w- llllilllllllllllllllllllllllllllllll

3 1293 01055 3331

This is to certify that the

thesis entitled

GENETIC PARAMETERS FOR PARTITIONED USES
OF ENERGY INTAKE ESTIMATED FROM FIELD
COLLECTED AND CALORIMETRIC DATA ON
THE SAME LACTATING HOLSTEIN CONS

presented by

PETER MALACHI SAAMA

has been accepted towards fulﬁllment
of the requirements for

 

 

M.S. degree in Animal Science

%>

1’

Major professor

 

Dateﬂé 1/ ”22

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

 

 

L: In a: my
‘1. a :1o‘-;:
:7 g
“t 1“- ‘1‘ H ’3’: f
Midis-vii ~ g» a a of“:
P? TM? M 1".l
263“ J" r ‘1"
__ HM

 

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

TOAVODFINESMJmonorbdonddodm.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Namath/o Action/Equal Opponmuy lndltution

owns-9.1

GMC PARAMETERS FOR PARTITIONED USES OF ENERGY INTAKE
FSTINIATED FROM FIELD COLLECTED AND CALORIMETRIC DATA ON
THE SAME LACTATING HOLSTEIN COWS

BY

PETERMALACHI SAAMA

ATHFSIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF SCIENCE

Department of Animal Science

1992

ABSTRACT
GENETIC PARAMETERS FOR PARTITIONED USES OF ENERGY INTAKE
ESTIMATED FROM FIELD COLLECTED AND CALORIMETRIC DATA ON
THE SAME LACTATING HOLSTEIN COWS
By

Peter Malachi Saama

Energy balance trials were conducted on 34 multiparous Holstein cows at the
University of New Hampshire, Durham, at wk 6, 10, and 14 postpartum during
19804985. Diet, parity, and season effects were found to be non-signiﬁcant sources
of variation in gross energy consumed, fecal, urinary, methane, heat, milk, and
maintenance energy, and tissue energy balance and their ratios. Milk energy as a
covariate was highly signiﬁcant in these variables except energy for heat production
and that for maintenance. The effect of maintenance interacted with periods. Field
estimates on gross efﬁciency were obtained from intake and production data recorded
during wk 5, 7, 9, 11, 13, and 15 postpartum. All energy and gross efﬁciency
estimates from ﬁeld data closely approximated measures of the same traits from
energy chamber data. This approximation was better at postpeak lactation. An
additional data set of 37 cows was procured from energy balance trials conducted
during 1987-1989. Animal models were used to estimate partial energy requirements
and genetic parameters for energy usage traits. Estimation and prediction was by a
derivative-free REML algorithm. Omitting animal effects did not affect solutions for
covariates. Genetic and phenotypic variations and heritability estimates in energy

intake variables, at postpeak lactation, were similar for chamber and ﬁeld data.

DEDICATION

To Beatrice Mary Namuyomba and Cyprian Kikunyi Bamwoze

ACKNOWLEDGIVIENTS

The author is grateful to Dr I.L. Mao for his friendship, guidance and counsel.
Dr. T.A. Ferris, Dr. R.S. Emery, and Dr. J.L. Gill are cited for their instructive
insights during the course of this work. The encouragement provided by Dr. M. C.
Dong, Dr. J. Jensen, Dr. G. Jeon, Dr. L. Gustavo, Ms. T. Moore, Ms. F.
Ngwerume, Dr. D. Krogmeier, Ms. G. Ferreira, and Mr. F. Grignola is greatly
appreciated. Extreme gratitude is extended to Dr. J.B. Holter for his critical review
of the manuscript and to F. J. Janicki and H. H. Hayes for their contributions to

collection of energy data.

