EFFECTS OF STOCKING RATE AND SUPPLEMENTATION ON PASTURE
UTILIZATION, COW PERFORMANCE, AND RUMEN ENVIRONMENT IN A PASTUREBASED AND AUTOMATIC MILKING SYSTEM
By
Christine Nieman

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Animal Science
2012

ABSTRACT
EFFECTS OF STOCKING RATE AND SUPPLEMENTATION ON PASTURE
UTILIZATION, COW PERFORMANCE, AND RUMEN ENVIRONMENT IN A
PASTURE-BASED AND AUTOMATIC MILKING SYSTEM
By
Christine Nieman
Integration of automatic milking systems (AMS) into Midwestern pasture-based systems
has been suggested as an alternative low cost and labor efficient dairy system, but proper
integration of this technology into a pasture-based dairy system is not well understood. The
objective of this thesis was to determine management guidelines for a pasture-based AMS
system by examining the combined effect of stocking rate and feeding strategies on cow
performance, AMS use, feed degradation, dairy cattle genotype, and pasture utilization. An
investigation lasting 8 weeks was conducted using a completely randomized design with a 2 x 2
factorial of two stocking rate systems ( high stocking rate (HSR 2.89 cows/ha); low stocking rate
(LSR 1.92 cows/ha)) and two genotypes of Holstein cattle, United States Holstein (USH) and
New Zealand Holstein Friesian (NZF). The HSR treatment received a partial mixed ration (PMR;
40% ground corn and 60% legume grass haylage) in variable amounts (average: 4.86 kg /cow; ±
0.14; mean and SEM) to supplement deficits in pasture growth rate and availability. Results
indicated no differences in milk production or AMS use between treatments, although
differences were found between genotypes. An in situ experiment showed no differences in feed
degradation, suggesting that the rumen environment was not affected by PMR supplementation.
No differences in pasture utilization or pasture growth were detected between the HSR and LSR
treatments. The results of this thesis suggest that dynamic supplementation in the HSR treatment
did not affect pasture utilization, cow performance, AMS use, or the rumen environment.

ACKNOWLEDGEMENTS

Thank you to my major professor, Dr. Utsumi, for all of his direction and guidance
during my Master’s program. His passion for pasture-based dairy systems and all of his hard
work and dedication is inspiring. He has been instrumental in the development of my
understanding of scientific research in animal science. I greatly appreciate his commitment and
patience in teaching. I am especially grateful for his persistence in teaching, particularly during
long discussions concerning fundamental principles of grazing and pasture utilization; I never
would have understood it without him. I am very thankful to him for giving me this opportunity.
I greatly value all the things I have learned from him and will continue to use them in the future.
Thank you to my committee members Dr. Beede, Dr. Hamilton, and Dr. Rowntree for
their guidance and assistance through my Master’s program.
Thanks again to Dr. Hamilton and also to David Weed for their help with laboratory
procedures and the use of their laboratory equipment. Also, thanks to the LTER lab at KBS, for
generously allowing me to use their equipment for sample processing.
Lastly, sincere thanks to my lab partner and fellow graduate student Kate Steensma for
all of her help and support with this project. I am very thankful that she was willing to spend her
time assisting me with pasture sampling, rain or shine.

iii

TABLE OF CONTENTS

LIST OF TABLES ..................................................................................................................... vii
LIST OF FIGURES .................................................................................................................. viii
LIST OF ABBREVIATIONS ..................................................................................................... x
GLOSSARY................................................................................................................................. xi
INTRODUCTION........................................................................................................................ 1
CHAPTER 1
LITERATURE REVIEW
Supplementation of Pasture-based Diets ........................................................................... 4
Pasture Management .......................................................................................................... 8
Stocking Rates and Utilization......................................................................................... 10
Genetic Merit ................................................................................................................... 12
Automatic Milking and Pasture-Based Systems .............................................................. 16
Factors Affecting Traffic and Voluntary Milkings in AMS ............................................ 17
Distance to Pasture ........................................................................................................... 18
Location of Water Source ................................................................................................ 19
Milking Frequency ........................................................................................................... 20
Production Efficiency of Pasture-Based Dairy Farms ..................................................... 23
Integration of AMS and Farm Efficiency ........................................................................ 26
APPENDIX .................................................................................................................................. 29
REFERENCES ............................................................................................................................ 32
CHAPTER 2
EFFECTS OF STOCKING RATE AND SUPPLEMENTATION ON THE
PERFORMANCE OF TWO GENOTYPES OF HOLSTEIN FRIESIAN CATTLE IN A
PASTURE-BASED AND AUTOMATIC MILKING SYSTEM
ABSTRACT ................................................................................................................................. 37
INTRODUCTION ....................................................................................................................... 39
MATERIALS AND METHODS ................................................................................................. 41
Animals, Experimental Design, and Treatments ............................................................. 41
Milking Barn and Pasture Layout .................................................................................... 42
Pastures and Grazing Management .................................................................................. 43
Feeds and Pasture Sampling ............................................................................................ 44
AMS Data Collection and Milk Samples......................................................................... 44
Pasture Measurements and Dry Matter Intake ................................................................. 45
Laboratory Analysis ......................................................................................................... 46
Statistical Analysis ........................................................................................................... 49
RESULTS .................................................................................................................................... 51
iv

Pasture and Forage Quality .............................................................................................. 51
Cow Performance and Milk Composition ....................................................................... 51
DISCUSSION .............................................................................................................................. 54
Cow Performance............................................................................................................. 54
Milking Frequency ........................................................................................................... 58
Genotype .......................................................................................................................... 62
CONCLUSIONS.......................................................................................................................... 64
APPENDIX .................................................................................................................................. 67
REFERENCES ............................................................................................................................ 80
CHAPTER 3
EFFECTS OF SUPPLEMENTATION OF A PARTIAL MIXED RATION ON THE
FERMENTATION AND RUMEN ENVIRONMENT OF DAIRY CATTLE IN A
PASTURE-BASED AND AUTOMATIC MILKING SYSTEM
ABSTRACT ................................................................................................................................. 82
INTRODUCTION ....................................................................................................................... 83
MATERIALS AND METHODS ................................................................................................. 85
Animals, Experimental Design, and Treatments ............................................................. 85
Pasture Layout, Barn, Feeds, and AMS Settings ............................................................. 85
Rumen Fluid Samples ...................................................................................................... 85
In Situ Degradation .......................................................................................................... 85
Laboratory Analysis ......................................................................................................... 88
Statistical Analyses .......................................................................................................... 88
RESULTS AND DISCUSSION .................................................................................................. 89
Rumen Evacuations and pH ............................................................................................. 89
In Situ Degradation .......................................................................................................... 90
CONCLUSIONS.......................................................................................................................... 93
APPENDIX .................................................................................................................................. 95
REFERENCES .......................................................................................................................... 100
CHAPTER 4
EFFECTS OF STOCKING RATE AND PARTIAL MIXED RATION
SUPPLEMENTATION ON FORAGE UTILIZATION AND OCCUPATION OF
AUTOMATIC MILKING SYSTEMS IN A PASTURE-BASED SYSTEM
ABSTRACT ............................................................................................................................... 102
INTRODUCTION ..................................................................................................................... 104
MATERIALS AND METHODS ............................................................................................... 106
Animals, Experimental Design, and Treatments ........................................................... 106
Pasture Layout, Barn, Feeds, and AMS Settings ........................................................... 107
Pasture Measurements and Growth Rate Calculations .................................................. 107
AMS Data Collection and Milk Samples....................................................................... 108
Statistical Analyses ........................................................................................................ 109
RESULTS .................................................................................................................................. 111
Pasture Measurements ................................................................................................... 111
Cow Performance........................................................................................................... 111
AMS Milking and Distribution Performance ................................................................ 112
v

DISCUSSION ............................................................................................................................ 114
Pasture Measurements ................................................................................................... 114
Intake and Feed .............................................................................................................. 116
AMS use and efficiency ................................................................................................. 117
CONCLUSIONS........................................................................................................................ 119
APPENDIX ................................................................................................................................ 121
REFERENCES .......................................................................................................................... 129
CHAPTER 5
GENERAL CONCLUSIONS AND FUTURE RESEARCH ............................................... 131

vi

LIST OF TABLES

Table 2.1. Ingredients and nutrient composition of feed pellet in AMS .......................................67
Table 2.2. Pasture and partial mixed ration quality.......................................................................68
Table 2.3. Milk production, milk quality, weight change, milking frequency, and milking refusal
of dairy cows managed with automatic milking systems and two different stocking
rates ..............................................................................................................................69
Table 2.4. Intake estimates and feed conversion efficiency by dairy cows in an automatic
milking system with two different stocking rates ........................................................70
Table 2.5. Diet composition and estimates of crude protein, acid detergent fiber, neutral
detergent fiber intakes by dairy cows in a pasture-based automatic milking system
with two different stocking rates..................................................................................71
Table 3.1. Effects of stocking rate and feeding system on rumen contents, estimated kinetic
parameters and amount of masticates collected from orchardgrass and ryegrass based
pastures ........................................................................................................................95
Table 3.2. Effects of stocking rate and feeding system treatment on in situ degradation
parameters of concentrates and partial mixed ration ...................................................96
Table 3.3. Effects of stocking rate treatment on in situ degradation of ryegrass and orchardgrass
pastures ........................................................................................................................97
Table 4.1. Pasture, grazing management, feed intake, milk performance and feed conversion in a
low stocking rate and high stocking rate pasture-based feeding system ...................121
Table 4.2. Cow performance in automatic milking systems managed with a low stocking rate or
high stocking rate pasture-based feeding system .......................................................122
Table 4.3. Automatic milking system performance in a low stocking rate and high stocking rate
pasture-based feeding system ....................................................................................122

vii

LIST OF FIGURES

Figure 1.1. Instantaneous and average growth rate of forage ....................................................29
Figure 1.2. Effects of interval on milk yield and milk duration.................................................30
Figure 2.1. Layout of the split freestall .....................................................................................72
Figure 2.2. Paddock layout showing paddocks used for the high stocking rate and low stocking
rate groups ...............................................................................................................73
Figure 2.3. Pre-grazing and post-grazing biomass of pastures grazed at high and low stocking
rate by dairy cows managed with AMS. .................................................................74
Figure 2.4. Milk production by New Zealand Friesian and United States Holstein cows .........75
Figure 2.5. Milking frequency of New Zealand Friesian and United States Holstein cows ......76
Figure 2.6. Milking distribution of cows grazed at high stocking rate or low stocking rate .....77
Figure 2.7. Milking interval of cows grazed at high stocking rate or low stocking rate ...........77
Figure 2.8. Milking distribution of New Zealand Friesian and United States Holstein cows ...78
Figure 2.9. Milking interval of New Zealand Friesian and United States Holstein cows .........78
Figure 3.1. Rumen pH of dairy cows grazing temperate pasture in either a low stocking rate
system supplemented with concentrates or a high stocking rate system
supplemented with concentrates and partial mixed ration .....................................98
Figure 3.2. Disappearance of ryegrass NDF in dairy cows grazing temperate pasture in either
a98 low stocking rate system supplemented with concentrates or a high stocking
rate system supplemented with concentrates and partial mixed ration. .................99
Figure 3.3. Disappearance of orchardgrass NDF in dairy cows grazing temperate pasture in
either a low stocking rate system supplemented with concentrates or a high stocking
rate system supplemented with concentrates and partial mixed ration. .................99
Figure 4.1 Pre-grazing and post-grazing pasture biomass for high stocking rate and low stocking
rate pasture-based feeding systems .......................................................................123
Figure 4.2. Pasture growth rate means both for high stocking rate and low stocking rate pasturebased feeding systems ...........................................................................................124

viii

Figure 4.3. Consumption of different feedstuffs in the high stocking rate treatment in a pasturebased feeding system .............................................................................................125
Figure 4.4. Consumption of different feedstuffs in a low stocking rate treatment in a pasturebased feeding system .............................................................................................125
Figure 4.5. Milk yield for high stocking rate and low stocking rate pasture-based feeding
systems ..................................................................................................................126
Figure 4.6. Milking frequency for high stocking rate and low stocking rate pasture-based feeding
systems ..................................................................................................................126
Figure 4.7. Milking refusals for high stocking rate and low stocking rate pasture-based feeding
systems ..................................................................................................................127
Figure 4.8. Milking distribution for high stocking rate and low stocking rate pasture-based
feeding systems .....................................................................................................127

ix

ABBREVIATIONS

ADF = Acid Detergent Fiber
AMS = Automatic Milking System
CP = Crude Protein
DM = Dry Matter
DMI = Dry Matter Intake
HSR = High Stocking Rate
IVTD = In Vitro True Digestibility
LSR = Low Stocking Rate
MF = Milking Frequency
MY = Milk Yield
NDF = Neutral Detergent Fiber
NZF = New Zealand Holstein Friesian
OG = Orchardgrass
PMR = Partial Mixed Ration
RG = Ryegrass
RPM = Rising Plate Meter
SEM = Standard Error of the Mean
SR = Stocking Rate
TMR = Total Mixed Ration
USH = United States Holstein

x

GLOSSARY

Associative Effect: Decreases in the rate or extent of feed degradation in the rumen due to
interactions between the fermentation of concentrates and forages.
Box Time: The total amount of time the cow occupies the AMS stall.
Fetch Cow: A cow with a milking interval of 12 hours or greater.
Forage Utilization: Amount of forage harvested per area.
Genetic Merit: Generally referred to as the level of genetic potential for a cow to produce milk.
Grazing Intensity: The amount of pasture consumed/amount of pasture offered.
Milk Time: The amount of time from which the teat cups successfully attach, to when the last
teat cup detaches after milking.
Milking Frequency: The average number of successful milkings per day.
Milking Interval: The time between successful milkings.
Milking Refusal: Cows visiting the AMS without meeting the criteria for milking (i.e., the
milking is not predicted to result in a harvest of greater than 9 kg or the cow is visiting before the
full length of the set milking interval)
Pasture Residual: Biomass left in the pasture after grazing.
Prep Time: The amount of time from which the cow enters the AMS stall, until successful teat
cup attachment.
Substitution Rate: Reduction in pasture intake per kg of concentrate or forage supplemented.

xi

INTRODUCTION

The increased cost of milk production in Michigan puts small- and medium-sized dairies
at a competitive disadvantage. The Michigan dairy industry, similar to the national trend, has
seen a decrease in the number of farms and an increase in the number of cows per farm and the
amount of milk per cow (USDA, 2007). The rising cost of fuel, feed, and labor contributes to the
loss of small to medium sized dairy farms (Dartt et al., 1999). However, farmers with small
operations have the option of adopting new management strategies, such as managementintensive rotational grazing that reduces feed and labor costs, and produces a variety of
ecosystem services. Adoption of new technology, such as automatic milking systems (AMS) also
decreases labor costs and can increase milk production (de Koning and Rudenburg, 2004);
however, the initial cost of investment is high.
Well-managed pasture-based dairies can yield a net income higher or comparable to
conventional confinement dairies. Pasture-based dairy profitability is a result of the increase in
labor and operating efficiency (Dartt et al., 1999). However, pasture utilization is also a main
factor for creating profit. Therefore, grazing dairies must also manage the pasture, such that
pasture growth and utilization are optimized.
Assuming all aspects of management are correct, the integration of an AMS into a
pasture-based dairy has potential to increase the labor efficiency of a small dairy farm and result
in higher overall farm efficiency and profitability. However, AMS integration into a pasturebased system is poorly understood. There is little information regarding management strategies
that will maintain milk production on pasture, maximize pasture utilization, and optimize
efficient use of the AMS.

1

Chapter 1 of this thesis introduces the reader to most relevant aspects of pasture-based
dairy management and AMS functionality. The literature review is centered on the potential
impacts of stocking rates, supplementation strategies, and cow genotypes on pasture-based
system production efficiency. This review then follows with a description of the operation and
functional aspects of the AMS and concludes with a conceptual framework intended to provide
general ideas for the proper integration of the AMS into Midwestern pasture-based systems. The
original research presented in Chapter 2 compares feed intake, milkings, milk production and
quality, and overall performance of dairy cows managed with AMS in two dairy systems. The
systems differ in grazing and feeding strategy; a forage-based supplemented high stocking rate
system vs. an un-supplemented low stocking rate system. Experimental cows used in this
experiment include United States Holstein and New Zealand Holstein Friesians, therefore a
comparative discussion between these two genotypes is also provided. The in situ research
presented in Chapter 3 describes the supplementation strategy effects on the rumen environment
and kinetics. The in situ experiment was designed to detect the occurrence of any associative
effects and pH change resulting from the proposed feeding strategies tested in Chapter 2. Finally,
the concluding study discussed in Chapter 4 uses a system-based approach to compare the forage
utilization, milk production, and AMS utilization of the grazing systems presented in Chapters 2
and 3. Guidelines regarding stocking rate management and dynamic supplementation are
outlined and their applications to manage forage utilization and milk production are described in
the final discussion.
The combination of low feed and labor costs with the higher management efficiency of
the AMS has potential to increase the profitability and production efficiency of small- to
medium-sized dairy farms. However, information regarding the proper AMS management

2

integration in a pasture-based systems is scarce and research needs to be conducted to help
farmers better understand the potential of AMS integration in grazing dairy systems. This thesis
aims to present new practical information to help fill gaps of knowledge focused on improving
milk production efficiency and sustainability in pasture-based systems, through proper AMS
incorporation and efficient pasture utilization.

3

CHAPTER 1

LITERATURE REVIEW

Supplementation of Pasture-based Diets
Low milk production associated with pasture-based dairy systems discourages the
practice of grazing in the United States. Several studies comparing the feeding of a total mixed
ration (TMR) with pasture-based diets show consistently higher milk production in TMR diets
(Kolver and Muller, 1998; Bargo et al., 2002a and 2002b). Kolver and Muller (1998) found milk
production per cow was 44.1 kg/d for cows fed a TMR compared to 29.9 kg/d in a diet of pasture
alone. The difference in milk production between TMR and pasture-based diets is often due to
lower dry matter intake in pasture-based diets (DMI; Kolver and Muller, 1998). The TMR
feeding system allows maximum feed intake and therefore higher milk production per cow. A
strategy to increase grazing cow DMI and milk production is to supplement the pasture diet with
harvested feeds (Kolver and Muller, 1998).
Concentrate supplementation on pasture diets increases total DMI and milk production.
Reis and Combs (2000) supplemented grazing dairy cows with 0, 5, and 10 kg shelled corn daily
and determined that both DMI (13.9, 17.7, and 19.8 kg DM/d, for 0, 5, and 10 kg DM
supplemented, respectively) and milk production increased with supplementation (21.8, 26.8,
and 30.4 kg/d for 0, 5, and 10 kg DM of supplement, respectively). Further, Bargo et al. (2002a)
reported milk production increases from 19.1 to 29.7 kg/d and intake from 18.3 to 24.1 kg/d
when supplementing 8.6 kg DM/d of corn at a low pasture allowance. Both studies indicate that
supplementation can increase milk per cow. However, as noted in these studies, as the level of

4

concentrate intake increases the amount of pasture intake decreases, causing an increase in
substitution rate of pasture by supplement.
Intake substitution occurs when pasture DMI decreases due to supplementation (Bargo et
al., 2003). High substitution rates may be caused by a combination of many factors. One
hypothesis suggests that supplements combined with pasture cause associative effects between
digestive and metabolic interactions and the increase in energy intake, which result in lower
pasture intakes (Bargo et al., 2003). A second hypothesis suggests that supplementation
decreases grazing time, and therefore decreases pasture intake (Reis and Combs, 2000; Bargo et
al., 2002a).
In grazing, associative effects can be caused by disruption of fermentation patterns
induced by the presence of a concentrate. The readily fermentable starch can cause a decrease in
pH and inhibits fiber degradation (Bargo et al., 2003), which may decrease intake and milk
production. Walker et al. (2001) demonstrated associative effects when increasing levels of
cereal grain-based supplements were fed to grazing dairy cattle. A curvilinear response in milk
production, which is an indication of associative effects, was reported. However, a curvilinear
response is not always found with increased concentrate supplementation (Reis and Combs,
2000; Bargo et al., 2002a and 2002b). Reis and Combs (2000) demonstrated that feeding
increasing levels from 0, 5, to 10 kg DM/d of ground shelled corn to grazing dairy cows did not
cause a curvilinear milk response. Using in situ dry matter digestibility trials, Reis and Combs
(2000) indicated no differences in fiber DM ruminal degradability or pH (average 6.68),
although there was a reduction in the potentially degradable fraction of the forage DM and an
increase in the undegradable fraction for the treatment that received 10 kg DM/d.

5

In an attempt to decrease associative effects or to complement a pasture-based diet,
forages have also been used to supplement grazing dairy cattle. Graf et al. (2005) supplemented
grazing dairy cows with hay at night in an attempt to increase rumination activity and stabilize
rumen pH and fermentation patterns. Hay supplementation in the evening led to decreases in
rumen pH for longer durations of time during the day compared to the control (no
supplementation). No increases in intake or milk production in response to the forage supplement
were reported. Similarly, Holden et al. (1995) found that supplementation of 2.3 kg/d corn silage,
plus 1 kg of corn-based grain per 4 kg or milk, did not affect milk production or DMI in cows
grazing orchardgrass dominated pasture. However, supplementing corn silage at the rate of 1 kg
per 4 kg of milk did not increase milk production or intake.
Forage may also be supplemented in the form of a partial total mixed ration (PMR)
(Soriano et al., 2001; Bargo et al., 2002b). Supplementing with a forage-based PMR has the
potential to increase energy intake, but discourages the formation of associative effects caused by
concentrate only supplementation. Soriano et al. (2001) compared three treatment groups: cows
consuming TMR only ration, cows supplemented with PMR in the morning and grazing on
pasture in the afternoon, and cows supplemented with PMR in the afternoon and grazing on
pasture in the morning. The researchers found PMR intakes for the half day PMR consumption
and half day grazing groups were between 65.8 and 76.3% compared to the TMR only treatment.
Yet, milk production in the two grazing groups was still comparable to the TMR-only treatment,
and was 28.2, 27.6, and 29.1 kg/d for the afternoon grazers, morning grazers, and TMR only,
respectively.
Bargo et al. (2002b) completed an extensive comparison between three diets, including
pasture plus concentrate (PC), a total mixed ration (TMR) and a combination of pasture and

6

TMR defined as partial TMR (pTMR). Bargo et al. (2002b) found that the pTMR treatment
group was moderate in DMI and milk production compared to the other groups. Dry matter
intake was found to be 21.6, 25.2, and 26.7 kg/d (P<0.05) and milk production was 28.5, 32, and
38.1 kg/d for the PC, pTMR, and TMR groups, respectively. Ruminal pH was not found to differ
among treatment groups (average pH of 5.87); rumen ammonia concentrations were significantly
higher in the PC treatment indicating that the inclusion of a pTMR increases the capture of
nitrogen. Bargo et al. (2002b) discovered that the pTMR group, although receiving a balanced
TMR, still experienced associative effects. This was seen in the in situ results where the
potentially degradable fraction of DM and the potentially degradable fraction of NDF were
reduced compared to in the PC treatment. Although supplementing with a pTMR may result in
some negative associative effects, the addition of pTMR into grazing cow diets allows cows to
maintain or increase feed intake and milk production (Soriano et al., 2001; Bargo et al., 2002b).
Hoffman et al. (1993) supplemented cow diets with varying amounts of grain over the
grazing season based on the quality and quantity of pasture. Cows received either 1 kg of grain
per 4 kg of milk, 1 kg of grain per 4 or 5 kg of milk based on pasture quality and availability, or
1 kg of grain per 3 kg of milk as control. Grain supplementation was reformulated based on
pasture quality, which was determined by chemical analysis of pasture samples collected one
week prior to the supplement reformulation. Over the 24 week trial no difference was seen in
milk production and only the control treatment of 1 kg of grain per 3 kg of milk gained body
weight. This suggests that grain supplementation based on changes in pasture quality and
quantity caused no difference in milk production. However, declining pasture availability over
the grazing season (approximately 1,800 kg/ha in May to approximately 1000 kg/ha in
September) in this study indicated the amount supplemented was not enough to compensate for

7

the limited pasture availability. Supplementation should therefore be used to manage growth
rates and future pasture availability, which manages pasture intake and thus helps to prevent
subsequent pasture deficits.
Pasture Management
Pasture supplementation can be used to add dietary energy and aid in pasture
management. Seasonal changes in pasture quality and quantity creates variability in the diet,
affecting nutrient intake. Supplementation can compensate for seasonal pasture variation, but
must interface with effective pasture management principles to ensure efficient pasture
utilization. Fulkerson et al. (2005) reported that supplementing grazing dairy cattle based on
pasture biomass can maintain milk production and post-grazing residuals. Maintaining desirable
post-grazing residual pasture biomass favors high pasture growth rates and prevents overgrazing
and eventual decline of pre-grazing biomass over the grazing season (Fulkerson et al., 2005).
Supplementation in pasture-based dairy farms should be adjusted to encourage optimum pasture
growth and utilization, and ultimately profitability (Gronow et al., 2010). However, the
supplementation studies mentioned prior were interested in the effects of pasture management on
animal production and not the effects on pasture growth rates and forage utilization. To optimize
pasture production and overall farm efficiency, focus must also be placed on pasture
productivity.
There are trade-offs between pasture growth and utilization. Plants require leaves for
photosynthesis. As the plant reaches potential maximum leaf area for biomass accumulation, it
will start to turn leaves over, or leaves will die and drop off the plant (senesce; Parsons et al.,
1998). To optimize utilization, plants need to be grazed before the older leaves in a tiller start to
senesce; however, pastures cannot be grazed so low that there is not enough leaf area to sustain

8

the required rate of photosynthesis. Overgrazing removes nearly all leaf tissue, and reduces rates
of photosynthesis and results in a decrease in biomass by requiring the plant to use reserves to
restore leaf tissue (Parsons et al., 1998). Balance needs to be made between leaving sufficient
leaf area behind for photosynthesis and growth versus pasture consumption.
Maximum pasture utilization is achieved when the sward is grazed at maximum average
pasture growth rate (see Parsons et al., 1998; Figure 1.1). Maximum average pasture growth rate
accounts for both biomass accumulated and the total time that has elapsed since previous
defoliation. This strategy takes advantage of the high growth rate in the beginning of regrowth,
but harvest is complete before the biomass reaches its ceiling yield and begins to senesce.
Importantly, the growth rate changes seasonally; during early spring growth rate peaks, and as
summer approaches growth rates slows. Supplementation and stocking rate management are
tools to manage forage utilization as pasture growth changes throughout the season.
The Parsons et al. (1998) review showed that in order to optimize (i.e., maximize) pasture
utilization, defoliation must occur at maximum average growth rate, which occurs after some
accumulation of biomass over time. Farmers often use fixed time, height, or target biomass
values to determine defoliation frequency. Fulkerson and Donaghy (2001) describe these
methods as inaccurate. Time is the least accurate method as environmental factors greatly affect
regrowth, and plant height is inaccurate because it does not fully reflect changes in biomass.
Biomass is described as the one of the most accurate measurement, but is also affected by
physiological and environmental factors. For example, the plant may not reach the target
biomass for defoliation because it is subject to environmental stress (i.e., water, nutrients,
temperature), or the plant may no longer continue to grow at a high rate if it has already reached
an advanced stage of physiological development or maturity (Fulkerson and Donaghy, 2001).

