d
ruminating compared with 314 and 460 min for eating and ruminating, respectively, for
multiparous cows in our study. Milk production (20 kg/d) and DMI (18 kg/d) also were

105

Table 6. Estimated cow numbers required for 80% probability of detecting significant
contrast differences between two treatment means for feeding variables for random and

 

 

 

 

 

 

complete block designs.1
13W
Contrast2 CRD3 RCB4
difference 1d 3d 5d 1d 3d 5d
------------ (no. of cows required)------------
Milk production, kg/d 3.3 268 260 256 160 148 148
DMI, kg/d 2.3 176 144 140 104 76 68
(Meal size, kg of DM .22 628 440 400 504 312 276
Eating bouts, /d 1.1 328 212 188 344 228 204
Eating time, min/d 30 208 172 164 208 168 164
Ruminating bouts, /d 1.4 216 156 144 172 112 100
Ruminating time, min/d 46 172 68 52 176 76 56 g
Total chewing time, min/d 76 108 56 44 104 56 44
Water intake, L/d 7. 8 388 348 340 204 160 152
Drinking bouts, /d 1.4 1104 960 932 1168 1024 996

 

1Assumes probability of Type I error = .05; treatment number = 4.
2Selected contrast difference equals 10% of variable mean.

3CRD = Completely random design; estimated 62 = Variation due to cow (Model

[1])-

4RCB = Randomized complete block design with parity as blocking factor; estimated

a: = Variation due to cow(parity) (Model [2]).

106

Table 7. Estimated cow numbers required for 80% probability of detecting significant
contrast differences between two treatment means for feeding variables for covariate and

Latin square designs.1

 

 

 

Dmmeasurcd
4

2 3
Contrast COV LS
difference 1d 3 d 5 d 1 d 3 d 5 d

------------ (no. of cows required)------------

Milk production, kg/d 3.3 32 20 20 8 4 4
DMI, kg/d 2.3 60 28 20 12 8 4
Meal size, kg of DM .22 256 64 28 52 20 12
Eating bouts, /d 1.1 180 64 40 32 12 8
Eating bout length, min 2.9 456 268 232 52 20 12
Eating time

min/d 30 84 48 40 12 8 4

min/kgofDM 1.5 196 112 92 24 12 8
Eating chews, /d 1880 156 84 72 24 8 8
Ruminating bouts, /d 1.4 148 92 80 20 8 8
Ruminating bout length, min 3.3 220 120 100 28 12 8
Ruminating time

min/d 46 164 60 40 32 12 8

min/kg of DM 2.1 200 80 60 36 12 8
Ruminating chews, /d 2920 240 76 44 44 16 12
Total chewing time

min/d 76 96 44 32 16 8 8

min/kg of DM 3.4 108 56 44 16 8 8
Total chews, /d 4800 176 76 56 28 12 8
Water intake, Ud 7.8 76 36 28 16 8 8
Drinking bouts, Id 1.4 324 180 152 40 16 12
Drinking rate, Umin .43 96 60 52 12 8 4

 

lAssumes probability of Type I error = .05; treatment number = 4.
2Selected contrast difference equals 10% of variable mean.

300v = Covariate design; estimated 63, = Variation due to cowlcovariate (Model [3]).
4L8 = Latin square crossover design with four cows and periods per square; estimated
6% = Variation due to day(cow); day(cow) represents the smallest estimate of LS 62 .

107
lower for cows in their study. After adjusting for differences in intake, time spent

chewing was more similar between these two studies.

Daily number and length of eating, drinking, and ruminating bouts depend greatly
on the chosen size of their MIBI, which is the minimum time interval between two similar
activities required for considering them separate events (Metz, 1975). In studies
measuring drinking (Andersson et al., 1984, MIBI = 30 s) and chewing activity
(Beauchemin and Buchanan-Smith, 1989, MIBI = 4 min; Harb and Campling, 1985,
MIBI = 2 min), shorter MIBI resulted in more daily bouts of shorter duration, compared
with those in our study. Many studies have been conducted to investigate diet and
management factors that affect feeding and chewing behavior. Dietary fiber concentrations
(Beauchemin and Buchanan-Smith, 1989; McLeod and Smith, 1989), forage particle size ‘
(Colenbrander et al., 1991; Norgaard, 1989), feeding time (Gordon, 1958), and housing
type (Colenbrander et al., 1991; Harb et al., 1985) have been examined.

Certain feeding variables assume many values within a day (e.g., meal size) that
may vary greatly. Consequently, daily means of these variables may be of little utility.
Experimental treatments may not affect the meanlvalues of these variables but rather alter
their temporal distribution within a day. To detect distributional differences due to
treatments, distributions must be similar across cows receiving the same treatment. In the
present study, intake occurred throughout the entire day, however, periods of greater and
lesser intake occurred that were consistent across cows (Figure 2). Distributional
consistency should allow detection of differences in future experiments when treatments
are applied.

Correlations within eating bouts between eating bout length and number of chews,
and within ruminating bouts between ruminating bout length and number of chews were
high and positive (r = .97, Table 2), indicating that either bout duration or number of
chews could be used to describe chewing activity. Correlations between bout duration and

number of chews may be smaller, however, when various treatments are applied. Eating

108

 

 

 

 

 

 

E 15‘

g . Milking l Milkian

"5 12'

'e .

.79.

2 9"

'5

g . h

5 5‘ “

c .

o

2 J t

8 3j
o'rr ‘I"T"l"tj1"l"lj
0 3 6 9 12 15 18 21 24

Hour of Day

Figure 2. Distribution of DMI within a day expressed as the mean of 12 cow means.

Bars represent the standard error of each mean. Feed was offered when cows were away
from stalls for milking.

109
activities were correlated more positively (P < .01) with the time interval between bout

and previous bout (r = .37) than with the time interval between bout and succeeding bout
(r = .14). For example, as time between meals increased, size and length of the meal that
followed this interval tended to increase. The higher correlation between meal size and
length of its pro-intermeal interval agrees with data of Metz (1975). These correlations
may exhibit diurnal variation (Metz, 1975) and vary among physiological states. Savory
(197 9) stated that under these conditions a satiety mechanism is controlling intake, in
contrast to a hunger mechanism initiating intake when meal size is correlated highly with
post-intermeal intervals. Such relationships suggest that cows in this study stopped eating
once a certain level of repletion was reached and may indicate that rumen fill was
determining meal size (Metz, 1975).

Ranges in mean daily milk production (23 to 44 kg, SD = 6.6) and DMI (15 to 27
kg, SD = 3.9) across the 12 cows were sufficiently large to compare cows varying in
production. Cows were at similar DIM, so production differences among cows were not
due to stage of lactation. Multiparous cows produced 8.8 kg/d of milk more than
primiparous cows and had greater DMI and water intakes. Not all of the multiparous
cows, however, produced and consumed more than the primiparous cows. High milk
production was accompanied by high DM and water intakes both within and across
parities. Smaller correlations among these variables were noted by others (Holter and
Urban, 1992; Murphy et al., 1983) perhaps because of less variation across cows or their
use of more diverse experimental conditions. Increases in daily DMI may occur because
of increases in number of meals consumed per day or increases in average meal size.
Correlations in Table 3 indicate that across all cows, meal size was related more to DMI (r
= .58) than was meal number (r = .35), a result also found by others (Metz, 1975;
Vasilatos and Wangsness, 1980).

Within parity (Table 4), DMI was related equally to meal size and number of meals

for multiparous cows and more related to number of meals than meal size for primiparous

l 10
cows. Multiparous cows ate larger meals by eating longer during each meal (r = .89).

High producing primiparous cows spent less time eating per unit of DMI (r = -.87) than
other primiparous cows, however, eating bout lengths were shorter so meal size was
unrelated to production and intake. Because rumen capacity tends to be smaller for
primiparous cows (Bines, 1976), their meal size may have been limited by rumen fill and
is supported by cow behavior results. These results suggest that different mechanisms
may be controlling individual meals and total daily DMI between cows of different parity,
or, perhaps, for cows differing in rumen capacity, BW, or growth requirements. Only six
observations within parity group were used so these relationships must be viewed with
caution. Other workers have reported that heifers with greater intake spent less time eating
per unit of intake (Deswysen et al., 1987 b), and dry cows exhibited no relationship
between eating rate and daily intake (Harb and Campling, 1985), which agrees with
results found in our study. Results of Bae et al. (1983) indicated that time spent eating per
unit intake of cell wall constituent was correlated negatively with BW in nonlactating cows
and steers. Because BW and milk production were correlated highly in our study, it
cannot be determined whether increased production across uniform BW cows would result
in decreased eating duration per unit of intake.

Relationships for rumination and total chewing were similar within parity groups;
therefore, only results of combined data will be discussed. Higher producing cows tended
to have fewer ruminating bouts per day, but each bout was significantly longer (Table 3),
resulting in only a tendency (r = .46) for these cows to ruminate longer each day. Total
time spent chewing per day was associated positively with milk production (I = .59).
Increases in chewing duration, however, were not proportional to increases in DMI
because ruminating and total chewing durations per unit of intake decreased as production
and intake increased. This is in agreement with reports by some workers (Deswysen et
al., 1987b; Harb and Campling, 1985) but in disagreement with others (Bae et al., 1981;
Beauchemin and Buchanan-Smith, 1989), who found similar chewing durations per unit

 

1 11
of fiber intake. The latter studies may differ in behavior because of restricted feeding

conditions. The amount of time spent chewing per unit of intake has been implicated as a
measure of mastication efficiency (Deswysen et al., 1987b; Welch, 1982). Increased
efficiency may be due to a shorter interboli time, a greater number of chews per unit time,
a lower proportion of pseudo-rumination, or more efficient regurgitation of long particles
(DeBoever et al., 1990). Efficiency also may be increased because of larger mouths,
larger tooth surfaces, stronger jaw muscles, or improved articulation of the jaw.
Elucidating why high producing cows process more feed during a given length of chewing
time should remain an active area of nutrition research. Whether similar relationships exist
at all stages of lactation remains in question and deserves study.

Variance component estimates were calculated to examine the relative importance
of independent sources of variation. Feeding behavior measurement is difficult and
expensive; therefore, knowledge about potential variation is useful for planning
experiments that minimize experimental error and replication. Differences between cows
were the greatest source of variation for most variables (Table 5), suggesting that reducing
experimental unit heterogeneity would be effective. Coefficients of variation across cows
ranged from 5 to 41%, and most variables ranged between 15 and 20%. Norgaard (1989)
and Harb and Campling (1985) observed similar variation across cows, but Vasilatos and
Wangsness (1980) detected larger variation (e.g., 30 to 40%), perhaps because of their
use of cows earlier in lactation. Blocking cows by parity in the present study did little to
reduce replicate heterogeneity. An alternative blocking factor, such as pretrial milk
production or intake, may be more effective. A pretrial covariate period reduced across-
cow variation more significantly. However, this period was extensive (5 d) and
immediately preceded the experimental collection period. The effect of a covariate period
probably is maximal under these conditions and likely will be smaller with a shorter,

earlier, or later period of measurement.

1 12
Days were considered as random samples within the main effect of cow; therefore,

the effect of day could not be tested for significance. Variation across days within cow
was substantial, especially for traits associated with rumination. Measuring behavior for
more than 1 d will decrease experimental error because a more accurate measure of the
replicate's true value is obtained. Only a limited decrease in error is possible by increased
sampling, and greatest gains occur with initial increases in sampling (Bemdtson, 1991).
Variables with large across day variance compared with across cow variance have the most
to gain from more days of measurement.

Estimates of experimental error facilitate calculation of replicates required for
adequate power and sensitivity in future studies measuring feeding behavior. A maximum
Type [I error rate of 20% commonly is used for replicate estimation (Bemdtson, 1991;
Gill, 1969). Biological significance of a 10% difference from a control mean probably
depends on the variable of interest. Variables with the largest variance components used
as error estimates required the most replicates to achieve desired power (Tables 6 and 7).
Gill (1969) determined sample sizes required for detection of differences in milk
production for various designs using more than 25,000 cow records. Replicates per
treatment were slightly greater than those calculated for milk production in the current
study because of different assumptions used in calculating experimental errors. Similarity
between the two studies suggests that our replicates were sufficiently diverse to represent a
more general cow population.

Many studies have monitored fwdin g behavior for only a single day (Colenbrander
et al., 1991; McLeod and Smith, 1989). Based on our data, l—d studies require a
minimum of 52 cows to detect 10% treatment differences for all feeding variables using a
Latin square design. Because experiments rarely utilize this many replicates, they will
have substantially less than 80% power. If differences are at least 20%, number of cows
required decreases to 20. At contrast differences of 40% of mean values, 8 cows are

required in Latin squares. Consequently, only when treatment differences are large will

113
statistical detection likely be achieved with l-d studies. Partial day measurements

extrapolated to complete day means (Norgaard, 1989) will require even larger differences
if small numbers of cows are used.

Replication required for completely random, block, and covariate designs exceeded
that typically available to researchers. Only with Latin square designs was sufficient cow
variation removed to obtain 80% power and 10% sensitivity while using a reasonable
number of cows. Twelve cows measured for 5 d in a Latin square design was the most
effective design for this feeding behavior study where mid-lactation cows received
medium quality diets.

CONCLUSIONS

Milk production was correlated positively with DM and water intakes both within
and across parities. For multiparous cows, production was related positively to meal size
and length of eating bouts, and unrelated to meal number and eating rate. For primiparous
cows, production tended to be related positively to meal number and eating rate, and
unrelated to meal size. These results suggest that different mechanisms may be controlling
individual meals and total daily DMI between cows of different parity, rumen capacity, or
BW. Milk production was correlated negatively with ruminating and total time spent
chewing per unit of intake, suggesting that differences may exist among cows for chewing
efficiency. Future experiments that involve fading behavior measurements with lactating
cows may attain sufficient statistical power and sensitivity when they use 12 cows

measmed for 5 d in a Latin square design.

CHAPTER 4

Intake Limitations, Feeding Behavior, and Rumen Function of Cows

Challenged with Rumen Fill from Dietary Fiber or Inert Bulk

ABSTRACT

Twelve multiparous, rumen cannulated cows (17 DIM) were fed 25 or 35% NDF
rations with or without added rumen inert bulk as water-filled plastic containers (500 ml
each) within 3, 4 x 4 Latin squares (21 (1 periods). Added bulk equaled 25% of pretrial
rumen volume. Objectives were to challenge early lactation cows with rumen fill in the
form of dietary NDF and rumen inert bulk to determine if intake limitations from
ruminal fill vary with NDF content of the diet, or if within-day feeding behavior changes
allow cows to adapt to higher fill. Inert bulk had no effect on DMI for cows fed 25%
NDF but decreased DMI for cows fed 35% NDF (18.7 vs. 16.6 kg/d). Production and
rumen fermentation results support the statement that fiber and inert bulk addition
created an energy deficit compared to the basal treatment. Volume of rumen digesta
plus inert bulk was similar for 35% NDF treatments regardless of inert bulk (mean = 102
L), and suggests that cows receiving high fiber diets have intakes limited by the physical
capacity of the reticulorumen. Changes in feeding behavior were insufficient to
maintain intake under conditions of high rumen fill, however, added fiber and bulk
altered rumination and reticular motility in a similar fashion, demonstrating that NDF
may have rumen-filling properties. Increased passage and rumination may help alleviate
fill. Future study is required to determine the effect of fiber digestibility and animal

characteristics on fill-limited intake.

114

l 15
INTRODUCTION

Maximum feed intake in early lactation cows establishes a foundation for high
milk production throughout the entire lactation. Cows do not consume adequate
amounts of energy during initial stages of lactation to support their energy requirements,
which supports the hypothesis that intake by cows during this period is not limited by
physiological control, but rather, by the physical limitation of insufficient reticulorumen
capacity. Minimizing rumen fill becomes more critical with larger energy deficits to
avoid excess loss of body energy reserves and metabolic disease. Once cows have
reached positive energy balance, physical fill may not be as important a control
mechanism for intake.

Several previous studies have added artificial forms of inert bulk to the rumen of
ruminants, including dairy cows, to examine the effect of reticuloruminal fill on intake
(Campling and Balch, 1961; Carr and Jacobson, 1967; Johnson and Combs, 1991, 1992).
It is a common misconception that if intake is limited by rumen capacity, intake should
decline in proportion to the amount of artificial fill added. Responses to fill, however,
have been variable. Campling and Balch (1961) observed linear decreases in intake (52
g of hay/L of fill) with non-lactating cows after additions of 23 and 34 L of inert bulk.
Johnson and Combs (1991) fed mixed diets to early lactation cows and detected
decreases in intake of 99 and 130 g/L of inert bulk, however these same workers found
no depression in intake with larger cows filled with similar levels of artificial bulk
(Johnson and Combs, 1992). Failure to achieve intake depression may be due to
submaximal volume in the reticulorumen prior to bulk addition because of other
dominating control mechanisms, or increased digesta disappearance from the rumen
after bulk addition (Waybright and Varga, 1991). Because intake is not likely controlled
by only one satiety factor at all times (Forbes, 1988), dilution of physical effects by other

control mechanisms may obscure detection of depressions due to fill.

116
Not all diets produce the same level of rumen fill. Blaxter et al. (1956)

calculated "ballast", or indigestible DM in the rumen using meal patterns, digestibility,
and passage data and found more digestible feeds resulted in less ballast. Feeds that
ferment and pass more quickly through the rumen are believed to be less filling because
they occupy space within the rumen for a shorter period of time (Aitchison et al., 1986).
The more slowly fermented, fibrous fraction of feedstuffs, as measured by NDF, has
been implicated as the component responsible for rumen fill (Mertens, 1987). Diets with
high NDF concentrations have been shown to promote greater rumen fill and less DMI
compared to lower fiber controls (Aitchison etal., 1986; Llamas-lamas and Combs,
1991). Cows receiving higher fiber diets may also be more susceptible to depressions in
intake upon addition of inert bulk to the rumen, illustrating that these diets may be under ‘
greater control by physical regulation of intake.

Under constraints of high rumen fill, cows may adjust their feeding behavior or
rumen activity to compensate for such fill with no decline in daily feed intake. For
example, they may increase the rate of passage of digesta from the rumen by increasing
extent or efficiency of rumination, allow larger particles to leave the rumen, increase
frequency or strength of reticular contractions (Okine et al., 1989), or eat smaller meals
with greater frequency (Baumont et al., 1990) in an attempt to keep the rumen at
maximum capacity as much as possible. Only by measuring within-day activity is it
possible to evaluate the effects of rumen fill on meal characteristics, rumination, and
rumen function. Use of high producing early lactation cows consuming high fiber diets
with large amounts of added inert bulk should increase the likelihood that capacity
effects will be observed.

Objectives of this study are to 1) challenge early lactation cows with rumen fill in
the form of dietary NDF and rumen inert bulk to determine if limitations to intake from
ruminal fill vary with NDF content; 2) measure feeding behavior and rumen activity to

determine if within-day behavior changes allow cows to adapt to higher fill conditions

l 17
and maintain normal daily intakes; and 3) compare fiber and bulk additions to determine

if dietary NDF has rumen filling characteristics.
MATERIALS AND METHODS
Experimental Design and Data Collection

Twelve multiparous Holstein cows in early lactation (17 :l: 6 DIM; mean :1: SD)
were blocked by calving date and assigned to one of three, 4 x 4 Latin squares balanced
for carry-over effects. Cows were ruminally cannulated 30 d prior to calving, were at
least 10 DIM when treatments commenced, and were not bred until treatments
concluded. Treatment periods were 21 d in length, with the last 7 d used for data and
sample collection. A 2 x 2 factorial arrangement of treatments was applied in each
square. Factors tested were concentration of dietary NDF and addition of rumen inert
bulk (RIB). Levels of NDF were 25% (LF) and 35% (HF) of diet DM; levels of RIB
were 0% (+0) and 25% (+B) of pretrial rumen volume. Inert bulk consisted of 500 ml
polyethylene containers filled with water and sealed with locking screw caps. Number
per cow ranged from 36 to 51 containers.

Diet ingredients were alfalfa and corn silages, ground shelled corn, soybean
meal, animal protein, minerals, and vitamins (Table 1). Values represent average
ingredient and nutrient compositions from across the entire experiment. Rations were
formulated to contain either 25 or 35% NDF, 18.5% CP, and provide at least .4 kg/cow
per d of rumen undegraded protein supplement. Alfalfa silage and corn silage were
offered in equal DM amounts within each diet. Differences in NDF content between
diets were achieved by altering forage:concentrate ratio. All concentrate ingredients
contained greater than 87% DM and were combined for each diet prior to the
experiment. Except for mineral concentrations, nutrient composition of the diets was
determined from chemical analysis of individual feed ingredients. Nutrient composition
of forages and concentrates is presented in Table 2. Rations were mixed once daily and

offered for ad libitum intake (10% refusal rate) as a TMR at 0700 and 1900 h when cows

118

Table 1. Injgedient and nutrient composition of 25% NDF, low fiber diet (LF) and 35%
NDF, high fiber diet (HF) offered to 12 cows during the first 14 wk of lactation.

 

 

 

 

 

 

LF HF
Ingredientscmmsition ------ (% of diet DM) -------
Alfalfa silage 22.6 40.1
Corn silage 22.7 40.7
Ground shelled corn 34.1 2.6
44% Soybean meal 16.4 12.9
Protein supplementl 1.6 1.8
Dicalciurn phosphate .8 .9
Calcium carbonate 1.0 .2
Trace mineral supplement .6 .6
Vitamin supplement2 .2 2
W Ii SD3 HF SD

DM, %

52'C oven 57.1 .9 44.9 1.1

Toluene 56.6 .9 44.2 1.0
NEL,4 Meal/kg of DM 1.76 1.61

(% of diet DM)

(III 92.0 .6 91.0 .2
NDF 25.7 .5 35.2 .9
ADF 14.0 .4 21.3 .6
Lignin 2.4 .2 3.7 .2
Indigestible NDF5 8.3 .3 13.8 .4
CP 18.2 .4 18.9 .3

RUP,4o5 % of diet CP 36.6 ND7 32.1 ND
Ca 1.3 ND 1.4 ND
P .47 ND .54 ND
Mg .24 ND .25 ND

 

1Contains 55% meat and bone meal, 25% blood meal, 10% fish meal, and 10%
feather meal.

2Contains 1733, 512, and 7.3 kIU/kg of DM of vitamins A, D, and B, respectively.

3Represents variation in chemical composition of samples composited weekly.

4Calculated from NRC (1989).

5Neutral detergent fiber residue after 120 h in vitro fermentation with buffered rumen
fluid.

6RUP = rumen undegraded protein.

7ND = not determined.

119

Table 2. Nutrient composition of forages and concentrates used to formulate low fiber
(LF) and high fiber (HF) diets.

 

 

 

 

 

Alfalfa Com m.
JflaL—silagc L_E HF
DM, %

52°C oven 46.6 35.5 88.1 89.6

Toluene 44.9 35.3

(% of DM)

OM 89.4 96.1 91.7 83.2
NDF 37.4 41.6 14.0 16.9
ADF 27.4 21.8 5.1 7.4
Lignin 6.3 2.3 . .8 1.1
Indigestible NDFl 18.9 14.1 1.5 2.8
CF 20.4 7 .3 21.8 40.5

 

1Neutral detergent fiber residue after 120 h in vitro fermentation.

were parlor-milked. Feed amounts offered and refused were recorded daily. Feed
ingredients and diets were sampled daily (.5 kg), frozen, and composited weekly. Feed
refused was sampled daily (12.5% of orts) during collection weeks and composited by
cow. Milk yield was recorded daily. Milk samples were obtained on d 16 and 17 of
each period for both milkings and individually analyzed for components. Cow BW was
determined at 2100 h (d 19) and 1700 h (d 21) each period and averaged. Cows were
visually scored for body condition (d 19) each period.

Fecal samples were obtained and frozen every 15 h from d 14 to 19 of each
period to represent defecation every 3 h throughout a 24 h day. Sampling commenced at
2000 h on d 14 and continued at 1100 h (d 15), 0200 h and 1700 h (d 16), 0800 h and
2300 h (d 17), 1400 h (d 18), and 0500 h (d 19). Cows were not disturbed to obtain
samples but were allowed to defecate naturally to prevent interruption of natural feeding
behavior that was being monitored concurrently. The eight samples for each cow were
composited on an equivalent wet basis. Fecal composites were split and analyzed for

nutrient composition and particle size.

120
Rumen digesta was manually removed from each cow via the cannula at 2100 h

on d 19 (2 h postfeeding) and 1700 h on d 21 (2 h prefeeding). Inert bulk, if present, was
removed with digesta and quantified to assure complete recovery. A 10% aliquot of the
digesta was separated during evacuation for sampling. Two digesta samples were
obtained. A 600 g sample of the complete digesta was taken for determination of DM
and nutrient composition. A second sample was strained through four layers of
cheesecloth to obtain 50 ml of fluid and combined with 10 ml of 10% (v/v) H2804 for
VFA analysis. Both digesta and fluid samples were frozen immediately (-20°C)
following collection. Total digesta weight and volume were determined. Following the
d 21 evacuation, RIB and rumen digesta were switched among cows to assist in
adaptation to the next treatment. Immediately prior to digesta removal, head space (gas)
volume within the dorsal rumen was estimated at both evacuation times by filling the
space with empty polyethylene containers and quantifying their number.

Feed disappearance, water intake, jaw movements, reticular contractions, and
ruminal pH were measured continuously from d 14 to 19 each period with monitors and
a computer data acquisition system described in Chapter 2. Data were written to files
every 5 s for each cow and variable. Monitors were calibrated each period, placed on the
cows 1 d prior to collection to allow for animal adaptation, and removed during the first
rumen digesta evacuation. Monitors were disconnected twice daily to permit cows to
leave the tie stalls for milking and exercise. Cows averaged 1.8 h/d away from feed and
monitoring equipment during which no behavior was recorded.

Sample and Statistical Analysis

Diets, dietary ingredients, orts, feces, and rumen digesta were analyzed for
nutrient concentrations. Samples were dried in a forced-air oven at 52°C for 72 h,
ground with a Wiley mill (l-mm screen), and analyzed in duplicate for ash, NDF, ADF,
lignin, and indigestible NDF. Ash was determined following sample ignition at 500°C

for 6 h. Analysis of NDF (Goering and Van Soest, 1970) was modified by addition of 4

121
ml of a 5% a-amylase solution (Tennamyl 120L, Novo Nordisk Bioindustrials, Inc.,

Danbury, CT) to each sample, substitution of triethylene glycol for 2-ethoxyethanol, and
omission of decahydronaphthalene and sodium sulfite. Neutral detergent residues were
sequentially analyzed for ADF and acid detergent sulfuric acid lignin (Goering and Van
Soest, 1970). Indigestible NDF was estimated as the NDF content of samples following
in vitro fermentation in buffered rumen media (Goering and Van Soest, 1970) for 120 h
without addition of pepsin. Cellulose content was calculated as ADF minus lignin
content; hemicellulose content was calculated as NDF minus ADF content. Diets,
dietary ingredients, orts, and feces were analyzed for CP (N x 6.25) using a modified
Kjeldahl (Hach) procedure (Watkins et al., 1987). Mineral analysis of diets was
performed using inductively coupled plasma emission spectroscopy (Northeast DHIA,
ltlraca, NY). Sample DM was determined by oven drying (52°C) for all samples and by
toluene distillation (AOAC, 1990) for forages, diets, and rumen digesta. Oven DM was
used in calculations for intake and nutrient digestibility. Concentrations of all nutrients,
except DM, were expressed as a percentage of DM determined from forced-air oven
drying at 105’C.

Milk samples were analyzed for lactose, fat, and protein content with infrared
spectroscopy by Michigan DHIA. Particle size of fecal NDF (oven dried at 105'C) was
determined by wet-sieving 40 g of feces following boiling (1 h) in 300 ml of neutral
detergent with .6 ml of 100% tennamyl 120L. Sieve sizes used were 600, 1200, and
2400 um. Data generated were percentage of total fecal NDF retained on each sieve and
percent < 600 um calculated by difference. Acidified rumen fluid was centrifuged at
13,000 x g for 30 min, decanted, and the supematent centrifuged at 26,000 x g for an
additional 30 min. Concentration of VFA in the supematent was determined using the
procedure of Siegfried et al. (1984) with HPLC. The column used was Aminex HPX-
87H, catalog number 125-0140, 300 x 7.8 mm (Bio-Rad Laboratories, Richmond, CA).

Column temperature was 50'C and solvent was .005 N H2804. Detection was by

122
refractive index (Waters 410, Millipore Corp., Milford, MA). Concentration of major

rumen VFA and their molar proportions were calculated.

Fecal output and apparent total tract digestibility of dietary nutrients were
calculated using indigestible NDF as an internal marker (Cochran et al., 1986).
Fractional rate of NDF passage from the rumen (kp) was calculated by dividing the
rumen pool size of indigestible NDF into the hourly intake of indigestible NDF, based
on the two-pool, first-order model of rumen fiber digestion of Waldo et al. (1972). This
model divides the total rumen fiber pool into a digestible pool and indigestible pool.
Fractional rate of digestible NDF digestion in the rumen (kd) was calculated as the
difference between total rate of escape of digestible NDF from the rumen and its rate of

passage (Allen and Mertens, 1988b):

k d _ (hourly digestible NDF intake)
(rumen pool size of digestible NDF)

Finally, apparent rumen digestibility of NDF as a percent of NDF intake was calculated
as:
Fraction of feed NDF that was digestible x {—L]
(kd + kp)
These calculations assume that rumen NDF pool sizes and fluxes are at steady state and
that the kp of digestible and indigestible NDF are equal. Apparent rumen digestibility of
NDF was also expressed as a percent of total tract NDF digestibility.

Eating, ruminating, and drinking activities, in addition to reticular motility, and
ruminal pH were determined from collected behavior data with previously developed
algorithms (Chapter 2). An IBM 3090 mainframe computer was used for data
interpretation (lntemational Business Machines Corporation, Arrnonk, NY). Feeding
behavior was summarized into variables each day as performed in Chapter 3, and
averaged by cow and period. Variables included number and length of eating bouts,

eating time, meal size, number and length of ruminating bouts, ruminating time, number

123
and size of drinking bouts, frequency of reticular contractions, and summary of rumen

pH, among others. Variables were screened to remove data when noted to be incorrect
as determined by manual observation of graphical displays during data collection.
Across the five activities monitored, 91% of the collected computer data were
considered acceptable.
All data were statistically analyzed using the general linear models procedure of
SAS (1985). The model used was consistent with replicated Latin squares:
Yijkl = 11 + Si + Cm) + Pkg) + T1+ ST“ + Eijkl

where 11 = overall mean,
8; = random effect of square (i = l to 3),
Cj(i) = random effect of cow within square (j = l to 4),
Pkg) = random effect of period within square (k = l to 4),
T1: fixed effect of treatment (1 = 1 to 4),
ST“ = interaction of square and treatment,
and
Eifid = residual, assumed normally distributed.