TABLE OF CONTENTS

List of Tables ....................................... vii
List of Figures ...................................... ix
1 Introduction ...................................... 1
2 Objectives ....................................... 4
3 Review of literature .................................. 5
3.1 Dietary sources of energy and ﬁber. ...................... 5
3.2 Protein ...................................... 5
3.2.1 Carbohydrate ................................. 6
3.2.2 Lipid ...................................... 7
3.3 Energy metabolism ................................. 8
3.3.1 Energy partitioning by indirect-calorimetry ............... 10
3.3.2 Energy use for maintenance ........................ 12
3.3.2.1 Sources of variation in maintenance energy ............ 13
3.3.2.2 Efﬁciency of energy utilization for maintenance ......... 13
3.3.2.3 Estimation of maintenance energy .................. 14
3.3.2.4 Genetic aspects of maintenance energy ............... 16
3.3.3 Energy use for milk production ...................... 16
3.3.3.1 Sources of variation in milk energy ................. 17
3.3.3.2 Efﬁciency of energy utilization for milk .............. 17
3.3.3.3 Estimation of milk energy ....................... 18
3.3.3.4 Genetic aspects of milk energy .................... 19
3.3.4 FJIergy use for live weight gain or loss ................. 20
3.4 Canonical correlation analysis. .......................... 20
3.5 (Co)variance component estimation. ...................... 21
3.5.1 ML estimates of (co)variance components ................ 22
3.5.2 REML estimates of (co)variance components .............. 23
3.5.2.1 Derivative-free type REML algorithms ............... 24
3.5.2.2 Limitations of DF-type algorithms .................. 25
4 Sources of variation in partitioned uses of energy intake in pluriparous
lactating Holstein cows in energy chamber .................... 26
4.1 Abstract ....................................... 26

4.2 Introduction ..................................... 26

4.3 Materials and methods ............................... 28
4.3.1 Experimental procedure and data ..................... 28
4.3.2 Model and analysis algorithm ....................... 29

4.4 Results and discussion ............................... 30

4.5 Conclusions ..................................... 37

5 Energy intake and gross efﬁciency comparisons from calorimetric and ﬁeld
data on the Same lactating Cows ......................... 39

5.1 Abstract ....................................... 39

5.2 Introduction ..................................... 40

5.3 Materials and methods ............................... 41
5.3.1 Experimental design and data ....................... 41
5.3.2 Analysis procedures ............................. 42

5.4 Results and discussion ............................... 43

5.5 Conclusions ..................................... 52

6 Comparisons of genetic parameters for energy intake estimated from energy
chamber and from ﬁeld collected data on the same lactating cows ..... 53

6.1 Abstract ....................................... 53

6.2 Introduction ..................................... 54

6.3 Materials and methods ............................... 56
6.3.1 Experimental procedure and data ..................... 56
6.3.2 Estimation of genetic parameters ..................... 57
6.3.3 Estimation of partial energetic efﬁciency ................. 59

6.4 Results and discussion ............................... 60
6.4.1 Genetic and phenotypic variation ..................... 60
6.4.2 Genetic and phenotypic correlations .................... 61
6.4.3 Partial energy requirements ......................... 63

6.5 Conclusions ..................................... 66

7 Summary ........................................ 68
8 Appendices ....................................... 70
9 References ....................................... 96

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

LIST OF TABLES

The effect of body weight on heat production ............. 15

Within period mean, SE, SD, and CV for energy measures
(n = 34) by selected weeks postpartum (pp) .............. 32

Overall and within period mean, SE, SD, and CV for energy
measures (11 = 34) by selected weeks postpartum (pp) ........ 34

Critical levels (P - value) for main effects and partial regression
coefﬁcients for covariates from within period model: energy
partition measures .............................. 35

Critical levels (P - value) for main effects and partial regression
coefﬁcients for covariates from within period model: energy
efﬁciency measures ............................. 36

Overall mean, SE, SD and CV for energy measures from energy
chamber and from ﬁeld data on the same 30 pluriparous Holstein
cows in early lactation ........................... 44

Mean, SE, SD and CV of energy measures from energy chamber
and estimates from ﬁeld data on the same 30 pluriparous Holstein
cows in early lactation ........................... 44

Overall mean, SE, SD and CV for efﬁciency measures from energy
chamber and estimates from ﬁeld data on the same 30 pluriparous
Holstein cows in early lactation ...................... 46

Mean, SE, SD and CV for efﬁciency measures from energy
chamber and estimates from ﬁeld data on the same 30 pluriparous
Holstein cows in early lactation ...................... 47

Product-moment correlations between ﬁeld and energy chamber
measures of energy intake from data on the same 30 pluriparous
Holstein cows in early lactation; values in parentheses are rank
correlations .................................. 48

Product-moment correlations between ﬁeld and energy chamber
measures of energy efﬁciency from data on the same 30 pluriparous
Holstein cows in early lactation; values in parentheses are rank
correlations .................................. 49

vii

Table 12.