9

In their review, Fulkerson and Donaghy (2001) described leaf number to be the best
plant-related indicator associated with regrowth. Perennial ryegrass (Lolium perenne) develops
three leaves per tiller before the oldest leaf starts to senesce, which sets the maximum grazing
interval for this plant species. Ryegrass grazed before two leaves retards regrowth and reduces
persistence and quality. However, when ryegrass is grazed at three leaves it has been shown to
increase net accumulation of biomass, persistence and quality. Setting a general goal for pasture
biomass is a useful strategy for grazing systems because future biomass can be estimated by the
average growth rate, however, everyday decisions need to made based on leaf number to ensure
that plants are not senescing (representing unused forage growth), decreasing pasture utilization.
Maintaining and optimizing pasture growth can be managed by varying stocking rates over the
season or by integrating supplementation.

Stocking Rates and Utilization
Stocking rate is the most important grazing management variable affecting milk
production per cow and per area in a pasture-based system. Increasing stocking rate decreases
milk production per cow, but increases milk production per area (Fales et al., 1995; Kennedy et
al., 2006; Macdonald et al., 2008). The decrease in milk production is due to the decrease in
pasture intake per cow (Macdonald et al., 2008). However, increasing stocking rates increases
pasture utilization, and the amount of pasture consumed/amount of pasture offered, often referred
to as grazing efficiency (Baudracco et al., 2011). The increase in grazing efficiency may be
caused by a decrease in grazing selectivity particularly in early spring when forage availability is
high (Baker and Leaver, 1986; Fales et al., 1995). Fales et al. (1995) reported that increasing the
stocking rate from 2.5, 3.25, and 4.0 cows/ha resulted in less rejected forage, with rejected

10

fractions decreasing from 43.6, 31.8, and 21.8% in spring for the low, medium, and high
stocking rates, respectively. It is critical to harvest the forage before it reaches its ceiling biomass
accumulation in early spring; otherwise it leads to decreased utilization due to senescence and
cattle selectivity (Fales et al., 1995). Proper stocking rates also increase overall sward quality
(Baker and Leaver, 1986, Fales et al., 1995) by encouraging uniform defoliation closer to the
plant growing points resulting in the promotion of new tillers (Baker and Leaver, 1986). The
short, uniform defoliation promotes new growth, which increases the total production of herbage
per area (Baker and Leaver, 1986, Macdonald et al., 2008).
Pasture production is a major factor of farm profitability (Gronow et al., 2010), therefore,
increasing pasture production is the main goal in most pasture-based dairies. A study conducted
in New Zealand by Macdonald et al. (2008) demonstrates that high pasture production and
utilization are possible with high stocking rates. The stocking rate treatments were 2.2, 2.7, 3.1,
3.7, and 4.3 cows/ha. Intake per cow decreased with increasing stocking rate from 5,438, 5,018,
4,575, 4,214, to 3,887 kg DM/cow per year, respectively, and milk production followed a similar
decreasing trend from 5,032, 4,351, 4,128, 3,616 to 3,448 kg/cow/year. However, as stocking
rates increased, the annual amount of forage consumed per hectare increased from 12,098,
13,785, 14,322, 15,609 to 16,597 kg of DM/ha per year, and milk production/ha increased from
11,071, 11,747, 12,796, 13,380 to 14,828 kg/ha, respectively. The increase in pasture consumed
per hectare led to greater amounts of pasture production per hectare, increasing pasture
utilization. Not only was milk production per hectare higher for the highest stocking rate system,
but interestingly these values were achieved on shorter lactation lengths and grazing days.
Lactation length was 291, 274, 258, 234, and 221 days for the treatments with 2.2, 2.7, 3.1, 3.7,
and 4.3 cows/ha, respectively. The lower pasture residuals led to a decrease in pasture pre-

11

grazing biomass, yet cows still produced more milk per hectare. Macdonald et al. (2008) was
able to demonstrate how high stocking rates may increase pasture growth and milk production
per hectare when all other aspects of pasture management (i.e. fertilization) are correct.
However, this strategy of producing the greatest amount of milk per hectare, which is successful
in New Zealand, may not be effective for Michigan, particularly when considering the genetic
merit of cows involved.

Genetic Merit
Genetic merit is generally referred to as the level of genetic potential for a cow to
produce milk and is an important factor for production in pasture-based dairy systems. Grazing
genotypes are generally categorized into two or three levels of genetic merit; a low genetic merit
or New Zealand grazing-based genetics, a high genetic merit or high production North American
genetics, and a medium genetic merit or high durability genotype (McCarthy et al., 2006; Horan
et al., 2005; Kennedy et al., 2003). High durability is a crossbred selected for milk production,
fertility, and muscularity traits. In recent years, the role of genetic merit in grazing systems has
been extensively examined. Countries with low genetic merit cows have been increasing their
production per cow by incorporating North American genetics, but now there are rising concerns
about reduced reproductive performance (Buckley et al., 2000; Fulkerson et al., 2001; Dillion et
al., 2006). Other studies examined the existence of genotype by environment (G x E) interactions
(Horan et al., 2005; Fulkerson et al., 2008). If a G x E interaction exists, and there is an
indication that certain genotypes perform better in certain feeding systems compared to others,
this information will guide farmers to choose their genetics accordingly.

12

The feeding systems that are generally tested for G x E interactions are often different
combinations of concentrate level and stocking rates. Buckley et al. (2000a) compared two
levels of genetic merit, high and medium, in three treatment groups including high stocking rate
with concentrate, high stocking rate plus twice the concentrate as the high stocking rate system,
and a high pasture allowance group with the same concentrate supplementation as the high
stocking rate system. No G x E interaction was detected, but cows of high genetic merit
consumed more pasture and total DM and produced more milk in each feeding system. A lack of
a G x E interaction for milk production traits has been seen in similar studies. No G x E
interaction in grazing-based experiments may suggest that there is not a substantial difference
between either the environmental conditions or the genotypes compared in the studies, or that G
x E interactions need more time to be detected.
Horan et al. (2005) compared three genetic genotype, including New Zealand Friesian,
high durability, and high producing North American cows across three feeding systems: a control
group (2.47 cows/ha and 368 kg concentrate/cow/yr), a high stocking rate (2.74 cows/ha and 364
kg concentrate/cow/yr), and a high concentrate diet (2.47 cows/ha and 1452 kg
concentrates/cow/yr). The results for milk production for the high pasture allowance, high
stocking rate, and high concentrate systems was 6,799, 6,283, and 7,877 kg/cow for the North
American cows, respectively and 6,039, 5,940, and 6,444 kg/cow for the New Zealand cows,
respectively. The differences in milk production between the two genotypes were 760, 343, and
1,433 kg/cow for high pasture allowance, high stocking rate, and high concentrate treatments,
respectively. These differences suggests that the high production North American cows out
performed in the three feeding systems and had a significantly higher milk yield response to
concentrate supplementation compared to the New Zealand cows. Similar differences in milk

13

production between high and low genetic merit genotypes are also reported in the studies
conducted by Kennedy et al. (2003) and McCarthy et al. (2007).
Change in body condition score (BCS) is another metric often used to compare animal
performance between different genotypes. Horan et al. (2005) reported above, BCS differed
among the three feeding systems. Cows in the high concentrate feeding system lost less body
condition from calving to nadir (0.51) compared with the high stocking rate and high allowance
feeding systems (0.64 and 0.72, respectively). There was also a significant difference between
genotypes for BCS. The New Zealand genotype had the highest BCS postpartum (3.37), at nadir
BCS (2.84) and at dry-off (3.13) compared to the high production cows which scored a 3.17,
2.45, and 2.68 BCS at postpartum, nadir and dry off, respectively. A consistently higher BCS in
New Zealand cows may suggest that this genotype is more suitable for production systems where
pasture availability and feed intake is low. Conversely, in pasture-based systems that can sustain
higher levels of pasture intake, a North American genotype may be more efficient.
The results from Horan et al. (2005) also indicate important differences between
genotypes. In this study, high production cows had a higher milk response to concentrate
compared to New Zealand genotypes, which may suggest that a pasture-only diet was not
sufficient to meet the energy requirements of the high production cows. This is also supported by
the larger decrease in BCS from post calving to nadir BCS, for high production cows (0.72)
compared to New Zealand cows (0.53). The lack of response in the New Zealand cows indicates
that they were closer to meeting their energy requirements and also a higher proportion of their
potential milk production. Kennedy et al. (2003) found that high production cows on forage
based diets are not capable of consuming more forage than cows of lower genetic merit, probably
because of the physical bulkiness of the grass or low rate of pasture intake. Therefore, in

14

situations where pasture intake is limiting, supplementation with digestible energy concentrates
increases the energy density of the diet and may increase DMI, potentially resulting in a greater
milk yield response. However, low genetic merit cows do not always show a high milk response
to concentrate supplementation because in most cases, pasture intake is not a limiting factor for
this genotype (Kennedy et al., 2003, McCarthy et al., 2006). Substitution rates are lower in high
producing cows because their energy requirements are not fully met with pasture only diets.
Therefore, adding concentrate supplementation to the diet of high producing cows generally
increases total dry matter intake compared to pasture only diets (Kennedy et al., 2003).
Genotype may also influence grazing behavior. McCarthy et al. (2006) investigated the
genetic effect (New Zealand, high durability, and high production) and feeding system (high
pasture allowance, high stocking rate, or high concentrate supplementation) on grazing behavior,
feed intake, and milk production. The high pasture allowance (2.47 cows/ha) and high stocking
rate (2.74 cows/ha) did not receive concentrate supplementation and the supplemented treatment
received 3.5 kg DM/day at a stocking rate of 2.47 cows/ha. McCarthy et al. (2006) reported that
the high durability genotype had the shortest grazing time, while the New Zealand genotype had
the longest grazing time and the high production genotype was intermediate. The high
production and New Zealand cows had the same number of bites per day, which were
significantly higher than the high durability cows. The high production cows had a significantly
higher biting rate than the New Zealand cows, whereas the high durability cows were
intermediate and not significantly different from the other genotypes.
These results are remarkable in two ways. The high biting rate by high producing cows
suggested a greater eating drive, and higher motivation to graze or capacity to chew and harvest
bites at the same time (Laca et al., 1994). The lower biting rate exhibited by the high durability

15

cows suggested that these cows exhibited a higher degree of grazing selectivity. The grazing
time for the high production and high durability cows decreased as stocking rate or pasture
allowance increased. At high pasture allowance, grazing time was 615 and 573 min for the high
durability and high production cows, respectively. At high stocking rate, grazing time was 553
min and 536 min for the high durability and high production cows, respectively. The longer time
spent grazing (631 min) in the high stocking rate system for the New Zealand cows, may indicate
that these cows have more drive to graze or willing to graze more selective compared to the other
genotypes.
The genotype comparison studies reported above were completed in countries other than
the U.S., namely Ireland and Australia (Kennedy et al., 2003; Horan et al., 2005; McCarthy et
al., 2006; Fulkerson et al., 2008). The main drive to investigate G x E interactions in pasturebased systems is the concern that North American genetics may have decreasing reproductive
performance, as a result of decreased pasture intake and BCS, as reported in Fulkerson et al.
(2008). Kolver and Muller (1998) established that North American Holsteins are energy limited
on pasture-based diets, and are not capable of reaching their genetic potential on pasture alone.
For this reason, there is increased interested in incorporating New Zealand genetics into pasturebased systems in the US, as these cows have lower energy requirements (Kennedy et al., 2003;
Horan et al., 2005; McCarthy et al., 2006). Further, there has been little research comparing
genotypes of Holstein in the US, and the feasibility and economics of New Zealand genetics or
New Zealand crosses in North American grazing systems has yet to be investigated.

Automatic Milking and Pasture-based Systems

16

There has been increased interest in investigating how the integration of AMS affects
farm efficiency in various types of dairy production systems. Garcia and Fulkerson (2005)
identified three potential benefits of integrating an AMS. The first benefit is an up to 20%
reduction in labor hours (de Koning and Rudenburg, 2004). The reduction in hours also leads to
a reduction in labor costs; Dijkhuizen et al. (1997) found an annual savings of up to $200/cow
was possible. An AMS allows the farmer to devote time to other aspects of farm and pasture
management, family, or recreation time. The second benefit was a potential increase in milk
production/cow (de Koning and Rudenburg, 2004). In confinement systems, this potential
increase in milk production is generally due to the increase in milking frequency from two to
three milkings/day (de Koning et al., 2002). Lastly, AMS can result in reduced stress on cows
and also gives the farmer access to data that can aid in management decisions.
Despite the extensive research conducted in AMS in recent years, there is little
information available to aid farmers in the adoption of AMS into a pasture-based system.
Grazing dairies must of manage the challenges of pasture-based systems with AMS, namely,
motivating cows to voluntarily and independently visit the AMS from pastures located away
from the milking barn.

Factors Affecting Traffic and Voluntary Milkings in AMS
Grain is generally offered in the AMS stall, as it has been shown to be the main motive
for cows to visit the AMS (Prescott, 1998). Jago et al. (2007) offered either 0 kg or 1 kg of
barley in the AMS to test system efficacy. While average milkings per day was not affected, the
supplemented group visited the selection unit more often, indicating that the supplementation
resulted in a higher motivation for cows to return to the barn. The supplementation treatment also

17

increased milk yield per cow by 1.12 kg. Sporndly and Wredle (2004) compared feeding grass
silage ad libitum versus feeding limited to 3 kg of grass silage, both in the barn, while cows also
grazed equal distances from the barn. No increase in milk production or milking frequency for
the ad libitum silage group was reported. Sporndly and Wredle (2004) suggest that milk
production did not increase in the ad libitum silage group due to adequate amounts of high
quality pasture available and most likely resulted in substitution of the pasture for silage.
Ketelaar-de Lauwere et al. (2000) also studied the effect of a low sward height (10.8-11
cm) or a high sward height (12.3-14.5 cm) on milking frequency in a pasture-based AMS farm.
The researchers found an increase in total milking visits for cows on the low sward heights (2.65
and 2.95 milkings/d for high and low sward height, respectively), although grazing time did not
differ. However, no information was given about the effects of these sward heights on intake or
milk production. Depending on milk production, additional milkings may not be necessary
because may not result in the harvesting of additional milk.

Distance to Pasture
Distance to pasture has been referred to as an obstacle to efficient AMS utilization in
pasture-based systems. Ketelaar-de Lauwere et al. (2000) demonstrated that a distance of 360
and 355 m versus 146 and 168 m between the AMS milking barn and the pasture did not affect
milking frequency. However, Sporndly and Wredle (2004) saw a decrease in milking frequency
and milk production when cows traveled 260 m vs. 50 m from the milking barn to the pasture.
The difference in milking frequency in Sporndly and Wredle (2004) may be caused by the
extremely close pasture distance of 50 m, compared to Ketelaar-de Lauwere et al. (2000) where a
difference in milking frequency between approximately 150 and 350 m from the milking barn

18

was not seen. Sporndly and Wredle (2004) also noted that the motivation to graze decreased over
time, as time spent grazing decreased from 20 to 9% for the long distance, while the cows that
were a short distance from the barn maintained grazing time at 23 and 21%, respectively.
Sporndly and Wredle (2004) concluded that cows alter their behavior as the season progressed.
Further, a constant distance longer than 260 m decreases grazing time, and presumably DMI;
although it is not known whether alternating close and far distances would have the same effect.
This information may be valuable to farms with larger herds, in which cows must travel to
distant pastures over 260 m but also have closer pastures available in their grazing rotation.

Location of Water Source
Water location has been investigated as a source of motivation for cows to visit the AMS.
Ketalaar-de Lauwere et al. (1999) found that pasture-based dairies with AMS and water only
available in the milking barn may be restricting water consumption. Cows on restricted pasture
access (12 hours) and full pasture access spent less time drinking water than cows with no
pasture access (20.3, 21.6, and 27.5 min, for the restricted, full, and no pasture access treatments,
respectively). Although, the values for the cows with pasture access are lower, it may be due to
luxury water consumption in the restricted treatments or grazing dairy cows are obtaining their
required water from the high water content of fresh pasture. Sporndly and Wredle (2005)
investigated the effects of offering water in the barn only versus offering water in both the
milking barn and in the pasture. The researchers found no difference in milking frequency, milk
production, or water intake. The group with water only offered in the milking barn drank 40% of
their daily water intake within 30 min of entering the barn and the group with water in the barn
and field drank 50% of their water in the field. The researchers concluded that water does not

19

motivate cows to visit the AMS, as milking frequency did not differ between groups. However,
cows in the treatment with water in the field and barn had spent more time outdoors and grazing,
indicating that cows in the treatment offering water in the milking barn only spent more time in
the barn. The researchers did not record grazing bouts, but it may be possible that cows in the
barn only treatment may have had a higher number of grazing bouts that did not result in
additional milkings. Cows receiving water in the barn only may also have gone to the barn to
drink and stayed there for an extended period of time without having additional motivations (i.e.
supplemental feed) to visit the AMS more frequently. Although offering water in the field
resulted in more time spent grazing on pasture, offering water in the pastures often results in
uneven distribution of nutrients or differences in pasture productivity as a result of uneven
grazing pressures, soil compaction and concentration of urine and feces near water source
(Mathews et al., 1994). It may also be important for farmers to consider the possibility of added
labor and installation cost for providing water in the field at least in small to medium sized
pasture-based systems with relatively short distances to pastures (< 350 m).

Milking Frequency
Increasing daily milkings generally increases milk production per cow (Stelwagen, 2001;
Andre et al., 2010,). Stelwagen (2001) indicated an increase from two to three milkings per day
resulted in an increase in milk production of 18%. The AMS has the potential to increase milk
production per cow without increasing labor costs. Daily milking frequency and duration are
restricted by the amount of time the AMS has available for milking, which subsequently defines
the AMS capacity. Other time considerations for the AMS are rinsing, cleaning, and idle time.
Idle time is necessary to prevent overcrowding and waiting time at the AMS.

20

Milk yield per milking depends on the preceding milking interval (time between
milkings; Hogeveen et al., 2001; Andre et al., 2010). Shorter intervals increase milk production
per cow per hour; a greater response is seen in high production cows (Hogeveen et al., 2001).
Hogeveen et al. (2001) found that intervals greater than 18 hours can result in decreased daily
milk production. Similarly, Guinard-Flamnet (2006) stated that daily milk yield displayed a
curvilinear pattern with increasing milking intervals from 8 to 24 hours. The challenge for any
AMS system is balancing milk yield, milk duration, milking intervals, and number of cows per
AMS stall.
Milk yield per cow on a daily basis increases as the length of the interval between
milkings decreases, however, the total milking duration per day increases exponentially (Andre
et al., 2010; Figure 1.2). The increase in milking duration is restricted by the capacity of the
AMS, which is dependent on the cleaning time, frequency of cleaning, and idle time (Andre et
al., 2010). Decreasing milking intervals must create a balance between milk yield and milk
duration. Particularly because the amount of milk collected must justify the amount of time a
cow spends in the AMS, the milking duration of a cow is affected by the time that the AMS
requires to clean, prepare and attach teat cups to the cow. Therefore, the ideal or optimal milking
interval should maximize the amount of milk production per unit of milking time and AMS
occupation (i.e. box time). The optimal milking interval is therefore expected to differ as milking
speed changes and/or as feed intake and milk secretion change between cow breeds, feeding
systems, and lactations. The milking interval will differ based on stage of lactation, parity, and
time of day.
Maintaining short milking intervals is especially challenging in pasture-based AMS. In a
pasture-based AMS farm, management focuses on motivating cows to voluntarily return to the

21

AMS from pasture. Compared to a confinement feeding system, cow behavior and herd activities
are more synchronized and gregarious on pasture (Ketelaar-de Lauwere et al., 2000), creating
poorly distributed milkings throughout the day and in some situations longer waiting periods for
milking in the AMS. Compared to confinement feeding systems, feeding patterns in a grazing
system are more gregarious and also more erratic or variable as they are increasingly affected by
environmental conditions, varying feed quality and quantity, distance to pasture, and distance to
shading and water sources.
As previously stated, short milking intervals result in higher milk production. However,
in a pasture-based AMS cows are physically farther away from the AMS and need to exert time
and energy to return to the barn. In addition, cows on pasture, compared to TMR diets, produce
less milk (Kolver and Muller, 1998; Bargo et al., 2003). Not only do pasture-based AMS have to
motivate cows to visit the AMS, but cows must have sufficient milk to collect. A pasture-based
AMS may expect to find longer milking intervals, resulting in higher milk production per AMS,
but not per cow. Jago et al. (2007) studied the effects of two intervals of milking permissions
from pasture, 6 h vs. 12 h, in a New Zealand grazing system. In this study the permission of a
cow to visit the AMS was controlled with the use of a selection unit in the pasture that
determined whether cows could leave the pasture to visit the AMS for milking or to stay in the
pasture. Jago et al. (2007) found milking intervals of 12.56 h and 16.90 h for 6 h and 12 h
interval treatments respectively, and fewer number of milkings per cow for the 12 h interval
(1.42 milkings/day) compared to the 6 h interval (1.91 milkings/day). The 12 h permission
treatment had a higher milk yield per milking (16.40 kg) compared to the 6 h treatment (11.96
kg), but milk yield per day was not significantly different between permission treatments. Also,
milk production of a low producing cow in a low input pasture-based system was not affected by

22

a longer milking interval. The decrease in milking frequency with the greater milking permission
interval may allow farmers to increase their stocking rate to compensate for the fewer milkings
per cow without affecting the production per cow.