A reduced model without square x treatment was used when this effect was not
significant (P > .10). Type III sums of squares were used to determine significance of
treatment effects. Separation of treatment means was performed using protected least
significant differences (Snedecor and Cochran, 1980). Preplanned orthogonal contrasts
were used to determine significance of the main treatment effects of fiber level and bulk
level, and their interaction. Model effects were declared significant when P < .05,
unless otherwise specified.
RESULTS

Production and Nutrient Digestibility

The addition of RIB appeared to have little adverse affect on general rumen
function or cow comfort during collection weeks. Some of the containers were

temporarily removed from five cows during various adjustment periods because of

124
difficulties in adaptation to newly added bulk. Within 4 d, 100% of the RIB was

returned to their rumens without further complications. All cows contained all their
added RIB for each collection period.

Milk production and nutrient intake data are presented in Table 3. Cows
averaged 33 kg/d of milk and 20 kg/d of DMI across the experiment. Milk, 4% FCM,
protein, and lactose production were significantly greater for LP compared to HF
treatments (P < .01). Low fiber treatments exceeded HF by 5.7, 3.8, .21, and .31 kg/d
for milk, 4% FCM, protein, and lactose production, respectively. Milk protein and
lactose concentrations were higher (P < .01) for LP compared to HF treatments,
however milk fat concentration was not different among any of the treatments. Cows
receiving LF consumed 5.1 kg/d more DM and 10.7 L/d more free water (P < .01) than
those receiving HF. Addition of RIB tended to decrease milk, 4% FCM, and lactose
production within both fiber levels and significantly decreased milk protein production
and concentration (P < .01). Average milk production was 34.2 kg/d for +0 treatments
compared to 32.1 kg/d for +B treatments. Interactions among levels of fiber and RIB
were present for intake data. Inert bulk had no effect on DM, NDF, or indigestible NDF
intakes for cows fed LF but decreased intakes for cows fed HF by 2.1, .73, and .29 kg/d,
respectively. Cow BW and body condition scores were not different among treatments,
however empty BW (BW - rumen digesta mass) was significantly lower for HF and +B
treatments (P < .01).

Apparent total tract digestibilities of DM, OM, NDF, and lignin were
significantly higher (P < .01) for LP compared to HF treatments (Table 4). Average
DM and NDF digestibilities were 71.0 and 47.9% for LF treatments and 65.4 and 44.6%
for HF treatments, respectively. Digestibility of OM was slightly less for cows receiving
RIB compared to no bulk addition. No differences among treatments were detected for
ADF, hemicellulose, cellulose, or CF digestibilities. Cows fed LF produced less fecal
NDF and indigestible NDF compared to cows fed HF. Addition of RIB lowered fecal

125

Table 3. Milk production and feed intake for 12 cows receiving low fiber (LF) or high
fiber (HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

 

 

Main effectl
Trcannsnt Janificancc.
mm; M+B HF+0 HF+B SE F B FxB
Production, kg/d
Milk 37.03 35.0a 31 .4b 29.2b 1.2 ** 1‘ NS
4% FCM 33.5a 31.8ab 29.9bc 27.9c 1.1 ** NS NS
Fat 1.25 1.19 1.15 1.08 .06 T NS NS
Protein 1.0721 0.96b 0.85c 0.77c .03 ** ** NS
Lactose 1.76a 1.64‘1 1.45b 1.33b .06 ** 1' NS
Milk composition, %
Fat 3.43 3.44 3.68 3.69 .16 NS NS NS
Protein 2.92a 2.75b 2.73b 2.66b .04 ** ** NS
Lactose 4.77a 4.703 4.62b 4.57b .03 ** * NS
Intake, kg/d
DM 22.83 22.7a 18.7b 16.6c .4 ** * *
NDF 5.7 9b 5.7 5b 6.523 5.79b . 15 * * *
Indigestible NDF2 1.88c 1.85c 2.56a 2.27” .06 ** * *
CP 4.14a 4.13a 3.58b 3.18c .08 ** * *
Water intake, L/d
Free water 82.2a 7 8.7ab 74.4b 65.10 2.7 ** NS
Total water3 99.03 95.421 96.9a 84.8b 2.8 * NS
BW, kg 593 598 595 587 3 NS NS 1
Empty Bw,4 kg 5188 5113b 5081’ 493c 3 ** ** NS
Body condition score5 2.21 2.21 2.29 2.08 .06 NS NS NS

 

aJ’ocValues in same row with different superscripts differ (P < .05).
1Main effect levels differ, “P < .01, *P < .05, TP < .10.
2Neutral detergent fiber residue after 120 h in vitro fermentation.
3Total water intake = Free water intake + water from feed intake.
4Empty BW = BW - rumen digesta mass.

5Scale of 1 to 5; l = thin, 5 = obese.

126

Table 4. Apparent total tract digestibility of nutrients, fecal output, fecal composition, and
fecal fiber particle size from 12 cows receiving low fiber (LF) or high fiber (HF) diets

without (+0) or with (+B) added rumen inert bulk.l

 

 

 

 

 

Main effect2
manner“ m
my; M+B HF+0 HF+B SE F B FxB
Nutrient digestibility, %
DM 71.5a 70.6a 65.7b 65.2b .4 ** 1' NS
(1% 72.921 71.6b 66.9c 66.5c .4 ** * NS
NDF 48.02‘ 47.82‘ 45.39 43.9b .8 ** NS NS
ADF 45.9 45.4 46.2 46.7 1.6 NS NS NS
Lignin 12.3a 12.2a «.6b 4.739 3.1 ** NS NS
Hemicellulose 46.4 44.7 45.1 43.6 1.2 NS NS NS
Cellulose 52.6 52.4 55.3 55.0 1.6 NS NS NS
CP 70.2 69.8 70.3 69.1 .6 NS NS NS
Fecal output, kg/d
DM 6.50a 6.69a 6.44a 5.80b . 18 * NS *
NDF 3.01b 3.00b 3.57a 3.255 . 10 ** NS NS
Indigestible NDF3 1.88c 1.85c 2.56a 2.27b .06 ** * *
Fecal composition
DM, % 16.0a 16.2a 14.15 13.8b .2 ** NS NS
NDF, % of DM 46.49 45.1b 55.4a 56.0a 6 ** N 8 NS
Indigestible NDF,
% of DM 29.0b 27.9b 39.7a 39.1a .4 ** * NS
Fecal NDF particle size ---—-(% of total fecal NDF)-----
<600 um 46.8be 45.6c 49.3ab 51.43 1.2 ** NS NS
600-1200 um 18.9ab 19.63 18.4be 18.0c .3 ** NS NS
1200-2400 um 13.8a 14.3a 12.1b 11.89 .3 ** NS NS
>2400 um 20.5 20.4 20.2 18.8 1.4 NS NS NS

 

atbtcValues in same row with different superscripts differ (P < .05).
1Calculated using indigestible NDF as an internal marker.

2Main effect levels differ, ”P < .01, *P < .05, 11’ < .10.

3Neutral detergent fiber residue after 120 h in vitro fermentation.

127
output of DM and indigestible NDF, but only for cows fed HF. As a percentage of total

fecal NDF, fecal NDF particle size was smaller (P < .01) for HP treatments compared to
LF, however no differences were observed between treatments for fecal particles greater
than 2400 pm. A greater percentage of fecal NDF was retained on both 600 and 1200
um screens for LF treatments while more fecal NDF was smaller than 600 um for HF
treatments. Added RIB did not alter particle size of fecal NDF.
Rumen Characteristics

Pre and postfeeding rumen digesta evacuation data, including VFA analyses,
were averaged to obtain values presented in Tables 5 and 6. Digesta volume (Table 5)
was largest with treatment HF+0 (98.8 L) and declined with lower dietary NDF content
and addition of RIB. Volume of digesta plus RIB was similar for both HF treatments
and LF+B (mean = 101.6 L) but was significantly less for treatment LF+0. Total rumen
volume, as measured by the sum of digesta, RIB, and head space volume, was not
different between LF and HF treatments but was significantly greater (P < .01) for
treatments with added RIB (110 L vs. 120 L). Low fiber treatments resulted in digesta
with higher DM and lower NDF contents (P < .01) compared with HF treatments.
Added RIB decreased digesta DM content (P < .01) but did not affect NDF content.

Rumen mass of wet digesta followed a similar pattern to rumen volume, with the
exception that treatment HF+B resulted in greater rumen wet mass compared to its
volume because of its higher moisture content (Table 5). Total rumen mass (wet digesta
+ R1B)was highest for HF+B (94.1 kg), similar for LF+B and HF+0 (86.9 kg), and
lowest for LF+0 (75.7 kg). Rumen DM mass decreased significantly (P < .01) upon
addition of RIB, however no difference was detected between levels of dietary fiber.
Masses of various components of rumen fiber all followed similar trends. Without RIB,
HF increased rumen NDF mass compared to LP (6.0 vs. 7 .0 kg). Addition of RIB
decreased (P < .01) rumen NDF to the same mass (4.6 kg) for both LF and HF

treatments. The greater decline in rumen fiber mass upon addition of RIB for HF

128

Table 5. Rumen digesta characteristics for 12 cows receiving low fiber (LF) or high fiber
(HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

 

Main effectl
Treatment Animate.
Variable LF+0 LF+B HF+0 HF+B SE F B FxB
Rumen volume, L
Digesta 89.8b 78.9c 98.83 82.7c 2.4 * ** NS
Digesta + inert bulk 89.8b 101.1a 98.82‘ 104.921 2.4 * ** NS
Digesta + inert bulk
~1-headspace2 108.0b 117.9a 111.5b 121.121 2.0 NS ** NS
Digesta composition
DM, %
52'C oven 14.3a 12.8b 13.09 10.6c .2 ** ** *
Toluene 12.7a 11 .3a 11.6a 8.7b .5 ** ** NS
NDF, % of DM 55.5b 54.4b 61.7a 61.421 .8 ** NS NS '
Rumen mass, kg
Wet digesta 75.79 65.3c 86.4a 71.9b 2.0 ** ** NS
Wet digesta
+ inert bulk 75.7c 87.4b 86.4b 94.1a 2.0 ** ** NS
DM 10.8a 8.3b 11.3a 7.79 .4 NS ** NS
(1V1 9.5a 7.2b 10.0a 6.79 .3 NS ** NS
NDF 6.09 4.5c 7.0a 4.7c .2 * ** NS
ADF 3.5b 2.6c 4.1a 2.8c .1 ** ** NS
Lignin .80b 60“ 1.01a .69c .03 ** ** T
Hemicellulose 2.69 1.9c 2.9a 1.9c . 1 NS ** ’r
Cellulose 2.69 1.9c 3.1a 2.1° .1 ** ** ’r
Indigestible NDF3 3.39 2.5c 4.2a 2.8c . 1 ** ** 1'
Density, kg/L .84b .83b .88a .87a .01 ** T NS

 

almValues in same row with different superscripts differ (P < .05).
1Main effect levels differ, "P < .01, *P < .05, TP < .10.

2Head space = Volume of gas between digesta and top of the rumen.
3Neutral detergent fiber residue after 120 h in vitro fermentation.

129

Table 6. Concentration and molar proportion of VFA in rumen fluid from 12 cows

recentgi‘rlrlgK low fiber (LF) or high fiber (HF) diets without (+0) or with (+B) added rumen
rnert .

 

 

 

 

 

 

Main effectl
Treannent 4%
m1: LE+Q LF+B HF+0 HF+B SE F B FxB
VFA concentrations (mmol/L)
Acetate 77.6a 73.8a 77.7a 68.6b 1.4 T ** T
Propionate 30.9a 26.4b 24.2b 19.5c 1 . 1 ** ** N S
Butyrate 13.0a 12.5a 12.9a 11.0b .3 ** ** *
Valerate 2.9a 2.5bc 2.7ab 2.3° . 1 T ** NS
Branched-chain2 5.8a 5.2b 5.4b 5.2b . 1 NS *"‘ 1‘
Total 130.2a 120.59 1229') 106.6° 2 .2 ** ** NS
VFA molar proportions ----------(% of total VFA)-----------
Acetate 59.8c 61.5b 63.2ab 64.4a .6 ** * NS
Propionate 23.53 21 .6b 19.7bc 18.29 .6 ** * NS
Butyrate 10.0 10.5 10.5 10.3 .2 NS NS NS
Valerate 2.2 2.1 2.2 2.1 .1 NS NS NS
Branched-chain 4.5b 4.3b 4.4b 5.0a . 1 * 1' **
Acetatezpropionate ratio 2.6c 3.0b 3.2b 3.6a . 1 ** ** NS

 

aJ’oCValues in same row with different superscripts differ (P < .05).
1Main effect levels differ, "P < .01, *P < .05, TP < .10.
2isobutyrate + isovalerate.

130
compared to LP tended to result in an interaction among levels of fiber and RIB (P =

.10). Density of rumen digesta was significantly higher (P < .01) for HF compared to
LF treatments.

Rumen fluid concentrations of all measured VFA were significantly lower (P <
.01) upon addition of RIB (Table 6). Total VFA concentrations averaged 126.6 mmol/L
and 113.6 mmol/L for +0 and +B treatments, respectively. Rumen propionate, butyrate,
and total VFA concentrations were higher for LF compared to HP treatments (P < .01).
Molar proportions of rumen VFA were also calculated (Table 6). Proportion of acetate
increased and propionate decreased upon addition of RIB or fiber in the diet, with fiber
level having a slightly greater effect than RIB. Consequently, acetatezpropionate ratio
increased significantly (P < .01) as additional fiber or RIB was included in the
treatment.

Rumen digestibility and digestion kinetics of NDF were calculated from pre- and
postfeeding rumen digesta evacuations and are presented in Table 7. Fractional passage ‘
rates of rumen NDF were similar between pre- and postfeedin g estimates and were
highest for treatment HF+B (.035 h'l) and lowest for treatment LF+0 (.024 h'l).
Additional dietary fiber and RIB both significantly (P < .01) increased passage of NDF
by .002 and .008 h'l, respectively. Fractional digestion rate of digestible NDF tended to
differ more between pre- and postfeeding estimates than did passage estimates.
Digestion rates 2 h prefeeding were not different between LF and HF treatments but
were greater for LP 2 h postfeeding (.046 vs. .038 h'l). Conversely, digestion rates were
greater (P < .01) for treatments with added RIB 2 h prefeeding (.037 vs. .049 H) but
were not greater 2 h postfeeding. Apparent digestibility of NDF in the rumen was
significantly higher (P < .01) for LP compared to HF treatments. Rumen digestibility
averaged 41.2 and 34% for LP and HF treatments, respectively, when expressed as a
percent of NDF intake, and 86.0 and 76.2% when expressed as a percent of total tract
NDF digestibility. Addition of RIB did not affect rumen digestibility of NDF.

131

Table 7. Rumen digestion kinetics of NDF and rumen apparent digestibility of NDF from
12 cows receiving low fiber (LF) or high fiber (HF) diets without (+0) or with (+B) added

rumen inert bulk.l

 

 

 

 

 

 

Main effect2
Treatment Janificanse.
Variable LEfl) LF+B HF+0 HF+B SE F B FxB
Digestion kinetics (h'l)
Fractional passage rate

2 h prefeeding .024d .033b .027c .037‘11 .001 ** ** NS

2 h postfeeding .025b .032a .026b .034a .001 T ** NS

Average .024c .032b .026c .0353 .001 ** ** NS

Fractional digestion rate .

2 h prefeeding .038b .052a .036b .0462‘ .003 NS ** NS

2 h postfeeding .0463 .0472' .034b .042ab .004 * NS NS

Average .0419c .049a .034c .0443" .003 * ** NS *
Rumen apparent digestibility

«««««« (% of NDF intake)---------

2 h prefeeding 40.7a 41 . 1a 34.6b 33.9b 1.0 ** NS NS

2 h postfeeding 42.7a 40.3a 34.1b 33.19 .9 ** 1’ NS

Average 41.6a 40.73 34.4b 33.6b .8 ** NS NS

---(% of total tract digestibility)---

2 h prefeeding 85.1a 86.13 76.39 77 .5b 2.2 ** NS NS

2 h postfeeding 89.0a 84.49 75.5c 75.4c 1.4 ** NS NS

Average 86.93 85.28 75.9b 76.59 1.5 ** NS NS

 

 

avadValues in same row with different superscripts differ (P < .05).

1Calculated using indigestible NDF as an internal marker and rumen pool size

estimates.

2Main effect levels differ, “P < .01, *P < .05, TP < .10.

132
Feeding Behavior

Daily eating, ruminating, and drinking activities of cows were summarized for
each treatment (Table 8). Number of eating bouts were not different across treatments
and averaged 11.6 bouts/d. Mean eating bout lengths were significantly shorter (P <
.01) for LP treatments (27.2 min) compared to HF treatments (32.6 min) that resulted in
shorter daily durations of eating for LP (301 vs. 352 min/d). Because number of eating
bouts were similar across treatments, treatment differences in mean meal size were
similar to differences for daily DMI (Table 3), including the presence of interaction
among levels of fiber and RIB. Treatment HF+B resulted in the largest number of
ruminating bouts (14.7 bouts/d) and bout number decreased (P < .01) as fiber or RIB
were removed. Ruminating bout lengths were shorter (P < .01) for LP treatments but
longer for +0 treatments. Consequently, time spent ruminating was less (P < .01) for LP
(395 min/d) compared to HF treatments (500 min/d), but no effect of RIB was detected.

Time spent eating, ruminating, or chewing per unit of DM or NDF intake
increased significantly upon inclusion of additional dietary NDF or RIB. Eating,
ruminating, and total number of chews per day were fewer for LP compared to HF
treatments (P < .01) and reflected time spent chewing for each activity. High fiber
treatments resulted in faster chewing rates during eating compared to LP treatments
(62.3 vs. 65.0 chews/min), while addition of RIB resulted in slower chewing rates during
ruminating (61.8 vs. 58.2 chews/min). Number of drinking bouts and daily time spent
drinking were greater for LF treatments (15.8 bouts/d; 16.6 min/d) compared to HF
treatments (12.8 bouts/d; 15.1 min/d).

Reticular contractions were expressed as total per day and frequency per min
during eating, ruminating, and idle periods (Table 9). No differences were detected
between treatments during eating for number of contractions, however, frequency of
contractions decreased as fiber was added (P < .01). During rumination, additional fiber
or RIB increased (P < .01) both the total number and frequency of contractions. An

133

Table 8. Eating, ruminating, and drinking activities for 12 cows receiving low fiber (LF)
or high fiber (HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

 

Main effectl
Trement Jennings.
MEL Jew—EM
Eating bouts,/d 11.9 11.4 11.1 11.8 .4 NS NS NS
Meal size
Mean
kg of DM 19”” 2.0a 1.8” 1.5c .2 ** NS *
kg of NDF .48” .50” .613 .50” .04 * NS *
SD,kgofDM 2.1 2.1 1.9 1.6
Maximum, kg ofDM 6.2 5.9 5.6 5.1 .4 NS NS NS
Eating bout length, min 25.9c 28.5”c 33.5al 31.63” 1.5 ** NS NS
Eating time
min/d 294” 308” 349a 3553 10 ** NS NS
min/kg of DM 13.5c 14.9c 19.0” 21.9a .4 ** ** NS
min/kg of NDF 53.9c 59.4” 55.7”c 64.9” 1.2 T ** NS
Eating chews, /d 18,618” 19,912” 22,783a 23,4308 660 ** NS NS
Eating chew rate,
chews/min 61.5” 63.1”” 64.83 65.12‘ 1.0 * NS NS
Water intake during
meals, L/d 50.4a 47.63 44.63” 40.5b 2.5 * NS NS
Ruminating bouts, /d 12.1c 13.4” 13.5” 14.72‘ .4 ** ** NS
Ruminating bout length,
min 32.10 31.1c 37 .23 34.8” .8 ** * NS
Ruminating time
min/d 383” 407” 4962‘ 503a 14 ** NS NS
min/kg of DM 17 .7d 19.7c 27.0” 30.88 .4 ** ** NS
min/kg of NDF 70.8c 78.2” 79.0” 91.1a 1.5 ** ** NS
Ruminating chews, /d 23,867” 24,363” 31,232a 29,93311 917 ** NS NS
Ruminating chew rate,
chews/min 60.8a 58.4” 62.7a 58.1” .8 NS ** NS
Total chewing time
min/d 677” 716” 846a 8583 20 ** NS NS
min/kg of DM 31.3‘l 34.6c 45.9” 52.7a .6 ** ** T
ming of NDF 125c 138” 135”c 156a 2 ** ** NS
Total chews, /d 42,485” 44,275” 54,0153 53,363a 1370 ** NS NS
Drinking bouts, /d 16.1a 15.4a 13.63” 12.0” .9 ** NS NS
Drinking bout size, L 5.6 5.4 6.0 5.7 .3 NS NS NS
Drinking time, min/d 16.7a 16.43 15.93” 14.3” .6 * NS NS
Drinking rate, Umin 4.9 4.7 4.8 4.5 .2 NS NS NS

 

av”r°vdValues in same row with different superscripts differ (P < .05).
1Main effect levels differ, "P < .01, *P < .05, TP < .10.

134

Table 9. Reticular contractions and ruminal pH for 12 cows receiving low fiber (LF) or
high fiber (HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

 

 

Main effectl
"1‘1”an Janificanse.
Variable M43 HF+0 HF+B SE F B FxB
Reticular contractions
Daily total, no./d
Eating 564 552 607 61 l 20 NS NS NS
Ruminating 427° 532” 578” 73021 17 ** ** NS
Idle 850° 764°” 661”° 607° 35 ** 1 NS
Total 1841” 1848” 1846” 1948a 22 * * *
Frequency, no./min .
Eating 1.93a 1.80”” 1.75” 1.72” .05 ** T NS
Ruminating 1.13° 1.32” 1.17° 1.47a .02 ** ** *
Idle 1.11 1.05 1.09 1.04 .03 NS NS NS
Ruminal pH
Daily pH
Mean 6.23° 6.48” 6.543” 6.693 .06 ** ** NS
Variance .24 .24 .21 .18
Minimum 5.52” 5,843 5.95a 6.14a .1 1 ** * NS
Maximum 6.73” 7.023 7.00a 7.07a .05 ** ** *
Range 1.21a 1.18a 1.06”” .92” .09 * NS NS
Hours below pH:
5.5 1.7a .2” .1” .0” .5 T 1' NS
6.0 7.08 2.8” 1.6” .8” 1.2 ** * NS
6.5 15.4‘1 11.13” 9.3”° 5.3° 1.7 ** * NS
7.0 23.9a 22.7a 23.2a 20.6” .7 T * NS
Start of eating 6.30c 6.53” 6.60”” 6.77a .06 ** ** NS
End of eating 6.20° 6.40” 6.503” 6.62a .06 ** * NS

Startofrumination 6.19° 6.42” 6.503” 6.641' .06 ** ** NS
End of rumination 6.25c 6.50” 6.573” 6.713 .07 ** * NS

 

av”r°Values in same row with different superscripts differ (P < .05).
1Main effect levels differ, "P < .01, *P < .05, TP < .10.

135
average difference of .10 contractions/min occurred between LF and HF treatments and

.24 contractions/min between +0 and +B treatments, with the addition of fiber having a
larger effect when RIB was present

Ruminal pH was measured continuously each period across 5 d with an electrode
placed through the rumen cannula and into the ventral sac. Ruminal pH was consistently
lower for LP compared to HF treatments and +0 compared to +B treatments, regardless
of summary variable used (Table 9). Mean pH was 6.23, 6.48, 6.54, and 6.69 for LF+0,
LF+B, HF+0, and HF+B treatments, respectively. Both daily minimum and maximum
pH values were lowest for LF+0 and highest for HF+B. The number of hours during the
day in which pH was below 5.5, 6.0, 6.5, and 7.0 were all highest for LF+0 and lowest
for HF+B. Hours below pH 6.5 demonstrated the most sensitivity to treatment
differences compared to other pH values chosen. Declines in pH between the start and
end of eating and increases in pH between the start and end of rumination averaged . 12
and .07 units, respectively, and were consistent across treatments.

DISCUSSION

The primary objective of this study was to determine if depressions in intake
because of RIB addition varied with NDF content of the diet. Addition of RIB
decreased intake with HF but not with LF, resulting in an interaction which supports our
general hypothesis that intake by cows receiving higher fiber diets is under greater
control by physical constraints of the rumen. Johnson and Combs (1992) conducted a
study with similar objectives using 4 cows at 53 DIM and 4 cows at 185 DIM at the start
of the experiment, fed 27 and 33% NDF diets with or without RIB (25% of rumen
volume). Milk and DMI responses to fiber were similar to our study, however, they
observed no depression in milk production or intake from RIB at either fiber level. They

suggested that their cows adapted to treatments over the 2 wk adjustment period.

136

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9 40' Milk production
\ A
g 35. A
E 30 3* + T'
1:: ' *3 l LF+0
G E] LF+B
g 25‘ DMI O HF+0
'8' £79,. 0 HF+B
3 201
3 e— A A
a 4. 9
5 15-1 3'”
E

10

1 2 3

Week of experimental period

Figure 1. Mean milk production and DMI by week of experimental period for 12
cows receiving low fiber (LF) or high fiber (HF) diets without (+0) or with (+B) added
rumen inert bulk.

Figure 1 illustrates a lack of adaptation over time for cows on our study. In a review of
literature (Chapter 1), depressions in DMI due to RIB addition were regressed against
several factors; intake of DM was reduced to a greater degree for animals in higher states
of production and for animals with smaller within species BW. Johnson and Combs
(1992) may not have detected intake depressions because cows were not in early
lactation throughout the duration of their study and were relatively large in size (BW =
705 kg).

The majority of the results support the general statement that fiber and RIB
addition created an energy deficit compared to the basal treatment (LF+0). Milk and
milk protein production and protein content decreased with both added fiber and bulk,
and milk fat content tended to be higher with added fiber. These results are consistent

with increases in dietary fiber concentration (Sutton, 1989). Dietary NDF concentrations

137
of 25 and 35% were selected for this study because they are considered to be below and

above optimal NDF concentrations for milk production (Allen and Mertens, 1988a;
NRC, 1989) and are near practical extremes for diets with these ingredients. No
differences were observed in BW or body condition between treatments, however, lower
empty BW for treatments with added fiber and RIB were indicative of energy deficit.
Cow BW was similar because of higher rumen digesta mass with fiber and RIB addition.
Three week experimental periods may have been too short to observe changes in
external condition. When RIB was added to early lactation cows in another study
(Johnson and Combs, 1991), milk production, milk composition, and BW were affected
in a similar manner.

Addition of RIB resulted in no difference in rurrrinal or total tract digestibility of
nutrients (except for OM) and is consistent with another study with cows (Johnson and
Combs, 1991) but not with one with sheep (Waybright and Varga, 1991). Waybright
and Varga (1991) inserted large quantities of RIB (up to 66% of rumen volume) that
depressed DM, OM, ADF, and starch digestibilities. In our study, treatments with
higher NDF content resulted in lower DM, OM, and NDF apparent total tract
digestibilities and a shift in NDF digestion from the rumen to the hind-gut. Higher NDF
concentrations with the same NDF source typically lower DM digestibility and increase
NDF digestibility (Woodford et al., 1986; Cameron et al., 1991) because of improved
conditions for fiber digesting microorganisms in the rumen. Changes in rumen
fermentation as measured by VFA concentrations and ruminal pH were almost identical
with fiber and RIB addition. Both produced lower total VFA, increased percent of VFA
as acetate, decreased percent of VFA as propionate, and increased rumen pH as
measured by numerous variables (Table 9). These changes are consistent with the milk
production and composition data from the current study and agree with results from
other studies (Johnson and Combs, 1991; Johnson and Combs, 1992). Bulk addition

may alter VFA production because of decreased rumen digesta volume and increased

138
dilution rate of rumen liquid (Peters et al., 1990), or because of the increased extent of

chewing per unit of DMI (Beauchemin, 1991a).

Insertion of inert bulk into the rumen to study physical aspects of intake control
is not a novel approach, however, the type of RIB used in this study may be. Previous
studies have inserted polyethylene cubes (Carr and Jacobson, 1967), expanded
polystyrene (Baumont et al., 1990), air-filled balloons (Mowat, 1963), water-filled trash
bags (Waybright and Varga, 1991), and water-filled balloons or tubes (Campling and
Balch, 1961; Grovum, 1979; Johnson and Combs, 1992). Smaller water-filled cartons
may be advantageous over large tubes because of their ability to flow among the digesta
and allow a more natural movement of the digesta within the reticulorumen. Cartons
were visually observed to be distributed throughout the entire reticulorumen digesta and
resulted in no deleterious health effects to the cows. Amount of RIB added in this study
(25% of rumen volume) was similar to that used in other studies with dairy cows
(Mowat, 1963; Johnson and Combs, 1991). Added bulk may have made rumination boli
more difficult to form, resulting in smaller boli and more frequent regurgitation;
however, fiber also increased reticular motility during rumination so perhaps not all the
increase in regurgitation frequency was because of RIB's non-physiological presence.

Changes in rumen volume upon fiber and RIB addition support the concept that
daily intake was limited by rumen fill. Digesta volume declined to a greater extent with
RIB on HF compared to LF treatments (16 vs. 11 L). Rumen volume expanded to
compensate for added RIB, but to a lesser degree for HF. Digesta plus bulk volume was
not significantly different for either bulk treatments and HF+0, suggesting that HF+0
was already limited by rumen capacity prior to bulk addition. Johnson and Combs
(1992) measured digesta volume decreases of only 5 L for both fiber levels and more
substantial expansion of the rumen to compensate for bulk addition. Total rumen
volume includes digesta, bulk, and gaseous head space volumes. Head space volume,

which may be a function of rate of fermentation and gas production, was significantly

139
smaller for treatment HF+0 such that no difference in total volume between fiber diets

was observed. Changes in head space volume may compensate for rumen fill.

Digesta plus bulk volume was compared between measurements taken 2 h
prefeeding and 2 h postfeeding (Figure 2) to examine possible control mechanisms for
the main meal following the PM feeding. Main meal size averaged 2.6, 3.1, 3.3, and 2.9
kg of DM for LF+0, LF+B, HF+0, and HF+B, respectively. Volumes were similar for
both determinations for treatment LF+0 but were larger postfeeding for the other three
treatments. These data suggest that different mechanisms dominated satiety for LF+0
compared to the other treatments. We propose that quickly responding physiological
controls had greater influence over meal size for LF+0 and that more slowly responding
physical controls had greater influence over meal size for the other treatments.