Table 13.

Table 14.

Table 15.

Table 16.

Table 17.

Canonical roots and correlations (R2,) between ﬁeld and calorimetric
energy intake measures on the same 30 lactating cows at wk 6, 10

and 14 postpartum .............................. 51

Canonical roots and correlations (R2,) between ﬁeld and calorimetric
gross efﬁciency measures on the same 30 lactating cows at wk 6, 10,

and 14 postpartum ............................. 51

Additive genetic and phenotypic SD and heritability for energy
measures traits at peak lactation ...................... 61

Estimates of phenotypic correlations (above diagonal) and genetic
correlations (below diagonal) between energy utilization traits . . . 62

Partial regression coefﬁcients (Coeff.) for covariates in animal models

at peak and postpeak lactation ....................... 65
Partial regression coefﬁcients (Coeff.) for covariates in multiple
regression models at peak and poku lactation ........... 65

viii

Figure 1.
Figure 2.
Figure 3.

Figure 4.

LIST OF FIGURES

Framework of energy metabolism in the lactating dairy cow ...... 9

General principle of open circuit apparutus ................. 11
Overall mean percentages for partitioning of energy intake ....... 31
Lactation curve for 30 pluriparous Holstein cows ............. 45

1 INTRODUCTION

The main purpose of raising livestock is to convert feeds into food products
desirable to humans. The animal industry plays a major role in maximizing food for
human consumption from the available feedstuffs and other inputs. The nutrients or
resources that are needed for an animal to achieve its potential have been described.
Emmans and Oldham (1988) suggest that a number of functions will become
important when resources and nutrients are limiting. They are: l) maintenance of
essential metabolic processes and tissue integrity; 2) maintenance of established
pregnancy; 3) maximal rate of secretion of milk constituents as determined by
genotype, stage of lactation, and perhaps age and parity; 4) achievement of an upper
limit to body protein mass at a particular stage of maturity; 5) achievement of a
desired fat mass in relation to protein mass or stage of maturity.

High peak milk yields and an extension of the high yield phase of lactation are
occurring in dairy cattle due to continued improvements in genetics, adoption of new
management and feed systems, and biotechnological advances. Nutrient requirements
of the cow increase with milk yield (NRC, 1989).

During early lactation, cows are often in negative energy and N balance because
maximal DMI does not occur until peak milk production. Therefore, cows mobilize
energy and protein from body tissue to support milk production. Energy and amino
acids are the two nutritional factors most likely to limit milk production. The

relationship between N and energy requirements for cows is complex because there

2

are two requirments to be met: one for the host animal and another for the ruminal
microbes.

The ultimate goal of any animal related agricultural enterprise is to provide the
most acceptable, properly balanced, and least expensive ration that can be fed to a
given species of animals. This is so because the provision of feed to animals
constitues a major component of farm expenditure. Since all of the organic nutrients
contained in a feed can serve as a source of energy, a measure of energy value of a
feed is desirable.

Several attempts at deﬁning an over-all energy unit of measurement which is
easily determined, accurate and readily reproducible have been made and continue to
be made by scientists in the US and elsewhere. Four different methods are used in the
US to determined the useful energy of livestock feeds. They are the determination of
total digestible nutients, digestible energy, metabolizable energy, and net energy. Net
energy is the most scientiﬁcally acceptable method of expressing the energy value of a
feed (Moe and Tyrrell, 1973). Whence, the energy value of a feed is deﬁned as the
increase in energy retention which occurs per unit increment of a feed given.

In this study, ”efﬁciency“ is deﬁned as the ratio of energy in the product to the
product of energy partitioning from which it was formed. Efﬁciency of energy use
and energy requirements have been identiﬁed more precisely. Progressively more
intensive experimentation has described physiological and biochemical bases for an
ever increasing body of knowledge concerning variation in energy use. In most cases,

the laboratory techniques are too expensive or impractical to be used on a routine

basis.

In the long run, methods to predict quantities of nutrients absorbed from the gut
will permit a more ﬂexible and accurate method of evaluating diets, predicting animal
performance and estimating the genetic potential for animal feed and energetic
efﬁciency.