Production Efficiency of Pasture-based Dairy Farms
Recently, system-based or farmlet studies determined the efficiency of whole farm
systems with varying supplementation amounts, stocking rates or combinations of both factors
(MacDonald et al., 2008;Farina et al., 2011). MacDonald et al. (2008) observed increases in milk
production and pasture utilization with increases in stocking rate up to 4.3 cows/ha. In this study,
milk production per cow was sacrificed but milk production per ha increased as stocking rate
increased, indicating pasture utilization importance.
Another strategy to increase milk yield per ha is to increase milk production per cow, a
practice which is particularly important in systems where stocking rate and forage utilization is
near maximum. Farina et al. (2011) investigated the differences in milk production efficiency
between farming systems increasing milk production per cow by feeding supplements, increasing
stocking rates, or a combination. The researchers incorporated four total treatments: 1) a low
stocking rate (2.5 cows/ha) with low milk production used as control (C); 2) a low stocking rate
with high production (HMY); 3) a high stocking rate (3.8 cows/ha) with low production (HSR);
4) high stocking rate with high production (HH). Milk production goals for low and high
production levels were 6000 and 9000 kg/cow/lactation, cows were fed supplements to reach
those production goals. The pasture management was based on pasture growth rate and
supplemented if necessary, ensuring that pastures met the pre and post grazing targets of 2400
and 1500 kg DM/ha to optimize pasture growth. This pasture management should allow for cows

23

on the high stocking rate systems to maintain moderate milk production, compared to systems in
which high stocking rate treatments are not supplemented, leading to decreased milk production
(Macdonald et al., 2008). Due to the use of this grazing management strategy, Farina et al.
(2011) did not find any difference between treatments for total pasture utilized.
Farina et al. (2011) established that the milk production per ha increased with both
increased stocking rate and that the combination of high stocking rates and supplementation. The
system with increased stocking rate (HSR) led to a proportional increase in milk production/ha of
49%. The system with the combination of both higher stocking rates and supplementation (HH)
led to a proportional increase in milk production/ha of 66%. Interestingly, the increase in
stocking rate did not affect cow performance at either high or low target levels of milk
yield/cow, although at each stocking rate level, milk yield/cow was higher for the high milk yield
groups (HH, HMY).
The high milk/cow systems maintained a higher mean BCS (P<0.0001), but this higher
BCS did not lead to increased reproductive performance, indicating that the cows were receiving
excess feed to the requirements for normal reproductive performance. The high BCS also
indicates a higher marginal efficiency of supplements for the HSR system. The lack of response
in milk production and increased body condition score may also indicate a low to moderate
genetic merit of the cows used in this study, which may also influence the feed conversion
efficiency of the farming system.
Optimizing feed conversion efficiency (FCE) in grazing dairy systems is challenging.
The FCE is generally evaluated as kg of energy-corrected milk (ECM) per kg DM consumed or
utilized by animals. Inclusion of the North American Holstein genotype in grazing systems has
potential to increase the FCE because of their high milk response when feed availability is not a

24

limiting factor (Kennedy et al., 2003; Horan et al., 2005). The North American Holstein
genotype was selected on high energy TMR diets, and their production on pasture is limited due
to behavioral and physiological constraints that limit intake (Kolver et al., 1998; Bargo et al.,
2002). As previously mentioned, Kennedy et al. (2003) determined that there is not much
difference between genetic merit in total pasture intake when North American and New Zealand
Friesians are offered unlimited amounts of high quality pasture. Pasture allowance, biomass and
structure determine the upper limit to total pasture intake in a grazing system. To take advantage
of the high genetic potential of the North American dairy cow, supplements need to be added to
increase energy intake and production, which will result in higher FCE. Proper pasture
management will also increase FCE by ensuring the availability and quality of forage, which can
impact milk yield and FCE (Beever and Doyle, 2001).
Beever and Doyle (2001) stated that the decrease in intake and milk production in grazing
dairy cows increase the maintenance cost and, therefore, decrease FCE. A supplement with high
metabolizable energy (ME) contents should increase milk production and FCE. However, at
some point increasing supplementation may lead to high substitution rates and/or a curvilinear
pattern for milk production and FCE due to negative associative effects. Tozer et al. (2004)
found significant differences between supplemented and un-supplemented groups in FCE for
milk yield but not for fat yield. Farina et al. (2011) also investigated the level of supplementation
on FCE and found that there was no difference between the high and low supplementation
groups.
In most dairy grazing systems, there is a relationship between stocking rate and FCE.
McCall and Clark (1999) described a trade-off between pasture utilization and FCE, concluding
that at higher stocking rate pasture utilization was higher, but FCE decreased. Tozer et al. (2004)

25

reported FCE for milk yield or milk fat did not decrease from pasture allowances of 40 kg
DM/cow/day vs. 25 kg DM/cow/day (total intake differed significantly by 2.9 kg DM). The
authors concluded that high genetic merit cows in grazing systems with low pasture allowance
convert pasture DMI into milk or milk components as efficiently as in a high pasture allowance
system. Farina et al. (2011) found similar results; no significant different in FCE with increasing
stocking rates. Farina et al. (2011) credited the lack of difference to the maintenance of constant
feed intake levels through a supplementation plan that covered pasture availability deficits. It
was also noted that supplements in a high stocking rate system resulted in a higher marginal feed
efficiency. However, the high yielding cow group had significantly higher BCS which was most
likely the cause of the lack of difference in FCE as energy was partitioned to body reserves and
not to milk production.

Integration of AMS and Farm Efficiency
To increase farm efficiency in pasture-based AMS dairies, one must balance all aspects
of the farm. Supplementation of cow diets with concentrates will increase milk production per
cow and FCE, but that requires very high levels of concentrate supplementation (greater than 10
kg of concentrates) and undesirable associative effects may occur (Reis and Combs, 2000; Bargo
et al., 2003). Ultimately, high levels of concentrate supplementation can decrease milk
production and FCE (Walker et al., 2011). Supplements, particularly forages, may cause
substitution and decrease pasture intake (Bargo et al., 2003). However, forage supplements may
be used to maintain post-grazing pasture residuals and improve pasture growth and utilization by
preventing overgrazing, which is particularly important during periods or seasons of low pasture
growth (Fulkerson et al. 2005). Increasing stocking rates may decrease pasture intake and milk

26

production and FCE per cow, but increases forage utilization and milk production per hectare
(Macdonald et al. 2008). In pasture-based systems with AMS, the increase in stocking rate may
increase AMS occupation and milking efficiency if the number of cows allotted to an AMS or
the stocking rate of the AMS is a limiting factor (Jago et al., 2007). New Zealand based genetics
may require less feed to reach their genetic potential, which improves BCS and reproduction
performance (Fulkerson et al., 2008), but North American Holsteins are high yielding and have
potential to increase milk production due to their higher marginal response to high quality
pasture and energy concentrates. Increasing milk production per visit or milking intervals in the
AMS will increase milking duration, possibly decreasing the number of total milkings per robot
per day (Andre et al., 2010), and potentially decreasing milk production per robot if no
adjustments in the AMS stocking rate (cows per AMS) are made (Jago et al., 2007). Farmlet
studies have the opportunity to investigate the consequences of stocking rates, grazing
management, supplementation and genetic merit on the overall efficiency of the farm. Further
investigation is required to determine the necessary conditions to maximize farm efficiency in
pasture-based systems using AMS. This could be a particularly important alternative to decrease
inputs and labor costs that may help small and medium sized dairy farms remain profitable in
Michigan.

27

APPENDIX

28

𝑊

𝑊 − 𝑊0
𝑡

𝑊2

𝑊1

Crop weight, W (t OM ha-1)

7

Instantaneous and average growth rate
(kg OM ha-1 day-1)

125

𝑑𝑊
𝑑𝑡

𝑊0
0

0
a

b

c

Time (t)

Figure 1.1. The change in pasture instantaneous growth rate (dW/dt), the weight of the forage
(W), and the average growth rate (W-W0)/t (based on Parsons 1998). Grazing at maximum yield
would occur at time ‘c’ from a single period of regrowth between W-W0. Harvesting the sward
at time ‘a’ would interrupt pasture growth when instantaneous growth rate was at maximum, but
the amount of forage harvested would be reduced. Maximum utilization is achieved when the
forage is harvested when the average growth rate (W-W0)/t is at maximum. Maximum average
plant growth rate is identified at time ‘b’.
For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this thesis.

29

Milk Yield (kg)

Milk Duration (min)

Milk Yield/Day
Milk Duration/Day

Interval (h)
Figure 1.2 Effects of milking interval on milk yield and milking duration per day (Andre et al.,
2010).

30

REFERENCES

31

REFERENCES

Andre, G., P. B. M. Berentsen, B. Engel, C. J. A. M. de Koning, and A. G. J. M. Oude Lansink.
2010. Increasing the revenues from automatic milking by using individual variation in
milking characteristics. J. Dairy Sci., 93: 942–953.
Baker, A.-M.C.and J. D. Leaver. 1986. Effect of stocking rate in early season on dairy cow
performance and sward characteristics. Grass and Forage Sci., 41: 333–340.
Bargo, F., L. D. Muller, E. S. Kolver, and J. E. Delahoy. 2003. Invited Review: Production and
digestion of supplemented dairy cows on pasture. J. Dairy Sci., 86: 1-42.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002a. Milk response to concentrate
supplementation of high producing dairy cows grazing at two pasture allowances. J.
Dairy Sci., 85: 1777–1792.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002b. Performance of high
producing dairy cows with three different feeding systems combining pasture and total
mixed rations. J. Dairy Sci., 85: 2960–2975.
Baudracco, J., N. Lopez-Villalobosa, L.A. Romerob, D. Scandolob, M. Macielb, E.A.
Comeronb, C.W. Holmesa, and T.N. Barrya. 2011. Effects of stocking rate on pasture
production, milk production and reproduction of supplemented crossbred Holstein–Jersey
dairy cows grazing lucerne pasture. Anim. Feed. Sci. Technol, 168: 131-143.
Beever, D. E. and P. T. Doyle. 2007. Feed conversion efficiency as a key determinant of dairy
herd performance: a review. Aust. J. Exp. Agric., 47: 645–657.
Buckley, F., Dillon, P., Rath, M. and Veerkamp, R.F., 2000. The relationship between genetic
merit for yield and live weight, condition score, and energy balance of spring calving
Holstein Friesian dairy cows on grass based systems of milk production. J. Dairy Sci., 83:
1878– 1886.
Dartt, B. A., J. W. Lloyd, B. R. Radke, J. R. Black, and J. B. Kaneene. 1999. A comparison of
profitability and economic efficiencies between management-intensive grazing and
conventionally managed dairies in Michigan. J. Dairy Sci., 82: 2412-2420.
de Koning, K., Y. van der Vorst, and A. Meijering. 2002. Automatic milking experience and
development in Europe. In Proc. First N. Am. Conf. on Robotic Milking, Toronto,
Canada. Wageningen Academic Publishers, Wageningen, the Netherlands. pp I1–I11.
Dijkhuizen, A. A., R. B. M. Huirne, S. B. Harsh, and R. W. Gardner. 1997. Economics of robotic
application. Computers Electronics Agric., 17: 111–121.

32

Dillon, P., D.P. Berry, R.D. Evans, F. Buckley and B. Horan. 2006. Consequences of genetic
selection for increased milk production in European seasonal pasture-based systems of
milk production. Livest. Sci., 99: 141-158.
Fales, S. L., L. D. Muller, S. A. Ford, M. O’Sullivan, R. J. Hoover, L. A. Holden, L. E. Lanyon,
and D. R. Buckmaster. 1995. Stocking rate affects production and profitability in a
rotationally grazed pasture system. J. Prod. Agric., 8: 88–96.
Farina, S. R., S. C. Garcia, W. J. Fulkerson and I. M. Barchia. 2011. Pasture-based dairy farm
systems increasing milk production through stocking rate or milk yield per cow: pasture
and animal responses. Grass Forage Sci., 66: 316-332.
Fulkerson, W. J. and D. J. Donaghy. 2001. Plant-soluble carbohydrate reserves and senescence –
key criteria for developing an effective grazing management system for ryegrass-based
pastures: a review. Aust. J. Exp. Agric., 41: 261–275.
Fulkerson, W. J., K. McKean, K. S. Nandra and I. M. Barchia. 2005. Benefits of accurately
allocating feed on a daily basis to dairy cows grazing pasture. Aust. J. Exp. Agric. 45:
331-336.
Fulkerson, W. J., S. C. Garcia. 2005. Opportunities for future Australian dairy systems: a review.
Aust. J. Exp. Agric., 45: 1041-1055.
Fulkerson, W. J., T. M. Davison, S. C. Garcia, G. Hough, M. E. Goddard, R. Dobos, and M.
Blockey. 2008. Holstein-Friesian dairy cows under a predominantly grazing system:
interaction between genotype and environment. J. Dairy Sci., 91: 826–839.
Graf, C. M., M. Kreuzer, and F. Dohme. 2005. Effects of supplemental hay and corn silage
versus full-time grazing on ruminal pH and chewing activity of dairy cows. J. Dairy Sci.,
88: 711-725.
Gronow, C. S., S. B. Little, and G. V. Cleary. 2010. Managing production system risk: An
application of Monte Carlo simulation to Tasmanian dairy farm data. Proceedings of the
4th Australasian Dairy Science Symposium. Location of the meetings, country. pages
Guinard-Flament, J., S. Lemosquet, E. Delamaire, G. Le Bris, P. Lamberton, C. Hurtaud. 2011.
Alteration of the nutrient uptake by the udder over an extended milking interval in dairy
cows. J. Dairy Sci., 94: 5458-5468.
Hoffman, K., L. D. Muller, S. L. Fales, and L. A. Holden. 1993. Quality evaluation and
concentrate supplementation of rotational pasture grazed by lactating cows. J. Dairy Sci.,
76: 2651-2663.
Hogeveen, H., W. Ouweltjes, C. J. A. M. de Koning, and K. Stelwagen. 2001. Milking interval,
milk production and milk flow-rate in an automatic milking system. Livest. Prod. Sci.,
72: 157–167.
33

Holden, L. A., L. D. Muller, T. Lykos, and T. W. Cassidy. 1995. Effect of corn silage
supplementation on intake and milk production in cows grazing grass pasture. J. Dairy
Sci., 78: 154–160.
Horan B., Dillon P., Faverdin P., Delaby L., Buckley F. and Rath M. 2005. The interaction of
strain of Holstein Friesian cows and pasture-based feed systems on milk yield, body
weight and body condition score. J. Dairy Sci., 88: 1231–1243.
Jago, J. G., K. L. Davis, P. J. Copeman, I. Ohnstad, and M. M. Woolford. 2007. Supplementary
feeding at milking and minimum milking interval effects on cow traffic and milking
performance in a pasture-based automatic milking system. J. Dairy Res., 74: 492–499.
Kennedy J., Dillon P., Delaby L. and Faverdin P. 2003. Effect of genetic merit and concentrate
supplementation on grass intake and milk production with Holstein Friesian dairy cows.
J. Dairy Sci., 86: 610–621.
Kennedy, E., M. O’Donovan, J. P. Murphy, F. P. O’Mara, and L. Delaby. 2006. The effect of
initial spring grazing date and subsequent stocking rate on the grazing management, grass
dry matter intake and milk production of dairy cows in summer. Grass Forage Sci., 61:
375–384.
Ketelaar-de Lauwere, C. C., A. H. Ipema, C. Lokhorst, J. H. M. Metz, J. P. T. M. Noordhuizen,
W. G. P. Schouten, and A. C. Smits. 2000. Effect of sward height and distance between
pasture and barn on cows’ visits to an automatic milking system and other behavior.
Livest. Prod. Sci., 65: 131–142.
Ketelaar-de Lauwere, C. C., A. H. Ipema, E. N. J. van Ouwerkerk, M. N. W. B. Hendriks, J. H.
M. Metz, J. P. T. M. Noordhuizen, W. G. P. Schouten. 1999. Voluntary automatic
milking in combination with grazing of dairy cows milking frequency and effects on
behavior. Appl. Anim. Behav. Sci., 64: 91-109.
Kolver, E. S., L. D. Muller, G. A. Varga, and T. J. Cassidy. 1998. Synchronization of ruminal
degradation of supplemental carbohydrate with pasture nitrogen in lactating dairy cows.
J. Dairy Sci., 81: 2017–2028.
Koning, K., and J. Rodenburg. 2004. Automatic milking: State of the art in Europe and North
America. In: Automatic Milking—A Better Understanding. A. Meijering, H. Hogeveen,
and C. J. A M. de Koning (Eds). Wageningen Academic Publishers, Wageningen, the
Netherlands. pp. 27–37
Laca, E.A., Ungar, E.D. & Demment, M.W. (1994). Mechanisms of handling time and intake
rate of a large mammalian grazer. App. Anim. Behav Sci., 39 : 3–19.
Leaver, J. D. 1985. Milk production from grazed temperature grassland. J. Dairy Res., 52: 313344.

34

MacDonald, K. A., J. W. Penno, J. A. S. Lancaster and J. R. Roche. 2008. Effect of stocking rate
on pasture production, milk production, and reproduction of dairy cows in pasture-based
systems. J. Dairy Sci., 91: 2151–2163.
Mathews, B. W., L. E. Sollenberger, V. D. Nair, and C. R. Staples. 1994. Impact of grazing
management on soil nitrogen, phosphorus, potassium, and sulfur distribution. J. Environ.
Qual., 23 :1006-1013.
McCall, D. G. and D. A. Clark. 1999. Optimized dairy grazing systems in the northeast United
States and New Zealand. II. System analysis. J. Dairy Sci., 82: 1808–1816.
McCarthy, S., B. Horan, M. Rath, M. Linnane, P. O’Connor and P. Dillon. 2006. The influence
of strain of Holstein-Friesian dairy cow and pasture-based feeding system on grazing
behaviour, intake and milk production. Grass Forage Sci., 62: 13–26.
Parsons, A. J. and D. F. Chapman. 1998. Grass for dairy cattle. Principles of grass growth and
pasture utilization. In: Grass for dairy cattle. (Eds). CAB International,Wallingford,
United Kingdom,. pp 283309..
Prescott, N. B., T. T. Mottram, and A. J. F. Webster. 1998. Relative motivations of dairy cows to
be milked or fed in a Y-maze and an automatic milking system. Appl. Anim. Behav. Sci.,
57: 23-33.
Reis, R. B., and D. K. Combs. 2000. Effects of increasing levels of grain supplementation on
rumen environment and lactation performance of dairy cows grazing grass-legume
pasture. J. Dairy Sci., 83: 2888–2898.
Soriano, F. S., C. E. Polan, and C. N. Miller. 2001. Supplementing pasture to lactating Holsteins
fed a total mixed ration Diet. J. Dairy Sci., 84: 2460–2468.
Sporndly, E., and E. Wredle. 2004. Automatic milking and grazing—Effects of distance to
pasture and level of supplements on milk yield and cow behavior. J. Dairy Sci., 87:
1702–1712.
Sporndly, E., and E. Wredle. 2005. Automatic milking and grazing—Effects of location of
drinking water on water intake, milk yield, and cow behavior. J. Dairy Sci. 88: 1711–
1722.
Stelwagen, K. 2001. Effect of milking frequency on mammary functioning and shape of the
lactation curve. J. Dairy Sci., 84(E. Suppl.): E204-E211.
Tozer, P. R., F. Bargo, and L. D. Muller. 2004. The effect of pasture allowance and
supplementation on feed efficiency and profitability of dairy systems. J. Dairy Sci., 87:
2902–2911

35

United States Department of Agriculture. 2007. Economic Research Service. Profits, costs, and
the changing structure of dairy farming. Accessed April 26, 2011.
http://www.ers.usda.gov/publications/err47/
Walker, G. P., C. R. Stockdale, W. J. Wales, P. T. Doyle and D. W. Dellow. 2001. Effect of level
of grain supplementation on milk production responses of dairy cows in mid–late
lactation when grazing irrigated pastures high in paspalum (Paspalum dilatatum Poir.).
Aust. J. Exp. Agric., 41: 1–11.

36

CHAPTER 2
EFFECTS OF STOCKING RATE AND DIET SUPPLEMENTATION ON THE
PERFORMANCE OF TWO GENOTYPES OF HOLSTEIN FRIESIAN COWS IN A
PASTURE-BASED AND AUTOMATIC MILKING SYSTEM
ABSTRACT

Despite the numerous investigations of Automatic Milking Systems (AMS) in recent
decades, management guidelines for the proper integration of this technology into a pasturebased system are not yet well developed. The study objective was to determine the effects of
stocking rate and supplementation strategies on milk yield, milk quality and milking frequency
of mid-lactation cows managed in an AMS pasture-based system. A farmlet-based experiment
lasting 8 wk was conducted using a completely randomized design with a 2 x 2 factorial
arrangement of two stocking rate systems and two genotypes of Friesian cows, United States
Holstein (USH) and New Zealand Friesian (NZF). At study onset, fourteen mid-lactation (154 ±
25 DIM; mean ± SD) USH (BW: 564 ± 38 kg; Parity: 1.43 ± 0.10; mean ± SD) and NZF (BW:
393 ± 27 kg; Parity: 1 ± 0; mean ± SD) cows were randomly assigned to one of two SR
treatments, a low stocking rate treatment (LSR; 1.92 cows/ha) and a high stocking rate treatment
(HSR; 2.89 cows/ha). The HSR treatment received a partial mixed ration (40% ground corn and
60% legume grass haylage) in variable amounts (average: 4.86 kg /cow ± 0.14; mean ± SEM) to
supplement deficits in pasture growth rate and availability across the 8 weeks of the study. In the
two stocking rate treatments cows received 1 kg of pellet concentrate (ME: 0.65 Mcal, CP:
17.61%, NDF: 42.81%) per 6 kg of milk in the AMS stall and 1.36 kg of grounded corn. Cow
performance responses included milk production and milking frequency and were analyzed as
repeated measures using a mixed model. Several significant differences (P < 0.05) existed
between genotypes, indicating important differences in genetic merit particularly for milk
37

production (15.3 and 27.9 g/cow per d, ± 1.3 SEM) for NZF and USH, respectively, P < 0.0001).
However, few differences existed between stocking rates, with no significant difference between
milk yield per cow (20.85 and 22.41 kg/cow per d, ± 1.11 SEM for LSR and HSR, respectively).
There was a significant time effect, as well as time by treatment interactions related to changes in
pasture availability and quality throughout the study. As predicted, few differences were detected
between stocking rates, but both the duration of grazing and the genotype resulted in strong
differences in milk production and use of the AMS system.
Keywords: pasture-based, stocking rate, genotype, automatic milking systems

38

INTRODUCTION

Despite the extensive research conducted in recent decades on AMS, few evaluations of
the performance of this technology have been conducted in a pasture-based system. This
information is still particularly important to develop appropriate management guidelines for the
successful integration of AMS in dairy systems where grazed pastures remain as the main dietary
component of cow diets.
Multiple management and animal related factors could potentially affect the efficiency of
AMS in pasture-based systems. The stocking rate of pastures and the stocking rate of the AMS
(cows per AMS stall) could influence cow traffic, milking frequency, and overall AMS
occupation. Pasture stocking rate and supplementation strategies are two grazing management
factors that have potential to affect pasture utilization, feed intake and milk production (Fales et
al., 1995; Farina et al., 2011).
Previous grazing studies show that the number of milking visits and total AMS
occupation (i.e. time that the AMS spent milking) are not strongly affected by the amount of
supplemental feed offered in the AMS milking barn (Sporndly and Wredle, 2004) or by levels of
concentrate supplementation delivered in the AMS (Jago et al., 2007). However, conclusions on
AMS performance developed from the comparative analysis of different pasture-based systems
are difficult due to the large differences in feeding and grazing practices, management of the
AMS, and cow genotype used.
Recent research has focused on the comparison of different genotypes of Holstein dairy
cattle and their adaptability to different feeding practices in a pasture-based system (Buckley et
al., 2000; Kennedy et al., 2003; Horan et al., 2005; Fulkerson et al., 2008). In most of these
studies, the genotype of cows was tested using conventional milking, therefore direct
39

extrapolations of cow performance to an AMS using free traffic and voluntary grazing systems
are difficult to conduct. The AMS milking efficiency could be significantly impacted by the
genetic potential of cows, through differences in milking intervals, milk yield and the overall
AMS use.
The objective of the present study was to determine the effects of stocking rate and
dynamic feeding of supplemental TMR on pasture utilization, feed intake, milk production and
milking distribution of two genotypes of Holstein Friesian cows in a dairy system using the same
AMS milking stalls. Specific predictions were that: 1) Dynamic supplementation would not
affect pasture utilization, total feed intake, or milk production, but would alter the visitation
patterns to the AMS; and 2) The United States Holstein genotype would out-perform the New
Zealand Friesian genotype in milk production.