Rumen NDF has been considered the component most closely associated with
the space-occupying properties of rumen digesta (Mertens, 1987) because it represents
the structural cell wall, and implies that neutral detergent solubles occupy no volume.
Rumen digesta volume was regressed against rumen pool size of NDF using each cow-
period mean (Figure 3). Every kilogram of rumen NDF was equivalent to 7.5 L of
rumen digesta volume, which is similar to the bulk volume of alfalfa cell wall (Van
Soest, 1982). Regression intercept was 45.9 L that represents the liquid pool in the
rumen not associated with fiber particles (model R2 = .56). Regression of volume on
other digesta components resulted in slightly improved model fit (DM: R2 = .63; OM:
R2 = .59; indigestible NDF: R2 = .57), suggesting that NDF may not completely
represent the fill properties of rumen digesta. Average RIB addition of 22.2 L to the HF
diet was equivalent to a 2.3 kg decline in rumen NDF (9.6 L/kg), a .73 kg/d decrease in
NDF intake, and a 2.1 kg/d decrease in DMI (94.6 g/d per L). Other studies have shown
DMI depressions from RIB addition ranging from no decline with cows (Johnson and
Combs, 1992) to 300 g/L with sheep (Egan, 1972); average depression in DMI from RIB
addition was 94 g/L (Chapter 1).

140

 

C I 2 h Prefeeding
11a. 2 h Postfeeding J

 

 

..s
0
0|

 

 

 

/

é

 

    
   

 

 
 

Digesta + inert bulk volume (L)

 

 

 

 

 

90- ,, /
.. PI
..- él

Figure 2. Rumen digesta plus inert bulk volume 2 h prefeeding and 2 h postfeeding

for 12 cows receiving low fiber (LF) or high fiber (HF) diets without (+0) or with (+B)
added rumen inert bulk.

141

  
   

 

 

110..
g 100i
3 903
2 r
g 804
5 70 ‘
E 1 l‘ = .75
a 60. I I
50: I Rumen volume = 45.9 + 7.5(Flumen NDF)
4o ' l ' l ' l ' I ' l ' I ‘ I fl l
2 3 4 5 6 7 8 9 10
Rumen NDF (kg)

Figure 3. Rumen digesta volume as a function of rumen NDF pool size for 12 cows
receiving low or high fiber diets without or with added rumen inert bulk.

Both fiber and RIB addition increased fractional passage rate of NDF from the
rumen which has been previously observed (Woodford et al., 1986; Johnson and Combs,
1992). Cows may be increasing passage from the rumen in an attempt to minimize the
effects of rumen fill on intake to maintain consumption of digestible energy. Fecal fiber
particle size was measured to determine if increased passage from the rumen forced
excretion of larger particles fi'om the rumen and digestive tract. No increase in size was
observed with added RIB, and added fiber resulted in a decrease in fiber particle size,
not an increase. This fiber response may be due to the greater extent of rumination and
chewing per unit of DMI for higher fiber diets (Welch, 1982; Shaver et al., 1988b).
Changes in fractional digestion rates of digestible NDF because of fiber or RIB addition
were in opposite directions and depended on whether they were measured 2 h pre or 2 h

postfeeding. Fiber addition decreased digestion rate 2 h postfeeding and was somewhat

In.

142
Rumen volume Max: 100 L

      

NDF intake Max = 6.5 kg/d

25 28.5% ’ 35
Dietary NDF content (% of DM)

Figure 4. Model of NDF and DM intake response to increases in dietary NDF
content. Intake of NDF increases until maximum rumen volume, at which point DMI
decreases. Based on a maximum rumen volume of 100 L, maximum NDF intake of 6.5
kg/d, and baseline DMI of 22.8 kg/d, the threshold for intake limitation is 28.5% NDF.
expected because of a larger meal of fiber for HF and the likelihood of non-steady state
conditions during measurements; addition of RIB increased digestion rate 2 h
prefeeding, perhaps as a function of decreased rumen pool size of NDF.

Based on intake and rumen parameters associated with this study, a conceptual
model of NDF and DM intake is proposed for diets of varying NDF concentration
(Figure 4). This model is similar to others proposed by Conrad et al. (1966) and Mertens
(1987). Intake of diets with low NDF content is not limited by rumen capacity. As NDF
content increases, intake of NDF and rumen volume increase until a maximum rumen
volume is reached. Assuming maximum NDF intake and rumen volume were obtained
with treatment HF+0, maximum rumen digesta volume, digesta NDF, and NDF intake,
are 100 L, 7.0 kg, and 6.5 kg/d, respectively. Using LF treatments as base DMI, the
threshold for intake limitations due to rumen capacity is 28.5% NDF in the diet. At

143
NDF concentrations less than 28.5%, rumen volume and digesta removal rates can

increase to adequately compensate for additions of dietary fiber. Above NDF
concentrations of 28.5%, volume and removal rates are maximal and can no longer
compensate, resulting in decreased DMI with increases in NDF content. Such values are
specific for cows at physiological states receiving diets of composition similar to this
study. Larger, lower producing cows likely have intake thresholds at higher NDF
concentrations. Under such a model, physical limits to intake may be reduced by
increasing maximum rumen volume, or by increasing rates of fiber removal from the
rumen via increases in the proportion of fiber that is digestible or increases in fractional
rates of digestion and passage. Further research is required to determine whether these
factors influence intake in such a manner.

Cows did not increase frequency of eating to maintain intakes during fiber and
RIB challenge. Changes in mean meal sizes were similar to differences in daily DMI
because of no differences in the number of eating bouts. Baumont et al. (1990) reported
that the number of eating and ruminating bouts increased when bulk was added to the
rumen of sheep and may have helped maintain a full rumen for as much of the day as
possible. Added fiber and RIB increased the percentage of small meals and decreased
the percentage of larger meals (Figure 5). Distribution of DMI within a day is illustrated
in Figure 6. Greater DMI occurred immediately following the fresh allocation of feed.
It was hypothesized that fiber and RIB addition would promote a more even distribution
of eating throughout the day, however, the opposite was observed, which is consistent
with the rumen volume data depicted in Figure 2. Satiety mechanisms for individual
meals may be more variable when rumen capacity is limiting daily intake. Number of
ruminating bouts increased with fiber and RIB addition and may reflect the necessity to
more frequently process rumen digesta to maximize digestive efficiency. Distribution of
rumination activity within a day was similar for all treatments (Figure 7). Eating,

ruminating, and total time spent chewing per unit of DM and NDF intake were higher

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E
0
E.
g I‘i I I II I 1m
2
0
G
0
G
g
"3'
3’.
Q
Q
E
2 25
C _ I'——'l
3 2° lHF-i-OI
§ 15-
'8
0
9
E I_I I I I I
§
G
n.
I I Fl I I I I I I
1 2 3 4 5 6 7 8 9 1O
Meal size (kg of DM)

Figure 5. Distribution of individual meal sizes expressed as a percentage of the total
number of meals within each treatment for 12 cows receiving low fiber (LF) or high
fiber (HF) diets without (+0) or with (+B) added rumen inert bulk.

145

30 PM Feeding AM Feeding

 

aa

 

     

LF+0
‘ I LF+B
HF+0
‘ - HF+B

25—

 

    

20-

 

 

 

 

  

15-

 

    

   

   
    

Percentage of total daily DMI

 

VIII/IIIIIIIIIIIA ‘

 

k\\\\\\\\\\\\\\\
VII/Illllllllll -
L\\\\\\\\\\\‘
Will/I’ll 3

   

0

-3 3-6 6-9 9-12 0-3 3-6 6-9 9-12
Hours since feed offered

Figure 6. Distribution of DMI within a day expressed as a percentage of total daily
DMI for 12 cows receiving low fiber (LF) or high fiber (HF) diets without (+0) or with
(+B) added rumen inert bulk. Bars within a time period with different letters differ
(P < .05).

146

 

      

 

 

 

 

 

 

PM Feeding AM Feeding
O
.5 7- ,,
a 4
E r .
E 6- :
a : t l, l
a 5.- l\ A
a . p .
"a" 4: /A\ \ l" ‘1
o .. " . .
=5 I i y i .\.
E 3: / x
g I u + LF+0
a 2: / -e— LF+B 3
a 1
§ ‘ —G— HF+B
G
D.

 

 

IIIIIIIIIITIIIIIIIIIIIfifi

123456789101112123456789101112
Hours since feed offered

Figure 7. Distribution of rumination within a day expressed as a percentage of total

time spent ruminating for 12 cows receiving low fiber (LF) or high fiber (HF) diets
without (+0) or with (+B) added rumen inert bulk.

 

147
with both fiber and RIB addition. These variables may more accurately reflect the

chewing stimulation of both effects compared to examination of differences based only
on time spent chewing. Increased chewing activity may have promoted greater rates of
ruminal digestion and passage. Similar results have been reported (W oodford et al.,
1986; Baumont et al., 1990).

Johnson and Combs (1992) monitored reticular motility and observed no
differences in the frequency of contractions due to level of fiber or RIB. Frequency of
contractions during eating may have decreased in our study because of slower rates of
DM consumption during eating. Eating is generally considered to increase the frequency
of reticular contractions (Ruckebusch, 1988). Increased frequency of contractions
during rumination may have resulted from increased rumen volume, which is believed to
increase frequency of contractions following feeding (Okine et al., 1989).

CONCLUSIONS

Addition of RIB decreased DMI for high producing early lactation cows
receiving 35% NDF diets but not for cows receiving 25% NDF diets. Volume of rumen
digesta plus RIB was similar for HF treatments whether or not RIB was present. Fiber
and RIB addition increased NDF passage from the rumen in a similar manner. These
data support the hypothesis that cows receiving high fiber diets have intakes limited by
the physical capacity of the reticulorumen. Changes in feeding behavior or rumen
function were insufficient to maintain intake under conditions of high rumen fill. Added
fiber and RIB altered aspects of rumination, rumen fermentation, and reticular motility in
a similar fashion, demonstrating that NDF has rumen-filling characteristics. A proposed
intake model suggests that expansion of rumen volume and increases in digesta passage
can compensate for increases in dietary fiber concentrations to maintain intake up to
28.5% total diet NDF for cows under conditions of this experiment. Further research is
required to determine the effect of fiber digestibility and animal characteristics on fill-
limited intake.

CHAPTER 5

Harvesting Alfalfa Forages with Similar Fiber Contents but
Different Fiber Digestibilities from Production-Scale Fields

ABSTRACT

Five alfalfa harvests were obtained during the 1992 growing season fi'om two
adjacent 8 ha fields and ensiled in bags. Objectives were to produce silages with similar
fiber concentrations but different fiber digestibilities in sufficient quantities for animal
studies, and examine the consistency of production-sized fields for these variables. Four
of the five silages were equal in NDF content (P > .25), averaging 41.2% NDF.
Metabolic losses of OM during silage fermentation indicated loss of both non-NDF OM
and potentially digestible NDF such that NDF content did not change between pre- and
post-ensiled samples. Rate of NDF digestion, as estimawd by in vitro rumen fluid
digestion, was not different across silages. Extent of NDF digestion, however, was
different between silages after 10, 24, and 120 h of fermentation. Two of the forages
differed consistently in NDF digestibility by 5% units, resulting in a 4% difference in
estimated NEL concentration. Higher forage digestibility may significantly improve
animal performance because of increased ration energy density or increased forage intake.
An experiment using these two forages with early lactation dairy cows would be useful to
assess the importance of fiber digestibility to ruminants with high energy requirements.

INTRODUCTION
Evaluation of forage quality and its importance to ruminant performance has been

of interest for many years (Crampton et al., 1960; Waldo and Jorgensen, 1981). It is

148

149
generally agreed that the goal of forage feeding is to maximize daily intake of digestible

nutrients. Such a goal requires either high forage intake, high forage digestibility, or both.
The non-cell wall portion of forages is rapidly solubilized in the rumen and almost 100%
digestible (Van Soest, 1965), consequently having little affect on forage intake when
limited by rumen fill and little affect on digestibility at constant dietary concentrations.
Conversely, concentration and availability of forage cell wall, or NDF, are the primary
determinants of the digestible energy content of forages (Goering and Van Soest, 1970)
and may promote differences in forage intake (Van Soest et al., 1978).

Concentration of forage NDF is negatively correlated to intake and digestibility
(Van Soest et al., 1978). In dairy cattle, dietary NDF content was related negatively to
both milk yield and intake (Briceno et al., 1987; Mertens, 1983), and positively to rumen
fill (Shaver et al., 1988a). Fiber content alone may not describe completely the affect of
fiber on intake and digestibility. Of increasing interest is NDF digestibility. Ruminants
are unique among food animals because of their ability to extract large amounts of energy
from forage NDF. The efficiency of rumen microbial fiber digestion depends, among
other things, on the intrinsic characteristics of the plant fiber that render it more or less
available to rumen fermentation. Greater fiber digestion has been suggested to improve
forage energy content and intake potential (Allen, 1991).

Studies that have examined fiber digestibility on animal performance are limited.
Diets similar in NDF content but different in rate and extent of NDF digestion were
compared and demonstrawd advantages to faster and more completely digestible NDF
(Varga et al., 1984). However, this study used different sources of NDF, including NDF
from by-product feeds, that confounded interpretation of results for fiber digestibility. A
preferred approach is to examine ruminant performance in response to a single forage
specie harvested from different environments. Environmental growing conditions can
significantly alter fiber composition and digestibility (Van Soest et al., 1978). An ideal

comparison would include two harvests of the same forage type with similar fiber

150
contents, similar protein contents, but different fiber digestibilities. Such a study would

isolate the effect of NDF digestibility to allow its evaluation independent of NDF content
and source. In vitro NDF digestibility has been shown to differ significantly between
small plots of alfalfa with similar fiber contents grown under varying conditions (Allen et
al. , 1991 ).

Finally, analyses of forage composition useful for ration formulation require that
samples not only reflect the true mean of the forage but also that the range of forage quality
represented by the samples is relatively small. For example, rations formulated using an
average forage NDF content of 40% vary widely in quality if the material from which the
sample was taken ranges from 30 to 50% NDF. Even relatively pure stands of a forage
specie may fluctuate greatly in large fields with diverse soil types and fertilization. Little is
known about the consistency of forage fiber digestibility within the same harvest,
especially for material from a large stand. Consistency needs to be demonstrated before
this characteristic of forage quality can be implemented in ration formulation.

Objectives of this study were to harvest production-sized fields of alfalfa at similar
fiber concentrations during the diverse conditions of a single growing season and in
sufficient quantities to permit their use in animal studies. Consistency of differences for in
vitro fiber digestibility of the resulting silages was to be examined. If significant
differences were obtained, forages would be reserved to evaluate the effect of fiber
digestibility on intake and in vivo digestibility in lactating dairy cows.

MATERIALS AND METHODS

One 16.2 ha field of pure-stand alfalfa (Big 10 cultivar) in its 2nd year of
production on the campus of Michigan State University, East Lansing was used for this
study. Two 8.1 ha plots (A and B) were generated by dividing the field in half (Figure 1).
Plot A forage was cut and removed on May 19, 1992. This harvest was not included in
the study and was removed to stagger regrowth and harvesting between plots A and B.
After May 19, samples from each stand were taken approximately every 3 d at 4 locations

 

151

 

 

 

 

 

 

 

 

 

112m 110m
216m
300m
'ssm
H—>
I
l
PlotA I
260
m : PlotB
I
l
l
344m

Figure 1. Dimensions of plots from which alfalfa forage was sampled and harvested.

152
within each plot. Samples were cut with a scissors at a height approximating that of

mechanical cutting (10 cm stubble), dried at 52’C, and ground through a Wiley mill (1-
mm screen). Samples were individually analde for NDF concentration using a
procedure previously described (Chapter 2).

The decision of when to cut and harvest each plot was made when projections
indicated harvested forage would contain 40% NDF. Standing forage was out between
0800 and 1200 h with a conventional sickle bar mower-conditioner. A total of six cuttings
were made (3 for each plot) throughout the summer following the initial clip of plot A.
Intervals between all cuttings averaged 19 d (range = 14 to 22 d). Four samples taken on
the day of cutting were separated to estimate leafzstem DM ratios and evaluated for stage of
maturity (mean stage by weight) according to procedures of Kalu and Pick (1981).
Forages were field wilted for 1 to 2 d to achieve target DM concentrations of 30 to 35%
and chopped between 1200 and 1700 h with a 12-knife forage harvester to achieve
medium fineness of chop. Every load of forage was sampled during packing into a 2.5 m
diameter silage bag (Ag-Bag Corporation, Blair, NE). A separate bag was used for each
harvest. Ensiling was completed in l d for each forage.

Pre-ensiled samples were dried immediately at 52'C in a forced-air oven. Two
months after the final harvest, post-ensiled samples were taken from four locations within
each bag and dried. Following grinding to 1 mm with a Wiley mill, both pre- and post-
ensiled forage samples were analyzed for ash, NDF, and NDF digestibility as estimated by
in vitro fermentation with buffered rumen fluid (Goering and Van Soest, 1970). Samples
were analyzed for NDF immediately following 0, 4, 10, 24, and 120 h of fermentation.
Rate of NDF digestion was estimated for each sample as the negative slope of the
regression between the natural logarithm of digestible NDF remaining vs. time at 4, 10,
and 24 h. This model assumes first order digestion with respect to digestible NDF
concentration. Indigestible NDF was defined as NDF remaining after 120 h of

fermentation and was subtracted fi'om the total NDF remaining at each time point to obtain

153
concentrations of digestible NDF. Post-ensiled samples were analyzed additionally for CP

content using the Hach procedure (Watkins et al., 1987). Statistical differences in forage
composition were determined with analysis of variance (SAS, 1985).
RESULTS AND DISCUSSION

Sampling of alfalfa plots began at least 8 (1 prior to the actual day of cutting (Figure
2). Concentrations of NDF ranged from 25 to 33% at first sampling and 38 to 42% at
cutting. Rate of NDF increase for all forages averaged .74% units/d and ranged from .52
to 1.02% units/d. The largest NDF increase between any two sampling days was 2.2%
units/d while the smallest was zero; variation in NDF accumulation appeared to be related
to daily temperature and light availability, although these variables were not measured.

Out of six cuttings, five forages were successfully ensiled (Table 1). Forage C was
destroyed by 6 d of continuous rain following cutting and was chopped onto the field as
fertilizer to allow continued plant regrowth. Average amount of forage ensiled for the five
successful harvests was 22 tonne of DM. Small but statistically significant differences
were present for stage of maturity (2 = late vegetative, 3 = early bud, 4 = late bud) and
leafzstem ratio. Prior to ensiling, forage DM was similar for A and E, and B and F.
Whole plant NDF content within forage was consistent such that mean differences as small
as 1.4% units were different (P < .05) across forages. Digestibility of NDF differed
significantly after 10 and 120 h of fermentation, but rate of NDF digestion was similar.

After ensiling, visual appraisal indicated that each forage silage was of excellent
quality (Table 1). Our objective was to obtain forages with similar NDF concentrations
after ensiling, which was achieved for forages B, D, E, and F (mean = 41.2% NDF; P >
.25). It should be nowd that more differences were likely if more replication had been
performed. Forage A will not be included in subsequent results and discussion. Several
challenges were encountered in the attempt to produce silages with equal NDF contents.
First, four spot samples taken immediately prior to cutting may not have represented the

average composition of the cut material. A main cause of this may have been the inability

154

 

 

 

 

 

 

 

A40
2\: q
C
.9
E
335'
§:/ /-. Forage
8~ ' —e—A
u.‘ /. +3
030‘ +0
2* —u—D
+E
, . .:. F
25..’. .r-..-..-

-18 -15 -12 -9 -6 -3 0
Days preharvest

Figure 2. Concentration of NDF in alfalfa forages sampled at various days prior to
harvest. Objectives were to obtain forage with 40% NDF at day 0.

155

Table 1. Characteristics of alfalfa forage harvested during the 1992 growing season
from two adjacent 8 ha fields.

 

 

 

 

 

 

 

 

Forage
A B C D E E
Field B A B A B A
Cutting number 1 1 2 2 3 3
Cutting date 6/8 6/27 7/11 7/27 8/ 17 9/8
Yield, tonne of DM 21.6 21.7 Lost due 20.7 23.6 20.8
to rain
Stage of maturity1 3.123” 2.72d 3.02”° 2.88“1 3.238
Leaf:stem ratio1 .57° .88” .983” 1.14a .963”
Pre-ensiled
n 12 10 . 9 1 l 10
DM, % 336° 40.93 35.9” 330° 39.7a
Ash, % of DM 9.8”c 96° 10.211 9.93” 9.6”°
NDF, % of DM
Whole plant 44.4a 40.8c 403° 38.9‘1 42.8”
Lear2 23.4 22.2 24.0 23.2 25.6
Stem2 55.0 54.5 58.5 56.8 56.2
NDF digestibility”, %
10 h 27.2” 27.7” 27.6” 30.13” 32.02‘
24 h 44.2 43.8 43.9 45.5 45.4
120 h 50.5”” 50.83” 49.0” 51.7a 50.73”
Rate of NDF
digestion, h'1 .098 .092 .105 .099 .102
Post-ensiled
n 4 4 4 4 4
DM, % 30.0” 39.6a 33.3” 32.5” 38.02‘
Ash, % of DM 105° 11.1” 11.1” 10.7”° 12.2a
CP, % of DM 218° 24.2” 26.2a 24.5” 24.6”
NDF, % of DM 45.8a 41.9” 40.4” 40.1” 42.4”
NDF digestibility, %
10 h 17 .02‘ 19.53” 18.32‘ 23.1° 22.0”c
24 h 41.3a 41.2a 39.6a 44.8” 41.4a
120 h 47.7”” 48.2”” 45.5a 51.0” 47 . 1a
Rate of NDF
digestion, h'1 .100 .095 .101 .104 .103
1n = 4.
2n = l.

3Determined from in vitro fermentation with buffered rumen fluid for length of time

indicated.

ar”v°tha1ues in same row with different superscripts differ (P < .05).

156
to predict cutting difficulties faced with lodged plants and the extent of remaining plant

stubble following cutting. Second, extent of leaf loss may have varied across harvests
depending on windrow density and width, leaf density, and drying conditions. Third,
harvested forages may have differed in their fermentation characteristics during ensiling,
resulting in variable losses of NDF and non-NDF substrate; the correlation between pre-
and post—ensiled NDF contents (r = .94) suggests that this problem was minimal. Finally,
the challenges presented by the occurrence of rain and the schedule of hired labor were
probably of greatest concern.

Silage DM was similar for B and F, and D and E (T able 1), with DM increasing 2
to 3% units fiom the first to last sample for each forage. In vitro digestion of NDF was
most rapid during initial periods of fermentation, which supports the concept of first-order
digestion kinetics. Nevertheless, none of the forages differed in fractional rate of
digestion (mean = .101 h”. Digestibility of NDF was consistent within each ensiled
forage to produce significant differences across B, D, E, and F after 10, 24, and 120 h.
Figure 3 illustrates consistency of NDF content and NDF digestibility for forages D and E.
Across all forages, the average coefficient of variation for NDF content within forage was
4.0%, and that for NDF digestibility was 5.2%, indicating that fiber concentration may be
slightly less variable within fields of alfalfa than digestibility. Allen et al. (1991) detected
a large range (25 to 55%) in alfalfa NDF digestibility after 30 h fermentations across
samples taken from numerous small plots. It was not known, however, whether
differences would be consistent across a field of a size typically found on farms. This
study suggests that differences in NDF digestibility are repeatable within an entire field of
alfalfa and may be useful in evaluating forage quality.

Forage analyses were compared between samples taken pre- and post-ensiling
(Table 2), averaged across all five forages harvested. Changes in composition reflect
metabolic processes associated with silage fermentation. Fermentation requires the

conversion of some silage OM to organic acids and heat. A reduction in forage DM

157

 

 

 

 

 

 

 

 

 

 

 

:5 55' 393139311911 % + ForageD e
3. 51- ' —e— ForageE
;§ 47-
173
8) 43-
E 391 I— —I /
g 35
2282031195!

43-
K? J
2’ 39- N 46
E
9 35
C
o
O
Li. W
D 43
Z

39

35l I I I I I I I I I l

12 3 4 5 6 7 8 91011
Load No.

Figure 3. Neutral detergent fiber concentrations and NDF digestibility for loads of
chopped alfalfa obtained pre- and post-ensiling from harvests D and E. Digestibility of
NDF was determined from 24 h in vitro fermentation with buffered rumen fluid.

158
Table 2. Composition of alfalfa forages before and after ensiling.l

 

 

 

 

Ire-ensiled Female! Significanm

DM, % 37.1 34.7 .01
Ash, % of DM 9.8 11.1 .02
NDF, % of DM 41.3 42.1 N82
NDF digestibility,3 %

10 h 28.9 20.0 .01

24 h 44.6 41.7 .02

120 h 50.5 47.9 .01
Rate of NDF

digestion, h'1 .099 .101 N S

 

1Average across all forages harvested.
2NS = P > .10.

3Determined from in vitro fermentation with buffered rumen fluid for length of time
indicated.
content and an increase in ash is characteristic of such a process (Table 2). Concentration
of NDF did not change significantly; however, NDF digestibility decreased substantially,
especially with earlier lengths of in vitro fermentation. These changes suggest that both
digestible NDF and non-NDF OM were consumed by microorganisms during silage
fermentation such that total NDF concentration did not change. Other work also has
detected fermentation of cell wall components during ensiling (Morrison, 1988). Studies
that seek to improve silage fermentation should not strive solely to reduce forage NDF, but
rather, strive to decrease indigestible NDF or increase the rate of NDF digestion in the
rumen.

Correlations among the various descriptors of forage quality across the five
harvested silages are listed in Table 3. Ratio of leaveszstems was related negatively to
NDF content and tended to be positively related to CP content and 10 h NDF digestibility.
Except for this tendency, differences in NDF digestibility were not related to differences in
maturity, 1eaf:stem ratio, or NDF content, a result similar to that reported by Allen et a1.

159

Table 3. Correlation matrix among descriptors of forage quality for all five forage
silages.l

 

 

 

Maturity L:S 2 CP% NDF% 10 h3 24 113 120 h 3
L:S -.28
CP% -.15 .79
NDF% .43 -.95 -.89
10 h -.08 .81 .35 -.59
24 h -.26 .44 -.20 -.24 .73
120 n -.44 .33 -.29 -.17 .62 .97
Rate4 .59 .44 .19 -.25 .54 .45 .22

 

lValues > l.87l are significant (P < .05).

2L:S = Leaf:stem.

3NDF digestibility (%) following indicated time of in vitro fermentation.
4Rate = Fractional rate of NDF digestion (h‘l).

(1991). It is speculated that differences in NDF digestibility between cuttings were due to
the interaction of light, temperature, and moisture during the vegetative growth period of
each forage. Vough and Marten (1971), as well as others, have demonstrated the effects
of these environmental factors on alfalfa yield and nutrient concentrations but have not
examined their effects on fiber digestibility. Van Soest et al. (197 8) summarized previous
studies and found that temperature and water availability increased, and light decreased,
forage cell wall content. Opposite relationships were found with DM digestibility. They
suggeswd that increased lignification of the cell wall because of more rapid plant growth
decreases the digestibility of the cell wall. Digestibility of cell wall also may vary
depending on the physical structure of individual fiber molecules and their interaction with
rumen microorganisms (Hatfield, 1993; Dehority, 1993).

Currently fiber digestibility is not measured routinely in forage analyses nor is it
accounted for when estimating the energy content of forages. A common method for
calculating forage energy is to determine its fiber concentration and then assume a direct

relationship between fiber and energy (Harlan et al., 1991). For example, if two alfalfa

160
silages have the same NDF and ADF concentration, they are assumed to contain equal

energy, regardless of their actual fiber digestibilities.

Forages D and E provide a unique opportunity to examine the potential error
involved by not considering fiber digestibility when determining energy content. Except
for crude protein content, D and E are identical in all respects other than NDF digestibility.
In addition, they are the only two forages that differ in NDF digestibility at every time of
fermentation. Based on energyzfiber relationships, both forages would be assigned an
energy value of 1.34 Mcal of NEng of DM, using equation [14] from Harlan et al.
(1991). Based on NDF digestibility at 24 h, forage E contains 2.1% units of TDN more
than forage D ([44.8 - 39.6] * 40.25% NDF), or .05 Mcal of NEng of DM (1% TDN =
.0245 Meal NEng of DM; NRC, 1989). Increased NDF digestibility may also increase '
forage intake for early lactation cows, whose intakes may be limited by rumen fill (Chapter
4). If undigested NDF is the nutrient responsible for rumen fill, a limit in ruminal capacity
suggests that forages are consumed to a constant undigested NDF load. For every
kilogram of forage intake, forage D contributes an additional 21 g of unferrnented NDF to
the rumen compared to forage E. Consequently, greater quantities of forage E can be
consunwd prior to reaching maximum rumen load.

The potential impact of fiber digestibility on animal performance can be examined
by incorporating the energy and intake differences between forages D and E into
theoretical diets for early lactation cows (Table 4). When formulawd to 27% NDF, the
diet of forage E contains .02 Mcal NEng of DM more than diet D and an intake
advantage of 1.6 kg/d of DM. The combined effect of energy density and intake potential
results in an NEL intake advantage of 3.1 Meal/d for diet E, which is equivalent to 4.5
kg/d of milk production. Actual response to increased fiber digestibility will probably be
equal to or less than this maximal estimate; increases in forage intake may increase passage
of digesta from the rumen that results in decreased rumen digestibility. Robinson and

McQueen (1992) examined theresponse of midlactation cows to differences in timothy

161

Table 4. Potential improvement in intake and milk production in early lactation cows
because of an increase in NDF digestibility between forages D and E.1

 

 

1112 E I. '1 .1.
Low High

Theoretical diet composition ----------- % of DM ------------
Forage D 54.7
Forage E 54.7
Concentrate2 45.3 45.3
NDF3 27.0 27.0
Indigestible NDF4 14.9 13.8
NEL, Mcal/kg DM 1.58 1.60

Theoretical cow performance
Indigestible NDF intake, kg/d 3.0 3.0
NDF intake, kg/d 5.4 5.9
DMI, kg/d 20.1 21.7
NEL intake, Mcal/d 31.7 34.8
Milk production (3.5% fat), kg/d 31.9 36.4

 

lAssumes maximal indigestible NDF intake of 3.0 kg/d because of rumen fill.