In general, there are two ways in which the net energy value of feed may be
estimated: a) by regression of total energy balance on DMI, and b) by assuming a
maintenance requirement in terms of either energy intake or energy balance and either
subtracting the metabolizable energy requirement for maintenance from the actual
metabolizable energy intake or adding the net energy required for maintenance to the
actual energy balance (Moe and Tyrrell, 1972). Whenever either of these approaches
is used, it is important to ask how accurate the approximations are and whether this
accuracy holds in other production conditions. An initial attempt to address these
issues was made by Walter and Mao ( 1989) when they compared estimates of energy
requirements from barn data with reported values. Consequently, the use of ﬁeld
collected data to predict enery requirements of dairy cattle was considered a plausible
source of data for generating databases for quantititative analyses. However
veriﬁeation of this assertion has not been made.

This study represents an unprecedented effort at making comparisons of various
measures energy utilization and energetic efﬁciency as well as genetic parameter
estimates for energy intake from ﬁeld collected and energy chamber data on the same

COWS.

2 OBJECTIVES

1) To examine sources of variation in partitioned uses of energy intake using
energy chamber data.

2) To compare energy intake and gross efﬁciency measures from calorimetric
and ﬁeld data on the same pluriparous lactating cows.

3) To compare genetic parameters for energy intake traits estimated from
respiration chamber data and ﬁeld data on the same primiparous lactating

COWS.

3 REVIEW OF LITERATURE

3.1 Dietary sources of energy and fiber
Adequate dietary energy and ﬁber are essential for high producing dairy cows
during early lactation. The three main dietary sources of energy and ﬁber are protein,

carbohydrate, and lipid.

3.2 Protein

Adequate intake of protein is nwded to provide the proper amount of total protein
to the small intestine for digestion and absorption. Due to limited body reserves
(Andrew, 1990), protein deﬁcient diets quickly inﬂuence the nutritional and
productive status of cows in early lactation (Wohlt, 1978' ; Wohlt, 1978”). The
quantity and quality of amino acids reaching the small intestine are the result of
microbial synthesis in the rumen and the extent to which feed proteins escape ruminal
degradation. However, the amount of protein supplied by microbial synthesis in the
rumen is not adequate to meet the needs of high producing cows (NRC, 1989). In
lieu of this, dietary undegradable protein (UIP) is often required (Satter, 1986;
Robinson, 1991).

Milk yield responses to increasing the UIP content of rations have been

observed (Cardoniga, 1988; Rogers, 1987); however, the amino acid proﬁle of by-
pass protein must be of high quality, supplying required nutrients (Satter, 1986;

Susmel, 1989). Effects of supplemental fat and UIP on milk yield were additive

6
(Ferguson, 1988; Wolht, 1991; Wilks, 1991). Fish meal has been reported to be less

degradable in the rumen than soybean meal (Sniffen, 1987; Atwal, 1992).

3.2.1 Carbohydrate
Carbohydrates are the major components in plant tissues and they comprise
up to 50% of the dry matter in forages, although higher concentrations (up to 80%)
may be found in some seeds, especially cereal grains (Church, 1982). The primary
function of carbohydrates in animal nutrition is to serve as a source of energy for
normal life processes. Type of dietary carbohydrate and level of carbohydrate intake
are factors which often determine level of performance of lactating dairy cattle.
Forages that are high in digestibility and that can be consumed in large amounts
are an essential diet component for high producing dairy cattle. Alfalfa is a widely
used source of energy, ﬁber, and protein for dairy cows. The high solubility and
degradability of alfalfa protein, however, may result in N wastage in the rumen.
Additional sources of feed protein, a portion of which will pass out of the rumen
undegraded, may be necessary to supplement the protein in alfalfa forage. Increasing
dietary crude protein (CP) levels from 13.8% to 17.5% by the use of cottonseed meal
(CSM) was beneﬁcial to cows consuming alfalfa-based diets in early lactation (Grings,
1991).
Because of its oil content, whole cotton seed (W CS) his considered a high energy
ingredient. Feeding WCS supplements increased yields (Anderson, 1979) and resulted

in higher milk fat percentage (DePeters, 1985) but depressed milk protein percentage

(Smith, 1981; DePeters, 1985).