40

MATERIALS AND METHODS

Animals, Experimental Design, and Treatments. The study was conducted at the
Michigan State University's Pasture Dairy Research Center (PDRC) located at the W.K. Kellogg
Biological Station, Hickory Corners, MI, during the months of July, August, and September of
2011. Protocols for animal handling and husbandry have been previously approved by the
Michigan State University's IACUC office. The PDRC has 64 ha of pasture growing on soils that
consist mostly of sandy loam of the Kalamazoo loam soil series (USDA-NRCS, 2009). The
mean annual precipitation ranges from 760 to 915 mm, the mean annual air temperature ranges
from 7 to 9ºC, and the frost-free period ranges from 140 to 150 d.
The study employed a completely randomized design with a 2 x 2 factorial treatment
arrangement of two genotypes and two stocking rate systems: a high stocking rate system (HSR,
2.89 cows/ha) and a low stocking rate system (LSR, 1.92 cows/ha). Fourteen United States
Holstein cows (USH) and 14 New Zealand Friesian cows (NZF) were randomly assigned at
study onset to one of two stocking rate treatments. A partial mixed ration (PMR) of 60% legume
grass haylage and 40% ground corn supplemented the cows’ diets in the high stocking rate
system. The amount of PMR fed was adjusted weekly and was used to supplement any deficit in
pasture growth rate (PGR) and availability. The supplementation criterion for target pasture
intake per cow was 14 kg DM/d or approximately 36.4 Mcal of ME from grazed pasture. Based
on this criterion, PMR supplementation occurred weekly if PGR and/or pasture availability for
grazing were lower than the expected values (PGR x SR). Target PGR values were 26.88
kg/ha/day (1.98 cows/ha x 14 kg/cow per d) or 40.46 kg/ha per d (2.98 cows/ha x 14 kg cow/d)
for the LSR and HSR treatments, respectively. Similarly, the total amount of PMR fed in a given

41

day was the deficit between pasture availability and the target pasture intake. The diet of cows in
the LSR treatment included 1 kg of pellet concentrate per 6 kg of milk (Table 2.1) fed in the
AMS and 1.36 kg of ground corn fed at the automatic Cosmix feeder (Lely Cosmix, Lely
Industries, N. V., Maassluis, Netherlands). The HSR treatment included the same pasture and
concentrate feeding system as LSR plus the addition of the PMR fed at the bunk. The PMR was
fed at 0500 h daily and consisted of 60% legume-grass haylage and 40% ground corn mixture
(Table 2.2). As indicated above, the PMR amount fed was adjusted weekly based on weekly
changes in pasture growth rate. Therefore the amount of PMR fed varied by week, but average
was 4.86 ± 0.14 kg DM/cow/day (mean ± SEM) during the study.
Milking Barn and Pasture Layout. The milking barn was a split free stall barn equipped
with two automatic milking systems (AMS, Lely Astronaut 3, Lely Industries, N. V., Maassluis,
Netherlands), two automated grain feeders (Lely Cosmix, Lely Industries, N. V., Maassluis,
Netherlands ), two automatic manure scrapers and free stalls equipped with double chamber
water beds. Based on the barn design, number of free stalls and feed rail space per side, this barn
had capacity to house and milk two separate herds of approximately 60 cows each in a free
traffic system.
Two automated gates (Lely Grazeway, Ley Industries, N. V., Maassluis, Netherlands)
and split two-way lanes controlled the free traffic of cows between the pastures and the milking
barn (Figure 2.1). The freestall allowed for free traffic; cows had free access to stalls, the AMS,
or pasture if the criteria for exit had been met. Pasture access was monitored by the automated
gate which allowed cows to leave the free-stall milking barn if they had been milked within a set
milking interval given by a minimum and maximum number of milkings per day, or if their
expected milk production was less than 9 kg, whichever applied first. Maximum and minimum

42

number of milkings per day were: 5 and 4 for cows < 30 DIM: 4 and 3 for cows between 31 and
30 d prior to dry off, and, 3 and 2 for cows with less than 29 days prior to dry off. Dry off
occurred at 256 DIM.
Pastures and grazing management. The LSR treatment had access to twenty-four 1
hectare strips (Figure 2.2). Sixteen of the 24 strips were a mix of 5 dominant plant species (OG)
including orchardgrass (Dactylis glomerata), tall fescue (Festuca arundinacea), alfalfa
(Medicago sativa), white clover (Trifolium repens), and red clover (Trifolium pratense). The
remaining 8 strips were a mix of 2 species (RG) including white clover and perennial ryegrass
(Lolium perenne). The HSR treatment had access to sixteen 1 hectare strips. The strips contained
the same OG and RG mixes as LSR, but there were 8 hectares of each pasture mixture in the
LSR treatment. Fresh pasture allocations were opened for grazing at 0500 h every morning.
Pastures were rotationally grazed based on approximate targets of 2,400 kg DM and of 1,600 kg
DM for pre- and post-grazed biomass, determined by a rising plate meter. When necessary,
pastures were grazed based on maturity, indicated by leaf number for grasses and flower
maturity for legumes, in order to harvest the mature material and stimulate regrowth.
The HSR treatment had access to a fresh 0.5 ha at any one time. A temporary fence was
used to prevent cows from grazing the second half of the 1 ha strip. On the second day of grazing
the temporary fence was removed, opening the new allocation allowing cows to back-graze. The
barn’s traffic exit system allowed for separating the cows' traffic to either north or south strips of
the dairy barn (Figure 2.2).
Allocations were switched every other day from north to south for the LSR treatment,
forcing cows to enter the barn before moving to the next allocation. The traffic was similar for
HSR, except that they switched from north to south every two days because they stayed on one

43

strip for two days. An attempt was made to balance distance to the dairy by alternating far and
near pasture for both treatments, however, pasture maturity and biomass took precedence over
distance. The PGR in the HSR and LSR treatment was measured weekly using a rapid pasture
meter (C-Dax rapid pasture meter, C-Dax, Agricultural Solutions, Ltd., Palmerston North, NZ)
Feeds and pasture sampling. A composite haylage sample was taken from the bunk and
ground corn samples were taken from the automatic feeder. A sample of each pasture species
was taken weekly with clippings from three 50 x 50 cm quadrats taken both before and after
grazing. Quadrat clippings were taken randomly, but spaced over the length of the entire grazing
strip. The first quadrat was placed approximately 50 steps from the end of the pasture, the
remaining quadrats were spaced approximately 50 steps from each other, resulting in one
clipping taken from each third of the one ha strip. Forage samples were dried in a forced air oven
for 48 h at 60°C and ground with a Christy mill through a 1 mm screen and stored for analysis of
CP, NDF, ADF, and in vitro true digestibility.
Each week and one week prior to the study onset, biomass was recorded in the HSR and
LSR treatments with a portable optical measurement system (C-Dax rapid pasture meter,
Agricultural Solutions, Ltd., Palmerston North, NZ ). The difference in biomass between weeks
divided by the number of days determined the PGR of each strip. An average PGR was used to
determine the amount of supplementation, so that the amount of PMR supplemented covered the
deficit in PGR. If a pasture had been grazed in between readings it was not included in
calculation of the average.
AMS Data Collection and Milk Samples. The AMS identified each cow by an assigned
collar equipped with electronic ID and collected information on body weight, milk production,
milking frequency, AMS visits, and amount of pellets distributed. Milk samples were taken three

44

times during the study and analyzed for fat, protein, and milk urea nitrogen at the DHIA lab in
Lansing, MI.
Pasture Measurements and Dry Matter Intake. An estimate of pasture pre- and postgrazing biomass was made with a rising plate meter (F300 plate meter, Farmworks, Palmerston
North, NZ). The difference between the pre- and post-grazing biomass (i.e., the disappearance of
pasture) was considered pasture harvested. The units for height measurements given by the plate
meter are “clicks”; 1 “click” equals 0.5 cm of pasture compressed by the plate meter. The plate
meter was calibrated in the previous year and the calibration factor to determine biomass (y) as a
function of RPM clicks (x) in the orchardgrass, tall fescue, white clover, alfalfa, and red clover
pasture was y = 92 x. The factor to determine biomass in the ryegrass and white clover pasture
was y = 78 x. Pellet and corn consumption were known through AMS the computer system. An
estimate of PMR intake per cow was made by dividing the total amount fed to the herd by the
total number of cows in that treatment.
During week 4 of the experiment an indigestible marker, Cr2O3, was used to determine
total dry matter intake (DMI). Twelve of the 14 experimental cows from each treatment were
orally dosed with 10 g of Cr2O3 once daily for 10 days. In the final five days of dosing fecal
grab samples were collected once daily at dosing. Fecal samples were dried at 55°C in a forced
air oven. Samples were ground through a 1 mm screen and the grinder was cleaned with
compressed air between samples to avoid cross contamination. Samples were combusted at
600°C for 2 h. The ash was dissolved in the crucible by a phosphoric acid-manganese sulfate
solution and potassium bromate over a hot plate until effervescence ceased, approximately 30
min. Once the sample was cool it was washed into a beaker with distilled water and then poured
into a 50-ml flask. Calcium chloride was added and the flask was brought to a volume of 50 ml
45

with distilled water. Standards were made following the same procedure, with the exception of
the use of a blank fecal sample and stock chromium added in concentrations of 0, 2.5, 5, 10, and
20 ppm. The standards were used to create a standard curve to derive the Cr content for the
samples. Flame atomic absorption spectroscopy (AAS; Model 3100, Perkin Elmer, Waltham,
MA) using an acetylene-air flame was used to measure Cr in the samples. The wavelength
setting was set at 358 nm and fuel:oxidant ratio was 3.5:2.5. The AAS recorded five readings per
sample and the average was used for calculations.
Fecal Cr concentration (ppm) was used to calculate fecal output by the equation: Cr
dosed per day (6.842 g/Cr ppm in sample x 100)/1,000,000. Total dry matter intake was
calculated by dividing the total fecal output by the percent indigestibility of the diet, determined
to be 26.64%. Indigestibility was determined by 48 in vitro true digestibility (IVTD) of all feeds
in the diet and their approximate amount in the diet. Total diet digestibility did not differ
between the two treatments; therefore the same value of indigestibility was used for both.
Laboratory Analysis. Neutral detergent fiber (NDF) analysis was completed with the
filter bag technique (Ankom 200 Fiber Analyzer, Ankom Technology, Macedon, NY). Filter
bags (Ankom F57) with pore size of 25 microns were weighed and filled with 0.5 g of sample
and sealed with a heat sealer (Model AIE-200, American International Electric Inc., Industry,
CA). Sample preparation and NDF determination were conducted in duplicate. A premixed
solution of NDF (30.0g Sodium dodecyl sulfate, USP; 18.61g ethylenediaminetetraacetic
disodium salt, dehydrate; 6.81g sodium borate; 4.56 g sodium phosphate dibasic anhydrous) and
10.0 ml triethylene glycol were added to 1 L of distilled water. Twenty four complete filter bags
and one blank were placed in the carousel and placed in the Ankom 200 vessel. The NDF
solution, 4 mL of alpha amylase, and 20g of sodium sulfite anhydrous were added to the vessel.

46

The unit was set for 1 hour and 15 min with agitation, at a temperature of 100°C. After
incubation, the unit and the samples were rinsed twice with water plus 4 mL of alpha amylase,
and rinsed once with only water, for a total of three 5 min rinses. Samples were dried in a forced
air oven at 60°C for 24 hours and reweighed. The equation for NDF determination was NDF% =
x 100, where W1 = bag tare weight, W2=sample weight, W3=dried weight of
bag with fiber after extraction process, and C1 = blank bag correction (final oven dried weight
divided by the original blank bag weight).
Acid detergent fiber (ADF) analysis was also completed with the filter bag technique
(Ankom 200 fiber analyzer, Ankom Technology, Macedon, NY). After NDF was completed on
the samples they were saved and kept for ADF analysis. An Ankom ADF liquid premix (20 g
cetyl trimethylammonium bromide to 1 L 1.00/V H2SO4) was used and duplicate samples were
heated to 100°C and agitated for 1 hour in the Ankom 200 vessel. After analysis, the samples
were rinsed in the vessel 3 times for 5 min with water. Samples were dried in a forced air oven at
60°C for 24 hours and reweighed. The equation to determine % ADF was same as the previous
equation given for NDF.
In vitro true digestibly (IVTD) was completed with the Ankom DAISYII Incubator
(Ankom Technology, Macedon, NY). The two buffer solutions for the procedure were made in
the laboratory. Buffer Solution A combined 10.0 g KH2PO4, 0.5 g MgSO4·7H2O, .05 g NaCl, 1
g CaCl2·2H2O, and 0.5g of reagent grade urea in 1 liter of distilled H2O. Buffer Solution B
combined 15 g Na2CO3 and 1 g Na2S·9H2O in 1 liter of distilled H2O. The buffers were
warmed to 39°C with a hot plate and combined in a 1:5 ratio. The amount of the combined buffer

47

solution added was adjusted to obtain a final pH of 6.8 at 39°C. A total of 1600 ml of the
combined A/B solution mixture was added to each digestion jar.
Two 2 L thermo bottles were warmed to 39°C with water and were not emptied until just
before collection of rumen inoculum. Two ruminally fistulated cows managed under the same
feeding systems as the experimental cows were used to collect approximately 4 L of rumen
inoculum. A filter attached to a rubber tube was inserted into the rumen. The rubber tube was
also attached to flask to capture the fluid. Another attachment on the plastic flask allowed for
another plastic tube fitted with a rubber hand pump to create a vacuum. Fluid was taken from
both cows and mixed in the thermos. From both cows approximately one fistful of rumen digesta
was added to each thermos.
The inocula were brought back to the lab within 30 min and added to a previously
warmed blender (39°C) and the fluid and the digesta were blended for approximately 30 sec.
While blending, the blender was purged with CO2. After blending, the inoculum was squeezed
through four layers of cheese cloth and 400 ml of filtered inoculum was added to each warm
digestion jar. The digestion jar was purged with CO2 for 30 s and samples were added before
sealing. Duplicate filter bags (Ankom F57, Ankom Technology, Macedon, NY) were weighed
and filled with 0.5 g of sample material and sealed with an impulse sealer (American
International Electric Inc., Model AIE-200, Industry, CA) before the IVTD procedure began.
Blanks and standards were also included in each jar to determine correction factors and indicate
possible variation between jars. The jars were placed in the incubator and rotated and heated to
39°C for 48 hours of incubation.
After incubation samples were soaked in cold water and rinsed by hand over a sink for
approximately 2 min each. Samples were dried in a forced air oven at 60°C for 24 hours and
48

reweighed. To calculate the % IVTD the following equation was used, % IVTD =
(

(

))

x100; where W1 is bag tare weight, W2 is sample weight, W3 is final
bag weight after in vitro incubation, and C1 is blank bag correction (final oven-dried
weight/original blank bag weight).
Nitrogen determination was conducted by combustion using an automated analyzer
(Costech Elemental Combustion System 4010 CN, Costech Analytical Techologies Inc.,
Valencia, CA). This system uses a helium carrier gas to bring combustion gases to a gas
chromatographic separation column and thermal conductivity detector. Therefore, the results
give the percentage N, which was multiplied by 6.25 to obtain the % CP. The analyzer uses a
small sample size of 2 to 5 mg, weighed using a micro balance and packed into tin containers for
analysis. Every run included a standard to detect variation between runs. CN analysis of samples
was conducted in triplicate.
Statistical Analysis. Forage quality, pre- and post-biomass, grazing intensity, and pasture
utilization were analyzed by least-squares ANOVA using the MIXED procedure of SAS
(Version 9.2, SAS Institute Inc.). Weekly measurements were analyzed as repeated measures.
The covariance structure of models was previously tested and the selection among UN, CS and
AR(1) covariance structures were determined based on the lowest Akaike’s Information
Criterion and Bayesian Information Criterion (Little et al., 1996). Kenward-Rogers was the
method used for estimation of degrees of freedom. Individual strips were considered
experimental units. The fixed effects were treatment, species, genotype x treatment, week,
treatment x week, species x week, and species x week x treatment. The random effect was the
individual strip within each species by treatment group.

49

Cow performance data were analyzed by least-squares ANOVA using the MIXED
procedure of SAS (Version 9.2, SAS Institute Inc.). Weekly data on milk yield, milking
frequency, and weight were analyzed as repeated measures. Following Little et al. (1996) the
covariance structure of models was previously tested and the selection among UN, CS and
AR(1) covariance structures determined based on the lowest Akaike’s Information Criterion and
Bayesian Information Criterion. Kenward-Rogers was the method used for estimation of degrees
of freedom. Individual cows were considered as the experimental unit for animal-related
variables, as is common in whole farm systems. For the animal related measurements the fixed
effects were treatment, genotype, genotype x treatment, week, treatment x week, genotype x
week, genotype x week x treatment. The random effect for the animal related measurements was
the individual cow within each genotype by treatment group.
Analyses of milking distribution and milking interval data were done using the
FREQUENCY procedure of SAS (Version 9.2, SAS Institute Inc.). In this procedure a Chisquared test was used to determine the frequency of milkings at a particular hour and the
frequency of milkings at a particular interval. All milking failures (i.e. unsuccessful teat cup
attachment) and the 7 milkings following a milking failure were removed from the analysis, as
they may not be considered normal milkings (Bach and Busto, 2005).

50

RESULTS

Pasture and forage quality. There were no stocking rate effects on pasture biomass.
However, significant variation across weeks was detected in the pre-grazing and post-grazing
pasture herbage mass (Figure 2.3). The difference over weeks is likely caused by changing
environmental conditions and precipitation that occurred at the trial onset. Environmental
conditions became more favorable, encouraging plant growth as indicated by sharp increases in
pasture biomass by week 3 of the study in both the HSR and LSR treatments (Figure 2.3).
Significant differences were detected between pasture species for CP, NDF, ADF, and %IVTD
(P < 0.05; Table 2.2).
Cow milk performance and milk composition. Milk production per day was significantly
affected (P = 0.03) by a stocking rate (SR) by genotype by week interaction (Figure 2.4), while
there was a marginal tendency (P = 0.09) for a SR by genotype by week interaction effect on MF
(Figure 2.5). Despite the differences in milk production and milking frequency, energy
corrected milk (ECM) production was not affected (P = 0.15) by a SR by week by genotype
interaction, but significant differences in ECM were detected between genotypes (Table 2.4).
Milk production showed a significant genotype by week interaction (P < 0.0001; Figure
2.4), while MF showed only a tendency (P = 0.06) to decrease in NZF cows by the end of the
study (Figure 2.5). A significant (P = 0.0032) treatment by week interaction (Table 2.4) was
seen for milk fat, likely explained by the increase in milk fat for the low stocking rate treatment
(LSR) and decrease in milk fat in the high stocking rate treatment (HSR). Milk urea nitrogen and
protein did not show any significant SR by week interactions. Milking frequency had a
significant treatment by week interaction (P = 0.0002). Milk production had a tendency towards

51

significance for a SR treatment by week interaction (P = 0.09), likely explained by a more steady
milk production by cows in the high stocking rate treatment receiving PMR supplementation.
The only milk composition variable that demonstrated a treatment by genotype
interaction was milk urea nitrogen (MUN; P = 0.0013); cows in the LSR-USH group (15.46 ±
0.65 mg/dl) had a higher MUN compared to the cows in the HSR-USH group (12.78 ± 0.66
mg/dl). Across stocking rate treatments, the MUN was 13.62 ± 0.46 mg/dl and 15.7 ± 0.46 mg/dl
for LSR and HSR, respectively. Across genotypes, MUN was 14.12 ± 0.46 mg/dl and 14.66 ±
0.46 mg/dl for USH and NZF, respectively. There was a marginal tendency (P = 0.08) for milk
fat percent to be higher in the LSR (4.04 ± 0.12%) than in the HSR treatment (3.75 ± 0.12%).
Significant differences in milk production and AMS milkings between genotypes existed
(Table 2.3). The USH cows had higher (P < 0.0001) milk production (27.92 ± 1.33 kg) than NZF
cows (15.34 kg ± 0.83 kg). The USH cows voluntarily milked at the AMS with higher (P =0.02)
frequency (2.51 ± 0.07 milkings/day) than NSF cows (2.25 ± 0.07 milking/day). Conversely,
milk fat and milk protein percentages were higher for NZF (milk fat: 4.21± 0.12%, milk protein
3.34 ± 0.09%) than USH cows (milk fat: 3.58 ± 0.12%, milk protein 3.05 ± 0.09%).
Significant differences in feed intake and feed conversion efficiency (FCE) were also
detected between the two genotypes (Table 2.4). Intake of pellet concentrate, forage intake
(pasture only or pasture plus PMR), and total feed intake were higher for USH than NZF cows.
On average, the USH cows consumed a total of 22.32 ± 0.82 kg of DM, whereas the NZF cows
consumed 16.30 ± 0.82 kg of DM. Interestingly, differences in feed intake disappear when intake
is corrected by differences in body weight (Table 2.4). Differences in milk production and feed
intake reported above resulted in significant differences in FCE between genotypes (Table 2.4).
The FCE was higher (P = 0.01) for USH (1.24 ± 0.06 ECM/kg DM/d) than NZF cows (1.01 ±

52

0.06 ECM/kg DM/d). Differences in CP, NDF, and ADF intake between genotypes (Table 2.5)
were also detected as a consequence of the differences in feed intake reported above.
Differences in milking distribution and milking intervals existed between SR treatments
and genotypes. As shown in figure 2.6, AMS occupation between 0300 and 0400 h was always
low regardless of the stocking rate treatment, but feeding of PMR at 0500 h in the HSR treatment
enticed cows to increase utilization of the AMS from 0500 to 0900 h. Cows in the LSR treatment
increased the use of the AMS from 0600 to 1000 h and maintained a steady use of the AMS unit
from 1100 to 2100 h. The PMR feeding in the HSR treatment also resulted in shorter intervals
and very few milkings exceeding 16 h (Figure 2.7). Conversely, milking intervals were more
variable for the LSR treatment (Figure 2.7)
Genotypes differed significantly in their use of the AMS, both for distribution of milkings
and milking intervals. Visitations for USH peaked at hours 0700 to 0900, 1600 to 1700, and
2300 to 2400. Conversely, visitations for NZF peaked at hours 1100 to 1300, 1900 to 2200, and
0500 to 0600 (Figure 2.8). There was also a difference in milking interval between genotypes.
Intervals for USH were shorter and had fewer milking intervals that exceeded 10 hours, while
NZF had longer milking intervals and more that exceeded 10 hours compared to USH (Figure
2.9).

53

DISCUSSION

Cow Performance. The main effect of week was significant for nearly all cow
performance variables, except milk fat % and kg fat /d. This time effect was related to changes in
soil water availability and resultant differences in forage growth and availability. At study onset,
low precipitation and low soil moisture negatively affected pasture growth rate and pasture
availability (Figure 2.3), thereby reducing pasture intake and milk production (Figure 2.4). As
the study progressed and soil moisture was replenished by rainfall, pasture biomass (Figure 2.3)
and milk production recovered.
All cow performance parameters were significant for the main effect of genotype, except
milk urea nitrogen (MUN). As expected USH had a higher milk, protein, and fat yield, while
NZF had higher % protein and fat. MUN was affected by a SR by genotype interaction. The
values for MUN in USH cows were higher in the LSR treatment (15.46 mg/dl) than in the HSR
(12.78 mg/dl) treatment, whereas MUN values for NZF cows were higher for the HSR (15.7
mg/dl) treatment than for the LSR (13.62 mg/dl) treatment. In similar studies un-supplemented
USH cows have, in general, higher levels of MUN. Bargo et al. (2002a) reported MUN levels of
13.9 and 11.6 mg/dl for un-supplemented and supplemented cows receiving 8.6 kg concentrate.
In another study, feeding pasture plus concentrate (PC) or pasture plus total mixed ration (PMR).
Bargo et al. (2002b) found MUN levels of 14.9 and 12.0 mg/dl for PC and PMR, respectively. In
both studies, MUN decreased with supplementation with either concentrates or total mixed
ration. This response was detected in the USH cows in the present study, but the opposite
occurred with the NZF cows. It is possible that the combination of the higher proportion of
ryegrass and white clover pastures containing higher levels of crude protein (Table 2.2), together

54

with lower eating drive and reductions in PMR intake by NZF cows, may have resulted in higher
pasture consumption and MUN levels. The PMR supplementation may decrease CP in diets as
compared to a pasture only diet, or the starch from the PMR (PMR was 40% grounded corn) may
have increased the capture of ammonia nitrogen in the rumen of USH cows. This mechanism of
nitrogen capture was probably absent in the NZF due to lower PMR consumption. The PMR
intake for USH and NZF cows was derived from average group estimates, as no individual PMR
intake measurements were collected
Near peak milk production was reached sooner in HSR (week 3) compared to the LSR
treatment (week 4; Figure 2.4). This increase in milk production was quicker for USH than NZF
cows, which indicates higher milk production response to PMR supplementation, particularly
when pasture availability was low (Week 1 to 4 of the study). Stockdale (1994) indicated
supplementation with corn silage resulted in a marginal response of 1.2 kg milk of milk for each
of the first 4 kg DM of silage offered, but no response in milk production was detected when
supplementation exceeded 4 kg of corn silage. The review by Phillips (1988) also showed
increases in total intake with forage supplementation but only if the forage supplement is of
higher quality than the pasture offered. Conversely, if forage quality of the supplement is lower
than the pasture offered, intakes may decrease. Phillips (1988) suggested that the use of a foragebased supplement was appropriate when low pasture allowance limits intake. This may be true
only for the first four weeks of this study, when a marginal milk production response to PMR
supplementation was detected. Once pasture availability recovered, the milk response appears to
be non-existent due to high pasture allowance and the similar quality between the PMR and
pasture forage (Table 2.2).