2Assumes concentrate composition of: NDF = 11% of DM, NDF digestibility =
67%.

3Diets formulated to contain 27 % NDF.

4Calculated using NDF digestibility after 24 h in vitro fermentation.

162
grass NDF digestibility and found increases in milk yield and in vivo NDF digestibility but

no increase in intake. Rumen fill for these cows may not have limited intake, resulting in
no improvement in intake with higher fiber digestibility.

Obtaining sufficient quantities of forages like D and E to examine the effect of fiber
digestibility on early lactation performance was an important objective of this work. These
forages were reserved for such an experiment. The experiment was conducted during the
spring of 1993 and results are reported in Chapter 6. Future work in this area should
strive to obtain large quantities of forage with greater differences in NDF digestibility.
Comparisons within various NDF concentrations, across different extents of NDF
digestion, and across different rates of NDF digestion would also be valuable. The
ultimate goal of this work is to improve estimation of forage quality as it relates to animal
performance and to formulate more accurate rations for ruminant animals.

CONCLUSIONS

Interactions between heat, light, and moisture during a growing season may alter
fiber formation in alfalfa throughout an entire field and make its digestibility distinctly
different from that of another cutting, variety, or location. Because fiber concentration and
fiber digestibility may be consistent for an entire field, average quality characteristics of the
alfalfa field are useful. Forage analyses differ between pre- and post-ensiled samples,
especially for fiber digestibility. Use of post-ensiled analyses are most relevant for
comparing fiber digestibilities of alfalfa silage. Differences in fiber digestibility may affect
currently used estimates of energy values for alfalfa and other forages. Higher forage
digestibility may significantly improve animal performance because of increased energy
density or increased forage intake.

CHAPTER 6

Enhanced Intake and Production of Cows Offered Ensiled Alfalfa with
Higher Neutral Detergent Fiber Digestibility

ABSTRACT

Two alfalfa silages with similar NDF concentrations (40%) but different NDF
digestibilities (40 vs. 45% after 24 h in vitro fermentation) were offered to 12
multiparous cows (13 DIM) in a 2 period balanced crossover experiment. Silages were
mixed with a shelled com-based concentrate to achieve diet NDF concentrations of 35%.
Objectives were to determine the effect of forage fiber digestibility on intake and
production of cows with intakes limited by rumen fill using treatments unconfounded by
differences in fiber source, concentrate, or forage:concentrate ratio. Milk production
(36.3 vs. 38.2 kg/d) and DMI (19.4 vs. 20.4 kg/d) were significantly greater with higher
NDF digestibility. Apparent in vivo DM and NDF digestibilities, total rumen VFA after
feeding, and molar proportion of propionate also tended to increase with higher NDF
digestibility. During the trial, diets differed not only in NDF digestibility (3% units) but
also in NDF content (1.8% units) because of loss of digestible NDF in the higher quality
forage; consequently, interpretation of results is difficult. If increases in NDF
digestibility prior to ensiling promote decreases in NDF content after ensiling, benefits
to harvesting forages with higher NDF digestibility may be substantial. Further
experimentation with forages with larger differences in fiber digestibility at the time of

feeding is required to more closely examine this effect.

163

164
INTRODUCTION

Formulation of diets for dairy cows based on NDF concentrations of the
ingredients has been advocated because of positive relationships between NDF content
and rumen fill and negative relationships between NDF content and energy
concentration (Mertens, 1985; 1987). Mertens (1987) indicated that cows consume
approximately 1.1% of their BW as NDF per day when intake is limited by rumen
capacity and suggested that this relationship was sufficiently consistent to predict DMI.
Comparing data across several experiments, Briceno et al. (1987) observed curvilinear
relationships between DMI and NDF content of the diet that differed by forage species,
indicating that not all forage NDF results in the same intake response. Indeed, intake of
NDF as a percent of BW has been shown to vary with DIM, FCM production, and NDF '
content of the diet (Rayburn and Fox, 1993). A recent attempt to improve intake
predictions using NDF found that additional information is needed on the effect of fiber
digestibility and rumen capacity on intake (Williams et al., 1989). When intake is
limited by rumen fill, increases in the fraction of NDF that is digested or increases in
rates of NDF digestion and passage may enhance clearance of fill from the rumen and
improve DMI (Chapter 4).

Rumen digestion of fiber is a dynamic process that involves microbial attachment
and fermentation of cell wall polysaccharides. The extent of fiber digestion (i.e., fiber
digestibility) at any time is a function of the proportion of fiber that is potentially
digestible, the rate of fiber digestion in the rumen, and the rate of fiber passage from the
rumen (Allen and Mertens, 1988b). Measurement of digestion rate, digestion lag,
potential extent of digestion, and passage rate is required to completely describe fiber
digestion quantitatively (Mertens, 1977). Several studies have provided evidence that
rate and potential extent of NDF digestion varies both across and within fiber sources
(Smith et al., 1972; Varga and Hoover, 1983; Robinson and McQueen, 1992). Smith et

al. (197 2) observed variation in rate and extent of digestion between numerous forage

165
species and within forages differing in maturity. Varga and Hoover (1983) detected

differences between forages and a number of concentrate and by-product feeds. More
recent studies have shown differences in digestibility within forage species (Llamas-
lamas and Combs, 1990; Allen et al., 1991; Allen et al., 1992; Robinson and McQueen,
1992). Interest in cell wall digestibility by ruminants has stimulated a recent
international symposium (USDA-ARS, 1993) that reviewed current knowledge of forage
fiber digestion.

Improvements in forage fiber digestibility may not only increase the energy value
of a particular forage, but also stimulate additional DMI (Gill et al., 1969). Crampton
and co-workers (Crampton et al., 1960; Donefer et al., 1960) developed a nutritive value
index for forages and found that relative forage intake with sheep was highly correlated
(r > .83) with cellulose digestibility after 12 h of in vitro incubation. Muller et al. (197 2)
fed sheep corn silage stover from a normal and brown mid-rib mutant variety that
contained similar NDF concentrations but different in vitro NDF digestibilities. In vivo
DM digestibility increased 11% and DMI increased 29% with improved fiber
digestibility. Intake by steers fed straw was related to rate and extent of DM, and likely
NDF, digestion (Orskov et al., 1988). Recent studies with lactating dairy cows have
shown variable responses to fiber digestibility (Keith et al., 197 9; Varga et al., 1984;
Miller et al., 1990; Llamas-lamas and Combs, 1990; Robinson and McQueen, 1992;
Feng et al., 1993). In only one study (Llamas-lamas and Combs, 1990) was DMI
significantly greater with higher fiber digestibility. Studies of Varga et al. (1984), Miller
et al. (1990), and Feng et a1. (1993) were based on rumen fill characteristics of numerous
feeds (Varga and Hoover, 1983) but are difficult to interpret because of the wide variety
of fiber and starch sources utilized. Studies of Keith et al. (197 9) and Robinson and
McQueen (1992) were not confounded by differences in fiber source and
forage:concentrate ratio. Both studies showed improved milk production with greater

fiber digestibility but no difference in DMI, perhaps because cows were not in early

is

166
lactation and did not have intakes limited by rumen capacity. If changes in intake are

shown to be related to fiber digestibility, it may be possible to more accurately predict
the intake potential of forages.

Perhaps a more appropriate experimental design would involve use of early
lactation cows with high nutrient requirements fed high fiber diets so intake would more
likely be limited by rumen capacity (Chapter 4). To isolate differences in forage fiber
digestibility in unconfounded treatments, offering forages of the same species that
contain similar fiber and protein concentrations but different fiber digestibilities would
be appropriate. Such forages might be obtained by harvesting forages under different
growing conditions, and taking advantage of variation in fiber composition associated
with differences in heat, light, and temperature (Van Soest et al., 1978).

Objectives of this experiment were to obtain alfalfa silages with similar fiber and
protein concentrations but differing fiber digestibilities, as measured by in vitro
fermentation, and determine the effect of forage fiber digestibility on intake and
production of early lactation dairy cows. A further objective was to measure feeding
behavior and rumen characteristics to determine if changes were consistent with
conditions of fill-limited intake.

MATERIALS AND METHODS
Forage Harvest and Selection

Two adjacent 8 ha fields of pure-stand alfalfa in their 2nd year of production on
the campus of Michigan State University, East Lansing were used for this study.
Regrowth between fields was staggered by early cutting of one field to provide a harvest
approximately every 3 wk throughout the 1992 growing season. Forage NDF
concentrations were monitored during plant growth; fields were cut and harvested when
forage NDF was predicted to eqiral 40%. Forages were chopped and stored as silage in
individual bags (Ag-Bag Corporation, Blair, NE). Objectives were to produce silages

167
with similar NDF and CP concentrations but different NDF digestibilities in sufficient

quantities for this lactation study, as described in Chapter 5.

Out of six cuttings, one was destroyed by rain and five were successfully ensiled.
Two months after the final harvest, post-ensiled samples were taken from four locations
within each bag and analyzed for ash, CP, NDF, and NDF digestibility at 10, 24, and 120
h as estimated by in vitro fermentation with buffered rumen fluid (Goering and Van
Soest, 1970). Four of the silages contained equal concentrations of NDF; among these,
two differed (P < .05) in NDF digestibility by 5% units at each time of in vitro
fermentation (Table 1). Consequently, these two forages were reserved for the lactation
study and designated as low digestible fiber (LDF) and high digestible fiber (HDF)
forages.
Experimental Design and Data Collection

Twelve multiparous Holstein cows in early lactation (13 :l: 6 DIM; mean 1: SD)
were grouped by date of calving and assigned to one of two treatment diets in a two
period balanced simple crossover design. Calving dates were distributed across 2.5 mo,
therefore groups of cows commenced treatment at different times to attain starting dates
at similar DIM. Cows were not bred until the experiment was completed. Treatment
periods were 28 d in length, with the last 10 d used for data and sample collection. Diet
ingredients were ground shelled corn, soybean meal, animal protein, minerals, vitamins,
and one of the two reserved alfalfa silages that differed in NDF digestibility (Table 2).
Values represent average ingredient and nutrient compositions from across the entire
experiment. Rations were formulated to contain 35% NDF, at least 22% CP, and
provide at least .5 kg/cow per (1 of the rumen undegraded protein supplement. Diets
containing 35% NDF were found to limit intake by early lactation cows because of
rumen fill (Chapter 4). All concentrate ingredients contained greater than 87% DM and

were combined for each diet prior to the experiment. Except for mineral concentrations,

168
Table 1. Characteristics of low digestible fiber (LDF) and high digestible fiber (HDF)
alfalfa silages prior to commencement of animal experiment.l

 

 

 

 

LDEsilasc HDch
Cutting date 7/27/92 8/17/92
Cutting number 2 3
Stage of maturity2 3.0 2.9
Leaf:stem DM ratio 1.0 1.1
DM, % 33.3 32.5
OM, % of DM 88.9 89.3
CP,a % of DM 26.2 24.5
NDF, % of DM 40.4 40.1
NDF digestibilityz, %

10 h3 18.3 23.1

24 h3 39.6 44.8

120 ha 45.5 51.0

 

1n = 4 samples per forage.

2Estimated as mean stage by weight (Kalu and Fle, 1981); 2 = late vegetative,
3 = early bud, 4 = late bud.

3Determined from in vitro fermentation with buffered rumen fluid for length of
time indicated.

aNutrient concentration between forages differs significantly (P < .05).

169

Table 2. Ingredient and nutrient composition of low digestible fiber (LDF) and high

digestible fiber (HDF) diets offered to 12 cows during the first 10 wk of lactation.

 

I 1' C . .
LDF alfalfa silage
HDF alfalfa silage
Ground shelled com
44% Soybean meal
Protein supplementl
Dicalciurn phosphate
Magnesium oxide

Trace mineral supplement

Vitamin supplement2

11' C ..

DM, %

OMa

NDFa

ADFa
Hemicellulose
Cellulosea
Lignina

CPa

RUP,4 % of diet CP

Ca
P
Mg

 

 

 

 

LllF HDF
--------(% of diet DM)-«««-
83.3
. .. 83.3
11.8 11.8
1.0 1.0
2.6 2.6
.9 .9
.02 .02
.33 .33
.03 .03
3
LDF SD HDF SD
37.0 1.8 38.0 1.1
(% of diet DM)
89.2 .2 89.6 .2
36.7 7 34.9 .6
26.6 5 24.6 .7
10.2 6 10.3 .5
20.4 4 19.2 .5
6.2 l 5.4 .2
23.9 .5 23.0 .3
27.3 ND5 27.5 ND
1.42 ND 1.37 ND
.58 ND .56 ND
.28 ND .26 ND

 

1Contains 55% meat and bone meal, 25% blood meal, 10% fish meal, and 10%

feather meal.

2Contains 1733, 512, and 7.3 kIU/kg of DM of vitamins A, D, and B, respectively.
3Represents variation in chemical composition of samples composited weekly (n =

8/diet).

4RUP = rumen undegraded protein, calculated from NRC (1989).

5ND = not determined.

aNutrient concentration between diets differs significantly (P < .01).

170
nutrient composition of the diets was determined from chemical analysis of individual

feed ingredients.

Rations were mixed once daily and offered for ad libitum intake (10% refusal
rate) as a TMR at 0800 and 2000 h, after cows were milked in their tie stalls. Feed
amounts offered and refused were recorded daily. Forages were sampled daily (.5 kg),
frozen, and composited weekly. Concentrate (.5 kg), diets (.5 kg), and feed refused
(12.5% of orts) were sampled daily during collection days, frozen, and composited
weekly. Milk yield was recorded daily. Milk samples were obtained on d 24 and 25 of
each period for both milkings and individually analyzed for lactose, fat, and protein
content with infrared spectroscopy by Michigan DHIA. Daily production of each
component was determined in addition to SCM, using the equation of Tyrrell and Reid
(1965). Cow BW was determined at 2200 h (d 24) and 1800 h (d 26) each period. Cows
were scored for body condition ((1 24) each period.

Fecal samples were obtained and frozen every 15 h from d 24 to 28 of each
period to represent defecation every 3 h throughout a 24 h day. Sampling commenced at
0900 h on d 24 and continued at 2400 h (d 24), 1500 h (d 25), 0600 h and 2100 h (d 26),
1200 h (d 27), and 0300 h and 1800 h (d 28). Cows were not disturbed to obtain samples
but were allowed to defecate naturally. The eight samples for each cow were
composited on an equivalent wet basis (150 g samples). Feed disappearance, water
intake, and jaw movements were measured continuously from d 19 to 24 each period
with monitors and a computer data acquisition system (Chapter 2). Chewing monitors
were placed on the cows 1 d prior to collection to allow for animal adaptation. Data
were written to files every 5 s for each cow and variable. Cows remained in their stalls
throughout the 5 (1 period, allowing for complete recording of all cow behavior.

Four of the twelve cows had a permanent rumen cannula and were used to
investigate rumen digesta characteristics. Rumen digesta was manually removed from

each cow via the cannula at 2200 h on d 24 (2 h postfeeding) and 1800 h on d 26 (2 h

17 1
prefeeding). A 10% aliquot of the digesta was separated during evacuation for sampling.

Two digesta samples were obtained. A 600 g sample of the complete digesta was taken
for determination of DM and nutrient composition. A second sample was strained
through four layers of cheesecloth; the resulting fluid was immediately measured for pH
and saved for VFA analysis. Both digesta and fluid samples were frozen immediately
(-20°C) following collection. Total digesta weight and volume were determined.
Digesta was returned to the rumen immediately following data collection.

Sample and Statistical Analysis

Diets, dietary ingredients, orts, feces, and rumen digesta were analyzed for
nutrient concentrations. Samples were dried in a forced-air oven at 52‘C for 72 h,
ground with a Wiley mill (l-mm screen), and analyzed in duplicate for ash, NDF, ADF,
lignin, and indigestible NDF. Ash and fiber analyses were performed using methods
previously described (Chapter 4). Indigestible NDF was estimated as the NDF content
of samples following in vitro fermentation in buffered rumen fluid (Goering and Van
Soest, 1970) for 120 h without addition of pepsin. Cellulose content was calculated as
ADF minus lignin content; hemicellulose content was calculated as NDF minus ADF
content. Diets, dietary ingredients, orts, and feces were analyzed for CP (N x 6.25) using
a modified Kjeldahl (Hach) procedure (Watkins et al., 1987). Mineral analysis of diets
was performed using inductively coupled plasma spectroscopy (Northeast DHIA, Ithaca,
NY). Oven DM was used in calculations for intake and nutrient digestibility.
Concentrations of all nutrients, except DM, were expressed as a percentage of DM
determined fiom forced-air oven drying at 105’C.

In addition to oven drying, forage DM was also determined by toluene
distillation (AOAC, 1990). Frozen silage samples were thawed and blended with 100 ml
H20 to extract water soluble components. Following centrifugation at 26,000 x g for 30
min, sample supematent was measured for pH and analyzed for organic acids using

HPLC (Siegfried et al., 1984). The column used was Aminex HPX-87H, catalog

172
number 125-0140, 300 x 7.8 mm (Bio-Rad Laboratories, Richmond, CA). Column

temperature was 50‘C and solvent was .005 N H2804. Detection was by refractive
index (Waters 410, Millipore Corp., Milford, MA).

Rates of digestible NDF digestion for LDF and HDF alfalfa silages and diets
were determined from in vitro fermentation. Four samples of each silage and diet were
incubated for 0, 3, 6, 12, 18, 24, 48, 72, and 120 h in buffered rumen fluid and analyzed
for NDF content as earlier described for indigestible NDF analysis. The fractional rate
of digestion was estimated for each sample as the negative slope of simple linear
regression between the natural logarithm of digestible NDF remaining vs. time. This
model assumes first order digestion with respect to digestible NDF concentration
(Waldo, 1972). Values for times prior to 6 h were omitted from regression analysis to
remove effects of fermentation lag from the rate estimate. Indigestible NDF was initially
defined as NDF remaining after 120 h of fermentation and was subtracted from the total
NDF remaining at each time point to obtain concentrations of digestible NDF. Estimates
of indigestible NDF were iteratively adjusted to maximize the coefficient of
determination for regression, as suggested by Mertens (1993). In vitro NDF digestibility
and true DM digestibility were calculated for each sample following 12, 24, and 120 h of
incubation. Digestibility of NDF was calculated as the percentage of prefermented NDF
that had been fermented by 12, 24, or 120 h. True DM digestibility was calculated as
non-NDF DM plus NDF that was fermented, as a percent of DM, by 12, 24, or 120 h.
This calculation assumes that neutral detergent solubles have a digestibility of 100%
(Van Soest, 1965).

Fecal output and apparent total tract digestibility of dietary nutrients were
calculated using indigestible NDF as an internal marker (Cochran et al., 1986). Eating,
ruminating, and drinking activities were determined from collected behavior data with

previously developed algorithms (Chapter 2). Feeding behavior was summarized into

173
variables as performed by in Chapter 3. Across the three activities monitored, 90% of

the collected computer data were considered acceptable.

Rumen fluid was centrifuged at 13,000 x g for 30 min, decanted, and the
supematent centrifuged at 26,000 x g for an additional 30 min. Concentration of VFA in
the supematent was determined with HPLC using the method earlier described for the
analysis of silage acids. Concentration of major rumen VFA and their molar proportions
were calculated. Fractional rate of NDF passage from the rumen (kp), fi'actional rate of
digestible NDF digestion in the rumen (Rd), apparent rumen digestibility as a percent of
NDF intake, and apparent rumen digestibility as a percent of total tract NDF digestibility
were calculated using indigestible NDF as an internal marker and rumen pool size
estimates as described in Chapter 4.

All data were statistically analyzed using the specify model procedure of IMP
(1991). The model used was consistent for balanced simple crossover designs with two
treatments and two periods (Gill, 1978):

Yijk=tt+Ci+Pj+Tk+Eijk

where 11 = overall mean, ~
C; = random effect of cow (1 = 1 to 12),
Pj = random effect of period (j = 1 to 2),
Tk = fixed effect of treatment (k = 1 to 2),
Egg = residual, assumed normally distributed.

For variables measured from rumen cannulated cows, which included digesta
composition and digestion kinetics, only four cows were utilized (i = 1 to 4).
Interactions among the effects of cow, period, and treatment were assumed to be
negligible and were not included in the model. Type III sums of squares were used to
determine significance of treatment effects. When declared non-significant, treatment

effects had probability values > .10.

 

 

 

174
RESULTS

Forage and Diet Composition

Although the NDF content of both alfalfa silages was similar in samples obtained
prior to the lactation study (Table 1), analyses of samples taken throughout the duration
of the study as the forages were being fed indicated that concentrations of all nutrients,
except DM and hemicellulose, were significantly higher (P < .01) for LDF (Table 3).
Mean NDF content of LDF was 1.8% units greater than HDF (40.6 vs. 38.8%) because
of higher concentrations of both cellulose and lignin. Prior to the study, NDF
concentrations were 40.4 and 40.1% for LDF and HDF silages, respectively; therefore,
NDF content decreased for HDF silage between the two sampling times while LDF
silage remained unchanged. Silage pH was lower for HDF and was consistent with its
higher total organic acid content of 11.1% compared to 9.5% for LDF. Additionally,
HDF contained a higher lacticzacetic acid ratio. No other organic acids were detected in
either forage.

Differences in forage composition were reflected in the mean nutrient
composition of both treatment diets (Table 2). Mean NDF contents were 36.7% for LDF
and 34.9% for HDF diets, a difference of 1.8% units. Actual NDF content for LDF was
greater than the formulated target of 35% because concentrate NDF contents were higher
than anticipated. In vitro digestion characteristics of LDF and HDF silages and diets
sampled during the study are presented in Table 4; digestion of silages is summarized in
Figure 1. At each fermentation time, unfermented NDF as a percent of DM was
significantly larger (P < .01) for LDF silages and diets compared to HDF. Thus, NDF
digestibility and true DM digestibility were lower for LDF silages and diets compared to
_ HDF at each fermentation time. After 24 h, NDF digestibility was 38.3% for diet LDF
and 41.3% for diet HDF, a difference of 3% units. True DM digestibility after 24 h was
77.1% for diet LDF and 78.9% for diet HDF, a difference of 1.8% units. Approximately

50% of the difference in true DM digestibility between diets was due to differences in

Tal

 

 

(
\lrllll
]

 

175

Table 3. Nutrient composition of forages and concentrate sampled during animal

experiment and used to formulate low digestible fiber (LDF) and high digestible fiber
(HDF) diets. ~

 

 

 

 

_.Alfalta.silasL_
LDF HDF Concentrate
n 17 17 8
DM, %
52°C oven 32.8 33.9 89.2
Toluene 32.8 33.8
pHa 4.87 4.58
(% of DM)
OMa 89.4 89.8 88.4
NDFa 40.6 38.8 16.8
ADI“l 31.2 29.2 3.4
Hemicellulose 9.4 9.6 13.4
Cellulosea 24.0 22.8 2.6
Lignina 7.2 6.4 .8
CPa 24.7 23.3 20.3
Silage acidsa
Lactic 6.2 8.5
Acetic 3.3 2.6
Lactic:acetic ratio 1.9 3.3

 

aNutrient concentration between forages differs significantly (P < .01).

176

Table 4. In vitro NDF digestion characteristics of low digestible fiber (LDF) and high
digestible fiber (HDF) alfalfa silages and diets sampled during animal experiment. 13

 

 

Alfalfa silage, Dig;
LDF HDF LDF HDF

NDF, % of DM 40.6 38.8 36.7 34.9
Unferrnented NDF, % of DM

12 h 32.2 28.7 28.1 26.1

24 h 25.4 23.0 22.9 21.1

120 h 22.1 20.2 19.5 17.5
NDF digestibility, % of NDF

12 h 21.7 25.6 24.2 27.4

24 h 38.3 40.2 38.3 41.3

120 h 45.4 47.9 46.9 49.8
True DM digestibility, %

12 h 67.8 71.3 71.9 73.9

24 h 74.6 77.0 77.1 78.9

120 h 77.9 79.8 80.5 82.5
Rate of NDF digestion, h‘1 .112 .114 .126 .112

 

lDetermined from in vitro fermentation with buffered rumen fluid for time indicated.
2Except for rate of NDF digestion, digestion parameters between treatments differ
significantly for both forages and diets (P < .01).

177

 
  
  

 

 

 

 

 

 

 

 

40

g —I— LDF Silage

E . —e— HDF Silage

2 35d

23 .

u_ .

g .

1: 30-

¢ .

E

9‘ .

E .

.2 254

C 1

D l
i I

20...,.......?...#I-l-?

0 24 . 48 72 96 120

Fermentation time (h)

Figure 1. Digestion of low digestible fiber (LDF) and high digestible fiber (HDF)
alfalfa silage NDF determined from in vitro fermentation in buffered rumen fluid for
time indicated.

178
alfalfa silage NDF content, and 50% due to differences in alfalfa silage NDF

digestibility. Fractional rates of digestible NDF digestion in vitro were not different for
silages or diets, averaging .113 h'1 across silages and .119 h'1 across diets.
Cow Response

Daily milk production was significantly lower (P < .02) for cows receiving diet
LDF compared to HDF (Table 5). These early lactation cows averaged 36.3 kg/d of
milk when offered LDF and 38.2 kg/d when offered HDF, a difference of 1.9 kg/d. Milk
fat, protein, and lactose contents, however, tended to be greater for diet LDF, resulting in
no significant difference between diets for SCM production. Production of milk fat and
protein were similar for both diets, averaging 1.47 and 1.02 kg/d, respectively. Lactose
production was .08 kg/d higher (P < .08) for diet HDF compared to LDF. No
differences between diets were observed for cow BW (mean = 616 kg) or body condition
score (mean = 2.1).

Table 5. Milk production and body characteristics of 12 early lactation cows receiving
low digestible fiber (LDF) or high digestible fiber (HDF) diets.

 

 

 

Probability >

Variable LDF HDF SE F vahL
Production, kg/d

Milk 36.3 38.2 5 .02

SCM 34.9 35.8 .6 NS1

Fat 1.46 1.48 .04 NS

Protein 1.00 1.03 .02 NS

Lactose 1.74 1.82 .03 .08
Milk composition, %

Fat 4.03 3.88 .1 1 NS

Protein 2.76 2.71 .03 NS

Lactose 4.82 4.78 .03 NS
BW, kg 616 616 3 NS
Body condition score2 2.1 2.1 .1 NS

1NS = P > .10.

2Scale of 1 to 5; l = thin, 5 = obese.

179
Mean DMI was 19.4 kg/d for cows offered diet LDF and 20.4 kg/d when offered

diet HDF (Table 6), a difference of 1.0 kg/d (P < .01). Intake of NDF was identical
between treatments and averaged 7.1 kg/d. Intake of indigestible NDF was higher (P <
.09) for diet LDF because this treatment contained higher concentrations of indigestible
NDF. Fecal DM and NDF composition was similar between treatments, but indigestible
NDF content of feces was significantly higher (P < .01) for diet LDF. Daily fecal
excretion estimates of DM and NDF tended to be greater for cows receiving diet LDF.
Apparent in vivo total tract digestibility of diet HDF was higher than diet LDF for the
nutrients of DM, OM, NDF, ADF, hemicellulose, and cellulose (P S .08). Mean
increase in digestibility across all nutrients was 4.2%. Similar digestibilities were
obtained if lignin was used as an internal marker to calculate fecal output but treatment
differences were slightly larger (6.9%, P < .01; Appendix Table 12).

Few differences in eating, ruminating, and drinking behavior were observed
(Table 7). Cows averaged 12.5 eating bouts/d that were 27.0 min in length, for a total
time spent eating of 327 min/d. Mean meal size of DM tended to be smaller for diet
LDF compared to HDF and resulted in lower daily DMI for diet LDF as previously
described. Cows receiving diet LDF ruminated during 13.7 bouts/d compared with 14.5
bouts/d for diet HDF (P < .04). Mean ruminating bout length, however, was longer (P
< .03) for diet LDF (40.3 min) compared to diet HDF (38.2 min), resulting in no
difference between treatments in daily time spent ruminating (mean = 546 min/d). Time
spent eating or ruminating per unit of DM tended to be higher for diet LDF; total time
spent chewing per unit of DM was significantly higher (P < .08) for diet LDF. Number
of chews and chewing rates were not different between treatments for any activity. No
drinking parameters were different between treatments.

Characterization of rumen digesta from four rumen cannulated cows also
indicated few differences between diets. Volume of rumen digesta was greater 2 h

postfeeding compared to 2 h prefeeding, however, no differences across treatments were

Table 6. Nutrient intake, fecal composition, fecal output, and apparent total tract

digestibility for 12 early lactation cows receiving low digestible fiber (LDF) or high

 

 

digestible fiber (HDF) diets.1
Probability >
M QDF HDF SE F valng
Intake, kg/d
DM 19.4 20.4 .2 <.01
NDF 7.1 7.1 . 1 NS2
Indigestible NDF3 3.8 3.6 . 1 .09
CP 4.6 4.7 .1 NS
Fecal composition
DM, % 14.4 14.1 .2 NS
NDF, % of DM 57.1 56.7 .2 NS
Indigestible NDF, % of DM 52.2 50.4 .4 .01
Fecal output, kg/d
DM 7.3 7.1 .2 NS
NDF 4.2 4.0 . 1 NS
Nutrient digestibility, %
DM 62.6 65.2 .5 <.01
(1V1 63.5 66.2 .5 <.01
NDF 41.6 43.4 .6 .08
ADF 40.2 41.6 .7 NS
Lignin -1.3 -5.5 1.3 .04
Hemicellulose 45 .2 47.5 . 8 .07
Cellulose 52.8 54.8 .7 .06
CP 73.8 74.8 .5 NS

 

lCalculated using indigestible NDF as an internal marker.

2NS =P > .10.

3Neutral detergent fiber residue after 120 h in vitro fermentation.

181

Table 7. Eating, ruminating, and drinking activities for 12 early lactation cows receiving
low digestible fiber (LDF) or high digestible fiber (HDF) diets.