Increased grain in the diet has been shown to be responsible for an increase in
milk production (Hoffman, 1991; Kesler, 1962), higher protein percentage (Y ousef,
1970). Nonetheless, some studies have shown a decrease in milk fat percentage as
grain increases (Donker, 1982; Macleod, 1983). Increasing energy in diets using
cereal grain supplements necessitates greater reductions in forage levels than if
supplemental fat is fed. However, production response to added fat primarily depends
on the nature of the diet, form of added fat, and availability of the fat to the rumen

microbes and to the animal postruminally (Chalupa, 1986; Jenkins, 1982).

3.2.2 Lipid

Energy requirements at peak lactation exceed the energy intake thus creating a
deﬁcit. Consequently feeding supplemental fat is utilized as a means of increasing the
ration energy density. Feeding supplemental fat increased milk yield (Hoffman,
1991; Palmquist, 1978). Cows fed supplemental fat also had higher BW, and weight
gain was signiﬁcant with time (Hoffman, 1991). Dietary fat supplements increase the
energy density of the diet as well as total tract apparent digestion of N (Ohaj uruka,
1991) but dietary fats can have a negative impact on milk protein (DePeters, 1987),
rumen fermentation and ﬁber digestibility (Palmquist, 1978; Palmquist, 1980). Thus,
fat supplements must be relatively inert in the rumen to reduce these detrimental
effects (Ferretti, 1990).

Calcium salts of long-chain fatty acids (Ca-LCFA) of palm-oil are chemically

8
bound dietary fats that do not adversely affect rumen fermentation (Chalupa, 1986) or

ﬁber digestibility (Schauff, 1989) in lactating cows. Strategic feeding of regimens
including use of Ca-LCFA (Kent, 1988; Schneider, 1988) have been used as a method
to alleviate a portion of the dietary energy deﬁcit experienced by early postpartum
dairy cows (Bauman, 1980; Coppock, 1985). The net energy of Ca-LCFA from palm-

oil has been determined in mature holstein cows (Andrew, 1990) .

3.3 Energy metabolism

Dietary energy can be partitioned several ways. The ﬂow of energy in the
lactating cow, as described by NRC (1989) is shown in Figure 1. Intake of dietary
energy is the gross energy (GE) of the food consumed. A substantial portion of GE is
lost from the animal as fecal energy (FE) and the difference (GE-FE) is termed the
apparent digestible energy (DE). Portions of the DE are voided as urinary energy
(UE) and gaseous energy in the form of methane (CH). The remainder of GE-FE-
UE is metabolizable energy (ME). An increase in heat production (HP) following
consumption of food is termed heat increment (HI) and includes heat of fermentation,
heat of product formation, thermal regulation, waste formation and excretion,
voluntary activity, and basal metabolism. The difference (ME-HI) is net energy (NE).
The NE can be recovered as a useful product such as maintenance (MNT), milk
energy (MKE), body gain or loss, and conceptus energy (CE). What is left over is
termed tissue energy balance (EB). This bioenergetics framework can be expanded to

include many of the intermediate steps of metabolism involved and each component

9
can be divided into component parts. For example the ability of the food consumed

to meet the NE requirement for maintenance is expressed as NE... It is important to
recognize that dietary energy is not used with equal efﬁciency for all physiological
functions. Approximate ranges for the efﬁciency use of ME are described by Moe et

a1. (1973).

 

 

 

GROSS ENERGY

 

 

 

- Fecal Bury

 

1

DIGESTIBLE
ENERGY

 

 

 

 

 

Urinary Energy
' Methane energy

 

7
METABOLIZABLE
ENERGY

-[ ...2::. ]

NET ENERGY

Body gala ENERGY
lam-«J -l 2....

Figure l . Framework of energy metabolism in the lactating dairy cow.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10
3.3.1 Energy partitioning by indirect-calorimetry

Many techniques used to study energy metabolism have been discussed in detail
recently by Blaxter (1989). In general, measurement of the overall energy
transformations in an animal in terms of free energy is not possible for technical
reasons. Therefore, measurements of energy exchanges in dairy cattle are made
simply in terms of the changes in heat on complete oxidation. The heat produced in
oxidation of food is measured by techniques of calorimetry. The heat of combution of
food, whether in vitro or in vivo, is carried out by the technique of calorimetry. For
instance, the heat of combustion of food is carried out by adiabatic bomb calorimetry.