55

In this study the supplemental PMR containing 60 % forage and 40% concentrate was
used to compensate deficits in forage growth and availability. This combined use of alfalfa
haylage forage and concentrates (Table 2.2) in a PMR may not cause the same effects as
supplementing the diet with either haylage or concentrates alone. Bargo et al. (2002) compared
treatments of pasture plus concentrate (PC) and pasture plus TMR (PMR) and found intakes to
be higher in the PMR treatment compared to the PC treatment (21.6 and 25.2 kg/d for PC and
PMR, respectively). Milk production was also higher in PMR than PC treatment (26.9 and 32.0,
for PC and PMR, respectively). It is important to note that the PMR diet in Bargo’s study
consisted of 15.5 kg/d of PMR. In the present study the amount of PMR fed averaged 4.8 kg/d.
This lower level of supplementation, together with the adjustments in PMR fed based on the
change in pasture availability and growth, may have explained the lack of response in both feed
intake (Table 2.4) and milk production (Table 2.3)
Milk production exhibited a significant triple interaction among SR, week and genotype,
but no main effect for SR was detected. This triple interaction is most likely the result of
genotype x environmental interactions, since PMR supplementation in HSR resulted in steeper
increases in milk production in the early weeks of the study only for USH cows. Other factors
may be contributing to the triple interaction. There is a small amount of milk response of the
NZF in HSR from supplementation and also a minimal response in NZF-LSR due to recoveries
in pasture quality and quantity over time. Due to the lower overall requirements of NZF cows,
dry matter intake and milk production were not affected by the low pasture allowance in the
beginning of the study. This finding is supported by Horan et al. (2005), in which
supplementation created a low milk response and increased substitution for similar NZF

56

genetics. The lack of response in milk production is believed to be related to the lower genetic
merit of this genotype of Holstein cattle (Horan et al., 2005).
Table 2.4 shows the estimated intakes for both SR and Genotype. The technique used
chromium oxide as an external marker, possibly resulting in a 10% overestimation of total dry
matter intake (Holden et al., 1994). As expected no significant differences were seen between SR
treatments. However, significant differences existed in estimated forage DMI between genotypes
(16.30 and 22.32 kg DM/cow per d for NZF and USH, respectively). Table 2.5 shows the
estimated ADF, NDF, and CP in the farmlets. Intakes for PMR for USH were estimated based on
the total amount fed divided by the number of cows. Intakes for PMR for NZF were calculated
based on the percent intake of PMR by USH. Therefore Table 2.5 shows rough estimates, but is
shown to give a general idea of the estimated quality of the diet and intakes in both stocking rate
systems.
An estimate of feed conversion efficiency (FCE; kg of ECM produced per kg of feed
DM) is given in Table 2.4. There were no significant differences between SR treatments, but
significant differences existed between genotypes (1.01 and 1.24 for NZF and USH,
respectively). The difference between genotypes is most likely due to the increased genetic
potential in USH, and perhaps the slight difference in parity between genotypes. It has been
proposed that there is a trade-off between stocking rates and FCE; this trade-off is quite often the
result of reduced production per cow at high stocking rates (McCall and Clark, 1999). Tozer et
al. (2004) proposed that diet supplements could be used to increase milk production per cow by
increasing intake. The increase in intake and milk production would help dilute maintenance
requirements of the cow and increase FCE (Beever and Doyle, 2007). The current study showed
no difference between SR, therefore supplementation added enough energy to compensate for

57

high stocking rates and maintained FCE. In the present study PMR supplementation was
strategically used to compensate deficits in pasture availability in the HSR system. Therefore,
adjustments in both quality and quantity of PMR fed have been used to minimize any adverse
substitution of pasture by supplement. This study therefore suggest that timely management of
PMR supplementation in HSR dairy systems could be used as a tool to manage feed intake and
milk production without severely affecting pasture utilization (i.e. total forage consumed per
area).
Milking frequency. Average milking frequency (MF) per day and number of refusals per
day are reported in Table 2.3 Frequency of voluntary milkings differed between genotypes, but
not between SR, as predicted. Interestingly, at the beginning of the study MF is nearly the same
for each group (2.6 milkings/day), except NZF HSR, which started at 2.4 milkings/day. Other
than the drop in MF in week five for USH HSR, the HSR treatment maintained MF, ending at
the nearly the same MF in which they started. The USH LSR group declined in MF, but after
week 4 remained relatively constant. However, NZF LSR decreased in MF almost linearly over
the 8 weeks (from 2.6 to 1.75 milkings/day). All groups began to decrease in MF by week 8,
possibly due to increase in DIM or perhaps due to change in environmental factors such as cooler
temperatures or shorter day length.
The data on milk production (Figure 2.4) and MF (Figure 2.5) suggest that milk
production was not directly related to increases in MF. As the study progressed, milk production
may have increased as a result of increased pasture quality and availability, because MF did not
exhibit important changes over time except for NZF cows in the LSR treatment. At study onset,
both MF and refusals were relatively high, probably because low pasture availability or high
temperatures during daylight enticed cows to seek more feed resources from the AMS (i.e., feed

58

pellets). As the study progressed and temperatures dropped and forage quality and quantity in the
pastures recovered, refusals begin to decrease, indicating that cows were spending more time
grazing on pasture. However, towards the end of the study (fall season) there was a decrease in
milk production for both USH and NZF cows in the LSR treatment, probably related to decreases
in MF. Interestingly, the NZF-LSR group started with a high MF and had a linear decline in MF
compared to the NZF-HSR group, yet MP was not greatly affected until week 7. This change in
MF may indicate that for NZF cows higher MF was not resulting in higher MP. An increase in
MF without an increase in MP reduces AMS efficiency, since the extra time spent in the AMS
did not result in any additional milk production. However, a large decrease in MF into fall
created a steep decrease in MP for the NZF-LSR cows. The PMR feeding may have given cows
more motivations to return to the barn and visit the AMS. Lack of additional PMR
supplementation in the barn together with lower levels of pellet supplementation in the AMS in
response to the lower milk production in later stages of lactation, may have caused the NZF cows
in the LSR treatment to visit the AMS less frequently.
The NZF-HSR cows may have maintained their MF throughout the study because the
PMR motivated them to return to barn and visits the AMS regularly. Although the results were
not significant, USH-LSR had numerically fewer refusals than NZF-LSR cows (0.89 and 1.53
refusals/day for USH-LSR and NZF-LSR, respectively). The refusals for the NZF-LSR group are
actually the same as the NZF-HSR group and not significantly different from USH-HSR (1.54
and 1.83 refusals/day for NZF-HSR and USH-HSR, respectively), but MF is still lower for NZFLSR. It appears that NZF-LSR were visiting the AMS, but not during the qualified time or at
regular intervals. The low refusals for USH-LSR show that they are not visiting the AMS often
enough to be refused, perhaps the higher eating drive of USH-LSR and lack of PMR

59

supplementation forced high producing Holsteins to graze longer hours. The low refusals for
USH-LSR may indicate that they are using their time more efficiently, in the pasture, or that their
grazing and visitation patterns differ from USH-HSR, due to the lack of PMR supplementation.
Milking interval also affects MF, MP, and AMS utilization. In the current study milking
interval did not differ between treatments (Table 2.3). Jago et al. (2007) examined the
management strategy of increasing milking interval for low producing New Zealand Holstein
Friesians. As it would be expected, decreases in MF were correlated with increases in MP per
milking. The MF of cows managed with milking intervals of 6 or 12 h was 1.92 and 1.43
milkings/day, respectively. Despite this difference in MF, the amount of milk produced was not
significantly different and was 22.78 and 23.27 kg/d for the cows in the 6 h and 12 h intervals,
respectively.
The NZF cows in the current study, all in their first parity, produced milk at an average
rate of 15.34 kg/cow per d, with a MF of 2.25 and a milking interval of 10.67 h. As MF
decreased for NZF LSR, MP did not decrease until late in the study, indicating that increases in
MF did not result in any additional milk. This observation is supported by Jago et al. (2007), in
which MP did not change with decreased MF for low producing NZF cows grazing ryegrass
pasture. Therefore, as genetic merit for milk production decreases, cows may require longer
milking intervals to maintain a high level of milk production per milking. The study by Jago et
al. (2007) suggested that an increase in number of cows and the use of prolonged milking
intervals between 16.9 and 18 h increased milk production per robot in grazing systems using
NZF or low-producing dairy cows.
The availability of PMR may have enticed cows in the HSR treatment to maintain their
MF. Sporndly and Wredle (2004) examined the effects of distance to pasture and level of

60

supplementation on MF and MP. The treatments in the study were near pasture (NP; 50 m from
the barn), a distant pasture (DP; 260 m from the barn), and a distant pasture plus ad libitum
silage supplementation in the barn (DP+S). Both NP and DP received a small amount of silage, 3
kg DM of silage per cow, in the barn, so it is not possible to determine if silage supplementation
increased MF since there was no un-supplemented treatment. However, the amount of
supplement offered did not affected MF. Further, similar to our study, cow behavior changed
over the grazing season. Early in the season, Jun 5 to Jul 13, the milking frequency was the same
for DP and DP+S (2.3 milkings/day, MP), although significantly different compared to NP (2.5
milkings/day). Milk production was not significantly different between DP and DP+S (26.4 and
26.1 kg/d, for DP and DP+S) but was significantly higher in NP (30.5 kg/day). Also, Sporndly
and Wredle (2004) found that in the DP and DP+S treatments, cows spent less time grazing and
showed reductions in milk production. This decrease in MP occurred despite the increase in MF
in DP (2.5 milkings/d), which may suggest an attempt of these cows to increase intake from the
feed delivered by the AMS unit. These results show that the cows’ motivation to travel the
distance to pasture changed over the grazing season and that offering ad libitum silage was no
more effective as offering 3 kg of silage DM in the barn to entice cows to visit the AMS
regularly. In the current study a change over time was also seen, indicating that in a system
where cows have more free choice, a better understanding of their behavior is required to
maintain milk production and AMS efficiency.
The addition of a PMR did change the distribution of milkings. This distribution was
similar for both SR treatments, but higher frequency of shorter milking intervals was observed
for the HSR treatment. This difference is most likely due to the feeding of the PMR at 0500 h,

61

motivating cows to return to the barn. There is a diurnal pattern shown by the reduced number of
milkings from 0200 to 0500 h.
The average milking interval appears to be slightly longer for LSR, however, the
interesting difference was the frequency of longer milking intervals. The percent of intervals
above 23 hours for LSR was 1.19 %, which was low, but HSR did not have any milkings over 23
h. Thus feeding PMR in the barn may decrease the frequency of long intervals, since the PMR
was motivating cows to return to the barn at a certain time every day. Perhaps, feeding and
milking patterns in the HSR treatment were more synchronized compared to the LSR treatment,
and this synchronization may have resulted in more regular milking intervals and milk
production.
Genotype. Genotypes role on milk production and AMS utilization was examined in this
study. The average MP for USH was significantly higher than NZF (27.92 and 15.34, for USH
and NZF, respectively), but this differential response was also affected by differences in parity.
Parity was 1.3 and 1 lactations for USH and NZF, respectively. The MP for NZF is lower than
other studies using cows with New Zealand genetics (Jago et al., 2007), most likely due to the
low parity of these cows. The interest in NZF stems from their lower maintenance requirements
and therefore better adaptation to pasture dairy systems where energy intake could be a limiting
factor for milk production. Early in this study herbage availability was low (less than 2000 kg
DM/ha), which may have caused the low MP in USH. However, NZF did not increase MP once
pasture conditions improved, suggesting that NZF cows had lower energy requirements and are
better suited for conditions where pasture availability per cow is lower. Conversely, a response in
MP was clearly seen in USH when pasture availability increased from week 1 to week 4 of the
study (Figures 2.3 and 2.4). The MP for USH is similar to other USH studies grazing high

62

producing Holstein cows (Bargo et al., 2002a, 2002b, Kolver and Muller, 1998).Despite this
increase in MP by USH cows, their potential maximum MP was not achieved in this study,
particularly because grazing behavior often sets an upper limit to feed intake even in a situation
where pasture availability and quality are high. In pasture-based systems using high producing
cows, controlled supplementation with concentrates based on MP, parity, days in milk and/ or
genotype could be used to improve MP performance of high producing dairy cows.
The hypothesis that different genotypes of Holstein may also utilize the AMS in different
ways was also tested. The AMS milking interval settings were based on expected milk
production and maximum and minimum number of daily milkings. This flexible milking
schedule allows dairy farmers to efficiently manage cow milking frequency in order to maximize
milk responses at each milking. For example, the expectation for early lactation cows was to
harvest at least 9 kg of milk per milking while achieving a high milking throughput with the
AMS. Decreasing the minimum limit of expected milk production per milking could have
increased milking frequency to 3 times a day. However, in our system of low producing NZF
cows a third milking may have been an inefficient use of the AMS, since it probably would not
result in additional milk. For NZF, longer milking intervals and higher stocking rates of cows per
AMS unit may be the most efficient use of the AMS, as was seen in Jago et al. (2007).
Differences in milking interval were detected between NZF and USH cows. Due to
differences in milk production, resulting in differences in concentrate intake from the AMS and
the pre-established milking settings (see Materials and Methods), NZF cows had longer milking
intervals than USH cows (Figure 2.8). Interestingly, the NZF cows had no milking interval
shorter than 7 h, whereas USH cows had around 9% of the milkings before 7 h (Figure 2.8).
There is a clear differential pattern of AMS milkings for both genotypes (Figure 2.9) suggesting

63

complementary uses of the AMS. While both NZF and USH cows had a low frequency of
milkings between 0200 and 0400 h, NZF cows appeared to use the AMS more opportunistically
by avoiding hours of high AMS use by USH cows. Probably, due to differences in body size,
USH cows behaved as a dominate genotype forcing NZF cows to use the AMS at hours of lower
visitations. There may be potential to exploit the adaptation of NZF to USH milking patterns.
Due to the low milk production and longer milking interval of NZF, their visitations to the AMS
may be more flexible. Since USH produces more milk, requiring shorter intervals and more
frequent milkings, they are less flexible. Although no grazing behavior data are included in this
study, it may be possible that the NZF cows altered their grazing times.
It is possible that USH may have filled the NZF milking distribution if the stocking rate
had been higher. Ketelarr-de Lauwere et al. (1996) found that cows with determined low
dominance values adapt their visits to the AMS by visiting at times when they would not have to
compete with dominant cows. This conclusion was supported by the timing of the visits to the
AMS with cows of higher dominance visiting the AMS more between 1200 and 1800 h, while
the lower rank cows made visits at a complementary time, outside of peak visitation hours for the
USH cows. It is possible that a combination of genotypes may utilize the AMS more efficiently,
due to their different milking requirements and milking distributions, but more research is
needed in this area to make a sound conclusion.
Conclusion. As used in this study, PMR supplementation did not affect MP or MF for
either genotype of dairy cows used. However, trends in decreased MP and MF for NZF LSR
indicate that supplementation may have motivated cows to return to the barn, which resulted in a
more stable MF. High stocking rates at both the pasture and AMS levels may be best for NZF,
because of their lower energy requirements. The low genetic merit of NZF also indicates that

64

their milking interval can be lengthened to increase AMS efficiency. Supplementing with a PMR
did not result in an increase in milk production except when pasture availability was low,
indicating that supplemental concentrates may be required to increase MP in this system.

65

APPENDIX

66

TABLES

Table 2.1 Ingredients and nutrient composition of pellet feed in the AMS
Ingredients
Soy Hills
Ground Corn
Sure Pro
Soybean Meal 47.5 Solv
Molasses
Salt (NaCl)
Tallow

%
62.3
15.0
9.5
6.1
3.1
2.5
1.5

Nutrient Composition
DM, %
CP %
RUP(%CP)
RDP (%CP)
RDP %
ME (MCal/kg)
Nel (MCal/kg)
ADF %
NDF %

89.45
17.61
14.12
85.88
15.12
0.65
0.42
32.68
42.81

67

Table 2.2
Differences in pasture species quality and partial mixed ration quality
Variable
CP, %
NDF, %
ADF, %
IVTD, %

1

RG
20.92
35.33
24.35
76.38

2

OG
SEM P value <
17.57 0.58 < .0001
50.29 1.2
< .0001
29.38 1.39 < .0001
71.11 0.68 < .0001

Partial Mixed Ration
CP,%
14.86
NDF,%
32.00
ADF, %
23.32
IVDMD, % 72.64
1

Ryegrass pasture

2

Orchardgrass pasture

68

Table 2.3 Milk production, milk quality, weight change, milking frequency, and milking refusal
of dairy cows managed with automatic milking systems and two different stocking rates.

LSR
USH
Milk, kg/d

3

1

HSR
4

NZF

USH

2

P value

NZF

SEM

SR x G

SR

G

27.24

14.46

28.6

16.22

1.89

0.9

0.33

<0.01

24.87

15.22

26.37

16.97

1.66

0.8

0.45

<0.01

2.99
0.8

3.35
0.48

3.12
0.89

3.34
0.54

0.09
0.05

0.42
0.74

0.52
0.14

<0.02
<0.01

3.59
0.97

4.50
0.65

3.57
1.01

3.92
0.64

0.16
0.06

0.1
0.69

0.08
0.8

<0.01
<0.01

15.46

13.62

12.78

15.7

0.66

<0.01

0.64

0.41

Weight Change , kg

9.05

3.04

13.16

3.19

1.45

0.32

<0.01

<0.01

Milking
8
Frequency /d

2.38

2.22

2.65

2.27

0.1

0.34

0.13

0.02

10.48

11.23

9.23

10.71

0.56

0.56

0.17

0.09

0.89

1.53

1.83

1.54

0.53

0.3

0.28

0.7

5

ECM , kg/d
Milk Protein
%
kg/d
Milk Fat
%
kg/d
6

MUN , mg/dl
7

9

Milking Interval , h
10

Refusals /d
1

Low Stocking Rate Treatment

2

High Stocking Rate Treatment

3

United States Holstein Friesian

4

New Zealand Holstein Friesian

5

Energy corrected milk

6

Milk urea nitrogen

7

Weight change per week

8

Milking frequency, average successful milkings per day

9

Milking interval (24 h/MF)

10

Denied milking by the AMS as a result of not meeting milking criteria, average refusals per
day
69

Table 2.4 Intake estimates and feed conversion efficiency by dairy cows in an automatic milking
system with two different stocking rates.
HSR
Variable

USH

Total intake, kg DM

3

1

LSR
4

2

P value
5

NZF

USH

NZF

21.99 a

16.25 b

22.65 a

16.35 b

1.26

0.81 0.75 <0.01

Pellet intake, kg DM

4.43 a

2.68 b

4.56 a

2.45 b

0.27

0.51 0.85 <0.01

Corn intake, kg DM

1.15 a

0.96 a

0.79 a

0.95 a

0.18

0.36 0.32

12.61 b

17.3 a

12.95 b

0.95

0.76

1.26 a

0.98 b

1.21 a

1.04 b

0.08

0.48 0.97

0.01

3.93 a

4.05 a

3.99 a

4.20 a

0.25

0.86 0.68

0.53

Forage Intake, kg DM 16.41 a
6

FCE

Total intake, %BW

7

SEM

SR x G

1

High stocking rate treatment

2

Low stocking rate treatment

3

United States Holstein Friesian

4

New Zealand Holstein Friesian

5

Genotype

6

Feed conversion efficiency; energy corrected milk kg/dry matter intake kg

7

Body weight

Means within a row with different subscripts differ (P < .05)

70

SR

G

0.95

0.5 <0.01

Table 2.5
Diet composition and estimates of crude protein, acid detergent fiber, and neutral detergent fiber
intakes by dairy cows in a pasture-based automatic milking system with two different stocking
rates.
HSR
3

1

LSR

2

P value

4

NZF
Variable USH
USH NZF
SE
SR x G
CP Intake, kg/d
Total
3.92
2.89
4.1
2.92
0.21
0.73
Pellet
0.84
0.51 0.86
0.46
0.05
0.51
Corn
0.12
0.1 0.08
0.1
0.02
0.36
7
PMR
0.77
0.56
NA
NA
0.008
NA
Pasture
2.19
1.73 3.38
2.36
0.19
0.17
NDF Intake, kg/d
Total
8.53
6.27 9.83
6.97
0.48
0.54
Pellet
2.04
1.23
2.1
1.13
0.12
0.51
Corn
0.13
0.11 0.09
0.11
0.02
0.36
PMR
1.5
1.08
NA
NA
0.02
NA
Pasture
4.87
3.85 7.64
5.73
0.42
0.3
%BW
1.56
1.1 1.77
1.19
0.09
0.49
ADF Intake, kg/d
Total
5.72
4.17
6.3
4.4
0.3
0.58
Pellet
1.56
0.94
1.6
0.86
0.09
0.51
Corn
0.05
0.04 0.03
0.04
0.01
0.36
PMR
1.06
0.77 NA8
NA
0.01
NA
Pasture
3.06
2.41 4.67
3.5
0.26
0.32
Diet, %
CP
17.81
17.81 18.1
17.88
0.12
0.35
NDF
38.79
38.56 43.47
42.56
0.42
0.42
ADF
26.02
25.62 27.91
26.92
0.31
0.36

SR

6

5

G

0.63
0.85
0.32

<0.01
<0.01
0.95

NA
<0.01

<0.01
<0.01

0.05
0.85
0.32
NA
<0.01
0.12

<0.01
<0.01
0.95
<0.01
<0.01
<0.01

0.19
0.85
0.32
NA
<0.01

<0.01
<0.01
0.95
<0.01
<0.01

0.13
<0.01
<0.01

0.32
0.19
0.04

1

High stocking rate treatment

2

Low stocking rate treatment

3

United States Holstein Friesian

4

New Zealand Holstein Friesian

5

Genotype effect

7

Partial mixed ration; values for PMR were
estimated by dividing the total amount of group
fed PMR by the number of cows in the herd.
Values of PMR intake for USH and NZF cows
were derived estimations of average PMR
intake of the herd expressed in a % body
weight basis

6

Stocking rate effect

71

AMS

Cosmix

AMS

Exit to pasture

Exit to pasture

FIGURES

Figure 2.1 Layout of the split freestall

72

Cosmix

HSR
LSR

Figure 2.2. Paddock layout showing paddocks used for the high stocking rate (HSR) and low
stocking rate (LSR) groups.

73

Pre- and post-grazing biomass of pasture
3000
Biomass (kg/ha)

2500
2000
1500

Pre

1000

Post

500
0

1

2

3

4

5

6

7

8

Week
Figure 2.3 Pre-grazing (Pre) and post-grazing (Post) biomass of pastures grazed at high and low
stocking rate by dairy cows managed with AMS. There was no stocking rate treatment effect (P
= 0.89 for pre and P = .07 for post), but significant differences across weeks of the study were
detected (P <0.0001 for both pre and post measurements). Bars denote standard errors of means
(SEM) for pre grazing (SEM = 11.44) and post grazing (SEM = 72.97) biomass

74

Milk Production
35

Milk Yield kg

30
25

NZF LSR
USH LSR

20

NZF HSR
USH HSR

15
10
1

2

3

4

5

6

7

8

Week
Figure 2.4. Milk production by New Zealand Friesian (NZF) and United States Holstein (USH)
dairy cows grazed at high stocking rate (HSR: 2.89 cows/ha) or low stocking rate (1.92 cows/ha)
in a pasture-based automatic milking system. There was a significant stocking rate x week x
genotype interaction (P < 0.03.0value). Bars denote standard error of means (SEM = 1.98).