 

 

Probability >

Variable LDF HDF SE F vain:
Eating bouts, /d 12.5 12.5 .4 NS1
Meal size, kg

DM 1.60 1.64 .05 NS

NDF .58 .57 .02 NS
Eating bout length, min 27.1 26.9 1.2 NS
Eating time

min/d 327 327 7 NS

min/kg DM 17.5 16.5 .4 NS

min/kg NDF 47.9 47.5 1.3 NS
Eating chews, /d 18,503 19,347 702 NS
Eating chew rate, chew/min 56.2 59.0 1.3 NS
Water intake during meals, Ud 45.7 49.0 3.0 NS
Ruminating bouts, /d 13.7 14.5 .3 .04
Ruminating bout length, min 40.3 38.2 .6 .03
Ruminating time

min/d 545 547 8 NS

min/kg DM 29.3 27.7 .7 NS

min/kg NDF 80.2 80.1 2.1 NS
Ruminating chews, [(1 32,969 34,009 1059 NS
Ruminating chew rate,

chews/min 59.6 61.5 1.2 NS
Total chewing time

min/d 872 874 13 NS

min/kg DM 46.7 44.2 .9 .08

min/kg NDF 128 127 3 NS
Total chews, /d 51,472 53,356 1576 NS
Water intake, Ud 90.3 92.9 1.5 NS
Drinking bouts, /d 16.8 17.7 .7 NS
Drinking bout size, L 6.1 5.6 .3 NS
Drinkingtime,min/d 21.1 23.4 1.6 NS
Drinking rate, Umin 4.6 4.3 .2 NS

 

1NS =P > .10.

182
observed (Table 8). Mean rumen volume was 78.3 L across all determinations. Digesta

NDF content tended to be higher for diet LDF and indigestible NDF content was 3.9%
units higher (P < .09) for cows offered diet LDF. Average rumen pool size of nutrients
was not different between treatments but tended to be greater for diet LDF compared to
HDF for all nutrients measured.

In a pattern similar to rumen volume, rumen fluid pH was higher 2 h prefeeding
compared to 2 h postfeeding (Table 9). Although not significant, rumen pH tended to be
higher for cows offered diet LDF (6.84 vs. 6.73). Differences in pH corresponded to a
tendency for lower total VFA concentrations for diet LDF compared to HDF; differences
between treatments 2 h postfeeding were significant (P < .03). The molar proportion of
propionate in rumen fluid was lower (P < .08) for diet LDF, while acetate tended to be
higher. Consequently, the acetatezpropionate ratio was higher (P < . 10) for diet LDF
(3.4) compared to diet HDF (3.0). No differences between diets were detected for in
vivo rumen NDF digestion kinetic and digestibility parameters (Table 10). Average
NDF kp was .028 and average digestible NDF kd was .058. Apparent digestibility of
NDF in the rumen averaged 32.4% of NDF intake and contributed 76.7% of total tract
NDF digestibility.

DISCUSSION

Current systems of forage analysis and ration formulation do not account for
differences in fiber digestibility (NRC, 1989). Fermentation of fiber in the rumen can
contribute a significant portion of a dairy cow's total digestible energy intake (Galyean
and Goetsch, 1993). Alfalfa forages with similar NDF concentrations have been shown
to range in NDF digestibility from 25 to 55% after 30 h of in vitro fermentation (Allen et
al., 1991), that is equivalent to a potential range in energy content of .33 Mcal of NEng
of DM for forage containing 45% NDF. Forage fiber digestibility may also influence
feed intake, especially if intake is limited by the physical capacity of the reticulorumen.

Increasing the fraction of fiber that is potentially digestible subjects a larger portion of

183

Table 8. Rumen digesta characteristics for 4 early lactation cows receiving low digestible
fiber (LDF) or high digestible fiber (HDF) diets.

 

 

 

Probability >
Variable LDF HDF SE F vain:
Volume, L
2 h prefeeding 75.5 74.8 2.8 NS1
2 h postfeeding 82.5 80.5 4.7 NS
Average 79.0 77 .6 3.8 NS
Composition2
DM, % 13.8 13.7 .1 NS
NDF, % of DM 67.1 64.6 1.0 NS
Indigestible NDF,3 % of DM 52.6 48.7 .9 .09
Mass,2 kg
Wet digesta 71.2 69.7 3.7 NS
DM 9.8 9.6 .4 NS
(1% 8.8 8.6 .4 NS
NDF 6.6 6.2 . 3 NS
ADF 4.7 4.3 .2 NS
Lignin 1.5 1.3 .1 NS
Hemicellulose 1.9 1.9 . 1 NS
Cellulose 3.2 3.0 .2 NS
Indigestible NDF 5.1 4.7 . 2 NS
Density,2 kg/L . 91 .90 .01 NS
1NS = P > .10.

2Average of pre and postfeeding samples.
3Neutral detergent fiber residue after 120 h in vitro fermentation.

184

Table 9. Rumen fluid pH and VFA concentration of digesta from 4 early lactation cows

receiving low digestible fiber (LDF) or high digestible fiber (HDF) diets.

 

 

 

Probability >
Xan'ahlp LDF IIDF SE F valua
pH
2 h prefeeding 6.93 6.78 .04 NS1
2 h postfeeding 6.74 6.69 .07 NS
Average 6.84 6.73 .06 NS
Total VFA concentration, mM
2 h prefeeding 137.5 139.6 5.1 NS
2 h postfeeding 137.5 148.3 1.4 .03
Average 137.5 143.9 3.2 NS
VFA molar proportions2 --(% of total VFA)--
Acetate 65.0 62.6 .6 NS
Propionate 19.0 20.7 .4 .08
Butyrate 8.5 8.7 .2 NS
Valerate 2.6 3.2 .1 NS
Branched-chain3 4.9 4.8 .2 NS
Acetaterpropionate ratio 3.4 3.0 . 1 .10
1N8 = P > .10.

2Average of pre and postfeeding samples.

3isobutyrate + isovalerate.

185

Table 10. Rumen digestion kinetics of NDF and rumen apparent digestibility of NDF
from 4 early lactation cows receiving low digestible fiber (LDF) or high digestible fiber
(HDF) diets.l

 

 

 

 

Probability >
Xariablc LDE HDF SE F vain;
Digestion kinetics (h'l)
Fractional passage rate

2 h prefeeding .028 .029 .001 NS2

2 h postfeeding .026 .028 .001 NS

Average .027 .029 .001 NS

Fractional digestion rate

2 h prefeeding .063 .063 .007 NS

2 h postfeeding .055 .054 ‘ .004 NS

Average .058 .058 .004 NS
Rumen apparent digestibility

------------ (% of NDF intake)----------

2 h prefeeding 32.2 33.8 1.5 NS

2 h postfeeding 31.4 32.4 . .3 NS

Average 31.7 33.1 .8 NS

----(% of total tract digestibility mm

2 h prefeeding 77.5 78.5 2.2 NS

2 h postfeeding 75.6 75.3 .9 NS

Average 76.5 76.9 .7 NS

 

lCalculated using indigestible NDF as an internal marker and rumen pool size
estimates.
2N3 = P > .10.

186
the rumen fiber pool to both digestion and passage, which may increase the turnover rate

of fiber within the rumen and allow for additional fiber consumption and DMI. If
improved fiber digestion increases both DMI and energy density of the forage, a
substantial increase in digestible DMI may be realiud, as shown by Muller et al., 1972.

The first objective of this study was to obtain two alfalfa forages with similar
NDF and protein concentrations but different NDF digestibilities to isolate the treatment
effect of fiber digestibility. Equal NDF contents of the forages would also allow
identical forage:concentrate ratios and one concentrate supplement to be used for both
dietary treatments. Previous attempts to examine forage fiber digestibility have been
unable to unconfound the effects of fiber digestibility, NDF content, fiber source, forage
protein content, concentrate composition, or forage:concentrate ratio (Varga et al., 1984;
Miller et al., 1990; Robinson and McQueen, 1992). Robinson and McQueen (1992)
most successfully isolated the effect of NDF digestibility using two different maturities
of timothy silage, however, this effect was still confounded with amount of a second
forage source and with source of concentrate.

Selected forages initially met requirements that enabled an unconfounded fiber
digestibility feeding trial (Table 1). Unfortunately, NDF content of HDF silage
decreased between initial sampling and forage feedout, that was probably due to either a
loss of cellulose between these two sampling periods (Table 3) or sampling bias during
initial forage screening. The change in NDF content for silage HDF was predominantly
due to a loss of digestible NDF (Figure 2). Loss of digestible fiber also was noted
during fermentation for both silages based on samples taken prior to ensiling (Chapter
5). Loss of fiber during silage fermentation has been previously reported (Morrison,
1988). Clear interpretations of the results of this study are diffith to make since both
NDF content and digestibility appeared to differ between diets at feedout. The
difference in NDF content, however, may be related directly to the higher concentration

of digestible fiber in HDF silage prior to the start of the trial. Consequently,

187

23

 

 

. Pre

22 Post 1

 

 

 

 

 

 

21

I

20

 

 

19

aw \\\ ,s\

DNDF ' INDF ' DNDF l DF
LDF HDF

 

NDF concentration (% of DM)

 

 

We
\\§_.

 

s\\\\\\ ‘

 

Z

 

 

Figure 2. Concentration of digestible NDF (DNDF) and indigestible NDF (INDF)
for low digestible fiber (LDF) and high digestible fiber (HDF) alfalfa silages following
120 h in vitro fermentation for samples taken prior to conunencement of the lactation
study (Pre) and samples taken during the lactation study (Post).

 

188
improvements in cow performance from feeding diet HDF may ultimately reflect the

advantages of high fiber digestibility.

Lower lignin concentration for HDF silage was expected, because lignification of
the plant cell wall is considered to be the primary limitation to its digestibility (Smith et
al., 1972; USDA-ARS, 1993). Slight differences in CF content between the forages and
diets was not a major concern. Both diets were considered to be in CP excess for the
level of milk production expected. Amount of rumen undegraded protein was estimated
to support 40 kg/d of milk, also above the expected milk production fi'om these energy-
lirnited diets.

Many of the cow responses to these diets are similar to those obtained with the
35% NDF diet fed in the trial reported in Chapter 4. Cows were estimated to be in
negative energy balance because mean energy output (35.4 Mcal/d of NE) was
equivalent to a dietary energy concentration of 1.80 Meal of NEng of DM, a value
likely to be an overestimate of the true energy content of these 83% alfalfa-based diets.
Changes in BW were not calculated because they do not accurately reflect changes in
body tissue for early lactation cows. During this period, changes in BW are confounded
with increases in rumen digesta mass. Although high forage diets were used, high milk
production was achieved, suggesting that silages were of excellent quality. Despite the
relative small difference between diets LDF and HDF for NDF concentration (1.8%
units) and NDF digestibility (3% units), significant increases in milk production were
obtained with diet HDF. Other studies investigating fiber digestibility have also noted
irnprovements in milk yield with higher digestibility (V arga et al., 1984; Robinson and
McQueen, 1992), although others have not (Miller et al., 1990; Feng et al., 1993;
McQueen and Robinson, 1993). Milk component concentrations tended to be higher for
diet LDF, therefore, component and SCM yields were not different between diets.
Rumen VFA suggest a trend toward increased microbial activity for diet HDF;

differences in acetatezpropionate ratio support the trend for higher milk fat content for

189
diet LDF. Varga et al. (1984) also observed higher milk fat content for a diet with lower

fiber digestibility.

The most significant result in this study was the improvement in DMI with diet
HDF. This increase was accompanied by increases in apparent in vivo DM and NDF
digestibilities. Digestible DMI was calculated for each diet (Figure 3); approximately
90% of the improvement in digestible DMI with diet HDF was due to increased DMI
and 10% was due to increased NDF digestibility. Almost all other studies with dairy
cows have detected no increase in DMI with increases in fiber digestibility (Varga et al.,
1984; Miller et al., 1990; Robinson and McQueen, 1992; Feng et al., 1993, McQueen
and Robinson, 1993), with the exception of Llamas-lamas and Combs (1990). Lack of
DMI increases may have been due to their use of cows whose intakes were not limited
by rumen fill. We expected that intake of NDF would be higher and intake of
indigestible NDF would be the same with diet HDF compared to diet LDF, if intake was
limited by the presence of indigestible fiber in the rumen. On the contrary, intakes of
NDF were equal while intakes of indigestible NDF tended to be lower with diet HDF
(Table 6). These results suggest that increases in DMI were due to the lower NDF
content of diet HDF.

Rumen digesta volume and rumen mass of NDF were similar between diets
(Table 8), a finding similar to that of others (Llamas-lamas and Combs, 1990; Robinson
and McQueen, 1992). Lower concentrations of indigestible NDF in diet HDF tended to
promote lower concentrations of indigestible NDF in rumen digesta, however this
decrease was not sufficient to significantly reduce rumen pool size of this component.
Rumen turnover time of NDF was not significantly faster for diet HDF (mean = 21.6 h),
consequently, changes in NDF digestibility did not increase clearance of NDF from the
rumen. No differences in fi'actional rates of NDF digestion in and passage from the
rumen were observed, and were not expected since these diets did not differ in in vitro

rates of NDF digestion. Higher NDF digestibility may have resulted in a greater

190
[j DNDF intake

DNDS intake

 

 

 

 

 

14- w 1.5-
124 296 3.08 124
10-1 ' O12

 

7 ‘ 0.94

0.6~

w.

—L
0
0|

 

'L\\

Digestible DMI (kg/d)
3"

0.3.

0 a 0 a

LDF HDF HDF-IF:l

 

 

 

 

 

 

\a

 

 

Figure 3. Contribution to digestible DMI from digestible NDF (DNDF) intake and
digestible neutral detergent solubles (DNDS) intake for low digestible fiber (LDF) and
high digestible fiber (HDF) diets offered to 12 early lactation cows.

191
digestible energy content for diet HDF, but appears to have had no affect on DMI.

Differences in NDF digestibility between diets were small, and rumen digesta data were
from only four cows, therefore, these results should be viewed as preliminary for
describing the effect of fiber digestibility on DMI of early lactation cows.

Diet HDF resulted in more rumination bouts of shorter duration compared to diet
LDF. Under conditions where intake was limited by rumen fill (Chapter 4), greater fill
also was associated with more rumination bouts of shorter duration. Time spent
chewing per unit of DM was higher for diet LDF, another characteristic of fill-limiting
diets. These data do not provide substantial support to argue that either diet was more
limited by rumen capacity than the other.

CONCLUSIONS

Ensiled alfalfa forages that differed in NDF digestibility by 5% units prior to
commencement of a study with early lactation cows were offered as part of a high fiber
ration. The alfalfa with higher fiber digestibility increased milk production and DMI by
1.9 and 1.0 kg/d, respectively. At the time of feeding, diets differed in both NDF content
(1.8% units) and NDF digestibility (3% units), perhaps due to loss of digestible NDF in
the higher quality forage during the trial. Equal NDF intake for these diets suggests that
DMI was higher because of lower NDF concentration in the higher quality forage. Both
DMI and NDF digestibility contributed to higher consumption of digestible DMI.

Isolating NDF digestibility as the sole treatment effect is difficult. If increases in
fiber digestibility prior to ensiling promoted decreases in fiber content after ensiling,
benefits to harvesting forages with higher fiber digestibility may be substantial. Further
experimentation is required before differences in fiber digestibility can be claimed to
have no effect on intake. Forages with larger differences in fiber digestibility at the time
of feeding to cows with intake limited by physical fill are required to more closely
examine this hypothesis. These forages may be difficult to obtain at only one location;

cooperation among researchers at different locations may be required.

4‘

EPILOGUE

This research project sought to fulfill several objectives relative to the influence
of forage quality on feed intake by cows in early stages of lactation. To adequately
measure these effects, a computerized data acquisition system to monitor feeding
behavior and rumen function was successfully developed and validated. Utilization of
current computer technology enhances one's ability to make multiple observations on
several animals simultaneously. The critical component of this system is the
interpretive algorithm that summarizes large streams of data into variables that can
easily be compared to evaluate treatment effects. At this time, five lactation studies
have been conducted using this system to monitor feeding behavior. Due to the
intensity in which variables can be measured, this, and similar systems like this, will
serve as a valuable tool to investigate interactions between feed quality and animal
response. Future experiments should seek to apply these techniques to free-roaming
cows in group housing.

Before applying this measurement system to critical lactation experiments, a
preliminary lactation study was conducted to examine the variation across cows and
across days for several feeding behavior variables. Little information has been reported
in the literature about such variation and how it might be used to design experiments to
maximize statistical power and inference. Significant variation between cows was
observed for many variables, including meal size, eating bouts, and total time spent
chewing. The mean coefficient of variation for all variables was 21%. To detect

contrast differences of 10% of means for all variables examined with 80% probability

192

193
required the use of 12 cows measured for 5 d in a Latin square design. Consequently,

all subsequent experiments were designed in such a fashion.

Two lactation studies with cows 2 wk into lactation were conducted using these
monitoring tools and additional measurements common to nutrition studies, such as
rumen digesta evacuation, fecal collections, and diet analysis, to evaluate the effect of
fiber concentration of the diet or extent of fiber digestibility on feed intake. In trial 1,
cows were challenged with rumen fill in the form of dietary NDF or inert bulk to
determine if filling characteristics of forage diets varied with fiber content, and if cows
adapted to rumen fill by altering their feeding behavior. Addition of inert bulk
decreased intake for cows receiving 35% NDF diets but not for cows receiving 25%
NDF diets. Volume of rumen digesta plus inert bulk was similar for high fiber
treatments whether or not inert bulk was present. Finally, fiber and inert bulk addition
increased NDF passage from the rumen in a similar manner. These data support the
hypothesis that cows receiving high fiber diets have intakes limited by the physical
capacity of the reticulo-rumen. Changes in feeding behavior or rumen function were
insufficient to maintain intake under conditions of high rumen fill. Added fiber and
inert bulk altered aspects of rumination, rumen fermentation, and reticular motility in a
similar fashion, demonstrating that NDF has rumen-filling characteristics.

In trial 2, cows were offered two diets that differed in the source of alfalfa silage
used. Two silages were harvested in the summer of 1992 that contained similar
concentrations of NDF (40%) but different digestibilities of NDF (40 vs. 45% after 24 h
in vitro fermentation). High forage mixed diets were fed to determine if higher NDF
digestibility resulted in increased intake and production. Higher fiber digestibility
increased milk production and intake by 1.9 and 1.0 kg/d, respectively. At the time of
feeding, however, diets differed in both NDF content (1.8% units) and NDF
digestibility (3% units), perhaps due to loss of digestible NDF in the higher quality
forage during the trial. Equal NDF intake for these diets suggests that DMI was

194
increased because of lower NDF concentration in the higher quality forage. Isolating

NDF digestibility as the sole treatment effect was difficult. If increases in fiber
digestibility prior to ensiling promote decreases in fiber content after ensiling, benefits
to harvesting forages with higher fiber digestibility may be substantial.

In summary, forage fiber did appear to limit voluntary feed intake in early
lactation dairy cows because of rumen fill. Future research is needed to determine what
other dietary factors influence the fill value of different diets and whether cows of
varying size, shape, stage of lactation, or some other trait vary in their ability to
accommodate rumen fill. Low fiber digestibility may have also limited intake by cows,
however, this treatment effect was confounded with fiber content of the diet. Forages
with larger differences in fiber digestibility at the time of feeding are required to more
closely examine this hypothesis. These forages may be difficult to obtain at only one

location; cooperation among researchers at different locations may be required.

LIST OF REFERENCES

LIST OF REFERENCES

Abrams, S.M., H.W. Harpster, P.J. Wangsness, J.S. Shenk, E. Keck, and J .L.
Rosenberger. 1987. Use of a standard forage to reduce effects of animal variation on
estimates of mean voluntary intake. J. Dairy Sci. 70: 1235.

Adams, GB, and J.M. Forbes. 1981. Additivity of effects of ruminal acetate and either
portal propionate or rumen distention on food intake in sheep. Proc. Nutr. Soc. 40:44A.

Agosin, E., B. Monties, and E. Odier. 1985. Structural changes in wheat straw
components during decay by lignin-degrading white-rot fungi in relation to
improvement of digestibility for ruminants. J. Sci. Food Agric. 36:925.

Aitchison, E.M., M. Gill, M.S. Dhahoa, and DR Osbourn. 1986. The effect of
digestibility and forage species on the removal of digesta from the rumen and the
voluntary intake of hay by sheep. Br. J. Nutr. 56:463.

Albright, J.L. 1993. Feeding behavior of dairy cattle. J. Dairy Sci. 76:485.

Alhadhrami, G., and J.T. Huber. 1992. Effects of alfalfa hay of varying fiber fed at 35
or 50% of diet on lactation and nutrient utilization by dairy cows. J. Dairy Sci. 75:3091.

Allen, MS. 1991. Carbohydrate nutrition. Page 327 in Veterinary Clinics of North
America: Food Animal Practice. Vol 7. W.B. Saunders, Philadelphia, PA.

Allen, M.S., and DR. Mertens. 1988a. Effect of diet fiber level and forage source on
intake and milk production of Holstein cows in early or late lactation. J. Dairy Sci.
71(Supp 1): 176. (Abstr.)

Allen, M.S., and DR. Mertens. 1988b. Evaluating constraints on fiber digestion by
rumen microbes. J. Nutr. 118:261.

Allen, M.S., O.B. Hesterman, and K.A. O'Neil. 1991. Relationships among alfalfa
maturity, fiber fractions, and in vitro fiber digestibility for first and subsequent cuttings.
Page B9 in Abstracts of Inter. S ymp. on Forage Cell Wall Structure and Digestibility.
USDA-ARS, US Dairy Forage Research Center, Madison, WI.

Allen, M.S., K.A. O'Neil, R.G. Dado, and R.A. Kohn. 1992. Variation in fiber content
and fiber digestibility of alfalfa and corn forages. Page 21 in Proceedings of lst Tri-
State Dairy Nutrition Conference. Purdue University, Fort Wayne, IN.

Aman, P. 1993. Composition and structure of cell wall polysaccharides in forages. Page

183 in Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D.
Hatfield, and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

195

196

Andersson, M., J. Schaar, and H. Wiktorsson. 1984. Effects of drinking water flow rates

and social rank on performance and drinking behaviour of tied-up dairy cows. Livest.
Prod. Sci. 11:599.

Anil, M.H., J .M. Forbes, and J .N. Mbanya. 1987. Additive effects of acetate,
propionate, and distention of the rumen on hay intake by lactating cows. J. Physiol.
386:61P.

Association of Official Analytical Chemists. 1990. Official Methods of Analysis. 15th
ed. AOAC, Washington, DC.

Bae, D.H., J.G. Welch, and A.M. Smith. 1981. Efficiency of mastication in relation to
hay intake by cattle. J. Anim. Sci. 52: 1371.

Bae, D.H., J .G. Welch, and BE. Gilman. 1983. Mastication and rumination in relation
. to body size of cattle. J. Dairy Sci. 66:2137.

Baile, C.A., and J .M. Forbes. 1974. Control of feed intake and energy balance in
ruminants. Physiol. Rev. 54:160.

Baile, C.A., and CL. McLaughlin. 1987. Mechanisms controlling feed intake in
ruminants: a review. J. Anim. Sci. 64:915.

Balch, CC. 1958. Observations on the act of eating in cattle. Br. J. Nutr. 12:330.

Balch, CC. 1971. Proposal to use time spent chewing as an index of the extent to which
diets for ruminants possess the physical property of fibrousness characteristic of
roughages. Br. J. Nutr. 26:383.

Balch, CC, and RC. Campling. 1962. Regulation of voluntary food intake in
ruminants. Nutr. Abstr. Revs. 32:669.

Bauman, D.E., S.N. McCutcheon, W.D. Steinhour, P.J. Eppard, and SJ. Sechen. 1985.
Sources of variation and prospects for improvement of productive efficiency in the
dairy cow: A review. J. Anim. Sci. 60:583.

Baumgardt, B.R., A.D. Peterson, and M. Clancy. 1973. Feeding behavior by sheep fed
complete diets. Fed. Proc. 32:899. (Abstr.)

Baumont, R., C.H. Malbert, and Y. Ruckebusch. 1990. Mechanical stimulation of
rumen fill and alimentary behaviour in sheep. Anim. Prod. 50:123.

Beauchemin, K.A. 1991a. Effects of dietary neutral detergent fiber concentration and
alfalfa hay quality on chewing, rumen function, and milk production of dairy cows. J.
Dairy Sci. 74:3140.

Beauchemin, K.A. 1991b. Ingestion and mastication of feed by dairy cattle. Page 439 in
Veterinary Clinics of North America: Food Animal Practice. Vol 7. W.B. Saunders,
Philadelphia, PA.

Beauchemin, K.A., and JG. Buchanan-Smith. 1989. Effects of dietary neutral detergent
fiber concentration and supplementary long hay on chewing activities and milk
production of dairy cows. J. Dairy Sci. 72:2288.

I;

197

Beauchemin, K.A., S. Zelin, D. Genner, and LG. Buchanan-Smith. 1989. An automatic
system for quantification of eating and ruminating activities of dairy cattle housed in
stalls. J. Dairy Sci. 72:2746.

Beauchemin, K.A., B.I. Farr, and L.M. Rode. 1991. Enhancement of the effective fiber
content of barley-based concentrates fed to dairy cows. J. Dairy Sci. 74:3128.

Bernard, J.K., and W.W. McNeill. 1991. Effect of high fiber energy supplements on
nutrient digestibility and milk production of lactating dairy cows. J. Dairy Sci. 74:991.

Berndtson, W.B. 1991. A simple, rapid and reliable method for selecting or assessing
the numbers of replicates for animal experiments. J. Anim. Sci. 69:67.

Bines, LA. 1971. Metabolic and physical control of food intake in ruminants. Proc.
Nutr. Soc. 30:116.

Bines, LA. 1976. Regulation of food intake in dairy cows in relation to milk production.
Livest. Prod. Sci. 3:115.

Bines, J .A., and A.W.F. Davey. 1970. Voluntary intake, digestion, rate of passage,
amount of material in the alimentary tract and behaviour in cows receiving complete
diets containing straw and concentrates in different proportions. Br. J. Nutr. 24: 1013.

Bines, J .A., P.A. Jones, and DJ. Napper. 1973. The effect of a long-term reduction in
rumen capacity on the intake of hay by a cow. Proc. Nutr. Soc. 32:75A.

Blaxter, K.L., N. Graham, and F.W. Wainman. 1956. Some observations on the
digestibility of food by sheep, and related problems. Br. J. Nutr. 10:69.

Brewer, B.A., B.R. Chapman, and C.J. Sniffen. 1977. Apparatus for measurement of
ingestive behavior in dairy cattle. J. Dairy Sci. 60:985.

Briceno, J.V., H.H. Van Horn, B. Harris, Jr., and C.J. Wilcox. 1987 . Effects of neutral
detergent fiber and roughage source on dry matter intake and milk yield and
composition of dairy cows. J. Dairy Sci. 70:298.

Broster, W.H. 1972. Effect on milk yield of the cow of the level of feeding during
lactation. Dairy Sci. Abstr. 34:265.

Brown, C.A., P.T. Chandler, and LB. Holter. 1977. Development of predictive
equations for milk yield and dry matter intake in lactating cows. J. Dairy Sci. 60: 1739.

Buxton, D.R., and MD. Casler. 1993. Environmental and genetic effects on cell wall
composition and digestibility. Page 685 in Forage Cell Wall Structure and Digestibility.
Jung, H.G., D.R. Buxton, R.D. Hatfield, and J. Ralph, eds. ASA-CSSA-SSSA,
Madison, WI.

Byatt, J.C., P.J. Eppard, L. Munyakazi, R.H. Sorbet, J.J. Veenhuizen, D.F. Curran, and
R.J. Collier. 1992. Stimulation of milk yield and feed intake by bovine placental
lactogen in the dairy cow. J. Dairy Sci. 75:1216.

198

Bywater, AC. 1984. A generalised model of feed intake and digestion in lactating
cows. Agric. Systems 13:167.

Cameron, M.G., G.C. Fahey, Jr., J .H. Clark, NR. Merchen, and LL. Berger. 1991.
Effects of feeding alkaline hydrogen peroxide-treated wheat straw-based diets on
intake, digestion, ruminal fermentation, and production responses by mid-lactation
dairy cows. J. Anim. Sci. 69:1775.

Campling, R.C., and CC. Balch. 1961. Factors affecting the voluntary intake of foods
by cows. 1. Preliminary observations on the effect, on voluntary intake of hay, of
changes in the amount of the reticulo-rumen contents. Br. J. Nutr. 15:523.

Campling, R.C., and M. Freer. 1966. Factors affecting the voluntary intake of food by
cows. 8. Experiments with ground, pelleted roughages. Br. J. Nutr. 20:229.

Campling, R.C., and C.A. Morgan. 1981. Eating behaviour of housed dairy cows-a
review. Dairy Sci. Abstr. 43:57.

Campling, R.C., M. Freer, and CC. Balch. 1961. Factors affecting the voluntary intake
of food by cows. 2. The relationship between the voluntary intake of roughages, the
amount of digesta in the reticulorumen and the rate of disappearance of digesta from the
alimentary tract. Br. J. Nutr. 15:531.

Carr, SB, and DR. Jacobson. 1967. Intraruminal addition of mass or removal of rumen
contents on voluntary intake of the bovine. J. Dairy Sci. 50:1814.

Carter, R.P., and W.L. Grovum. 1990. A review of the physiological significance of
hypertonic body fluids on feed intake and ruminal function: salivation, motility, and
microbes. J. Anim. Sci. 68:2811.

Chacon, E., T.H. Stobbs, and R.C. Sandland. 1976. Estimation of herbage consumption
by grazing cattle using measurements of eating behavior. I. Br. Grassl. Soc. 31:81.

Chandler, P.T., and H.W. Walker. 197 2. Generation of nutrient specifications for dairy
cattle for computerized least cost ration formulation. J. Dairy Sci. 55: 1741.

Chemey, D.J.R., and J.H. Chemey. 1991. Low-lignin, brown-midrib genotypes and
their potential for improving animal performance. Page 13 in Proceedings of Cornell
nutrition conference for feed manufacturers. Ithaca, NY.

Chemey, J.H., D.J.R. Chemey, and DR. Mertens. 1988. Fiber composition and
digestion kinetics in grass internodes as influenced by particle size. J. Dairy Sci.
7 1:21 12.

Chemey, D.J.R., J.H. Cherney, and R.F. Lucey. 1993. In vitro digestion kinetics and
quality of perennial grasses as influenced by forage maturity. J. Dairy Sci. 76:790.

Chesson, A. 1988. Lignin-polysaccharide complexes of the plant cell wall and their
effect on microbial degradation in the rumen. Anim. Feed Sci. Technol. 21:219.

Clark, P.W., and LE. Armentano. 1993. Effectiveness of neutral detergent fiber in
whole cottonseed and dried distillers grains compared with alfalfa haylage. J. Dairy Sci.
76:2644.

199

Clarke, T., P.C. Flinn, and A.A. McGowan. 1982. Low cost pepsin-cellulase assays for
prediction of digestibility of herbage. Grass For. Sci.37:147.