75

Milking Frequency
Milking Frequency (milkings/day)

3.1
2.9
2.7

2.5
NZF LSR

2.3

USH LSR

2.1

NZF HSR

1.9

USH LSR

1.7
1.5
1

2

3

4

5

6

7

8

Week
Figure 2.5. Milking frequency of New Zealand Friesian (NZF) and United States Holstein
(USH) dairy cows grazed at high stocking rate (HSR: 2.89 cows/ha) or low stocking rate (1.92
cows/ha) in a pasture-based automatic milking system. There was a trend for a stocking rate x
week x genotype interaction (P =.09). Bars denote standard error of means (SEM = 0.13).

76

Percent of Milkings

Milking Distribution
8
7
6
5
4
3
2
1
0

HSR
LSR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Figure 2.6. Milking distribution of dairy cows grazed at a high stocking rate (HSR: 2.89
cows/ha) or a low stocking rate (1.92 cows/ha) in a pasture-based automatic milking system.
There was a significant interaction for stocking rate by hour (P < 0.05).

Milking Interval

Percent of Milkings

25
20
15
HSR

10

LSR
5
0
1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728
Hour

Figure 2.7. Milking interval of dairy cows grazed at a high stocking rate (HSR: 2.89 cows/ha) or
a low stocking rate (1.92 cows/ha) in a pasture-based automatic milking system. There was a
significant interaction for stocking rate by interval (P < 0.05).

77

Percent of Milkings

Milking Distribution
9
8
7
6
5
4
3
2
1
0

USH
NZF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour

Figure 2.8. Milking distribution of New Zealand Friesian (NZF) and United States Holstein
(USH) dairy cows grazed at a high stocking rate (HSR: 2.89 cows/ha) or a low stocking rate
(1.92 cows/ha) in a pasture-based automatic milking system. There was a significant stocking
hour x genotype interaction (P = 0.05).

Milking Interval

Percent of Milkings

25
20
15
USH

10

NZF
5
0
1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627
Interval

Figure 2.9 Milking interval of New Zealand Friesian (NZF) and United States Holstein (USH)
dairy cows grazed at a high stocking rate (HSR: 2.89 cows/ha) or a low stocking rate (1.92
cows/ha) in a pasture-based automatic milking system. There was a significant stocking
genotype x interval interaction (P = 0.05).

78

REFERENCES

79

REFERENCES

Bach, A., and I. Busto. 2005. Effects on milk yield of milking interval regularity and teat cup
attachment failures with robotic milking systems. J. Dairy Res. 72:101–106.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002a.Milk response to concentrate
supplementation of high producing dairy cows grazing at two pasture allowances. J.
Dairy Sci. 85:1777–1792.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002b. Performance of high
producing dairy cows with three different feeding systems combining pasture and total
mixed rations. J. Dairy Sci. 85:2960–2975.
Beever, D. E., P. T. Doyle. 2007. Feed conversion efficiency as a key determinant of dairy herd
performance: a review. Aust. J. Exp. Agric., 47:645–657.
Buckley, F., P. Dillon,, M. Rath, and R. F. Veerkamp. 2000. The relationship between genetic
merit for yield and live weight, condition score, and energy balance of spring calving
Holstein Friesian dairy cows on grass based systems of milk production. J. Dairy Sci. 83,
1878– 1886.
Fales, S. L., L. D. Muller, S. A. Ford, M. O’Sullivan, R. J. Hoover, L. A. Holden, L. E. Lanyon,
and D. R. Buckmaster. 1995. Stocking rate affects production and profitability in a
rotationally grazed pasture system. J. Prod. Agric. 8: 88–96.
Farina, S. R., S. C. Garcia, W. J. Fulkerson and I. M. Barchia. 2011. Pasture-based dairy farm
systems increasing milk production through stocking rate or milk yield per cow: pasture
and animal responses. Grass Forage Sci. 66:316-332.
Fulkerson, W. J., T. M. Davison, S. C. Garcia, G. Hough, M. E. Goddard, R. Dobos, and M.
Blockey. 2008. Holstein-Friesian dairy cows under a predominantly grazing system:
interaction between genotype and environment. J. Dairy Sci., 91:826–839.
Horan B., Dillon P., Faverdin P., Delaby L., Buckley F. and Rath M. 2005. The interaction of
strain of Holstein Friesian cows and pasture-based feed systems on milk yield, body
weight and body condition score. J. Dairy Sci., 88:1231–1243.
Jago, J. G., K. L. Davis, P. J. Copeman, I. Ohnstad, and M. M. Woolford. 2007. Supplementary
feeding at milking and minimum milking interval effects on cow traffic and milking
performance in a pasture-based automatic milking system. J. Dairy Res. 74:492–499.
Kennedy J., Dillon P., Delaby L. and Faverdin P. 2003. Effect of genetic merit and concentrate
supplementation on grass intake and milk production with Holstein Friesian dairy cows.
J. Dairy Sci. 86:610–621.
80

Ketelaar-de Lauwere, C. C. 1996. The influence of social hierarchy on the time budget of cows
and their visits to an automatic milking system. Appl. Anim. Behav. Sci. 49:199-211.
Kolver, E. S., L. D. Muller, G. A. Varga, and T. J. Cassidy. 1998. Synchronization of ruminal
degradation of supplemental carbohydrate with pasture nitrogen in lactating dairy cows.
J. Dairy Sci. 81:2017–2028.
Little, R. C., Milliken, G. A., Stroup, W. W., and R. D. Wolfinger. 1996. SAS System for Mixed
Models. SAS Institute Inc.: New York.
McCall, D. G., D. A. Clark. 1999. Optimized dairy grazing systems in the northeast United
States and New Zealand. II. System analysis. J. Dairy Sci. 82:1808–1816.
Phillips, C. J. C.1988. The use of conserved forage as a supplement for grazing dairy cows.
Grass Forage Sci. 43:215-230.
S.A.S. Institute, 2008. SAS User's Guide: Statistics. SAS Institute Inc., Cary, NC.
Sporndly, E., and E. Wredle. 2004. Automatic milking and grazing—Effects of distance to
pasture and level of supplements on milk yield and cow behavior. J. Dairy Sci. 87:1702–
1712.
Stockdale, C. R. 1994. Persian clover and maize silage, 1. Silage as a supplement for lactating
dairy cows offered herbage of different quality [Trifolium resupinatum]. Aust. J. Agric.
Res. 45:1751-1765.
Tozer, P. R., F. Bargo, L. D. Muller. 2004. The effect of pasture allowance and supplementation
on feed efficiency and profitability of dairy systems. J. Dairy Sci. 87:2902–2911.

81

CHAPTER 3
EFFECTS OF SUPPLEMENTATION OF A PARTIAL MIXED RATION ON THE
FERMENTATION AND RUMEN ENVIRONMENT OF DAIRY COWS IN A PASTUREBASED AND AUTOMATIC MILKING SYSTEM
ABSTRACT

Supplementing pasture-based diets of dairy cattle has the potential to create negative
associative effects in the rumen, decreasing both the rate and extent of degradation of forages in
the rumen. The objective of this study was to characterize the rumen environment and to
determine the residual effect of feeding variable amounts of a total mixed ration on the in situ
degradation of pastures and concentrates. Six ruminally cannulated Holstein cows (parity = 1.6 ±
0.49, DIM 148 ± 13; mean ± SD) were randomly assigned to one of two stocking rate groups; a
low stocking rate (LSR) grazing system consisting of a diet containing grazed pasture plus
concentrates, and a high stocking rate (HSR) grazing system receiving the same based diet as the
LSR plus variable amounts of a partial mixed ration (PMR; 60% alfalfa grass haylage, 40%
ground corn) adjusted for weekly changes in pasture availability. There was no significant
difference (P = 0.61) in rumen pH between feeding systems. The in situ technique revealed no
significant differences in the degradation rate of DM and NDF. Therefore, it is concluded that
feeding variable weekly amounts of a PMR ration to fill deficits and weekly changes in forage
availability in a pasture-based system did not substantially change the rumen environment or
depress pasture degradation rates when PMR is fed at relatively low amounts.
Keywords: in situ, pasture-based, grazing, partial mixed ration.

82

INTRODUCTION

Supplementation of concentrates and forages in pasture-based diets has been extensively
studied and reviewed (Phillips, 1988; Kolver and Muller, 1998; Reis and Combs, 2000; Bargo et
al., 2002a, 2002b, 2003). However, only a few of these studies investigated the effects of feeding
mixed rations on the rumen environment and forage degradation rates in grazing cattle (Bargo et
al., 2002b).
Reis and Combs (2000) reported a decrease in the forage soluble fraction when
supplementing 10 kg DM of concentrates. Similarly, Bargo et al. (2002b) found decreases in the
disappearance of the potentially degradable fraction when supplementing with 15 kg DM of a
partial mixed ration (PMR), indicating associative effects. However, both of these studies
involved a large amount of supplementation and therefore may not be comparable to diets in
which the ration consists of 60 to 70% pasture. A diet high in pasture, which is vulnerable to
variable forage availability and quality and variable use of forage and concentrate supplements,
may affect the rumen environment and thus pasture forage degradation. No studies exist to my
knowledge examining this potential effect in a pasture-based dairy system.
This study aims to determine the effect of a variable PMR feeding strategy on the rumen
environment in a pasture-based and automatic milking system (AMS). In this system, cows have
the opportunity to make choices more individually and therefore may have more temporal (i.e.
diurnal and day to day) and individual variation in feeding rates on both pasture forage and
supplemental forage and concentrates, in contrast to a more controlled or forced feeding system.
However, despite the individual feed choices involved in this particular system, fermentation
patterns and feed degradation may be similar across animals due to the similarity of the diet the

83

cows select. The objective this study was to determine if supplementation at high stocking rates
caused associative effects that resulting in decreased rate or extent of disappearance of DM, CP,
or NDF. I predicted that no differences in the ruminal environment will be detected between
treatments.

84

MATERIALS AND METHODS

Animals, Experimental Design, and Treatments. Six ruminally cannulated US Holstein
cows with average parity of 1.6 (± 0.49 SD) and average days in milk (DIM) of 148 (± 13) days
were randomly assigned to one of two stocking rate (SR) treatment groups in a completely
randomized design. The low stocking rate (LSR) treatment had 1.92 cows/ha and the high
stocking rate (HSR) treatment had 2.89 cows/ha. Supplementation for HSR consisted of a partial
mixed ration (PMR) of 60% legume-grass haylage and 40% ground corn. The PMR feeding was
conducted as described in Chapter 2. Due to illness unrelated to the feeding treatments, one cow
was removed from the HSR.

Pasture Layout, Barn, Feeds, and AMS settings. All details regarding pasture layout,
barn, feeds and AMS settings are the same as in Chapter 2.

Rumen Fluid Samples. Rumen fluid samples were taken over a 48 h period during the in
situ incubation trial, resulting in samples times of 0830, 1530, 1830, and 2030 h. Fluid samples
were taken with a filtered syringe from different compartments of the rumen in order to get a
representative sample. The pH was recorded with a pH meter immediately after rumen fluid was
collected.

In situ Degradation. The in situ incubation technique (Orskov and McDonald, 1979) was
used to determine the presence of associative effects in the rate and extent of degradation of
pasture and concentrates in the rumen.

85

The material incubated in rumens of the cows in the LSR treatment diet included 4
ingredients: ground corn, concentrate based pellets, and masticated orchardgrass and ryegrass
pasture. The HSR feeds were the same as in LSR, with the addition of a legume-grass haylage as
the PMR feed mixed with grounded corn (60% haylage and 40% corn). Following Islas and
Soto-Navarro (2011), masticate samples were collected from the 6 ruminally fistulated cows
using the rumen evacuation method. Briefly, after a 2 h fasting the rumen content was
completely evacuated and the digesta plus liquid was stored in large plastic buckets, weighed,
and sealed with plastic cover. After this initial evacuation was completed, cows were allowed to
graze on either Ryegrass or Orchardgrass pastures for 2 hours. Cows were not allowed to drink
during the pasture grazing period. After grazing, cows were re-evacuated completely and
masticated contents stored in large plastic buckets. Contents of the first evacuation were then
returned into the evacuated rumen. The masticate samples were dried in a forced air oven at
60°C for 48 h and stored for use as incubation material in the in situ trial.
All feed ingredients used for incubations were chopped using a blender and then sieved
through a 1.5 mm screen, therefore all incubated particles were less than 1.5 mm. Nitrogen free
polyester bags (10 x 20 cm) with a 50 micron pore size (Ankom Technology, Macedon, NY)
were used. Bags were filled with 5 g of dry feed material and sealed with a heat sealer (Model
AIE-200, American International Electric Inc., Industry, CA). Bags were incubated for 2, 5, 12,
24, 36 and 48 h using duplicate bags for the 2, 5, 12, and 24 h incubation and triplicate bags for
the 36 and 48 h incubation. A blank bag was included at each incubation hour to correct for
changes in bag weight. All incubated bags were contained in a netted nylon bag closed to one
end with a plastic zip tie. Bags were incubated in reverse order starting with the 48 h incubation.
All bags were removed at the same time.

86

After incubation, bags were immediately soaked in cold water and hand rinsed once.
Bags were then placed in a washing machine and were rinsed in cold water for 15 minutes. After
rinsing, bags were placed in a forced air oven, dried for 48 hours at 60°C and stored. Crude
protein and neutral detergent fiber were determined using same techniques described in Chapter
2 and the disappearance of dry matter, CP and NDF calculated.
The disappearance of dry matter and CP was analyzed with an exponential model
-ct

following Orskov and McDonald (1979). The equation used was p = a + b (1-e ); where p is the
actual degradation over time (t), the parameter a is the intercept of the degradation curve at time
zero and represents the soluble fraction, the parameter b represents the potentially degradable
fraction relative to the rate constant of degradation indicated by the parameter c. With this
model, the degraded fraction is a + b and the undegradable fraction is 100 - (a + b). If the value
of a is positive then there is a component which is degraded instantaneously and or a component
which is soluble enough to escape by soaking and washing; whether parameter a represents rapid
degradation or washing losses can be determined with control bags (Orskov and McDonald,
1979). There were no control bags included in this study, therefore, parameter a included both
rapid degradation and possible washing losses. When a negative value for a is obtained, an
initiation period for degradation to start is required, this is referred to as lag time (Mertens and
Loften, 1980).
The NDF results demonstrated a lag time in degradation and were therefore analyzed
with a different exponential model following Mertens and Loften (1980). The equation used was
FR = P e

-kd(t-L)

+ U, where FR is the amount of NDF remaining at incubation time t, P is the

potentially digested fraction (100-U) at the fractional rate of degradation kd (%), U is fraction
undigested at 48 h (%), L is the lag time (h) for NDF degradation, and t is time of incubation.
87

Laboratory Analysis. Crude protein and NDF of the incubated material and residue
remaining after incubations were determined by same procedures described in Chapter 2.

Statistical analysis. Evaluation for centrality and dispersion of the data conducted prior
to the ANOVA analysis detected a consistent bias for one of the cows in the HSR treatment.
After study completion, this cow was diagnosed by the veterinarian with an unexplained uterus
illness not related to the feeding treatment and data from this cow were therefore removed from
the final analysis.
Rumen evacuation data and in situ degradation parameters were analyzed by the method
of least-squares ANOVA for a completely randomized design, using the Mixed Models
procedure of SAS (Version 9.2, SAS Institute Inc.). The Kenward-Rogers method was used to
estimate the degrees of freedom for comparing effects of stocking rates. Rumen evacuation data
were analyzed with a model with stocking rate as the independent fixed effect. Repeated
measures analysis was used to evaluate the effect of stocking rate on rumen degradation and pH
obtained in the in situ trial. Time of incubation (or time of day equivalent) was the repeated
measure and was modeled with a compound symmetry, autoregressive order 1 or unstructured
covariance structure based on the lowest Akaike and Bayesian Information Criterion test (Little
et al. 1996). In these analyses, individual cannulated cows were considered the experimental unit
and entered in the model as a random effect.

88

RESULTS AND DISCUSSION

Rumen evacuations and pH. The rumen evacuations allowed me to describe rumen fill,
retention time, and total clearance and to obtain unbiased masticated forage samples consumed
by cows for in situ incubations. No differences were detected between the two feeding systems
for any evacuation data collected (Table 3.1). However, absolute values for dry fill were
numerically lower and values for retention time and clearance rate were numerically higher for
HSR.
During the 48 hour in situ period, pH was not significantly different (P = 0.61) between
HSR and LSR at all sampling hours, but pH differed significantly (P = 0.008) over the course of
the day. Regardless of this difference, it is important to note that across all hours, pH was never
below 5.50, a condition which may be considered an indicator of acidosis in a supplemented
pasture-based system. However, the average pH for both treatments was slightly lower than the
ranges described in other grazing trials using similar pastures. Bargo et al. (2003) reviewed
recent literature describing pH values for grazing dairy cattle and reported average values of 6.10
(range: 5.75 to 6.29) for cows supplemented with 1.1 to 10 kg of concentrates. In this study, both
LSR and HSR cows received concentrates at a rate of 1 kg of pellet per 6 kg of milk (average 4.5
and 4.6 ± 0.3 kg for HSR and LSR, respectively) plus a constant amount of 1 ± 0.2 kg of ground
corn. The difference between treatments was the PMR (60 % forage, 40% ground corn) feeding
at an average rate of 5 ± 0.1 kg/cow/day during the study (see Chapter 2 for more details). Thus,
pH values recorded for HSR and LSR could be considered normal for supplemented grazing
cattle based on Bargo et al. (2003). Furthermore, the additional feeding of approximately 2 kg of

89

corn in the PMR mixture was not sufficient to cause a drop in rumen pH compared to LSR
receiving no PMR.
There is not a well-established reference for average pH values in grazing dairy cows
supplemented with PMR, particularly under commercial situations where PMR could be used to
stabilize variations in feed intake patterns and rumen pH resulting from diurnal and temporal
variations in pasture availability and consumption. Yet, feeding variable amounts of PMR over
time may create residual effects in rumen pH and degradation. Bargo et al. (2002) found no
significant differences in rumen pH between diets containing pasture plus concentrate (5.89) and
pasture plus partial total mixed ration (5.88), values that are very similar to the average rumen
pH values found in this study (5.98 and 5.89, for LSR and HSR, respectively). Evidently, the
amount of PMR fed in HSR (representing an addition of 2 kg in the diet) was not enough to
cause a constant decrease in pH over the course of the day and in detectable differences in pH
between treatments.

In situ degradation. The in situ rumen degradation data for the supplements are shown in
Table 3.2. Ground corn and pellets were incubated in both the HSR and LSR treatments, while
the PMR containing pasture haylage and ground corn was incubated separately only in the HSR
treatment. There were no detectable differences between the two stocking rate and feeding
systems in all degradation parameters tested. Therefore, these data support the hypothesis that
PMR supplementation does not affect degradation kinetics of concentrate supplements when low
levels of PMR are used as a complementary feed resource in pasture-based systems. As
expected, the rate of concentrate degradation was relatively high in this study (more than 16%
per hour for ground corn). However, we recognize that inferences on the net efficiency of

90

digestion of supplements are limited because neither the extent of concentrate degradation nor
total tract digestion was measured in this study. It is possible that the efficient use of highly
degradable concentrate supplements in pasture systems is limited by a very high rate of passage
for supplements through the tract. In pasture dairy systems, the rumen of cows grazing high
quality pasture of low DM and NDF content often suffers from low DM fill (fast turnover), with
the absence of a stratified contents in the rumen and lack of stable mat, as compared to TMR
feeding systems (Bargo et al. 2002b). This lack of structure in the rumen may prevent the
efficient retention of small concentrate particles that may adhere to the ventral sac of the rumen
and resist extensive degradation. Although not reported here, significant amounts of grain
concentrate have been removed from the rumen ventral sac during full rumen evacuations, which
suggests that the extent of degradation of concentrates was probably low. More controlled
studies are needed to determine actual efficiencies of concentrate degradation and digestion in
feeding systems using high quality pasture as the main feed resource.
As expected, the in situ technique revealed no significant differences in pasture
degradation between feeding systems (Table 3.3). However, important numerical differences in
most NDF degradation parameters (particularly lag time) were observed. It is possible that the 5
experimental cows used in the study did not provide enough statistical power to detect a
significant difference between feeding systems.
The values for dry matter disappearance found in this study are comparable to other
studies where a similar in situ technique was used. Hongerholt and Muller (1996) incubated
freeze dried pastures samples dominated by orchardgrass and Kentucky bluegrass, ground
through a 4-mm screen, and found values of 24.4%, 60.3%, and 6.7%/h for the soluble (a),
potentially degradable (b), and the rate of degradation (c) of dry matter. Van Vuuren et al. (1992)

91

also used a similar procedure, freeze drying samples ground to 3 mm, to incubate ryegrass
pasture and found degradation rates for dry matter of 40.8%, 53.1%, and 6.1%/h for the fractions
a, b, and c, respectively.
Hongerhold and Muller (1996) recorded values of 19.0, 75.3, and 5.6%/h for the fractions
a, b, and c for the CP degradation in orchardgrass based pastures. The values for CP for
Orchardgrass may be more similar to Hoffman et al. (1993), in which a result of 46.1 to 51.5 %
was found for the a fraction. However, in that study, sample processing was somewhat different
and cows were on a silage based diet. Van Vuuren et al. (1992) found CP degradation values of
47.2%, 47.5%, and 9.2%/h for fractions a, b, and c This value for c was much higher than what
was recorded in the current study. Differences in pasture maturity, botanical composition and
pre-processing of plant material through mastication may explain differences between studies.
Pasture dry matter and CP degradation are greatly influenced by plant maturity, creating
important variations over the life cycle of plant species (Hoffman et al., 1993). Differences
between the current study and the studies reported above may also occur because many of the
studies reviewed were completed indoors on a TMR or silage based diet, as for example in
Hongerhold and Muller (1996) and Hoffmann et al. (1993).
Figures 3.2 and 3.3 show the NDF degradation patterns of the two different pasture types.
Although differences between the two stocking rate treatments were not statistically significant,
more favorable fermentation conditions might have resulted in faster disappearance in the HSR
treatment. This conclusion is based on the trends seen in the data for the in situ degradation of
pasture, in which the degradation rate was 1 to 2 units higher for HSR vs. LSR (Table 3.3). The
difference may be caused by differences in passage rates. If the passage rate in HSR cows was
slower, possibly due to higher levels of indigestible NDF in the PMR, this may provide more

92

time for microbial population growth and attachment, which would increase NDF degradation
rates. Perhaps the PMR is creating synergistic effects on the rate of degradation of NDF by
increasing particle retention times. However, without a significant difference in degradation rates
this idea is difficult to support in this study.
The soluble fraction was similar in both orchardgrass and ryegrass based pastures (range
27 to 32 %), but the degradable fraction was about 10 units greater in ryegrass pastures. The
lower NDF and higher concentration of NSC (not determined) in ryegrass compared to
orchardgrass pastures (see table 2.2 of Chapter 2) may explain this difference.

Conclusion. The results from this study indicate that low levels of PMR supplementation
of a pasture-based diet do not change rumen degradation rates or cause other negative associative
effects in the rumen. Furthermore, when PMR is used to complement the deficit in availability of
high quality pasture, it may have synergistic effects on the extent of concentrate digestion and
NDF degradation. Overall, no significant differences were seen in measures of rumen
degradation between stocking rate and feeding systems, but more controlled research is
necessary to determine the actual occurrence of potential synergistic effects of PMR feeding.

93

APPENDIX

94

TABLES

Table 3.1. Effects of stocking rate and feeding system on rumen contents, estimated kinetic
parameters and amount of masticates collected from orchardgrass and ryegrass based pastures.
LSR

Variable
3

1

HSR

2

SEM

P Value

Estimated Intake , kg

19.50

19.12

0.82

0.75

Wet Fill, kg
Dry Fill, kg

64.40
8.41

70.59
8.72

8.6
2.19

0.62
0.92

DM Content, %

12.90

12.00

1.80

0.72

8.91

9.52

-

Clearance Rate of DM , %/h

0.11

0.11

-

Masticates*
Masticate Wet Fill, kg

OG

RG

42.57

14.43

SEM
4.78

Masticate Dry Fill, kg

5.05

2.21

0.72

0.06

Pasture intake rate, kg/h

2.52

1.10

0.36

0.06

4

Retention Time of DM , h
5

1

7

8

NA

6

NA
P Value
0.02

Low stocking rate treatment. Received a pasture plus concentrate diet.

2

High stocking rate treatment. Received a partial mixed ration supplementation in addition to

pasture plus concentrate.
3

Estimated Intake, derived from a parallel trial using fecal output estimates and 48 h digestibility

using chromium oxide as external marker.
4

Retention time was derived from estimated intake as [(DM intake/24h)/rumen DM content].

5

Clearance Rate was derived from estimated intake as [rumen DM content/(DM intake/24h)

6

NA, not applicable.

7

Orchard grass, fescue, alfalfa, red clover and white clover pasture mixture.

8

Ryegrass and white clover pasture mixture.

*Masticates were collected after the full evacuation of the rumen following a 2 h grazing session.

95

Table 3.2. Effects of stocking rate and feeding system treatment on in situ degradation
parameters of concentrates and partial mixed ration
1

Parameter
Ground Corn
Dry Matter
a
b
c, %/h
Crude Protein
a
b
c, %/h
Pellet
Dry Matter
a
b
c, %/h
Crude protein
a
b
c, %/h
Partial Mixed Ration
Dry Matter
a
b
c,%/h
Crude Protein
a
b
c, %/h
Neutral Detergent Fiber
Lag
P
Kd/h

HSR

2

LSR

3

SEM

P Value

0.25
0.70
0.16

0.30
0.64
0.17

0.03
0.04
0.01

0.26
0.32
0.65

0.20
0.77
0.12

0.28
0.69
0.12

0.05
0.05
0.01

0.31
0.31
0.85

0.31
0.86
0.03

0.29
0.75
0.038

0.03
0.06
0.01

0.63
0.26
0.30

0.27
0.71
0.05

0.09
0.81
0.11

0.09
0.05
0.04

0.22
0.19
0.30

0.34
0.42
0.07

-

0.02
0.01
0.02

NA
NA
NA

0.58
0.27
0.1

-

0.03
0.02
0.04

NA
NA
NA

1.67
0.39
0.05

-

0.46
0.08
0.02

NA
NA
NA

1

4

Parameters were a: soluble fraction, b and P: potentially degradable
fraction, c and Kd: fractional rate constant of degradation, Lag, lag time of
degradation.
2

High stocking rate treatment

3

Low stocking rate treatment

4

Not applicable because the comparison was limited by design.

96

Table 3.3. Effects of stocking rate treatment on in situ degradation of ryegrass and orchardgrass
pastures
Treatments
1

Parameter
Ryegrass
DM
a
b
c, %/h
CP
a
b
c, %/h
NDF
Lag
P
kd/h
Orchardgrass
DM
a
b
c, %/h
CP
a
b
c, %/h
NDF
Lag
P
Kd, %/h

HSR

2

LSR

3

P Value
SEM

SR

4

0.28
0.61
0.07

0.29
0.63
0.06

0.02
0.06
0.01

0.72
0.77
0.32

0.22
0.72
0.08

0.21
0.76
0.06

0.05
0.07
0.007

0.84
0.70
0.09

2.06
0.21
0.06

2.79
0.13
0.05

0.37
0.16
0.01

0.23
0.71
0.31

0.31
0.57
0.08

0.34
0.54
0.07

0.03
0.03
0.02

0.47
0.45
0.51

0.31
0.64
0.08

0.35
0.62
0.06

0.03
0.03
0.01

0.34
0.61
0.23

1.35
0.21
0.08

0.23
0.22
0.06

0.56
0.06
0.02

0.22
0.96
0.51

1

Parameters were a: soluble fraction, b and P: potentially degradable
fraction, c and Kd: fractional rate constant of degradation, Lag, lag time
of degradation.
2

High stocking rate treatment

3

Low stocking rate treatment

4

Stocking rate effect

97

FIGURES

pH
6.4
6.3

6.2
pH

6.1
6
5.9
5.8
5.7
5.6
0730

0930

1130

1330

1530

1730

1930

2130

Hour

Figure 3.1 Rumen pH of dairy cows grazing temperate pasture in either a low stocking rate
system supplemented with concentrates or a high stocking rate system supplemented with
concentrates and partial mixed ration. There was no stocking rate and feeding system effect ( P =
0.61), but significant differences were detected among hours of the day (P < 0.01). Bars denote
standard errors of means (SEM= 0.09).

98

Disappearance, %

Disappearance of Ryegrass NDF
90
80
70
60
50
40
30
20
10
0
-10

HSR
LSR

2

5

12
24
Hours of incubation

36

48

Figure 3.2. Disappearance of ryegrass NDF in dairy cows grazing temperate pasture in either a
low stocking rate system supplemented with concentrates (LSR) or a high stocking rate system
supplemented with concentrates and partial mixed ration (HSR). There was no stocking rate and
feeding system effect (P = 0.45, SEM= 4.23).

Disappearance, %

Disappearance of Orchardgrass NDF
90
80
70
60
50
40
30
20
10
0

HSR
LSR

2

5

12
24
Hours of incubation

36

48

Figure 3.3. Disappearance of orchardgrass NDF in dairy cows grazing temperate pasture in
either a low stocking rate system supplemented with concentrates (LSR) or a high stocking rate
system supplemented with concentrates and partial mixed ration (HSR). There was no stocking
rate and feeding system effect (P = 0.60, SEM= 4.71).

99

REFERENCES

100

REFERENCES

Bargo, F., L. D. Muller, E. S. Kolver, and J. E. Delahoy. 2003. Invited Review: Production and
digestion of supplemented dairy cows on pasture. J. Dairy Sci.86:1-42.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002a.Milk response to concentrate
supplementation of high producing dairy cows grazing at two pasture allowances. J.
Dairy Sci. 85:1777–1792.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002b. Performance of high
producing dairy cows with three different feeding systems combining pasture and total
mixed rations. J. Dairy Sci. 85:2960–2975.
Hoffman, P. C., S. J. Sievert, R. D. Shaver, D. A. Welch, and D. K. Combs. 1993. In situ dry
matter, protein, and fiber digestion of perennial forages. J. Dairy Sci. 76:2632–2643.
Hongerholt, D. D., and L. D. Muller. 1998. Supplementation of rumenundegradable protein to
the diets of early lactation Holstein cows grazing grass pasture. J. Dairy Sci. 81:2204–
2214.
Islas, A., and S. A. Soto-Navarro. 2011. Effect of supplementation of dried distillers grains with
solubles on forage intake and characteristics of digestion of beef heifers grazing smallgrain pasture. J. Anim. Sci. 89:1229-1237.
Kolver, E. S., L. D. Muller, G. A. Varga, and T. J. Cassidy. 1998. Synchronization of ruminal
degradation of supplemental carbohydrate with pasture nitrogen in lactating dairy cows.
J. Dairy Sci. 81:2017–2028.
Little, R. C., Milliken, G. A. Stroup, W. W., and Wolfinger, R. D. (1996), SAS system for mixed
models, Cary, NC: SAS Institute.
Mertens, D. R., and J. R. Loften. 1980. The effects of starch on forage fiber digestion in vitro. J.
Dairy Sci. 63:1437–1446.
Ørskov, E. R., and I. McDonald. 1979. The estimation of protein degradability in the rumen from
incubation measurements weighed according to rate of passage. J. Agric. Sci. (Camb.)
92:499–503.
Phillips, C. J. C. 1988. The use of conserved forage as a supplement for grazing dairy cows.
Grass Forage Sci. 43:215-230.
Reis, R. B., and D. K. Combs. 2000 Effects of increasing levels of grain supplementation on
rumen environment and lactation performance of dairy cows grazing grass-legume
pasture. J. Dairy Sci. 83:2888–2898.
S.A.S. Institute, 2008. SAS User's Guide: Statistics. SAS Institute Inc., Cary, NC.
101

Van Vuuren, A. M., F. Krol-Kramer, R. A. van der Lee, and H. Corbijn. 1992. Protein digestion
and intestinal amino acids in dairy cows fed fresh Lolium perenne with different nitrogen
contents. J. Dairy Sci. 75:2215–2225.

102

CHAPTER 4
EFFECTS OF STOCKING RATE AND PARTIAL MIXED RATION
SUPPLEMENTATION ON FORAGE UTILIZATION AND OCCUPATION OF
AUTOMATIC MILKING SYSTEMS IN A PASTURE-BASED SYSTEM
ABSTRACT
Integration of automatic milking system (AMS) into a Midwestern USA pasture-based
system is a potential alternative low cost and labor efficient dairy system. However, guidelines of
management in critical areas such as stocking rate, supplementation, and cow traffic are not yet
well understood. The objective of this study was to determine the combined effects of stocking
rate and feeding strategies on pasture utilization, milk yield, and interactions with the overall
occupation and milking efficiency of an AMS. A total of 80 United States Holstein and 14 New
Zealand Friesian cows divided into equal proportions in two herds balanced for parity (2.03 ±
0.1; mean ± SD), days in milk (227 ± 10 d; mean ± SD) were assigned to a low stocking rate
treatment (LSR; 1.92 cows/ha) and a high stocking rate treatment (HSR; 2.89 cows/ha)
implemented over 24 and 16 ha of temperate pasture, respectively. The feeding system for the
LSR treatment included pasture plus addition of pellet (weighted average = 4.2 ± 0.3 kg/cow per
d) fed at variable rates in the AMS at 1 kg per 6 kg of milk and ground corn (weighted average
1.0 ± 0.2 kg/cow per d) fed to individual cows at a fixed rate in automatic feeders. The same
pasture and concentrate feeding systems plus the feeding of variable weekly amounts of a partial
mixed ration (PMR; 60% forage, 40% concentrate) were used in the HSR treatment. The
objective of the PMR feeding system was to adjust for deficits in pasture availability resulting as
the difference between prescribed daily pasture consumption (SR x 14 kg DM/cow) and the

103

average pasture growth rate in the HSR treatment. No significant differences in average daily
pasture utilization were detected (alpha 0.05). Daily pasture utilization was 870.9 and 734.5± 70
kg DM/ha (mean ± SEM), respectively. Supplementation with PMR at low levels (average 4.86
± 0.14 kg/cow per d) did not result in significant differences in herd milk production. However,
daily milking distribution patterns by hour shifted (P < 0.01) as a consequence of PMR feeding
in the HSR treatment. Results supported the prediction that proper PMR supplementation could
be used as a tool to efficiently manage forage utilization in highly stocked pasture-based systems
that are aiming to improve the AMS occupation and milkings.

Keywords: automatic milking systems, pasture-based dairy, stocking rate, partial mixed ration

104

INTRODUCTION
Pasture forage utilization is a main driver for profitability and represents a major
management priority in a grazing dairy farm (Gronow et al., 2010). Past research showed that
proper stocking rate (animals/area) management generally increases pasture utilization
(Stockdale and King, 1980, MacDonald et al., 2008). Even defoliation, high stocking density
grazing (animals/area/time), maintenance of proper plant residues and avoidance of lenient
grazing leading to low quality forage are often cited as major benefits in properly stocked
grazing dairy systems (Parsons and Chapman, 2000). Reduction of substitution rates of grazed
pasture by supplements is another benefit in a properly stocked grazing system (Fales et al.,
1995).
Concurrently, as the utilization of grazed pasture increases, milk production per area also
increases, but milk production per cow often decreases (Stockdale and King, 1980; MacDonald
et al., 2008). This decrease in milk production is the result of lower feed intake and conversion
efficiencies (Little et al. 2011). To prevent the decrease in milk production per cow, managers
often rely on supplements (Reis and Combs 2000, Bargo et al., 2002a, 2002b, 2003). However,
improper supplementation with forages or concentrates could lead to a high substitution rate of
pasture and low levels of pasture utilization (Baker and Leaver, 1986, Fales et al., 1995).
Proper stocking rate management could also affect the milking efficiency and the timing
of occupation of automatic milking systems (AMS). Low density of animals per milking stall,
uneven milking distribution or reduced milkings could be some of the potential undesirable

105

effects in improperly stocked AMS grazing systems (Jago et al., 2007). However, no studies
have confirmed these observations carefully.
The present study tested the effect of stocking rates and variable feeding of a partial
mixed ration as a tool to manage forage utilization and milkings in an AMS grazing dairy
system. The hypothesis was that timely adjustments in PMR feeding rates could be used to
prevent undesirable decreases in pasture growth rate and forage utilization in a high stocking rate
system and also aimed to improve the occupation for AMS milking.

106

MATERIALS AND METHODS
Animals, Experimental Design, and Treatments. Eighty United States Holstein and 14
New Zealand Friesian cows divided in equal proportion to two herds balance for parity (2.03 ±
0.1), days in milk (227 ± 10 days), were assigned to a low stocking rate (LSR; 1.92 cows/ha) or a
high stocking rate treatment (HSR; 2.89 cows/ha) implemented over 24 and 16 ha of temperate
pasture, respectively.
Two pasture systems were included in both stocking rate treatments: a high biodiversity
pasture mixture (OG) including orchardgrass, tall fescue, alfalfa, white clover and red clover and
a low biodiversity pasture mixture (RG) composed of white clover and perennial ryegrass. Both
pasture mixtures were subdivided in paddocks of 1 ha for rotational grazing using a controlled
traffic system (See Chapter 2 for more details). The LSR treatment had access to twenty four 1
hectare paddocks (See Chapter 2, Figure 2.2), sixteen paddocks containing OG pasture and the
remaining 8 paddocks containing RG pasture. The HSR treatment had access to sixteen 1 ha
paddocks, with half of the paddocks containing OG pasture and the other half containing RG
pasture. Fresh pasture allocations were opened for grazing at 0500 h. The HSR treatment had
access to one half of a paddock per day. A temporary fence was used to prevent cows from
grazing the second half of the paddock. On the second day of grazing in the same paddock, the
temporary fence was removed and cows accessed the second half of the paddock. The LSR
system grazed 1 paddock per day. Rotational grazing of paddocks was based on approximate
targets of 2400 and of 1600 ± 200 kg DM for pre-grazing and post-grazing pasture biomass,

107

respectively. When necessary, pastures were grazed based on stage of maturity using the number of
extended leaf per tiller in grasses and flower percentage in legumes.

The feeding system for the LSR treatment included grazed pasture plus addition of
concentrate pellet (weighted average 4.2 ± 0.3 kg/head) feed in the AMS at a rate of 1 kg per 6
kg of milk and ground corn (weighted average 1.0 ± 0.2 kg/head) fed at a fixed rate in automatic
feeders. The same pasture and concentrate feeding systems plus the feeding of a partial mixed
ration (PMR; 60% pasture haylage, 40% concentrate) was used in the HSR. The objective of the
PMR feeding in the HSR system was to adjust for deficits in pasture availability resulting from
the difference between the prescribed pasture consumption (SR x pasture intake/head) and the
average pasture growth rate in the system. The prescribed pasture consumption in the HSR
treatment was 15 kg DM/head or approximately 36.4 Mcal of ME from grazed pasture (2.42
Mcal ME/kg DM).

Pasture Layout, Barn, Feeds, and AMS settings. All details regarding pasture and
supplemental chemical composition, farm layout, milking barn design, traffic and AMS milking
settings are reported in Chapter 2.

Pasture Measurements and Growth Rate Calculations. Indirect estimates of the pasture
pre-grazing and post-grazing biomass in each grazed paddock were made with an electronic
rising plate meter (RPM; F300, Farmworks, New Zealand), which calculates the average pasture
biomass in a given paddock (y) as a function of the compressed pasture height (x) recorded in

108

"click" units (1 click = 0.5 cm of compressed pasture height). Thirty "pluck" measurements of
pre-grazing and post-grazing pasture were conducted alongside transects covering the entire area
of a grazed paddock. The linear equation used to estimate pre-grazing and post-grazing biomass
in OG pastures was y = 92 x. The linear equation used to estimate the pre-grazing and postgrazing biomass in RG pastures was y = 78 x. For the purpose of this study, the difference
between the pre-grazing and post-grazing biomass or pasture harvested by animals was defined
as forage utilization.
The pasture growth rate (PGR) was determined weekly with a rapid pasture meter (CDax, Farmworks, New Zealand). The PGR was the weekly difference in the average pasture
cover of paddocks divided by7 d. The average PGR in the HSR systems was compared with the
target PGR or PGR needed to determined amount of PMR feeding as indicated above. For
example, the HSR farmlet stocked at 2.89 cows/ha and managed with a target pasture intake of
15 kg DM of pasture/cow required a PGR of 43.4 kg DM/ha/d. If the PGR determined weekly
was lower than this required amount, the resulting deficit divided the stocking rate in the farmlet
was defined as the amount of PMR to be fed. For example, if the PGR for the HSR was 30
kg/ha/d, the deficit between the PGR required (43.4 kg DM/ha/d) and the current PGR (30 kg
DM/ha/d) divided 2.89 cows/ha was equal to the 4.62 kg DM/cow of PMR fed that week.

AMS Data Collection and Milk Samples. The AMS identified each cow by an assigned
electronic collar ID and collected information on milk production, milking frequency, AMS
visits, and amount of ground corn and pellet consumed.

109

Statistical Analysis. Pellet and corn intake, AMS and milk performance data were
analyzed by method of least-squares ANOVA using the Mixed Models procedure of SAS
(Version 9.2, SAS Institute Inc.). Dependent variables were analyzed with a model including the
independent fixed factors stocking rate, week and the interaction stocking rate by week. The
week effect was entered in the model as a repeated measure variable and the random using a
compound symmetry, autoregressive order 1 or unstructured covariance structure based on the
lowest Akaike and Bayesian Information Criterion test (Little et al. 1996). In this analysis
individual cows were considered as the experimental unit for the animal-related variables and
entered in the model as a random effect, as it is commonly used in whole farmlet grazing trials
(Bargo et al. 2003). Pasture related data including pasture intake were analyzed similarly. In this
analysis the individual 1 ha paddocks were used as the experimental unit and unit of replication.
The model of analysis included the fixed effects stocking rate (farmlet), week and the treatment
by week interaction. The individual paddock was considered as a random factor in the model.
The selection for best fit of covariance structure was conducted as described above.
Nonparametric analysis of milking distributions during hours of the day and milking
intervals were conducted with a Chi-squared test using the FREQUENCY procedure of SAS
(Version 9.2, SAS Institute Inc.). This nonparametric method was preferred over a parametric
procedure because the skewness of the data clearly departed from robust normal distributions.
This analysis was used to determine the frequency of milkings at a particular hour of the day and
the frequency of milkings at a particular interval. Filtered data were used for this analysis in

110

order to exclude AMS milking failures and the 7 milkings following a given AMS milking
failure, as these milking events may not be considered normal records (Bach and Busto, 2005).
Mean and standard deviation values were used to describe the performance of AMS in
the two stocking rate systems. No analysis of variance was conducted on these variables, since
only 1 AMS unit per stocking rate treatment was available in the study.

111

RESULTS
Pasture Measurements. Pasture measurements; pre-grazing biomass, post-grazing
biomass, and pasture utilization (difference between pre-grazing and post-grazing biomass) were
not different between stocking rate (SR) treatments, but there was a marginal tendency for higher
post-grazing herbage mass residuals in LSR (Table 4.1). Higher grazing intensity and lower
forage harvested per cow were detected in HSR (Table 4.1). Significant variation across weeks
was found in the pre-grazing and post-grazing herbage mass of pasture for both treatments
(Figure 4.1). This difference in biomass was likely explained by the changing environmental
conditions and rainfall events occurring at the beginning of the study (data not shown).
Favorable growing conditions developed shortly after the start of the study causing a sharp
increase in pasture growth rate by week 3 of the study. Both SR treatments increased
dramatically in pasture biomass; there was no detectable difference in pasture growth rate
between treatments (Figure 4.2; P > 0.05).
Cow performance. Most cow performance variables were affected by a stocking rate by
week interaction (Table 4.2). Milk production in HSR increased consistently across weeks of the
study, whereas milk production in LSR achieved a maximum plateau by week 4 of the study
(Figure 4.5). Other milk production related variables followed similar variations. Preparation
time (time for teat cleaning plus teat cup attachment by the AMS) decreased as the study
progressed, most likely due to the change in udder conformation as a consequence of higher milk
production. On average, the preparation time was 36 s faster for the HSR treatment (Table 4.2).
Milking frequency (MF) remained around 2.2 milkings/d in the HSR treatment, but in the LSR

112

treatment MF decreased at a faster rate to less than 2 milkings/d by week 5 of the study (Figure
4.6). Better pasture quality and availability may have reduced the motivation of cows to return to
the barn for milking in the LSR treatment, since no PMR was fed in this treatment as compared
to the HSR treatment. Thus, PMR feeding in the HSR treatment may have resulted in a more
steady motivation for cows to visit the AMS and to milk at more regular intervals. As MF
decreased, milking interval increased over the course of the study, explained by a significant
stocking rate treatment by week interaction (Table 4.2).
Weight increase was significantly higher in the HSR treatment (P < 0.0037), but did not
show a significant SR x week interaction (Table 4.2). Refusals conducted by the AMS were
significantly higher in the HSR treatment (Table 4.2), which may be explained by the higher
motivation by cows to visit the AMS when PMR was fed in the milking barn.
No differences in pellet and ground corn supplement intake were detected between
stocking rates or across weeks (Figure 4.3 and 4.4). Total intake was lower from week 1 to 3 in
both stocking rate treatments, but total intake was higher in the LSR treatment by week 4 (Figure
4.3 and 4.4).

AMS milking distribution and performance. The PMR feeding in the HSR treatment
affected the diurnal pattern of milkings at the AMS (Figure 4.8). Frequency of milkings was low
from 0200 to 0500 h and from 2200 to 2300 h in both stocking rate treatments. Throughout
daylight hours cows exhibited two differential milking cycles, one in the morning (0500 to 1200
h) and one during the afternoon and evening hours (1500 to 2100 h), probably related to

113

environmental factors (i.e., temperatures), differential PMR feeding and diurnal patterns of
grazing. The PMR feeding at 0500 h increased the frequency of milkings from 0500 to 0900 h in
the HSR treatment. Conversely, the morning peak for AMS milkings in the LSR treatment
occurred between 0700 and 1200 h. Comparative descriptions of the AMS performance of the
two treatments are presented in Table 4.3. Underutilization of the AMS resulted from the low
number of cows used in the study. On average, both SR treatments had a 52% effective
occupation for milking; leaving a significant percent of free time that could be used for milkings
if the SR of the AMS (cows/AMS) increased. Total milk yield, total milkings, and refusals
means were numerically similar for both SR treatments. The numbers of fetched cows in each
treatment were similar, but the frequency of fetched cows was about 10 times lower during the
PM hours (Table 4.3).

114

DISCUSSION
Pasture measurements. The pre-grazing and post-grazing pasture measurements were
not affected by dynamic supplementation (Table 4.1 and Figure 4.1). The pre-grazing biomass
changed by week, due to environmental factors, yet the post-grazing biomass was maintained in
both SR systems as a consequence of the grazing management guidelines applied. There was a
tendency for a lower post-grazing biomass in the HSR treatment, which accompanied a higher
grazing intensity. This difference was probably due to the higher stocking density for grazing in
HSR, as the same numbers of cows were allocated to half of the daily grazed area used in LSR.
As expected no differences in forage growth rate and daily forage utilization per area were
detected between treatments, despite the predicted significant difference in the daily amount of
pasture harvested (i.e., intake) per cow. This result was remarkable because grazing management
guidelines were strategically designed to maintain desirable post-grazing residuals, regardless of
stocking rate. The goal was to implement a grazing guideline that favored pasture growth rate, in
order to achieve a high level of forage utilization per area regardless of the stocking rate system.
The results demonstrate that PMR feeding based on detectable changes in pasture growth rate
could be used as a tool to manage forage utilization in dairy grazing systems aiming to increase
overall milk production from increases in stocking rate. Forage utilization is the main driver of
profitability in pasture-based dairies (Gronow et al., 2010) and the major factor determining the
level of economic risk in pasture-based feeding systems (Little et al., 2011).
It is generally known that within rational ranges, increased stocking rates generally
increase pasture utilization (MacDonald et al., 2008). The increase in pasture utilization is

115

generally due to the avoidance of lenient uneven grazing and maintenance of desirable grazed
residuals stimulating high quality regrowth. However, as it has been seen previously, high
stocking rates may eventually cause overgrazing, when stocking rate exceeds the carrying
capacity of the pasture system. Overgrazing of pasture decreases plant community resilience and
causes degradation; and can lead to an early cessation of the grazing season (MacDonald et al.,
2008). Generally the increase in forage utilization with higher SR results in an increase in a
decrease in milk production per cow (MacDonald et al, 2008). This was not the case in this
study, as milk production per cow was not significantly different between stocking rate systems,
although there was also no difference in pasture harvested per area.
It also is commonly observed that supplementation, particularly with forage, increases
substitution rates (Bargo et al., 2003), which decreases pasture intake and may decrease pasture
utilization if adjustments in stocking rate are not made. However, if substitution is managed, as it
was proposed with the dynamic use of supplemental PMR, it can serve as an important grazing
management tool. In this study, substitution created by PMR feeding was used to ensure the
maintenance of an approximate residual of 1200 kg DM/ha in the HSR treatment. Therefore,
supplementation amounts were evaluated weekly to cover only the deficit in pasture growth. The
PMR was not designed to increase total intake or energy consumption in this study, but was
designed to promote pasture regrowth and subsequent pasture utilization.
The concept of dynamic supplementation in a pasture-based feeding system was also
used by Farina et al. (2011). In that study, the equation, pasture growth rate/stocking rate was
also used to determine the amount of feed available per cow from pasture. The difference of

116

target intake and the actual feed available determined the amount of feed needed from
supplements. The current study included the same approach, determining the pasture growth rate
(PGR) and estimating the amount of PMR fed based on the deficit in PGR. Farina et al. (2011),
using the concepts of dynamic supplementation, also did not find any differences in pasture
utilization when farmlets of high and low stocking rates were compared.