Cleale, R.M., and LS. Bull. 1986. Effect of forage maturity on ration digestibility and
production by dairy cows. J. Dairy Sci. 69:1587.

Cochran, R.C., D.C. Adams, J.D. Wallace, and ML. Galyean. 1986. Predicting
digestibility of different diets with internal markers: Evaluation of four potential
markers. J. Anim. Sci. 63:1476.

Colenbrander, V.F., D.L. Hill, M.L. Eastridge, and DR. Mertens. 1986. Formulating
dairy rations with neutral detergent fiber. 1. Effect of silage source. J. Dairy Sci.
69:2718.

Colenbrander, V.F., C.H. Noller, and R.J. Grant. 1991. Effect of fiber content and
particle size of alfalfa silage on performance and chewing behavior. J. Dairy Sci.
74:2681.

Conrad, HR. 1966. Symposium on factors influencing the voluntary intake of herbage
by ruminants: Physiological and physical factors limiting feed intake. J. Anim. Sci.
25:227.

Conrad, H.R., A.D. Pratt, and J.W. Hibbs. 1964. Regulation of feed intake in dairy
cows. 1. Change in importance of physical and physiological factors with increasing
digestiblity. J. Dairy Sci. 47:54.

Crampton, E.W., E. Donefer, and LE. Lloyd. 1960. A nutritive value index for forages.

J. Anim. Sci. 19:538.

Custodio, A.A., R.W. Blake, P.F. Dahm, T.C. Cartwright, G.T. Schelling, and CE.
Coppock. 1983. Relationships between measures of feed efficiency and transmitting
ability for milk of Holstein cows. I . Dairy Sci. 66: 1937.

Dado, R.G., and MS. Allen. 1993. Intake limitations from rumen fill for cows
challenged with high fiber diets and inert rumen bulk. J. Dairy Sci. 76(Supp 1):210.
(Abstr.)

De Jong, A. 1986. The role of metabolites and hormones as feedbacks in the control of
food intake in ruminants. Page 459 in Control of digestion and metabolism in
ruminants. Milligan, L.P., W.L. Grovum, and A. Dobson, eds. Prentice-Hall, Inc.,
Englewood Cliffs, NJ.

DeBoever, J .L., H. Andries, D.L. DeBrabander, B.G. Cottyn, and F.X. Buysse. 1990.
Chewing activity of ruminants as a measure of physical structure-A review of factors
affecting it. Anim. Feed Sci. Technol. 28:281.

Dehority, BA. 1993. Microbial ecology of cell wall fermentation. Page 425 in Forage
Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D. Hatfield, and J.
Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Demment, M.W., and P.J. Van Soest. 1983. Body size, digestive capacity, and feeding
strategies of herbivores. Wirrrock International, Morrilton, AK.

’5

 

200

DePeters, EL, and NB. Smith. 1986. Forage quality and concentrate for cows in early
lactation. J. Dairy Sci. 69:135.

Deswysen, A.G., W.C. Ellis, K.R. Pond, W.L. Jenkins, and J. Connolly. 1987a. Effects
of monensin on voluntary intake, eating and ruminating behavior and ruminal motility
in heifers fed corn silage. J. Anim. Sci. 64:827.

Deswysen, A.G., W.C. Ellis, and KR. Pond. 1987b. Interrelationships among voluntary
intake, eating and ruminating behavior and ruminal motility of heifers fed corn silage. J.
Anim. Sci. 64:835.

 

Donefer, E., E.W. Crampton, and LE. Lloyd. 1960. Prediction of the nutritive value
index of a forage from in vitro rumen fermentation data. J. Anim. Sci. 19:545.

Dulphy, JP. 1971. Influence du poids vif et du niveau d'ingestion sur le comportement
alimentaire et merycique du mouton. Ann. Zootech. 20:477.

Egan, AR. 1972. Nutritional status and intake regulation in sheep. VII. Control of

voluntary intake of three diets and the responses to intraruminal feeding. Aust. J. Agric.
Res. 23:347.

Ehle, ER. 1984. Influence of feed particle density on particulate passage from rumen of
Holstein cow. J. Dairy Sci. 67:693.

Erdman, R.A., T.W. Moreland, and W.R. Stricklin. 1989. Effect of time of feed access
on intake and production in lactating dairy cows. J. Dairy Sci. 72:1210.

Erlinger, L.L., D.R. Tolleson, and C.J. Brown. 1990. Comparison of bite size, biting
rate and grazing time of beef heifers from herds distinquished by mature size and rate of
maturity. J. Anim. Sci. 68:3578.

Fahey, G.C., Jr., L.D. Bourquin, E.C. Titgemeyer, and D.G. Atwell. 1993. Postharvest
treatment of fibrous feedstuffs to improve their nutritive value. Page 715 in Forage Cell
Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D. Hatfield, and J. Ralph,
eds. ASA-CSSA-SSSA, Madison, WI.

Feng, P., W.H. Hoover, T.K. Miller, and R. Blauwiekel. 1993. Interactions of fiber and
nonstructural carbohydrates on lactation and ruminal function. J. Dairy Sci. 76: 1324.

Fisher, D.S., J.C. Burns, and KR. Pond. 1987 . Modeling ad libitum DM intake by
ruminants as regulated by distention and chemostatic feedbacks. J. Theor. Biol.
126:407.

Forbes, J.M. 1969. The effect of pregnancy and fatness on the volume of rumen
contents in the ewe. J. Agric. Sci., Camb. 72:119.

Forbes, J .M. 1977. Development of a model of voluntary food intake and energy
balance in lactating cows. Anim. Prod. 24:203.

Forbes, J.M. 1980. A model of the short-term control of feeding in the ruminant: Effects
of changing animal or feed characteristics. Appetite 1:21.

201
Forbes, J .M. 1986. The Voluntary Food Intake of Farm Animals. Butterworths, London.

Forbes, J .M. 1988. Metabolic aspects of the regulation of voluntary food intake and
appetite. Nutr. Res. Rev. 1:145.

Forbes, J.M., D. Jackson, C.L. Johnson. P. Stockill, and BS. Hoyl. 1986. A method for
the automatic monitoring of food intake and feeding behavior of individual cows kept in
a group. Res. Dev. Agric. 3:175.

Freeman, A.E. 1975. Genetic variation in nutrition of dairy cattle. Page 19 in The Effect
of Genetic Variation on Nutritional Requirements of Animals. Natl. Acad. Sci.,
Washington, DC.

Freer, M., and R.C. Campling. 1963. Factors affecting the voluntary intake of food by
cows. 5. The relationship between the voluntary intake of food, the amount of digesta in
the reticulo-rumen and the rate of disappearance of digesta from the alimentary tract
with diets of hay, dried grass or concentrates. Br. J. Nutr. 17:79.

Freer, M., and R.C. Campling. 1965. Factors affecting the voluntary intake of foods by
cows. 7. The behavior and reticular motility of cows given diets of hay, dried grass,
concentrates and ground, pelleted hay. Br. J. Nutr. 19:195.

Friend, T.H., and CE. Polan. 1974. Social rank, feeding behavior, and free stall
utilization by dairy cattle. J. Dairy Sci. 57:1214.

Friend, T.H., C.E. Polan, and M.L. McGilliard. 1977. Free stall and feed bunk
requirements relative to behavior, production, and individual feed intake in dairy cows.
J. Dairy Sci. 60:108.

Gallouin, F., and M. Focant. 1980. Bases physiologiques du comportement alimentaire
chez les ruminants. Reprod. Nutr. Develop. 20: 1563.

Galyean, M.L., and AL. Goetsch. 1993. Utilization of forage fiber by ruminants. Page
33 in Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D.
Hatfield, and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Gibson, JP. 1984. The effects of frequency of feeding on milk production of dairy
cattle: an analysis of published results. Anim. Prod. 38:181.

Gill, J.L. 1969. Sample size for experiments on milk yield. J. Dairy Sci. 52:984.

Gill, J .L. 197 8. Design and Analysis of Experiments in the Animal and Medical
Sciences. Iowa State Univ. Press, Ames.

Gill, M.S., and ME. Castle. 1983. The effects of the frequency of feeding concentrates
on milk production and eating behavior in Ayrshire dairy cows. Anim. Prod. 36:79.

Gill, S.S., H.R. Conrad, and J.W. Hibbs. 1969. Relative rate of in vitro cellulose
disappearance as a possible estimator of digestible dry matter intake. J. Dairy Sci.
52: 1687.

202

Glenn, BR, and DR. Waldo. 1993. Cell wall degradation in the ruminant-session
synopsis. Page 603 in Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R.
Buxton, R.D. Hatfield, and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Goering, H.K., and P.J. Van Soest. 1970. Forage fiber analysis (apparatus, reagents,
procedures, and some applications). Agric. Handbook No. 379. ARS-USDA,
Washington, DC.

Gordon, J .G. 1958. The effect of time of feeding upon rumination. J. Agric. Sci.
(Camb.) 51:81.

Grant, R.J., and SJ. Weidner. 1992. Effect of fat from whole soybeans on performance
of dairy cows fed rations differing in fiber level and particle size. J. Dairy Sci. 75:2742.

Grant, R.J., V.F. Colenbrander, and DR. Mertens. 1990. Milk fat depression in dairy
cows: Role of particle size of alfalfa hay. J. Dairy Sci. 73:1823.

Grovum, W.L. 1979. Factors affecting the voluntary intake of food by sheep. 2. The
role of distention and tactile input from compartments of the stomach. Brit. J. N utr.
42:425.

Grovum, W.L. 1987. A new look at what is controlling food intake. Page 1 in Feed
intake by beef cattle. Symp. Proc., Oklahoma State Univ. Stillwater.

Grummer, R.R., A.L. Jacob, and LA. Woodford. 1987. Factors associated with
variation in milk fat depression resulting from high grain diets fed to dairy cows. J.
Dairy Sci. 70:613.

Harb, M.Y., and R.C. Campling. 1985. Variation among pregnant, non-lactating dairy
cows in eating and ruminating behaviour, digestibility and voluntary intake of hay.
Grass Forage Sci. 40:109.

Harb, M.Y., V.S. Reynolds, and R.C. Campling. 1985. Eating behaviour, social
dominance and voluntary intake of silage in group-fed milking cattle. Grass Forage Sci.
40:1 13.

Harlan, D.W., J.B. Holter, and H.H. Hayes. 1991. Detergent fiber traits to predict

productive energy of forages fed free choice to nonlactating dairy cattle. J. Dairy Sci.
74: 1337.

Hatfield, RD. 1993. Cell wall polysaccharide interactions and degradability. Page 285
in Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D. Hatfield,
and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Heinrichs, A.J., and HR. Conrad. 1987. Measuring feed intake patterns and meal size
of lactating dairy cows. J. Dairy Sci. 70:705.

Holter, J.B., and W.B. Urban, Jr. 1992. Water partitioning and intake prediction in dry
and lactating Holstein cows. J. Dairy Sci. 75:1472.

Hooper, A.P., and LG. Welch. 1985. Effects of particle size and forage composition on
functional specific gravity. J. Dairy Sci. 68:1181.

 

 

 

203

Hyer, J.C., J .W. Oltjen, and M. L. Galyean. 1991. Development of a model to predict
forage intake by grazing cattle. J. Anim. Sci. 69:827.

Illius, A.W., and LI. Gordon. 1991. Prediction of intake and digestion in ruminants by a
model of rumen kinetics integrating animal size and plant characteristics. J. Agric. Sci.
(Camb.) 116:145.

Illius, A.W., and MS. Allen. 1993. Assessing forage quality using integrated models of
intake and digestion by ruminants. In Proceedings of National Conference on Forage
Quality, Evaluation, and Utilization. Lincoln, NE. (In press)

Jarrige, R., C. Demarquilly, J .P. Dulphy, A. Hoden, J. Robelin, C. Beranger, Y. Geay,
M. Journet, C. Maltere, D. Micol, and M. Petit. 1986. The INRA "fill unit" system for
predicting the voluntary intake of forage-based diets in ruminants: A review. J. Anim.
Sci. 63: 1737.

J aster, E.H., and MR. Murphy. 1983. Effects of varying particle size of forage on
digestion and chewing behaviour of dairy heifers. J. Dairy Sci. 66:802.

IMP® User's Guide. 1991. Version 2.0.5. SAS Inst, Inc., Cary, NC.

Johnson, J.W., and JD. Sutton. 1968. Continuous recording of the pH in the bovine
rumen. Br. J. Nutr. 22:303.

Johnson, TR, and D.K. Combs. 1991. Effects of prepartum diet, inert rumen bulk, and
dietary polyethylene glycol on dry matter intake of lactating dairy cows. J. Dairy Sci.
74:933.

Johnson, TR, and D.K. Combs. 1992. Effects of inert rumen bulk on dry matter intake
in early and midlactation cows fed diets differing in forage content. J. Dairy Sci.
75:508.

Jones, G.M. 197 2. Chemical factors and their relation to feed intake regulation in
ruminants: a review. Can. J. Anim. Sci. 52:207.

Journet, M., and B. Remond. 197 6. Physiological factors affecting the voluntary intake
of feed by cows: a review. Livest. Prod. Sci. 3:129.

Jung, H.G. 1989. Forage lignins and their effects on fiber digestibility. Agron. J. 81:33.

Jung, H.G., and DA. Deetz. 1993. Cell wall lignification and degradability. Page 315 in
Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D. Hatfield,
and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Kalu, B.A., and G.W. Fick. 1981. Quantifying morphological development of alfalfa for
studies of herbage quality. Crop Sci. 21:267.

Kawas, J.R., N.A. Jorgensen, and J.L. Danelon. 1991. Fiber requirements of dairy
cows-optimum fiber level in luceme—based diets for high producing cows. Livest. Prod.
Sci. 28:107.

204

Keith, E.A., V.F. Colenbrander, V.L. Lechtenberg, and L.F. Bauman. 1979. Nutritional
value of brown midrib com silage for lactating dairy cows. J. Dairy Sci. 62:788.

Kertz, A.F., B.K. Darcy, and LR. Prewitt. 1981. Eating rate of lactating cows fed four
physical forms of the same grain ration. J. Dairy Sci. 64:2388.

Kertz, A.F., L.F. Reutzel, and GM. Thomson. 1991. Dry matter intake from parturition
to midlactation. J. Dairy Sci. 74:2290.

Kesler, E.M., and S.L. Spahr. 1964. Physiological effects of high level concentrate
feeding. J. Dairy Sci. 47:1122.

Ketelaars J.J.M.H., and BL Tolkarnp. 1992. Toward a new theory of feed intake
regulation in ruminants. 1. Causes of differences in voluntary feed intake-critique of
current views. Livest. Prod. Sci. 30:269.

Law, S.E., and EM. Sudweeks. 197 5. Electronic transducer for rumination research. J.
Anim. Sci. 41:213.

Leek, B.F. 1969. Reticulo-ruminal mechanoreceptors in sheep. J. Physiol. 202:585.
Lehmann, F. 1941. Die lehre vom ballast. Zeitschrift fur Tiern. Futterrnittelk. 5:155.

Lippke, H. 1986. Regulation of voluntary intake of ryegrass and sorghum forages in
cattle by indigestible neutral detergent fiber. J. Anim. Sci. 69: 1459.

Llamas-Lamas, G., and D.K. Combs. 1990. Effect of alfalfa maturity on fiber utilization
by high producing dairy cows. J. Dairy Sci. 73: 1069.

Llamas-Lamas, G., and D.K. Combs. 1991. Effect of forage to concentrate ratio and
intake level on utilization of early vegetative alfalfa silage by dairy cows. J. Dairy Sci.
74:526.

Luginbuhl, J.-M., K.R. Pond, and LC. Burns. 1987. A simple electronic device and

computer interface system for monitoring chewing behavior of stall-fed ruminant
animals. J. Dairy Sci. 70:1307.

Marinucci, M., B.A. Dehority, and S.C. Loerch. 1992. In vitro and in vivo studies of
factors affecting digestion of feeds in synthetic fiber bags. J. Anim. Sci. 70:296.

Mason, 8., LA. Shelford, E. Vink, and LJ. Fisher. 1991. Feeding behaviour of lactating
cows measured with a computerized forage feeding system. J. Dairy Sci. 74(Suppl.
1):219. (Abstr.)

McArthur, J.M., and J.B. Miltrnore. 1968. Continuous reading of the in vivo rumen pH
in fistulated cattle. Can. J. Anim. Sci. 48:237.

McGechan, MB. 1990. A review of losses arising during the conservation of grass
forage: part 2, storage losses. J . Agric. Eng. Res. 45:1.

McLeod, M.N., and B.R. Smith. 1989. Eating and ruminating behaviour in cattle given
forages differing in fibre content. Anim. Prod. 48:503.

 

 

205

McQueen, RE, and RH. Robinson. 1993. Feed intake and production of dairy cows as
influenced by dietary levels of rumen fermentable neutral detergent fiber. J. Dairy Sci.
76(Supp 1):209. (Abstr.)

Mertens, DR. 197 3. Application of theoretical mathematical models to cell wall
digestion and forage intake in ruminants. Ph.D. Dissertation. Cornell Univ. Ithaca, NY.

Mertens, DR. 1977. Dietary fiber components: relationship to the rate and extent of
ruminal digestion. Fed. Proc. 36:187.

Mertens, DR. 1983. Using neutral detergent fiber to formulate dairy rations and
estimate the energy content of forages. Page 60 in Proc. Cornell Nutr. Conf. Feed
Manuf., Syracuse, NY.

Mertens, DR. 1985. Factors influencing feed intake in lactating cows: from theory to
application using neutral detergent fiber. Page 1 in Proc. Georgia Nutr. Conf., Atlanta,
GA. ‘

Mertens, DR. 1987. Predicting intake and digestibility using mathematical models of
ruminal function. J. Anim. Sci. 64:1548.

Mertens, DR. 1993. Rate and extent of digestion. Page 13 in Quantitative Aspects of
Ruminant Digestion and Metabolism. J.M. Forbes and J. France, eds. CAB International

Mertens, D.R., and LO. Ely. 197 9. A dynamic model of fiber digestion and passage in
the ruminant for evaluating forage quality. J. Anim. Sci. 49: 1085.

Metz, J.H.M. 1975. Time patterns of feeding and rumination in domestic cattle. Meded.
Landbouwhogesch., Wageningen 75:12.

Metz, J.H.M., and G. Borel. 1975. Equipment for recording feeding, drinking,
rumination and standing-lying behavior of cattle. Meded. Landbouwhogesch.,
Wageningen 75:4.

Miller, B.G., and RB. Muntifering. 1985. Effect of forage:concentrate on kinetics of
forage fiber digestion in vivo. J. Dairy Sci. 68:40.

Miller, R.H., N.W. Hooven, Jr., J.W. Smith, and ME. Creegan. 1972. Feed
consumption differences among lactating cows. J. Dairy Sci. 55:454.

Miller, T.K., W.H. Hoover, W.W. Poland, Jr., R.W. Wood, and W.V. Thayne. 1990.
Effects of low and high fill diets on intake and milk production in dairy cows. J. Dairy
Sci. 73:2453.

Miner, J.L. 1992. Recent advances in the central control of intake in ruminants. J.
Anim. Sci. 70:1283.

Minson, DJ. 1966. The apparent retention of food in the reticulo-rumen at two levels of
feeding by means of an hourly feeding technique. Br. J. Nutr. 20:765.

Montgomery, ML and B.R. Baumgardt. 1965. Regulation of food intake in ruminants.
2. Rations varying in energy concentration and physical form. J. Dairy Sci. 48: 1623.

206

Mooney, CS, and MS. Allen. 1993. Effectiveness of whole fuzzy cottonseed NDF
relative to alfalfa silage NDF at two lengths of cut. J. Dairy Sci. 76(Suppl 1):247.
(Abstr)

Morrison, I.M. 1988. Influence of some chemical and biological additives on the fibre
fraction of luceme on ensilage in laboratory silos. J. Agric. Sci. (Camb.) 111:35.

Mowat, D.N. 1963. Factors affecting rumen capacity and the physical inhibition of feed
intake. PhD Dissertation. Cornell University, Ithaca, NY.

Muller, L.D., V.L. Lechtenberg, L.F. Bauman, R.F. Barnes, and CL. Rhykerd. 1972. In
vivo evaluation of a brown midrib mutant of zea mays L. J. Anim. Sci. 35:883.

Murphy, M.R., C.L. Davis, and G.C. McCoy. 1983. Factors affecting water
consumption by Holstein cows in early lactation. J. Dairy Sci. 66:35.

National Research Council. 1987. Predicting Feed Intake of Food-Producing Animals.
Natl. Acad. Sci., Washington, DC.

National Research Council. 1989. Nutrient Requirements of Dairy Cattle. 6th rev. ed.
Natl. Acad. Sci., Washington, DC.

Nelson, W.F., and L.D. Satter. 1990. Effect of stage of maturity and method of
preservation of alfalfa on production by lactating dairy cows. J. Dairy Sci. 73:1800.

Nelson, W.F., and L.D. Satter. 1992a. Impact of alfalfa maturity and preservation
method on milk production by cows in early lactation. J. Dairy Sci. 75: 1562.

Nelson, W.F., and L.D. Satter. 1992b. Impact of stage of maturity and method of
preservation of alfalfa on digestion in lactating dairy cows. J. Dairy Sci. 75: 1571.

N ocek, J.E. 1992. Feeding sequence and strategy effects on ruminal environment and
production performance in first lactation cows. J. Dairy Sci. 75:3100.

Nocek, J .E., and D.G. Braund. 1985. Effect of feeding fiequency on diurnal dry matter
and water consumption, liquid dilution rate, and milk yield. J. Dairy Sci. 68:2238.

Norgaard, P. 1989. The influence of physical form of ration on chewing activity and
rumen motility in lactating cows. Acta Agric. Scand. 39:187.

O'Connell, J., P.S. Giller, and W. Meaney. 1989. A comparison of dairy cattle
behavioural patterns at pasture and during confinement. Irish J. Agric. Res. 28:65.

Okine, E.K., G.W. Mathison, and R.T. Hardin. 1989. Effects of changes in frequency of
reticular contractions on fluid and particular passage rates in cattle. J. Anim. Sci.
67:3388.

Orskov, E.R., G.W. Reid, and M. Kay. 1988. Prediction of intake by cattle from
degradation characteristics of roughages. Anim. Prod. 46:29.

Osbourn, D.F., R.A. Terry, G.E. Outen, and SB. Cammell. 1974. Proc. XII
International Grassland Congress, 111:374. Moscow.

207

Penning, PD. 1983. A technique to record automatically some aspects of grazing and
ruminating behaviour in sheep. Grass Forage Sci. 38:89.

Penning, P.D., G.L. Steel, and R.H. Johnson. 1984. Further development and use of an
automatic recording system in sheep grazing studies. Grass Forage Sci. 39:345.

Peters, J.P., J.B. Paulissen, and J.A. Robinson. 1990. The effects of diet on water flux

and volatile fatty acid concentrations in the rumen of growing beef steers fed once
daily. J. Anim. Sci. 68:1711.

Phipps, R.H., J.A. Bines, R.F. Weller, and J. Thomas. 1984. Complete diets for dairy
cows: the effect of energy concentration and change in energy concentration of a
complete diet on intake and performance of lactating dairy cows. J. Agric. Sci. 103:323.

Pienaar, J .P., C.Z. Roux, P.J.K. Morgan, and L. Grattarola. 1980. Predicting voluntary
intake on medium quality roughages. S. Afr. J. Anim. Sci. 10:215.

Pitt, RE. 1990. A model of cellulase and amylase additives in silage. J. Dairy Sci.
73:1788.

Poore, M.H., J.A. Moore, RS. Swingle, T.P. Eck, and W.H. Brown. 1991. Wheat straw
or alfalfa hay in diets with 30% neutral detergent fiber for lactating Holstein cows. J.
Dairy Sci. 74:3152.

Poore, M.H., J .A. Moore, RS. Swingle, T.P. Eck, and W.H. Brown. 1993. Response of

lactating Holstein cows to diets varying in fiber source and runrinal starch degradability.
J. Dairy Sci. 76:2235.

Quicke, G.V., O.G. Bentley, H.W. Scott, and AL. Moxon. 1959. Cellulose digestion in
vitro as a measure of the digestibility of forage cellulose in ruminants. J. Anim. Sci.
18:275.

Ralph, J., and R.F. Helm. 1993. Lignin/hydroxycinnamic acid/polysaccharide
complexes: synthetic models for regiochemical characterization. Page 201 in Forage
Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D. Hatfield, and J.
Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Rayburn, EB, and D.G. Fox. 1993. Variation in neutral detergent fiber intake of
Holstein cows. J. Dairy Sci. 76:544.

Reid, J.T. 1961. Problems of feed evaluation related to feeding of dairy cows. J. Dairy
Sci. 44:2122.

Reid, R.L., G.A. Jung, and W.V. Thayne. 1988. Relationships between nutritive quality
and fiber components of cool season and warm season forages. J. Anim. Sci. 66: 1275.

Remond, B. 1988. Evolution du poids du contenu du reticulo-rumen chez les vaches
laitieres au tours de deux premiers mois de la lactation. Reprod. Nutr. Develop. 28:109.

Riley, J.L., and HM. Cook. 1974. A small inexpensive radiotelemetering capsule for
measuring rumen motility. Cornell Vet. 64:225.

 

208

globinzsonleiH. 1989. Dynamic aspects of feeding management for dairy cows. J. Dairy
ci. 7 :l .

Robinson, P.H., and R.E. McQueen. 1992. Influence of rumen fermentable neutral
detergent fiber levels on feed intake and milk production of dairy cows. J. Dairy Sci.
75:520.

Robinson, P.H., J.G. Fadel, and S. Tamminga. 1986. Evaluation of mathematical
models to describe neutral detergent residue in terms of its susceptibility to degradation
in the rumen. Anim. Feed Sci. Technol. 15:249.

Ruckebusch, Y. 1988. Motility of the gastro—intestinal tract. Page 64 in The Ruminant
Animal, Digestive Physiology and Nutrition. D.C. Church, ed. Prentice Hall,
Englewood Cliffs, NJ.

Sarwar, M., J.L. Frrkins, and M.L. Eastridge. 1992. Effects of varying forage and
concentrate carbohydrates on nutrient digestibilities and milk production by dairy cows.
J. Dairy Sci. 75:1533.

SAS® Language Guide for Personal Computers, Version 6 Edition. 1985. SAS Inst,
Inc., Cary, NC.

Savory, C.J. 1979. Feeding behaviour. Page 277 in Food Intake Regulation in Poultry.
K. N. Boorman, and B. M. Freeman, ed. Longman, Edinburgh, Engl.

Sharma, B.K., R.A. Erdman, and J.B. Reeves, III. 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. J. Dairy Sci. 71:3509.

Shaver, R.D., L.D. Satter, and NA. Jorgensen. 1988a. Impact of forage fiber content on
digestion and digesta passage in lactating dairy cows. J. Dairy Sci. 71:1556.

Shaver, R.D., A.J. Nytes, L.D. Satter, and NA. Jorgensen. 1988b. Influence of feed
intake, forage physical form, and forage fiber content on particle size of masticated
forage, ruminal digesta, and feces of dairy cows. J. Dairy Sci. 71:1566.

Siegfried, B.R., H. Ruckemann, and G. Stumpf. 1984. Method for the determination of
orgainc acids in silage by high performance liquid chromatography. Landwirtschaftliche
Forschung 37:298.

Smith, V.R. 1941. In vivo studies of hydrogen ion concentrations in the rumen of the
. dairy cow. J. Dairy Sci. 24:659.

Smith, L.W., H.K. Goering, and CH. Gordon. 197 2. Relationships of forage
compositions with rates of cell wall digestion and indigestibility of cell walls. J. Dairy
Sci. 55:1140.

Snedecor, G.W., and W.G. Cochran. 1980. Statistical Methods. 7th ed. The Iowa State
Univ. Press, Ames.

Stern, MD. and C.J. Ziemer. 1993. Consider value, cost when selecting nonforage
fiber. Page 14in Feedstuffs. January 11. Miller Publishing Co., Minnetonka, MN.

209

Stevens, CE, and AF. Sellers. 1968. Rumination. Page 2699 in Handbook of
Physiology. Section 6. Alimentary Canal. C. F. Cole, ed. Am. Physiol. Soc.,
Washington DC.

Sutton, JD. 1989. Altering milk composition by feeding. J. Dairy Sci. 72:2801.

Suzuki, S., H. Fujita, and Y. Shinde. 1969. Change in the rate of eating during a meal

and the effect of the interval between meals on the rate at which cows eat roughages.
Anim. Prod. 11:29.

Theander, O., and E. Westerlund. 1993. Quantitative analysis of cell wall components.
Page 83 in Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R. Buxton, R.D.
Hatfield, and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Thorton, R.F., and DJ. Minson. 1973. The relationship between apparent retention time
in the rumen, voluntary intake and apparent digestibility of legume and grass diets in
sheep. Aust. J. Agric. Res. 24:889.

Tyrrell, HF, and J.T. Reid. 1965. Prediction of the energy value of cow's milk. J. Dairy .
Sci. 48:1215.

Ulyatt, M.J., G.E. Waghom, A. John, C.S.W. Reid, and J. Monro. 1984. Effect of intake
and feeding frequency on feeding behaviour and quantitative aspects of digestion in
sheep fed chaffed luceme hay. J. Agric. Sci. (Camb.) 102:645.

USDA-ARS. 1993. Forage Cell Wall Structure and Digestibility. Jung, H.G., D.R.
Buxton, R.D. Hatfield, and J. Ralph, eds. ASA-CSSA-SSSA, Madison, WI.

Vadeveloo, J., and W. Holmes. 197 9. The prediction of the voluntary feed intake of
dairy cows. J. Agric. Sci.(Camb.) 93:553.

Van Soest, P.J. 1965. Symposium on factors influencing the voluntary intake of herbage
by ruminants: voluntary intake in relation to chemical composition and digestibility. J.
Anim. Sci. 24:834.

Van Soest, P.J. 1967 . Development of comprehensive system of feed analyses and its
application to forages. J. Anim. Sci. 26:119.

Van Soest, P.J. 1982. Nutritional Ecology of the Ruminant. Cornell University Press,
Ithaca, NY.

Van Soest, P.J., D.R. Mertens, and B. Deinum. 1978. Preharvest factors influencing
quality of conserved forage. J. Anim. Sci. 47:712.

Van Soest, P.J., J.B. Robertson, and BA. Lewis. 1991. Methods for dietary fiber,
neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J.
Dairy Sci. 74:3583.

Varga, G.A., and W.H. Hoover. 1983. Rate and extent of neutral detergent fiber
degradation of feedstuffs in situ. J. Dairy Sci. 66:2109.

 

 

210

Varga, G.A., E.M. Meisterling, R.A. Dailey, and W. H. Hoover. 1984. Effect of low
and high fill diets on dry matter intake, milk production, and reproductive performance
during early lactation. J. Dairy Sci. 67:1240.

Vasilatos, R., and P.J. Wangsness. 1980. Feeding behavior of lactating dairy cows as
measured by time-lapse photography. J. Dairy Sci. 63:412.

Von Keyserlingk, M.A.G., and G.W. Mathison. 1989. Use of the in situ technique and
passage rate constants in predicting voluntary intake and apparent digestibility of
forages by steers. Can. J. Anim. Sci. 69:973.

Vough, L.R., and G.C. Marten. 1971. Influence of soil moisture and ambient
temperature on yield and quality of alfalfa forage. Agron. J. 63:40.

Waldo, DR. 1986. Effect of forage quality on intake and forage-concentrate
interactions. J. Dairy Sci. 69:617.

Waldo, D.R., and NA. Jorgensen. 1981. Forages for high animal production: nutritional
factors and effects of conservation. J. Dairy Sci. 64: 1207.

Waldo, D.R., L.W. Smith, and 15.1.. Cox. 1972. Model of cellulose disappearance from
the rumen. J. Dairy Sci. 55:125.

Wangsness, P.J., L. Chase, A. Peterson, T. Hartstock, D. Kellnel, and B. Baumgardt.
1976. System for monitoring feeding behavior of sheep. J. Anim. Sci. 42: 1544.

Watkins, K.L., T.L. Veum, and GP. Krause. 1987. Total nitrogen determination of
various sample types: a comparison of the Hach, Kjeltec, and Kjeldahl methods. J.
Assoc. Offic. Agric. Chem. 70:410.

Waybright, T.R., and G.A. Varga. 1991. Effect of water-filled bags in the rumen of
wethers on ruminal digesta kinetics and total tract nutrient digestibility. J. Anim. Sci.
69:2157.

Weiss, W.P., G.R. Fisher, and GM. Erickson. 1989. Effect of source of neutral
detergent fiber and starch on nutrient utilization by dairy cows. J. Dairy Sci. 72:2308.

Welch, LG. 1982. Rumination, particle size and passage from the rumen. J. Anim. Sci.
54:885.

Welch, J .G., and A.M. Smith. 1969. Influence of forage quality on rumination time in
sheep. J. Anim. Sci. 28:813.

Welch, J .G., and A.M. Smith. 1970. Forage quality and rumination time in cattle. J.
Dairy Sci. 53:797.

Williams, C.B., P.A. Oltenacu, and C.J. Sniffen. 1989. Application of neutral detergent
fiber in modeling feed intake, lactation response, and body weight changes in dairy
cattle. J. Dairy Sci. 72:652.

Wolff, E. 1856. Die naturgesetzlichen Grundlagen des Ackerbaues. 3rd ed. Otto
Weigand. Leipzig.

 

 

211

Woodford, J .A., N.A. Jorgensen, and GP. Barrington. 1986. Impact of dietary fiber and
physical form on performance of lactating dairy cows. J. Dairy Sci. 69:1035.

Woodford, S.T., and MR. Murphy. 1988. Effect of forage physical form on chewing

activity, dry matter intake, and rumen function of dairy cows in early lactation. J. Dairy
Sci. 71:674.

Wright, P.L., and RB. Grainger. 1970. Diurnal variation in rumen volume and
metabolite concentrations. J. Dairy Sci. 53:785.

Wyrwicka, W., and C. Dobrzecka. 1960. Relationship between feeding and satiation
centers of the hypothalamus. Science. 132:805.

Yungblut, D.H., J.B. Stone, G.K. Macleod, D.G. Grieve, and EB. Burnside. 1981.

Development of feed intake prediction equations for lactating dairy cows. Can. J. Anim.
Sci. 61:151.

APPENDIX

Table 1. Continuous computer acquisition of feed and water intakes, chewing, reticular
motility, and ruminal pH of cattle: computer board assignments in Labtech Notebook
software setup.

 

1 Software board

 

Wham—3551mm FM

0 DAS-8 DAS-8 water 11, 12

1 Exp-16: #4 Exp-l6, gain=1 pH 1-12

2 Exp~16z #2, #3 Exp-16, gain=50 chewing 1-12
motility 1-12

3 Exp-l6: #1 Exp-16, gain=200 load 1-12

4 CTM-05: #1 CTM-05 water 1-5

5 CTM-OS: #2 CTM-05 water 6-10

 

1See Appendix Figure 1 for location of terminal boards.

212

 

213

Table 2. Continuous computer acquisition of feed and water intakes, chewing, reticular
motility, and ruminal pH of cattle: channel characteristics of algorithm functions for data

acquisition run under Labtech Notebook.1

 

Channel Name: Time

Channel No. - l Buffer size - 64
Channel type - Time No. of iterations - 1
Channel units - Secs No. of stages - 1
Time origin - Elapsed time Sampling rate, Hz - 0.200
Format - SSSSS.SSS Stage duration, sec - 2,000,000
Mode - Cumulative Start/ stop method - Immediate

Channel Name: Water 1

 

 

Channel No. - 2 Buffer size — 128
Channel type - Counter No. of iterations - 1
Channel units - Counts No. of stages - 1
Interface device - 4:CI'M-05 Sampling rate, Hz - 0.200
Interface channel #- 1 Stage duration, sec - 2,000,000
Mode - Cumulative Start / stop method - Immediate

Channel Name: Load 1

 

Channel No. - 14 Offset constant - variable
Channel type - Analog input Buffer size - 128
Channel units - Lbs. No. of iterations — 1
Interface device - 3:DAS8/Expl6 No. of stages - 1
Interface channel # - 16 Sampling rate, Hz - 0.200
Input range - 21:0.025V Stage duration, sec - 2,000,000
Scale factor - variable Start/ stop method - Immediate

Channel Name: Chew raw 1

 

Channel No. - 26 Offset constant - 0
Channel type - Analog input Buffer size - 128
Channel 1111118 - Volts No. of iterations - 1
Interface device - 2:DAS8/Exp16 No. of stages - 1
Interface channel # - 48 Sampling rate, Hz - 4.0
Input range - :l:0.1V Stage duration, sec - 2,000,000

Scale factor - 100 Start/ stop method - Immediate

Table 2. (cont)

Channel Name: Chew derivative 1

Channel No. - 38
Channel type - Calculated
Channel units - Volts
Operation - dx/dt
X input channel - 26
Y input channel - o o 0

Parameter, r - 1
Scale factor - 1

Channel Name: Chew count 1

Channel No. - 50

 

Channel type - Calculated
Channel units - Chews
Operation - Counter
X input channel - o o 0
Y input channel - o o 0

Parameter, r - 1
Scale factor - 1

214

Offset constant -
Buffer size -

No. of iterations -
No. of stages -

Sampling rate, Hz -
Stage duration, sec -
Start/ stop method -

Offset constant -
Buffer size -

No. of iterations -
No. of stages -

Sampling rate, Hz -
Stage duration, sec -
Start/ stop method -

Trigger channel -

Analog trigger value -

Analog trigger polarity -

No. of samples to save pretrigger -

Channel Name: Chews l

 

Channel No. - 62
Channel type - Calculated
Channel unllts - Chews
Operation - dx/dt
X input channel - 50
Y input channel - . . 0
Parameter, r - 0
Scale factor - 1

Channel Name: Motility raw 1
Channel No. - 74

 

Channel type - Analog input
Channel unllts - Volts
Interface device - 2:DAS8/Expl6
Interface channel # - 32
Input range - 10. 1V
Scale factor - 200

0
128
l
l
4.0

2,000,000
Immediate

128
4,000,000

1.0, 2.0

O, 0.5

Imm., On edge
e e e, 38

. . o, -0.20

e e e, LOW

. . ., o

Offset constant - 0

Buffer size -
No. of iterations -
No. of stages -

128
l
1

Sampling rate, Hz - 0.200
Stage duration, sec -
Start/ stop method -

Offset constant -
Buffer size -

No. of iterations -
No. of stages -

2,000,000
Immediate

0
128
1
1

Sampling rate, Hz - 4.0
Stage duration, sec - 2,000,000
Start/ stop method - Immediate

215
Table 2. (cont.)

Channel Name: Motility average 1

 

Channel No. - 86 Offset constant - 0
Channel type - Calculated Buffer size - 128
Channel units - Volts No. of iterations - 1
Operation - Block ave. No. of stages - 1
X input channel - 74 Sampling rate, Hz - 1.0
Y input channel - o o 0 Stage duration, sec - 2,000,000
Parameter, r - 3 Start/ stop method - Immediate

Scale factor - 1

Channel Name: Motility derivative 1

 

 

Channel No. - 98 Offset constant - 0
Channel type - Calculated Buffer size - 128
Channel units - Volts No. of iterations - 1
Operation - dx/dt No. of stages - 1
' X input channel - 86 Sampling rate, Hz - 1.0
Y input channel - o o 0 Stage duration, sec - 2,000,000
Parameter, r - 0 Start/ stop method - lrnmediate
Scale factor - 1
Channel Name: Contraction count 1
Channel No. - 110 Offset constant - 1
Channel type - Calculated Buffer size - 128
Channel units - Contractions No. of iterations - 2,000,000
Operation - Counter No. of stages - 2
X input channel - o o 0 Sampling rate, Hz - 1.0, 2.0
Y input channel - o o 0 Stage duration, sec - 0, 0.5
Parameter, r - 1 Start/ stop method - Irnm., On edge
Scale factor - l Trigger channel - o o o, 98

Analog trigger value - o 0 o, -0.10
Analog trigger polarity - o o 0, Low
No. of samples to save pretrigger - o o o, 0

Channel Name: Contractions 1

 

Channel No. - 122 Offset constant - 0
Channel type - Calculated Buffer size - 128
Channel units - Contractions No. of iterations - 1
Operation - dx/dt No. of stages - 1
X input channel - 110 Sampling rate, Hz - 0.200
Y input channel - o o 0 Stage duration, sec - 2,000,000
Parameter, r - 0 Start / stop method - Immediate

Scale factor - 1

. 216
Table 2. (cont)

Channel Name: pH 1

Channel No. - 134 Offset constant - variable
Channel type - Analog input Buffer size - 64
Channel units - pH No. of iterations - 1
Interface device - lzDASS/Expl6 No. of stages - 1
Interface channel # - 64 Sampling rate, Hz - 0.200
Input range - :l:5V Stage duration, sec - 2,000,000
Scale factor - variable Start/ stop method - Immediate

 

18emp designed for acquiring all five activities for one cow. Channel references
specified for stall #1 only.

217

Table 3. Continuous computer acquisition of feed and water intakes, chewing, reticular
motility, and ruminal pH of cattle: channel assignments for a complete acquisition run
under Labtech Notebook. Setup is designed for monitoring all five activities for 12 stalls.

 

 

Output
Channel no. Reference to

Mun—m BOWL. channel 111:?
1 Time -- -- -- Yes
2 Water 1 4 l -- Yes
3 Water 2 4 2 -- Yes
4 Water 3 4 3 -- Yes
5 Water 4 4 4 -- Yes
6 Water 5 4 5 -- Yes
7 Water 6 5 l -- Yes
8 Water 7 5 2 -- Yes
9 Water 8 5 3 -- Yes
10 Water 9 5 4 -- Yes
11 Water 10 5 5 -- Yes
12 Water 11 0 0 -- Yes
13 Water 12 0 1 -- Yes
14 Load 1 3 16 -- Yes
15 Load 2 3 l7 -- Yes
16 Load 3 3 18 -- Yes
17 Load 4 3 19 -- Yes
18 Load 5 3 20 -- Yes
19 Load 6 3 21 -- Yes
20 Load 7 3 22 -- Yes
21 Load 8 3 23 -- Yes
22 Load 9 3 24 -- Yes
23 Load 10 3 25 -- Yes
24 Load 11 3 26 -- Yes
25 Load 12 3 27 -- Yes
26 Chew raw 1 2 48 -- No
27 Chew raw 2 2 49 -- No
, 28 Chew raw 3 2 50 -- No
29 Chew raw 4 2 51 -- No
30 Chew raw 5 2 52 -- No
31 Chew raw 6 2 53 -- No
32 Chew raw 7 2 56 -- No
33 Chew raw 8 2 57 -- No
34 Chew raw 9 2 58 -- No
35 Chew raw 10 2 59 -- No
36 Chew raw ll 2 60 -- No
37 Chew raw 12 2 61 -- No
38 Chew derivative 1 -- -- 26 No

21 8
Table 3. (cont)

 

 

Output
Channel no. Reference to
Mun—Name Emilia—amend channel—Ed.
39 Chew derivative 2 -- -- 27 No
40 Chew derivative 3 -- -- 28 N o
41 Chew derivative 4 -- -- 29 No
42 Chew derivative 5 -- —- 30 N o
43 Chew derivative 6 -- -- 31 No
44 Chew derivative 7 -- -- 32 No
45 Chew derivative 8 -- -- 33 No
46 Chew derivative 9 -- -- 34 No
47 Chew derivative 10 -- -- ' 35 No
48 Chew derivative 11 -- -- 36 No
49 Chew derivative 12 -- -- 37 N o
50 Chew count 1 -- -- 38 N o
51 Chew count 2 -- -- 39 No
52 Chew count 3 -- -- 40 No
53 Chew count 4 -- -- 41 No
54 Chew count 5 -- -- 42 No
55 Chew count 6 -- -- 43 No
56 Chew count 7 -- -- 44 No
57 Chew count 8 -- -- 45 No
58 Chew count 9 -- -- 46 No
59 Chew count 10 -- -- 47 No
60 Chew count 11 -- -- 48 No
61 Chew count 12 -- -- 49 No
62 Chews 1 -- -- 50 Yes
63 Chews 2 -- -- 51 Yes
64 Chews 3 -- -- 52 Yes
65 Chews 4 -- -- 53 Yes
66 Chews 5 -- -- 54 Yes
67 Chews 6 -- -- 55 Yes
68 Chews 7 -- -- 56 Yes
69 Chews 8 -- -- 57 Yes
70 Chews 9 -- -- 58 Yes
71 Chews 10 -- -- 59 Yes
72 Chews 11 -- -- 60 Yes
73 Chews 12 -- -- 61 Yes
74 Motility raw 1 2 32 -- No
75 Motility raw 2 2 33 -- No
76 Motility raw 3 2 34 -- No
77 Motility raw 4 2 35 -- No
78 Motility raw 5 2 36 -- No
79 Motility raw 6 2 37 -- No

 

219
Table 3. (cont)

 

 

 

Output
Channel no. Reference to
JEWEL—Nam BOWL

80 Motility raw 7 2 40 -- No
81 Motility raw 8 2 41 -- No
82 Motility raw 9 2 42 -- No
83 Motility raw 10 2 43 -- No
84 Motility raw 11 2 44 -- N o
85 Motility raw 12 2 45 -- No
86 Motility average 1 -- -- 74 No
87 Motility average 2 -- -- 75 No
88 Motility average 3 -- -- 76 No
89 Motility average 4 -- -- 77 No
90 Motility average 5 -- -- 78 No ‘
91 Motility average 6 -- -- 79 No
92 Motility average 7 -- -- 80 No
93 Motility average 8 -- -- 81 No
94 Motility average 9 -- -- 82 No
95 Motility average 10 -- -- 83 N o
96 Motility average 11 -- -- 84 No
97 Motility average 12 -- -- 85 No
98 Motility derivative 1 -- -- 86 No
99 Motility derivative 2 -- -- 87 No
100 Motility derivative 3 -- -- 88 No
101 Motility derivative 4 -- -- 89 No
102 Motility derivative 5 -- -- 90 No
103 Motility derivative 6 -- -- 91 No
104 Motility derivative 7 -- -- 92 No
105 Motility derivative 8 -- -- 93 No
106 Motility derivative 9 -- -- 94 No
107 Motility derivative 10 -- -- 95 No
108 Motility derivative 11 -- -- 96 No
109 Motility derivative 12 -- -- 97 No
110 Contraction count 1 -- -- 98 No
111 Contraction count 2 -- -- 99 No
112 Contraction count 3 -- -- 100 No
113 Contraction count 4 -- -- 101 No
1 14 Contraction count 5 -- -- 102 No
115 Contraction count 6 -- -- 103 No
116 Contraction count 7 -- -- 104 No
117 Contraction count 8 -- -- 105 N o
118 Contraction count 9 -- -- 106 No
119 Contraction count 10 -- -- 107 No

120 Contraction count 11 -- -- 108 No

220
Table 3. (cont)

 

 

Output
Channel no. Reference to
immune—blame WM
121 Contraction count 12 -- -- 109 No
122 Contractions 1 -- -- 110 Yes
123 Contractions 2 -- -- 1 11 Yes
124 Contractions 3 -- -- 112 Yes
125 Contractions 4 -- -- 113 Yes
126 Contractions 5 -- -- 114 Yes
127 Contractions 6 -- -- 115 Yes
128 Contractions 7 -- -- 116 Yes
129 Contractions 8 -- -- 117 Yes
130 Contractions 9 -- -- 118 Yes
131 Contractions 10 -- -- 1 19 Yes
132 Contractions 11 -- -- 120 Yes
133 Contractions 12 -- -- 121 Yes
134 pH 1 1 64 -- Yes
135 pH 2 1 65 -- Yes
136 pH 3 l 66 -- Yes
137 pH 4 l 67 -- Yes
138 pH 5 1 68 -- Yes
139 pH 6 1 69 -- Yes
140 pH 7 1 70 -- Yes
141 pH 8 1 71 -- Yes
142 pH 9 1 72 -- Yes
143 pH 10 1 73 -- Yes
144 pH 11 1 74 -- Yes
145 pH 12 l 75 -- Yes

 

221

Table 4. Continuous computer acquisition of feed and water intakes, chewing, reticular
motility, and ruminal pH of cattle: feeding behavior data interpretation computer
program for summarizing individual cow data into bouts of eating, ruminating, and
drinking.1

 

*********************************************************************o
9

* File Name: Chewlntpsas;

* Date 1/20/93: Feeding behavior interpretive program, for IBM 3090-mainframe;
* Data form: For individual cows measured continuously over several days;

* Input files: Files generated by DA system and LABTECH NOTEBOOK;

* Output files: Bout summaries for contractions, eating, ruminating, and drinking;
********#********$**************************************#************;

OPTION LS=l32§

CMS FILEDEF DATA DISK S lPlCOWl PRN A;
CMS FILEDEF CONTSUM DISK CSSlPlCl DAT A;
CMS FILEDEF EATPER DISK EPSlPlCl DAT A;
CMS FILEDEF RUMPER DISK RPS lPlCl DAT A;
CMS FILEDEF DRNKPER DISK DPS lPlCl DAT A;

DATA ONE I’READ DATA FILE‘I;
INFILE DATA FIRSTOBS=8;
INPUT @2 TIME WMETR L1 CHEW PH Cl;
1F Ll<-10 THEN Ll=-5000;

DATA TWO /*BEGINNING LOAD ESTABLISHMENT”;
SET ONE(KEEP=L1); IF _N_=l THEN TOT =50;
IF L1=-5000 THEN DO; TOT=0; Ll=0; L2=0; END;
TOT-+1; IF TOT>5;L2+L1; IF TOT=11;
L2=ROUND(L2/6,. 1); DROP TOT Ll; OUTPUT;

DATA TWO A POETS THE LAST MEAL FOR SOME COWS‘l;
SET TWO END=LAST; OUTPUT; 1F LAST THEN DO; L2=-5000; OUTPUT;
END;
DATA THREE /*BEGINNING LOAD ESTABLISHMENT”;
SET ONE(KEEP=L1);
BY L1 NOTSORTED;

IF L1=—5000 AND FIRST.L1=1 THEN DO; TOT =0; SET TWO_A; END;
IF Ll=~5000 AND FIRST.L1=0 THEN DO; TOT =0; L1=L2; END;

IF _N_=1 THEN TOT=5; TOT+1; IF 1<TOT<6 THEN L1=L2;

DROP L2 TOT;

222
Table 4. (cont)

DATA FOUR /*LOAD, WATER, MOVEMENT”;
MERGE ONE(DROP=L1 C1) THREE;
WATERS=(WMETR-LAG(WMETR))*3.78/50 /*WATER IN LITERS”;

L2=LAG(L1); L3=LAG(L2); L4=LAG(L3); L5=LAG(L4);
L6=LAG(L5); L7=LAG(L6); L8=LAG(L7); L9=LAG(L8);
L10=LAG(L9); Ll 1=LAG(L10); L12=LAG(L11); L13=LAG(L12);
L14=LAG(L13); L15=LAG(Ll4); L16=LAG(L15); L17=LAG(L16);
L18=LAG(L17); L19=LAG(L18); L20=LAG(L19); L21=LAG(L20);
L22=LAG(L21);

STDEV=STD(OF Ll-L22);

DROP WMETR L2-L22;

DATA FIVE /*EATING AND RUMINATING DIFFERENTIATION”;
SET FOUR(KEEP=STDEV); ,
IF STDEV<1 THEN DO; RTYPE=1; ETYPE=0; END;
ELSE DO; RTYPE=O; ETYPE=1; END;

DATA SIX /*LOAD, WATER, CHEWS, PH DURING PAST 30 SEC.*/;
SET FOUR(DROP=STDEV);
SET FIVE(FIRSTOBS=12);
LOAD+L1;
TCHEWS+CHEW;
IF WATERS>O THEN DO; CHEW=0; RTYPE=0; ETYPE=0; COUNT+1; END;
IF RTYPE=1 THEN RCHEW=CHEW; ELSE [F ETYPE=1 THEN ECHEW=CHEW;
RCHEWS+RCHEW; ECHEWS +ECHEW;
TOT-l-l;
WATER+WATERS; PHS-l-PH;
IF TOT=6; LOAD=LOAD/6; COUNT=COUNTI12; PHS=PHS/6;
0DROP L1 CHEW TOT PH WATERS RTYPE ETYPE RCHEW ECHEW;
UTPUT;
TCHEWS=0; LOAD=0; TOT=0; WATER=0; COUNT=0; RCHEWS =0;
ECHEWS=O;
PHS=0;

DATA SEVEN /*3 MIN. MOVING SUM or CHEWS”;
sa'r SIX(KEEP=RCHEWS ECHEWS);
R1=LAG(RCHEWS); R2=LAG(R1); R3=LAG(R2);
R4=LAG(R3); R5=LAG(R4); R6=LAG(R5);
RCHEW=SUM(OF R1-R6);
E1=LAG(ECHEWS); E2=LAG(E1); E3=LAG(E2);
E4=LAG(E3); E5=LAG(E4); E6=LAG(E5);
ECHEW=SUM(OF El-E6);
DROP RCHEWS R1-R6 ECHEWS E1-E6;

5

 

223
Table 4. (cont)

DATA EIGHT /*CONTRACTIONS DURING LAST 30 SEC”;
SET ONE(KEEP=C1);
=LAG(C1); C3=LAG(C2); C4=LAG(C3); C5=SUM(OF C1-C4);
IF C3>0 THEN C6=C5; ELSE C6=O;
C7=LAG(C6); C8=LAG(C7); C9=SUM(C7,C8);
IF C9>0 THEN C10=O; ELSE C10£6;
IF C10>l THEN C1 l=l; ELSE Cl l=0;
DROP Cl-C10;
TOT+-1; CONT+C11;
IF TOT=6; DROP TOT C11; OUTPUT;
CONT=0; TOT=0;

DATA TEN I’DAILY SUM OF CONTRACTIONS”;
IF _N_= 1 THEN DO; FILE CONTSUM; PUT 'DAY' ' ' 'CONTRCT'; END;
MERGE SIX(KEEP=TIME) EIGHT;

TOT+1; CONTRACT +CONT;
IF TOT=2880; TIME=TIME/86400;

DROP TOT CONT;
FILE CONTSUM LRECL=200;
PUT TIME CONTRACT; TOT=0; CONTRACT =0;
DATA ELEVEN /*CHEWING TOTALS; 30 SEC FILE”;

MERGE SIX SEVEN(FIRSTOBS=4) EIGHT;

IF RCHEW>75 AND RCHEWS>8 THEN RUMNATIN=1; ELSE RUMNATIN=0;
IF ECHEW>75 AND ECHEWS>8 THEN EATING=1; ELSE EATING=0;
RUMCHEWS=RUMNATIN*RCHEWS;

EATCHEWS=EATING*ECHEWS;
ICHEWS=TCHEWS-(RUMCHEWS-i-EATCHEWS);

DROP TCHEWS RCHEWS RCHEW ECHEWS ECHEW;

DATA TWELVE /*TOTAL CHEWING EVERY FIVE MIN.” ;
SET ELEVEN(DROP=LOAD COUNT WATER PHS CONT);
TOT+1; RUM-i-RUMNATIN; EAT+EATING;
RCHEW+RUMCHEWS; ECHEW-i-EATCHEWS; ICHEW-t-ICHEWS;
IF TOT=10; TIME=TIME/60;
RUMNATIN=RUM/2; EATING=EAT/2;
DROP TIME TOT RUM EAT RUMCHEWS EATCHEWS ICHEWS;
OUTPUT; TOT=0; RUM=0; EAT=O;
RCI-IEW=0; ECHEW =0; ICHEW=0;

DATA THIRTEEN /"'5 MIN SUMS OF CHEWING”;
SET TWELVE(KEEP=RCHEW ECHEW);
R1=RCHEW; R2=LAG(R1); R3=LAG(R2); RSUM=SUM(OF Rl-R3);
E1=ECHEW; E2=LAG(E1); E3=LAG(E2); ESUM=SUM(OF E1-E3);
DROP RCHEW Rl-R3 ECHEW El-E3;

 

224
Table 4. (cont)

DATA FOURTEEN ”EATING, RUMINATING CORRECTION”;
MERGE TWELVE THIRTEEN(FIRSTOBS=2);
IF ESUM<RSUM THEN DO; RUMNATIN=EATING+RUMNATIN;
EATING=O; RCHEW=ECHEW+RCHEW; ECHEW=0; END;
ELSE DO; EATING=EATING+RUMNATIN;
RUMNATIN=O; ECHEW=ECHEW+RCHEW; RCHEW=0; END;
IF RUMNATIN>5 THEN RUMNATIN=5; IF EATING>5 THEN EATING=5;
IDLE=5-(RUMNATIN+EATING);
DROP ESUM RSUM;

DATA FIFTEEN /*EXPANSION OF CHEWS FROM 5 MIN TO 30 SEC”;
SET FOURTEEN(KEEP=RUMNATIN EATING);
RUM=RUMNATIN; EAT=EATING; DROP RUMNATIN EATING;
OUTPUT; OUTPUT; OUTPUT; OUTPUT; OUTPUT;
OUTPUT; OUTPUT; OUTPUT; OUTPUT; OUTPUT;

DATA SIXTEEN /*CORRECT ION OF 30 SEC TOTALS FOR CHEWS”; .
MERGE ELEVEN FIFTEEN;
IF RUM>EAT THEN DO; RUMNATIN=RUMNATIN+EATING; EATING=0;
RUMCHEWS=RUMCHEWS+EATCHEWS; EATCHEWS=0; END;
ELSE DO; EATING=EATING+RUMNATIN; RUMNATIN=0;
EATCHEWS=EATCHEWS+RUMCHEWS; RUMCHEWS=O; END;
IF STDEV<1 THEN STDEV=0;
IF RUMNATIN>1 THEN RUMNATIN=1;
IF EATING>1 THEN EATING=1§
DROP COUNT RUM EAT;

DATA SEVENTEN(KEEP=LI) I’STDEV OF LOAD FLAGGING”;
SET SIXTEEN(KEEP=LOAD STDEV); BY STDEV NOTSORTED;
IF STDEV=0 AND FIRST.STDEV=1;
L1=LOAD;

DATA EIGHTEEN(KEEP=L2) /*STDEV OF LOAD FLAGGING”;
SET SDCTEEN(KEEP=LOAD STDEV); BY STDEV NOT SORTED;
IF STDEV=0 AND LAST.STDEV=1; TOT+-1;
L2=LOAD; IF TOT>1;

DATA NINETEEN l’STABLE LOADS BEFORE AND AFTER MEALS”;
SET SIXTEEN; BY STDEV NOTSORTED;
IF STDEV=O AND FIRST.STDEV=1 THEN SET SEVENTEN;
IF STDEV=0 AND LAST.STDEV=1 THEN SET EIGHTEEN;

 

 

225
Table 4. (cont)

DATA TWENTY ”RUNNING SUMS FOR EAT AND RUM PERIODS”;
SET NINETEEN(KEEP=EATING RUMNATIN);
R1=RUMNATIN; R2=LAG(R1); R3=LAG(R2); R4=LAG(R3);
R5=LAG(R4); R6=LAG(R5); R7=LAG(R6); R8=LAG(R7);
R9=LAG(R8); R10=LAG(R9); R1 1=LAG(R 10); R12=LAG(R1 l);
R13=LAG(R12); Rl4=LAG(Rl3); R15=LAG(R14);

Rl6=SUM(OF Rl-RIO); R17=SUM(OF Rl-RlS);

E1=EATING; E2=LAG(E1); E3=LAG(E2); E4=LAG(53);
E5=LAG(E4); E6=LAG(E5); E7=LAG(E6); E8=LAGCE7);
E9=LAG(E8); E10=LAG(E9); E11=LAG(E10); E12=LAG(E1 l);
E13=LAG(E12); E14=LAG(E13); E15=LAG(E14);
E16=EATING; E l7=SUM(OF El-EIS);

DROP RUMNATIN Rl-RlS EATING El-ElS;

DATA TW ONE /*EATING PERIODS”;
MERGE TWENTY(FIRSTOBS=16 KEEP=E17)
TWENTY(FIRSTOBS=2 KEEP=E16)
NINETEEN(KEEP=TIME L1 L2 WATER EATING EATCHEWS CONT PHS);
IF El7=. THEN El7=El6;
WATR+WATER; EAT+EATING/2; CHEWS+EATCHEWS; CONTR+CONT;
IF El7=0 THEN E18=0; ELSE IF El6=l THEN E18=l; ELSE E18=2;
IF E18=LAG(E18) THEN El9=2; ELSE El9=E18;
IF El9=2 THEN DELETE;
IF El9=LAG(El9) THEN E20=2; ELSE E20=El9;
IF E20=2 THEN DELETE;
DROP WATER EATING EATCHEWS E16-El9 CONT;

DATA TW _TWO I’EATING PERIOD VARIABLES”;
SET TW_ O;NE STARTIME=(LAG(TIME))/86400;
STOPTIME=TIMEl86400;

DUR=ROUND(((STOPTIME-STARTIME)* 1440),. l);
EATCHEWS=CHEWS~LAG(CHEWS);

EATING=EAT-LAG(EAT);

STARTPH=ROUND(LAG(PHS),.01); STOPPH=ROUND(PHS,.01);
WATER=ROUND(WATR-LAG(WATR),. l); CONT=CONTR- LAG(CONTR);
STRTLOAD=ROUND(LAG(L1),. l); STOPLOAD=ROUND(L2,.1);

IF ___N =1 THEN DELETE; IF E20=l THEN DELETE;
CHEWRATE=ROUND(EATCHEWS/EATING,.1);

DROP TIME CHEWS EAT WATR L1 L2 E20 CONTR PHS;

226
Table 4. (cont)

DATA TW THREE /"‘CLEAN UP DLOADS, PRINT”;
IF _N_ =T THEN DO; FILE EATPER; PUT 'STARTIME' ' ' 'STOPTIME' ' '
'DURATION' ' ' 'EATCHEWS' ' ' 'EATING' ' ' 'CHEWRATE' ' ' 'STARTPH' ' '
'STOPPH' ' ' 'WATER' ' ' 'CONTRACTIONS' ' ' 'STARTLOAD' ' ' 'STOPLOAD'
' ' 'DLOAD' ’ ' 'TIMELAST'; END;
SET TW_TWO; LLOAD=LAG(STOPLOAD); DIF=LLOAD-STOPLOAD;
IF DIF">0 THEN DLOAD=STRTLOAD-STOPLOAD; ELSE DLOAD=DIF;
IF DLOAD<.1 THEN DELETE;
TIMELAST=ROUND(((STARTIME-LAG(STOPTIME))*1440),. 1);
DROP LLOAD DIF;
STARTIME=ROUND(STARTIME,.0001);
STOPTIME=ROUND(STOPTIME,.0001);
FILE EATPER LRECL=200;
PUT STARTIME STOPTIME DUR EATCHEWS EATING CHEWRATE
STARTPH
STOPPH WATER CONT STRTLOAD STOPLOAD DLOAD TIMELAST;

DATA TW FOUR /*RUMINATING PERIODS”;

MERGE TWENTYCFIRSTOBS=16 KEEP=R17)

TWENTY(FIRSTOBS=11 KEEP=R16)

NINETEEN(KEEP=TIME WATER RUMNATIN RUMCHEWS CONT PHS);

IF Rl7=. THEN Rl7=Rl6;

WATR+WATER; RUM-I-RUMNATIN/Z; CI-IEWS-l-RUMCHEWS;
CONTR+CONT;

IF Rl7=0 THEN R18=0; ELSE IF R16=10 THEN Rl8=l; ELSE Rl8=2;

IF R18=LAG(R18) THEN R19=2; ELSE R19=R18;

IF R19=2 THEN DELETE;

IF R19=LAG(R19) THEN R20=2; ELSE R20=R19;

IF R20=2 THEN DELETE;

DROP WATER RUMNATIN RUMCHEWS Rl6-Rl9 CONT;

DATA TW FIVE PRUMINATING PERIOD VARIABLES”;

IF _N_ =-1 THEN DO; FILE RUMPER; PUT 'STARTIME' ' ' 'STOPT'IME' ' '
'DURATION' ' ' 'TIMELAST‘ ' ' 'RUMCHEWS' ‘ ' 'RUMNATIN' ' ' 'CHEWRATE' ' '
'STARTPH' ' ' 'STOPPH' ' ' 'WATER' ' ' 'CONT'; END;

SET TW_FOUR; STARTIME=(LAG(TIME))l86400; STOPT'IME=TIME/86400;

DUR=ROUND(((STOPTIME-STARTIME)*1440),. l); TIMELAST=LAG(DUR);

RUMCHEWS=CHEWS-LAG(CHEWS);

RUMNAT1N=RUM-LAG(RUM);

STARTPH=ROUND(LAG(PHS),.01); STOPPH=ROUND(PHS,.01);

WATER=ROUND(WATR-LAG(WATR),. l); CONT=CONTR-LAG(CONTR);

IF _N_=l THEN DELETE; IF R20=l THEN DELETE;

IF _N_=3 THEN TIMELAST=.;

CHEWRATE=ROUND(RUMCHEWS/RUMNATIN,. 1);

DROP TIME CHEWS RUM WATR R20 CONTR PHS;

STARTIME=ROUND(STARTIME,.0001);

STOPTIME=ROUND(STOPTIME,.0001);

FILE RUMPER LRECL=200; .