Intake and Feed. Pasture intake shown in Table 4.1 was derived from differences
between the pre-grazing and post-grazing biomass as estimated by the rising plate meter.
Therefore, inferences on pasture intake are subject to inaccuracies associated with this indirect
method to estimate herbage mass. As expected with the study design, pasture intake rates were
significantly higher in LSR (15.9 kg DM/cow/day) than HSR (9.8 kg DM/cow/day). Intakes for
corn and pellets were not significantly different. Interestingly, RPM estimates of pasture
consumption in LSR (15.9 kg DM/cow per d) and total forage consumption (14.66 kg
DM/cow/d; resulting from the combination of pasture intake (9.8 kg DM/cow/d) and average
PMR fed (4.86 kg DM day)) in the HSR treatment was very close to the 16.7 and 15.8 kg
DM/cow per d of total forage intake derived from weighted averages (average adjusted by the
proportion of United States Holstein and New Zealand Friesian cows in farmlets) of forage
consumption estimated with the use of Chromium Oxide in Chapter 2. This similarity between
methods suggests the possibility to obtain accurate estimations of pasture intake from the use of
a RPM. This information is highly valuable for commercial pasture-based systems, where the

117

lack of accurate methodologies to estimate pasture consumption often preclude proper feeding
management.

AMS use and efficiency. The following paragraph briefly describes findings on MY and
MF; a more detailed discussion for with mid-lactation cows can be found in Chapter 2. Table 4.2
shows the milk performance of the two herds managed with either LSR or HSR feeding systems.
As shown by a significant stocking rate treatment x week interaction, milk production increased
more steadily in HSR as a consequence of PMR feeding. The use of the AMS appeared to be
affected more by changes in environmental conditions and pasture availability than by qualitative
and quantitative differences between the stocking rate treatments. Both MF and refusals shifted
during the study as the motivation to graze abundant high quality pasture increased across the
study, particularly in LSR. Supplementation of PMR may have created a stronger motivation for
cows to visit the AMS at a regular basis regardless of the detected changes in pasture growth and
availability. Ketelaar-de Lauwere (2000) found similar results and concluded that at lower sward
heights cows paid more total visits (milkings plus refusals) to the AMS. As MF decreases, the
milking interval increased, and averaged 14.06 and 13.12 h for LSR and HSR. The average
interval length was quite high, but included the entire herd and did not exclude cows in late
lactation with lower milk production. These results also show that feeding a PMR in the barn did
not affect MY or MF compared to the pasture plus concentrate feeding system used in LSR.
These findings suggest that a pasture-based AMS could maintain acceptable MF and MP without
providing extra forage supplementation in the milking barn.

118

Because treatments in this study were not replicated in more than 1 AMS unit, only
means and standard deviations of AMS performance were collected and computed. The AMS
performance was numerically similar between the two stocking rate systems, reflecting the lack
of effects on milk performance variables recorded on individual cows. On average, AMS units
milked cows in approximately 50% of the time as a consequence of under stocking. The average
percent of time spent milking, often referred as occupation time, was 51.2 and 52.7% for LSR
and HSR, respectively. The target occupation of an AMS is recommended to be 85% (Andre et
al., 2010) because it is the maximum amount of time that the AMS has available for milking if a
reasonable amount of extra time is left for maintenance, repairs and cleaning, or to increase
milking opportunities for timid cows and/or naive first lactation cows in transition. However,
during this study, the AMS was underutilized due to the small herd size (47 cows/AMS).
The distribution of milkings shifted over the course of the day between the two stocking
rate treatments, but as discussed above, AMS occupation and total milk production were not
affected. Cows in HSR visited the AMS earlier in the morning compared to the cows in the LSR.
The increase in early milking frequency was likely explained by the daily feeding of PMR at
0500. Thus, feed manipulations in the milking barn could be used to entice higher frequency of
AMS visitations at particular times of the day. Splitting the feeding of a PMR in the milking barn
as opposed to feeding a PMR in the pasture to reduce cow traveling efforts may have the
potential to increase efficiency in the AMS occupation through improvements in the distribution
of milkings across the day.

119

Conclusion. As used in this study, dynamic supplementation with low levels of PMR in a
highly stocked pasture-based feeding system did not affect pasture growth rate, forage
utilization, MY, MF or overall occupation of an underutilized AMS stall (low number of
cows/AMS). Dynamic supplementation allowed for the maintenance of proper pasture residuals,
thereby improving condition for fast pasture regrowth. Supplementing with a PMR did not result
in a marginal response in milk production per cow or per AMS stall, except when pasture
availability was low. Feeding a PMR in the barn influenced cow traffic through the system and
the frequency of milkings at certain hours of the day, indicating that PMR feeding could be used
to improve the distribution of milkings in the AMS.

120

APPENDIX

121

TABLES

Table 4.1. Pasture, grazing management, feed intake, milk performance and feed conversion in a
low stocking rate and high stocking rate pasture-based feeding system.
Treatments
Pasture Management Variable
Pre-grazing herbage mass, kg DM/ha
Post-grazing herbage mass, kg DM/ha
Grazing intensity, %
Forage consumed, kg DM/cow/day
Forage utilization, Kg DM/ha/day
Total milk production*, kg/d
Total intake*, kg/d
FCE*, kg milk/kg feed
1

1

LSR
1930
1211
36
15.90
734
20.85
19.12
1.12

P<
2

HSR
1952
1058
43
9.8
871
22.41
19.50
1.12

SEM
125
66
1.7
1.1
70
1.11
0.82
0.06

3

SR
0.89
0.08
< 0.01
< 0.01
0.13
0.33
0.75
0.97

Low stocking rate treatment

2

High stocking rate treatment

3

Stocking rate effect
*Values from Chapter 2 (mid-lactation cows), intake determined by indirect marker
method with chromium oxide

122

Table 4.2 Cow performance in automatic milking systems managed with a low stocking rate or
high stocking rate pasture-based feeding system.
Treatments
Variable
Milk, kg/day
Weight Change, kg
Milking Frequency/d
Milking Interval, h
Refusals/d
Box Time, min
Milk Time, min
Prep Time, min
1

1

P Value
2

LSR
20.67
0.98
2.01
14.06
0.89
15.14
11.99
3.13

HSR
21.56
2.00
2.17
13.12
1.28
15.57
12.04
3.49

SEM
1.47
0.24
0.08
0.62
0.11
0.76
0.69
0.13

3

SR
0.67
<0.01
0.17
0.29
0.01
0.69
0.96
0.06

4

W
<.0001
<.0001
<.0001
<.0001
<.0001
0.0005
<.0001
<.0001

5

SR x W
<0.01
0.11
0.03
0.02
0.07
0.08
0.01
0.03

Low stocking rate treatment

2

High stocking rate treatment

3

Stocking rate effect

4

Week effect

5

Stocking rate by week interaction

Table 4.3 AMS performance in a low stocking rate and high stocking rate pasture-based feeding
system
Variable
Number of Cows
Total Milk, kg
Percent Milking Time
Percent Free Time
Total Milkings
Refusals
4

Fetch cows am
5

Fetch cows pm
1

1

2

3

LSR
46.27
1036.07
51.24
41.91
96.82
43.00

HSR
45.05
1084.57
52.73
40.41
103.48
61.93

SD
2.86
97.49
3.66
4.01
9.29
42.35

11.06

10.34

4.57

1.21

1.21

2.23

Low stocking rate treatment

2

High stocking rate treatment

3

Standard deviation

4

Cows with milking interval over 12 h at 0500 h

5

Cows with milking interval over 12 h at 1500 h
123

FIGURES

Pre- and Post-grazing Biomass of Pasture
3000

Biomass (kg/ha)

2500
2000

1500

Pre

1000

Post

500

0
1

2

3

4

5

6

7

8

Week
Figure 4.1 Pre-grazing (Pre) and post-grazing (Post) pasture biomass for high stocking rate and
low stocking rate pasture-based feeding systems. There was no difference in pre-grazing and
post-grazing biomass between stocking rate treatments, but pasture pre-grazing and post-grazing
biomass differed across weeks of the study (P < 0.0001). Bars denote the standard error of means
(SEM = 117.44 and 72.97 for pre-grazing and post-grazing biomass, respectively)

124

Pasture Growth Rate
70

GR kg DM/day

60
50

40
30
20
10
0
1

2

3

4

5

6

7

8

Week

Figure 4.2 Pasture growth rate (GR) means both for high stocking rate and low stocking rate
pasture-based feeding systems. There was no significant difference between stocking rate
treatments, but growth rate did differed across weeks of the study (P < 0.0001). Bars denote the
means and error bars are the standard errors of the means (SEM = 4.1)

125

Feed Consumption HSR
30
25

DMI kg/d

20
PMR
15

Pasture
Corn

10

Pellet
5
0
1

2

3

4

5

6

7

8

Week
Figure 4.3 Consumption of different feedstuffs in the high stocking rate treatment (HSR) in a
pasture-based feeding system. Bars denote the standard error of means.

Feed Consumption LSR
30

DMI kg/d

25
20
Pasture

15

Corn

10

Pellet
5
0
1

2

3

4

5

6

7

8

Week
Figure 4.4 Consumption of different feedstuffs in a low stocking rate treatment in a pasturebased feeding system. Bars denote the standard error of means.

126

Milk Yield
27

Milk Yield, kg

25
23
21

HSR

19

LSR

17
15
1

2

3

4

5

6

7

8

Week
Figure 4.5 Milk yield for high stocking rate (HSR) and low stocking rate (LSR) pasture-based
feeding systems. There was no significant difference in milk yield between stocking rate
treatments, but milk yield differed across weeks of the study (P < 0.0008). Bars denote the
standard error of means (SEM = 1.79)

Milkings Per Day

Milking Frequency
2.4
2.3
2.2
2.1
2
1.9
1.8
1.7
1.6
1.5

HSR
LSR

1

2

3

4

5

6

7

8

Week
Figure 4.6 Milking frequency for high stocking rate (HSR) and low stocking rate (LSR) pasturebased feeding systems. There was no significant difference in milking frequency between
stocking rate treatments, but milking frequency differed across weeks of the study (P < 0.03).
Bars denote the standard error of means (SEM = 0.1)
127

Refusals
3

Refusals per day

2.5
2
1.5

HSR

1

LSR

0.5
0
1

2

3

4

5

6

7

8

Week
Figure 4.7 Milking refusals for high stocking rate (HSR) and low stocking rate (LSR) pasturebased feeding systems. There was a significant difference in milking refusals between stocking
rate treatments (P = 0.01). Milking refusals also differed across weeks of the study (P < 0.0001).
Bars denote the standard error of means (SEM = 0.11).

Milking Distribution
8

Percent of Milkings

7
6
5
4

HSR

3

LSR

2
1
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour

Figure 4.8 Milking distribution for high stocking rate (HSR) and low stocking rate (LSR)
pasture-based feeding systems. There was a significant difference in milking distribution
between stocking rate treatments (P < 0.0001).

128

REFERENCES

129

REFERENCES

Andre, G., P. B. M. Berentsen, B. Engel, C. J. A. M. de Koning, and A. G. J. M. Oude Lansink.
2010. Increasing the revenues from automatic milking by using individual variation in
milking characteristics. J. Dairy Sci. 93:942–953.
Bach, A., and I. Busto. 2005. Effects on milk yield of milking interval regularity and teat cup
attachment failures with robotic milking systems. J. Dairy Res. 72:101–106.
Baker, A.-M.C., J. D. Leaver. 1986. Effect of stocking rate in early season on dairy cow
performance and sward characteristics. Grass and Forage Sci. 41:333–340.
Bargo, F., L. D. Muller, E. S. Kolver, and J. E. Delahoy. 2003. Invited Review: Production and
digestion of supplemented dairy cows on pasture. J. Dairy Sci.86:1-42.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002a.Milk response to concentrate
supplementation of high producing dairy cows grazing at two pasture allowances. J.
Dairy Sci. 85:1777–1792.
Bargo, F., L. D. Muller, J. E. Delahoy, and T. W. Cassidy. 2002b. Performance of high
producing dairy cows with three different feeding systems combining pasture and total
mixed rations. J. Dairy Sci. 85:2960–2975.
Fales, S. L., L. D. Muller, S. A. Ford, M. O’Sullivan, R. J. Hoover, L. A. Holden, L. E. Lanyon,
and D. R. Buckmaster. 1995. Stocking rate affects production and profitability in a
rotationally grazed pasture system. J. Prod. Agric. 8: 88–96.
Farina, S. R., S. C. Garcia, W. J. Fulkerson and I. M. Barchia. 2011. Pasture-based dairy farm
systems increasing milk production through stocking rate or milk yield per cow: pasture
and animal responses. Grass Forage Sci. 66:316-332.
Gronow, C. S., S. B. Little, and G. V. Cleary. 2010. Managing production system risk: An
application of Monte Carlo simulation to Tasmanian dairy farm data. Proceedings of the
4th Australasian Dairy Science Symposium.
Jago, J. G., K. L. Davis, P. J. Copeman, I. Ohnstad, and M. M. Woolford. 2007. Supplementary
feeding at milking and minimum milking interval effects on cow traffic and milking
performance in a pasture-based automatic milking system. J. Dairy Res. 74:492–499.
Ketelaar-de Lauwere, C. C., A. H. Ipema, C. Lokhorst, J. H. M. Metz, J. P. T. M. Noordhuizen,
W. G. P. Schouten, and A. C. Smits. 2000. Effect of sward height and distance between
pasture and barn on cows’ visits to an automatic milking system and other behavior.
Livest. Prod. Sci. 65:131–142.

130

Little, S. 2011. Performance, profit and risk in pasture-based dairy feeding systems - Findings
from the TasMilk60 study. Dairy Australia's Grains2Milk program, Dairy Australia Ltd.,
Victoria, Australia. p 64.
Little, R. C., Milliken, G. A., Stroup, W. W., and R. D. Wolfinger. 1996. SAS System for Mixed
Models. SAS Institute Inc.: New York.
MacDonald, K. A., J. W. Penno, J. A. S. Lancaster, and J. R. Roche. 2008. Effect of stocking
rate on pasture production, milk production, and reproduction of dairy cows in pasturebased systems. J. Dairy Sci. 91: 2151–2163.
Parsons, A.J., and Chapman, D.F. 1998. Principles of grass growth and pasture utilization. In
Cherney, J.H., and Cherney D.J.R (Eds.), Grass for dairy cows. CABI publishing,
Wallingford, UK, pp. 283-310.
Reis, R. B., and D. K. Combs. 2000. Effects of increasing levels of grain supplementation on
rumen environment and lactation performance of dairy cows grazing grass-legume
pasture. J. Dairy Sci. 83:2888–2898.
S.A.S. Institute, 2008. SAS User's Guide: Statistics. SAS Institute Inc., Cary, NC.
Stockdale, C. R. and K. R. King. 1980. The effects of stocking rate and nitrogen fertilizer on the
productivity of irrigated perennial pasture grazed by dairy cows. Aust. J. Exp. Agric.
Anim. Husb. 20:529-536.

131

CHAPTER 5
GENERAL CONCLUSIONS AND FUTURE RESEARCH

There is growing interest in integrating automatic milking systems (AMS) and pasturebased dairying. However, guidelines outlining the proper management of these systems are
scarce. Our objective was to help fill the information gaps regarding the operation and efficiency
of a pasture-based dairy with AMS using free traffic flow and voluntary grazing with contrasting
stocking rates, supplementation strategies and cow genotypes.
All chapters of this thesis focused on two typical pasture-based feeding systems of the
Midwestern United States. This thesis compared a low stocking rate (LSR, 1.92 cows/ha) system
and a high stocking rate (HSR, 2.89 cow/ha) system fed partial mixed ration (PMR). This thesis
also compared the performance of two genotypes of dairy cows (New Zealand Friesian, NZF;
United States Holstein, USH) often used in Midwestern pasture-based dairy systems.
Our study showed that the increase in stocking rate from 1.92 to 2.89 cow/ha with
dynamic supplementation was a potential strategy to increase milk production per area and use of
the AMS. Dynamic feeding of PMR in the HSR system was used to balance the supply-demand
of forages based on expected changes in pasture growth rate (PGR). Supplementing only the
deficit in PGR helped maintain desirable grazing residuals without sacrificing pasture utilization.
This study showed that managing pasture residuals with dynamic PMR supplementation did not
affect PGR, which makes it a desirable option for dairies that want to increase stocking rate and
milk production per cow and/or per area without sacrificing forage utilization. This feeding
practice could be relevant for the proper integration of AMS in small pasture-based dairy farms
and may introduce greater flexibility to manage AMS occupancy and stocking rate (cows/AMS

132

stall) without sacrificing forage utilization. Future research could focus on further increases in
stocking rate and supplementation in order to increase milk production per area, while also
increasing milk production per cow.
Chapter 2 showed that a rising plate meter (RPM) gave an accurate measurement of
pasture intake. The RPM was used to record pre- and post- grazing biomasses, which in turn
were used to estimate pasture intake per cow. An estimate of pasture dry matter intake (DMI)
was determined by an external marker technique using chromium oxide. When the estimated
forage consumption by the external marker chromium oxide was compared to the RPM, the
values were very close (HSR: 10.18 ± 0.61 and 9.8 ± 1.10 kg DM; LSR: 14.8 ± 0.66 and 15.9 ±
1.10 kg DM; for chromium oxide pasture intake estimates and RPM pasture intake estimates,
respectively). This indicates that the RPM is a valuable tool for determining pasture DMI, which
is affordable and could be used in any grazing dairy.
Chapter 2 also highlighted the importance of dairy cow genotype in pasture-based
systems. Our study showed that USH had a greater genetic potential for milk production (MP;
27.9 ± 0.8 and 15.3 ± 1.3 kg for USH and NZF, respectively) accompanied by a greater level of
feed intake (22.3 ± 0.8 and 16.3 ± 0.8 kg DM for USH and NZF, respectively). The USH
genotype outperformed NZF cows in both HSR and LSR treatments, however a genotype by
time interaction suggested important differences between genotypes as pasture availability
changed in the study. At study onset, pasture biomass was low and NZF were nearly at their
maximum milk production. These results indicated that when pasture biomass per cow is limited,
either by environmental conditions, as in this case, or by SR in a very high SR system, NZF may
be the preferred genotype. Conversely, when pasture availability per cow was high, USH had
higher milk production, regardless of SR. This indicates both that USH performs better when

133

pasture availability is high and that HSR did not restrict intake. Further research could continue
to examine practical genotype by environment interactions that aid farmers in genotype
selection, based on the production system they choose.
It was observed in other studies that supplementation creates a higher marginal response
in cows of higher genetic merit. In this study, a PMR supplement with a similar quality to
pasture was used. The purpose of the PMR was to adjust for the deficit in PGR and therefore did
not result in significantly greater milk production. However, if a supplement higher in energy
was used, a significant marginal response in milk production may have been detected. Increasing
supplementation in the form of high energy concentrates may be a good option for this system
because the AMS and Cosmix feeders allow for supplementation of individual cows, perhaps
determined by genotype, stage of lactation, or parity. Using this technology for precision feeding
(i.e., feeding based on milk production level, genotype, or stage of lactation) may help increase
milk production per cow and, in turn, feed conversion efficiency (FCE).
Feed conversion efficiency in pasture-based systems may be improved by refining
pasture management practices. Supplementation and stocking rate can affect FCE due to low
pasture availability per cow in high stocking rate systems. However, our study showed that with
dynamic PMR supplementation, there is no significant difference in intake or milk yield,
preventing a decrease in FCE across stocking rates (1.12 ± 0.06 for both LSR and HSR). Our
study also showed that FCE was affected by genotype (1.00 ± 0.06 and 1.24 ± 0.06, for NZF and
USH, respectively). Assuming feed intake is not severely limited cows of greater genetic merit
should achieve higher FCE, since they have a higher marginal response to concentrates and
greater genetic potential for high milk production. Further research is required to better

134

understand the relationships between stocking rates, supplementation, and FCE in order to
increase the efficiency of pasture based systems.
Although supplementation may have potential to improve FCE in a pasture-based feeding
system, it may also affect the rumen environment and forage degradation in the rumen. The
study reported in Chapter 3 showed that low levels of PMR (40% ground corn, 60% haylage)
supplementation did not depress rumen pH or reduce degradation rates of pasture DM, CP, or
NDF. In this study there were potential indications that the PMR helped to slow the rate of
passage. This may have been due to greater levels of indigestible NDF in the haylage, slowing
passage rate and increasing the extent degradation and possibly creating a more favorable
environment for rumen microbes increasing the rate of degradation. Validation of these
indications should be researched in the future in order to continue to increase the utilization of
nutrients in a pasture based diet.
Although there may be some benefits to increasing supplementation, the effects of
supplementation on barn traffic and AMS must also be considered. Chapters 2 and 4 both
included important information on AMS utilization. Results in both chapters demonstrated that
dynamic supplementation did not increase milking frequency (MF). This indicates that additional
supplementation (other than in the AMS) was not required to achieve moderate frequency of
visitations. However, results in Chapter 2 did show, that as the grazing season extended into the
fall mid-lactation cows were less willing to visit the AMS, in which, feeding a PMR may be an
extra motivation to return to the barn. Understanding how feeding affects traffic flow to the AMS
is vital to running an efficient system and further research could aid in determining the effects of
other these feeding strategies on AMS occupation.

135

Although the PMR in the HSR did not significantly increase milking frequency (2.3 ±
0.07 and 2.5 ± 0.07 for LSR and HSR, respectively), there was a shift in distributions of milking
across the day causing HSR to increasing their number of milking visitations around the time of
PMR feeding (0500 h). The effect of the PMR on milking distribution is important because it
may be a possible tool used to increase milkings at times of the day when visitations are low. It
also is interesting to note the differences in visitations between genotypes, as seen in Chapter 2.
It is possible that a combination of genotypes may utilize the AMS better than one genotype; not
just considering the distribution of milkings but the differences in milking interval.
The NZF group had longer milking intervals which may allow for more flexibility in the
time of milkings for those cows, allowing USH cows with shorter milking intervals a priority for
milking. This combination of cows with different milk production and expected milking intervals
may allow for more efficient utilization of the AMS and also increase milk production for cows
of greater genetic merit, without affecting milk production in lower genetic merit cows. More
research is required to understand the effects of milking interval on milk production and how to
manipulate milking interval in this system to potentially increase milk production and system
efficiency.
Overall, the results of this research show that dynamic PMR supplementation maintains
milk production in both NZF and USH and helps optimize pasture utilization. The PMR used in
dynamic supplementation is a suitable substitute for pasture and did not cause negative
associative effects at the rate provided, but also did not significantly increase milk production. In
conclusion, dynamic PMR supplementation is a viable tool for pasture-based AMS dairy farms
because it does not affect the rumen environment, and may improve the efficiency of AMS

136

utilization by offering opportunities to improve AMS stocking rate (cow/AMS) and occupancy,
without sacrificing pasture utilization.

137