PUT STARTIME STOPTIME DUR TIMELAST RUMCHEWS RUMNATIN
CHEWRATE STARTPH STOPPH WATER CONT;

227
Table 4. (cont)

DATA TW SIX /*WATER FLAGGING’I;
SET ELEVEN(KEEP=TIME WATER COUNT);
IF WATER>0 THEN F1=1; ELSE Fl=0;
F2=LAG(F1); F3=LAG(F2); F4=LAG(F3); F5=LAG(F4);
F6=LAG(F5); F7=LAG(F6); F8=LAG(F7); FLAG=SUM(OF F1-F8);
IF FLAG>0 THEN FLAG=1;
WATERS+WATER; WTIMES+COUNT; DROP WATER COUNT F1-F8;

DATA TW_SEVEN I’DRINKING PERIODS”;
SET TW_SIX; BY FLAG NOTSORTED;
IF FIRST.FLAG=1;
STARTIME=(LAG(TIME)-30)/86400; STOPTIME=(TIME-240)/86400;
DUR=ROUND(((STOPTIME-STARTIME)*1440),. l);
DWATER=ROUND(W ATERS-LAG2(W ATERS),.01);
WTME=ROUND(WT11V1ES-LAG2(WTIMES),.01);
DKRATE=ROUND(DWATER/WTIME,.01);
IF WATERS=0 THEN DELETE; IF FLAG=1 THEN DELETE;
DROP TIME WATERS WTIMES FLAG;

DATA TW EIGHT I‘DRINKING BOUT FIX, PRINT”;

IF _N_ =I THEN DO; FILE DRNKPER; PUT 'STARTIME' ' ' 'STOPTIME' ' '
'DURATION' ' ’ 'DWATER' ' ' 'WTIME' ' ' 'DKRATE' ' ' 'TIMELAST'; END;

SET TW_SEVEN; 1F DWATER<.09 THEN DELETE;

'I‘IMELAST=ROUND(((STARTIME-LAG(STOPTIME))*1440),. 1);

STARTIME=ROUND(STARTIME,.0001);

STOPTIME=ROUND(STOPTIME,.0001);

FILE DRNKPER LRECL=200;

PUT STARTIME STOPTIME DUR DWATER WTIME DKRATE TIMELAST;

RUN;

 

lReticular contractions and ruminal pH are included in bout activities. Data input
are cow files obtained from feeding behavior data acquisition system and Labtech
Notebook. Data output are bout summaries in ASCII format. Current configuration is
set for running on IBM 3090 mainframe computer using the Statistical Analysis System
(SAS) software program.

 

228

Table 5. Intake limitations, feeding behavior, and rumen function of cows challenged
with rumen fill from dietary fiber or inert bulk: probability values for effects used to
describe variation in experimental variables.l

 

Source of variation

Period Cow Square
within within

 

 

 

Yarialzll: Squaw
(Probability >F‘ ,
Production
Milk .0006 .0004 .5663 .0002 ns
4% FCM .0001 .0091 .2187 .0027 ns
Fat .0001 .2167 .2080 .0144 ns
Protein .0008 .0001 .3385 .0010 ns
Lactose .0031 .0001 .4222 .0009 ns
Milk composition
Fat .0481 .5036 .4783 .0045 ns
Protein .0026 .0003 .0393 .0235 ns
Lactose .0001 .0001 .0020 .0001 ns
Intake
DM .0065 .0001 .0019 .0222 .0360
NDF .001 1 .0039 .0124 .0516 .0338
Indigestible NDF .0002 .0001 .0086 .0531 .0520
CP .0091 .0001 .0011 .0232 .0298
Water intake
Free water .0043 .0009 .0059 .0001 ns
Total water .0043 .0089 .0026 .0002 .0912
BW .0001 .1565 .0045 .0001 ns
Empty BW .0001 .0001 .0046 .0001 ns
Body condition score .0001 .1672 .0044 .0001 ns
Nutrient digestibility
DM .0003 .0001 .3164 .0708 ns
OM .0003 .0001 .5276 .1097 ns
NDF .0236 .0019 .8119 .6939 ns
ADF .0614 .9571 .8700 .1803 ns
Lignin .0031 .0170 .0162 .6360 ns
Hemicellulose .0215 .4481 .6518 .8898 ns
Cellulose .0848 .4238 .9305 .1570 ns
CP 0125 .4261 .0023 .2973 ns
Fecal output
DM .0070 .0140 .0107 .0248 .0763
NDF .0021 .0017 .0426 .0757 .0965
Indigestible NDF .0002 .0001 .0086 .0531 .0520
Fecal composition
DM
NDF

Indigestible NDF

 

Table 5. (cont)

229

 

 

Source OM
Period Cow Square
within within by
1mm; Squats—WWW
(Probability > F)
Fecal NDF particle size
<600 um .0781 .0133 .0719 .2493 ns
600-1200 11m .1536 .0027 .0533 .0501 ns
1200-2400 pm .7260 .0001 .3868 .6490 .0927
>2400 11m .1037 .7985 .1536 .4637 ns
Rumen volume
Digesta .01 19 .0001 .0008 .0001 ns
Digesta + inert bulk .0136 .0013 .0012 .0001 ns
Digesta + inert bulk
+ head space .0098 .0005 .0003 .0001 ns
Digesta composition
M
52’C .0001 .0001 .0410 .0001 .0034
Toluene
NDF .0013 .0001 .3677 1073 ns
Rumen mass
Wet digesta .0083 .0001 .0010 .0003 ns
Wet digesta + inert bulk .0116 .0001 .0014 .0002 ns
DM .0010 .0001 .0135 .0055 ns
OM .0002 .0001 .0130 .0034 ns
NDF .0001 .0001 .0694 .0015 ns
ADF .0001 .0001 .041 l .0020 ns
Lignin .0001 .0001 .0137 .0026 ns
Hemicellulose .0002 .0001 .0845 .0009 ns
Cellulose .0001 .0001 .0492 .0008 ns
Indigestible NDF .0001 .0001 .0355 .0009 ns
Density .0002 .0001 .0458 .0087 ns
VFA concentrations
Acetate .0072 .0006 .0554 .0038 .01 10
Propionate .0656 .0001 .0683 .0009 ns
Butyrate .0001 .0001 .7400 .0001 ns
Valerate .0041 .0128 .1057 .0878 ns
Branched-chain .0001 .0093 .01 15 .0004 ns
Total .0048 .0001 .0150 .0001 .0273
VFA molar proportions
Acetate .0785 .0001 .6122 .0057 ns
Propionate .4357 .0001 .2146 .0070 ns
Butyrate .0046 .3420 .5485 .1446 .0717
Valerate .0307 .6570 .2204 .6007 ns
Branched-chain .0001 .0025 .0189 .0164 .0629
Acetatezprop. ratio .1606 .0001 .3034 .0022 ns
Fractional passage rate
2 h prefeeding .0035 .0001 .1088 .0001 .0497
2 h postfeeding .0087 .0001 .2350 .0161 ns
Average .0012 .0001 .2532 .0002 ns

 

 

 

230
Table 5. (cont.)

 

 

 

 

 

, Source of yafiation
Period Cow Square
within within
latiable SQWHLM
(Probability >F‘ ,
Fractional digestion rate
2 h prefeeding .0916 .0005 .3734 .0074 ns
2 h postfeeding .2924 .0620 .8788 .1041 us
Average .1 191 .0044 .7265 .0235 ns
Rumen apparent digestibility
% of NDF intake
2 h prefeeding .1922 .0001 .6478 .1899 ns
2 h postfeeding .3317 .0001 .61 12 .031 l .0904
Average .1327 .0001 .4917 .0527 .0890
% of total tract digestibility
2 h prefeeding .0117 .0060 .9009 .5620 ns
2 h postfeeding .0007 .0001 .1893 .0026 .0089
Average .0007 .0001 .5278 .0895 .0767
Eating bouts .0001 .4806 .6676 .0015 ns
Meal size
Mean
DM .0001 .0002 .0057 .0002 .0678
NDF .0001 .0079 .0367 .0037 ns
Maximum 4940 .4234 .0672 .2094 ns
Eating bout length 0366 .0058 .6060 .0002 ns
Eating time
min/d .1324 .0004 .3142 .0020 ns
min/kg of DM .0008 .0001 .0084 .0005 ns
min/kg of NDF .0001 .0018 .0147 .0011 ns
Eating chews .0040 .0001 .0360 .0001 .0925
Eating chew rate .0079 .0509 .0009 .0001 ns
Water intake during meals .8727 .0603 .0056 .0001 ns
Ruminating bouts .7748 .0005 .0748 .0001 ns
Ruminating bout length .0001 .0001 .0731 .0001 ns
Ruminating time _
min/d .0001 .0001 .1408 .0440 ns
min/kg of DM .0001 .0001 .0001 .0037 .0122
min/kg of NDF .0002 .0001 .0004 .0048 .0084
Ruminating chews .0001 .0001 .0340 .0001 ns
Ruminating chew rate .0001 .0006 .0461 .0001 ns
Total chewing time
min/d .0005 .0001 .1783 .0106 ns
ming of DM .0001 .0001 .0001 .0004 .0178
min/kg of NDF .0001 .0001 .0005 .0022 .0223
Total chews .0001 .0001 .0200 .0001 ns
Drinking bouts .5299 .0219 .2975 .0002 ns
Drinking bout size .4697 .5547 .6073 .0001 ns
Drinking time .0001 .0759 .9973 .0001 ns
Drinking rate .0001 .3139 .5454 .0001 ns

23 1
Table 5. (cont)
Source of vadafimr

Period Cow Square
within within by

 

 

 

 

Mariam: 39W
(Probability > F)
Reticular contractions
Daily total
Eating .1567 .1070 .3526 .0127 ns
Ruminating .0297 .0001 .0756 .001 1 ns
Idle .0005 .0003 .6604 .0078 ns
Total .0005 .0051 .0170 .0001 ns
Frequency
Eating .9441 .0182 .2244 .0051 ns
Ruminating .0001 .0001 .0038 .0001 ns
Idle .0392 .3944 .7775 .0189 ns
Ruminal pH
Daily pH
Mean .0002 .0004 .7380 .0282 ns
Variance
Minimum .01 10 .0028 .6474 .0839 ns
Maximum .0002 .0005 .6630 .0232 ns
Range .5522 .097 1 .7520 .201 1 ns
Hours below pH
5.5 .2022 .0530 .5882 .41 10 ns
6.0 .0075 .0049 .5463 .0861 ns '
6.5 .0003 .0023 .7189 .0099 ns
7.0 .0016 .0170 .1725 .3775 .0301
Start of eating .0003 .0004 .7514 .0455 ns
End of eating .0003 .0004 .6953 .0648 ns
Start of rumination .0006 .0005 .7482 .0268 ns
End of rumination .0021 .0012 .9251 .0584 ns

 

1The statistical model used was appropriate for Latin square designs. A reduced
model without square x treatment was used when this effect was not significant (P >

.10).

 

232

Table 6. Intake limitations, feeding behavior, and rumen function of cows challenged
with rumen fill from dietary fiber or inert bulk: apparent total tract digestibility of

nutrients using acid detergent lignin as an internal marker from 12 cows receiving low
fiber (LF) or high fiber (HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

filament.—
XaliablL LF+0 LF_B__HEflL_I-1_F_B+ +
Digestibility, %
DM 66.8 65.9 65.2 63.3
OM 68.5 67.2 66.5 64.6
NDF 39.8 39.6 44.6 40.8
ADF 38.0 37.6 45.5 43.7
Indigestible NDF1 -16.4 -16.0 -1.5 -5.7
Hemicellulose 38. 1 36.2 44.5 40.5
Cellulose 45.8 45.6 54.6 52.4
CF 65.2 65.1 70.0 67.4
Fecal output, kg/d
DM 7.60 7.83 6.49 6.11
NDF 3.49 3.50 3.60 3.43
Lignin .52 .51 .67 .61

 

lNeutral detergent fiber residue after 120 h in vitro fermentation.

 

233

Table 7. Intake limitations, feeding behavior, and rumen function of cows challenged
with rumen fill from dietary fiber or inert bulk: rumen digesta parameters 2 h prefeeding
for 12 cows receiving low fiber (LF) or high fiber (HF) diets without (+0) or with (+B)
added rumen inert bulk.

 

 

 

 

 

Main effectl
Treaflnsnt _sisnifieanre
mam; MAB HF+0 HF+B SE F B FxB
Rumen volume, L
Digesta 89.9a 75.4” 95.63 78.6” 2.4 ‘l ** NS
Digesta + inert bulk 89.9” 97.63 95.63” 100.8a 2.4 i * NS
Digesta + inert bulk
+head space 108.4° 115.7%!b 110.1bc 119.381 2.0 NS ** NS
Digesta composition
DM, %
52'C oven 14.3a 12.4” 12.9” 104° .2 ** ** NS
Toluene 12.33 10.8” 11.9” 81° . 6 * ** NS
NDF, % 57.3” 56.8” 63.53 63.6a 1.0 ** NS NS .
Rumen mass, kg
Wet digesta 76.0” 63. 1° 83.63 683° 1.8 ** ** NS
Wet digesta
+ inert bulk 760° 85.33” 83.6” 90.5a 1.8 ** ** NS
DM 10.8a 7.8” 10.8a 7.2” . 3 NS ** NS
(WI 9.62‘ 6.8” 9.5a 6.3” .3 NS ** NS
NDF 6.2” 4.4° 6.9a 4.6° .2 * ** NS
ADF 3.6” 2.6° 4.1a 2.8° . 1 ** ** NS
Lignin .81” .60° .99a .68° .03 ** ** NS
Hemicellulose 2.6a 1.9” 2.8‘1 1.8” . 1 NS ** NS
Cellulose 2.6” 1.9° 3.0a 2.0° . 1 * ** NS
Indigestible NDF2 3.4” 2.4° 4.1a 2.7° . 1 ** ** '1
Density, kg/L .85” .84” .88a .87a .01 ** NS NS

 

a~”o°Values in same row with different superscripts differ (P < .05).
1Main effect levels differ, ”P < .01, *P < .05, TP < .10.
2Neutral detergent fiber residue after 120 h in vitro fermentation.

234

Table 8. Intake limitations, feeding behavior, and rumen function of cows challenged
with rumen fill from dietary fiber or inert bulk: rumen digesta parameters 2 h postfeeding
for 12 cows receiving low fiber (LF) or high fiber (HF) dliets without (+0) or with (+B)
added rumen inert bulk.

 

 

 

 

 

Main effectl
Treannem _Simificancc.
18031219 MW
Rumen volume, L
Digesta 89.8” 82.4” 101.92‘ 86.8” 2.8 ** ** NS
Digesta + inert bulk 89.8” 104.6a 101.93 108.921 2.8 ** ** NS
Digesta + inert bulk
+ head space 107.6” 120.23 112.9” 122.9a 2.4 NS ** NS
Digesta composition
DM, %
52°C oven 14.3a 13.2” 13.2” 107° .3 ** ** *
Toluene 12.5a 11.9a 11.6a 8.3” .7 ** * ’r
NDF, % 53.7” 51.9” 59.9a 59.2‘1 .9 ** NS NS
Rumen mass, kg
Wet digesta 75.3” 67.4° 89.221 75.5” 2.6 ** ** NS
Wet digesta
+ inert bulk 753° 89.6” 89.2” 97 .73 2.6 ** ** NS
DM 10.78 8.8” 11.98 8.2” .4 NS ** 1
(WI 9.43 7.7” 10.44 7.1” .4 NS ** T
NDF 5.8” 4.6° 7.1a 4.9c .3 ** ** i
ADF 3.3” 2.6° 4.2a 2.9”° .2 ** ** T
Lignin .79” .61° 1.024 .70”° . 1 ** ** ’r
Hemicellulose 2.5” 2.0° 2.9a 2.0° . 1 1' ** *
Cellulose 2.5” 1.9c 3.1a 2. 1° .1 ** ** ’r
Indigestible NDF2 3.3” 2.5° 4.2a 2.9° . 1 ** ** *
Density, kg/L .84” .82” .88a .87a .01 ** NS NS

 

a~”t°Values in Same row with different superscripts differ (P < .05).
1Main effect levels differ, ”P < .01, *P < .05, TP < .10.
2Neutral detergent fiber residue after 120 h in vitro fermentation.

235

Table 9. Intake limitations, feeding behavior, and rumen function of cows challenged
with rumen fill from dietary fiber or inert bulk: concentration and molar proportion of

VFA in rumen fluid 2 h prefeeding from 12 cows receiving low fiber (LF) or high fiber

(HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

 

Treannent
Xan’able + +
VFA concentrations (mmol/L)
Acetate 77.83 71.7” 78.121 67.4b
Propionate 3 1.93 26.2” 23.3b 17.7c
Butyrate 13.08 12.28 12.68 10.9b
Valerate 2.8a 2.3” 2.4b 1.9c
Branched-chain2 5 .6a 4.9” 5. 1” 4.9”
Total 131.1a 117.3b 121.5” 102.9c
VFA molar proportions --------(% of total VFA)-------.---
Acetate 59.4c 61.5b 64.38 65.68
Propionate 24.2‘1 21.8” 19. 1c 17 . 1°
Butyrate 10.0 10.6 10.4 10.6
Valerate 2.1a 1.9b 2.0ab 1.8b
Branched-chain 4.3” 4.2” 4.2” 4.93
Acetatezpropionate ratio 2.6d 3.0° 3.4” 3.93

 

 

L—I—L-Laoobx

Main effectl
_sianificancs.
NS 8* NS
#81: #81! NS

81! #81! NS
118* *4! NS
NS . NS
*4! *4! NS
#11: Ill NS
slut! 4! NS
NS NS NS

1 * NS

all * Ii!
*8 118* NS

 

a~”v°Va1ues in same row with different superscripts differ (P < .05).
1Main effect levels differ, "P < .01, *P < .05, TP < .10.

2isobutyrate + isovalerate.

236

Table 10. Intake limitations, feeding behavior, and rumen function of cows challenged
with rumen fill from dietary fiber or inert bulk: concentration and molar proportion of
VFA in rumen fluid 2 h postfeeding from 12 cows receiving low fiber (LF) or high fiber
(HF) diets without (+0) or with (+B) added rumen inert bulk.

 

 

 

 

 

 

Main effect1
Treatment _rLisrlificanse.
Variable L + + Fx
VFA concentrations (mmol/L)
Acetate 77.4a 76.0a 77.3a 69.7” 1.8 NS * NS
Propionate 29.9a 26.63” 25.2” 21.2° 1.2 ** ** NS
Butyrate 13.0a 12.8a 13.19 11.1” .3 * ** **
Valerate 2.9 2.7 3.0 2.7 .1 NS NS NS
Branched-chain2 6.0 5.5 5.7 5.6 .2 NS 8 NS
Total 129.2a 123.6a 124.3a 110.4” 3 .4 * ** NS
VFA molar proportions ---------(% of total VFA)---------
Acetate 60.1” 61.68” 62.18 63.18I 7 * 1 NS
Propionate 22.9a 21.43” 20.2”c 192° .6 ** * NS
Butyrate 10.0 10.4 10.6 10.0 .2 NS NS 1
Valerate 2.2 2.2 2.4 2.4 1 * NS NS
Branched-chain 4.7” 4.5” 4.7” 5.2a l * NS *
Acetatezpropionate ratio 2.7° 3.0”c 3.13” 3.3a 1 ** * NS

 

at”r°Values in same row with different superscripts differ (P < .05).
1Main effect levels differ, “P < .01, *P < .05, TP < .10.
2isobutyrate + isovalerate.

 

237

Table 11. Enhanced intake and production of cows offered ensiled alfalfa silage with
higher neutral detergent fiber digestibility: probability values for effects used to describe
variation in experimental variables.l

 

Soruce of variation

 

 

 

Karim; Cow Period Treatment...
(Probability > F)
Production
Milk .0001 .4086 .0192
SCM .0001 .8857 .3086
Fat .0007 .7694 .7229
Protein .0001 .0852 .1960
Lactose .0001 .7149 .0837
Milk composition
Fat .0371 .3982 .3666
Protein .0021 .1 147 .2329
Lactose .0052 .2255 .3966
BW .0001 .0049 .9587
Body condition score .0001 .0219 .5995
Intake
DM .0001 .0001 .0081
NDF .0001 .0001 1.000
Indigestible NDF .0009 .0001 .0872
CP .0001 .0001 .7817
Fecal composition
DM .0072 .8248 .2594
NDF .0041 .5080 .1875
Indigestible NDF .0373 .1993 .0143
Fecal output
DM .0005 .0001 .4147
NDF .001 l .0001 .3438
Nutrient digestibility
DM .7791 .3760 .0050
OM .8073 .4300 .0037
NDF .9503 .8626 .0778
ADF .8501 .4910 .1964
Lignin .6780 .0344 .0409
Hemicellulose .2096 .3427 .0688
Cellulose .6457 .9446 .0633
CP .1768 .3554 .207 1
Eating bouts .0687 3856 .9188
Meal size
DM .0113 .0010 .5527
NDF .0122 .0007 .6014
Eating bout length .0504 .2212 .9280
Eating time
min/d .0009 .2571 .9885
min/kg of DM .0057 .0028 .1132
min/kg of NDF .0160 .0041 .8358

 

238
Table 11 (cont)

 

 

 

 

 

 

 

Source Of W
MariablsL Low Pencil Treatment__
(Probability > F)
Eating chews .0686 .4947 .4151
Eating chew rate .0079 .9751 .1758
Water intake during meals .0042 .0574 .4430
Ruminating bouts .0009 .0700 .0415
Ruminating bout length .0008 .5795 .0255
Ruminating time
min/d .0030 .0255 .8738
min/kg of DM .0008 .0033 .1199
min/kg of NDF .0021 .0035 .9561
Ruminating chews .0622 .1842 .5033
Ruminating chew rate .2210 .6125 .2826
Total chewing time
min/d .0036 .0463 .9153
min/kg of DM .0016 .0011 .0753
min/kg Of NDF .0058 .0017 .9067
Total chews .0863 .2312 .4175
Water intake .0001 .0002 .5676
Drinking bouts .0007 .3897 .4190
Drinking bout size .0006 .8102 .2369
Drinking time .0121 .6340 .3343
Drinking rate .0014 .5331 .3470
Rumen digesta volume
2 h prefeeding .0679 .0215 .8657
2 h postfeeding .2698 .0651 .7937
Average .1497 .0404 .8199
Rumen digesta composition
DM .2780 .3846 .7536
NDF .7512 .4120 .2134
Indigestible NDF .6813 .6100 .0885
Rumen digesta mass
Wet digesta .1797 .0566 .7995
DM .1059 .0400 .7166
OM .1018 .0392 .6940
NDF .1865 .0732 .5162
ADF 1832 .07 17 .4082
Lignin 1714 .0733 .2785
Hemicellulose 1993 .0820 .8778
Cellulose .1856 .0696 .4768
Indigestible NDF .1587 .0651 .3250
Rumen digesta density .6464 .0474 .3333
Rumen pH
2 h prefeeding .0927 .1355 .1213
2 h postfeeding .2215 .1605 .6142
Average .1691 .1482 .2983
Total rumen VFA concentration
2 h prefeeding .5460 .1150 .7995
2 h postfeeding .0547 .0634 .0322

239
Table 11 (cont)

 

 

 

 

 

 

Sourgg gt yan'arigp
Mariam: COW Period Treatment__
(Probability > F)
Average .3951 .0995 .2957
Rumen VFA molar proportions
Acetate .7105 .4615 .1264
Propionate .4729 .3571 .0787
Butyrate .1306 .0770 .5769
Valerate .5459 .1856 .1 178
Branched-chain .2819 .2613 .7642
Acet:PrOp. ratio .6079 .8229 .1007
Fractional passage rate
2 h prefeeding .5037 .8039 .3753
2 h postfeeding .6915 1.000 .2929
Average .6915 .8039 .3753
Fractional digestion rate
2 h prefeeding .6642 .8741 .9457
2 h postfeeding .1916 .6364 .8253
Average .4253 .7142 1.000
Rumen apparent NDF digestibility
% of NDF intake
2 h prefeeding .8236 .8264 .5070
2 h postfeeding .0309 .4880 .1116
Average .3624 .9562 .3573
% of total tract digestibility
2 h prefeeding .8130 .9207 .7679
2 h postfeeding .0997 .2062 .8270
Average .1292 .2910 .6936

 

1The statistical model used was appropriate for balanced simple crossover designs.

 

240
Table 12. Enhanced intake and production of cows offered ensiled alfalfa silage with
higher neutral detergent fiber digestibility: apparent total tract digestibility of nutrients

using acid detergent lignin as an internal marker for 12 early lactation cows receiving low
digestible fiber (LDF) or high digestible fiber (HDF) diets.

 

 

Probability >
MIL LDF HDF SE F vain:
Fecal output, kg/d
DM 7.2 6.8 . 1 .008
NDF 4.1 3.8 . 1 .003
Nutrient digestibility, %
DM 63.0 67.0 .2 .001
(M 63.9 68.0 .3 .001
NDF 42.3 46.3 . .4 .001
ADF 40.9 44.6 .4 .001
Indigestible NDF1 1.1 5.0 1.2 .05
Hemicellulose 45.8 50. 1 .9 .006
Cellulose 53.4 57.0 .5 .001
CP 74.2 76.1 .3 .001

 

1Neutral detergent fiber residue after 120 h in vitro fermentation.

 

241

 

 

 

 

 

 

 

 

 

 

 

Terminal Boards Computer

Waterflow Meter pH Transmitter

 

l—-fi '— l— l—I
Exp-16 Exp- 16 Exp-l6
Expi 16 CTM 05 CTM 05
Qx
T

 
    

 

 

é;
........ WW wmm
n K ‘
Load Cell ' $1
\ i a
Pressure w
Transducer
V—L t3
pH

Pressure Electrode
‘ Transducer

\

      
   

 

 

 

 

 

 

 

 

 

 

Figure 1. Schematic of complete data acquisition system for continuous monitoring
of feed intake, water intake, chewing activity, reticular motility, and ruminal pH of
cannulated cows at Michigan State University Dairy Teaching and Research Farm.