The method of indirect calorimetry (Bursztein, 1989) provides a unique process
by which the type and rate of subtrate oxidation and heat production are measured in
vivo starting from gaseous exchange measurements. The use of gas exchange for
indirect calorimetry is based on assumptions that go back to the investigations of
Lavoisier in the late 18th century (Holmes, 1985). The standard gas equations are
reviewed by Bursztein (1989). These gas equations treat 02 and CD; as ideal gases.
Johnson (1980) argues that this is incorrect for O2 and N2, but is only partially true
for C02, thus introducing a small error. A more important potential error is related to
water vapor, where the expired air is assumed to be saturated and complete drying is
required for use of the gas equations, although neither of these conditions may be
totally correct. Respiration apparatus for indirect calorimetry are of two main types:
open circuit and closed circuit.

In the open circuit respiration apparatus, outdoor air is passed through the

ll

chamber of the instrument and the changes in its oxygen, methane and carbon dioxide
content are measured. The total amounts of carbon dioxide and methane produced and
of oxygen consumed can be determined if the amount of air which passes through the
apparatus and the incremental changes in gas concentrations are known (or ﬁxed).

The general principle of an open circuit apparatus is illustrated in Figure 2.

 

Outdoor

air ﬂows

  
    
 
  

through the

ANIMAL CHAMBER

—0 All AIR 9—-

chamber, a

sample of it

SPIROIIETER
A

being taken

 

 

 

continuously
Figure 2. General principle of open circuit apparutus (Blaxter,

and stored in 1989).

the spirometer A. The ﬂow through the chamber is measured and sample is taken and

stored in spirometer B. A further sample is deﬂected through an absorption system to

determine the proportion of C02 and CIL. Analyses of the air samples in the gas

meter provide a measure of the O2 consumption. Formulae for performing routine

calculations of respiration trial data in dairy cattle have been presented (Flatt, 1961).
In the closed circuit system (Blaxter, 1989), air is circulated continuously through

absorbents which remove carbon dioxide and water vapor; the air frwd of these gases

returns to the chamber. A fall in the pressure in the whole apparatus occurs as a

12
result of the absorption of oxygen by the animal, and oxygen admitted to the system

in proportion to this fall in pressure. By weighing the absorbents, the amount of
carbon dioxide produced can be measured directly, and the amount of oxygen can be
measured either by weight or volume. Technical problems with closed circuit systems,
as discussed by Wainman and Blaxter (1958), present some practical difﬁculties.
Implementation of indirect calorimetry apparutus by Armsby (1904), Ritzman and
Benedict (1929), Ritzman and Colovos (1932), Mitchell et a1. (1932), Kleiber (1936),
USDA-ARS, as described by Flatt et al. ( 1958), and others has provided a more
complete and scientiﬁcally sound basis of feed evaluation as well as a more thorough
knowledge of the energy requirements of dairy cattle and a basic understanding of

factors affecting the energy metabolism of dairy cattle.

3.3.2 Energy use for maintenance
The cow has certain obligatory needs for nutrients, which by deﬁnition must be

met to maintain life and functional processes. The partitioning of nutrients to various
body tissues involves two types of regulation, homeostatis and homeorhesis (Bauman
and Currie 1980). Homeostatic control involves maintenance of physiological
equillibrium or constant conditions in the internal environment. This includes
regulation to maintain constancy of body temperature (Kennedy, 1967) and the intake
of food and partitioning of nutrients in the absorptive and postabsorptive periods
(Tepperman and Tepperman, 1970). The co-ordination of metabolism in various

tissues to support a physiological state, such as pregnancy, is under homeorhetic

13
control (Kennedy, 1967). The informative study by Mertz and Van den Bergh (1977)

illustrates the relationship between homeostasis and homeorhesis.

3.3.2.1 Sources of variation in maintenance energy

Ritzman and Benedict (1938) found that the basal metabolism of their cows was
rather variable. Brouwer, et a1. (1961) found evidence for variation in the
maintenance requirement per 500 kg of BW of a cow. Signiﬁcant variation exists in
the maintenance requirement of animals when comparisons are made across a range of
species and ages (Reid, 1974; Reid et al., 1980). This includes differences due to type
of diet and physiological state, which are known to affect maintainenance
requirements (Garrett and Johnson, 1983). Signiﬁcant effects of breed, breed size,

age, and feeding level on maintenance were observed by Taylor et a1. (1986).

3.3.2.2 Efficiency of energy utilization for maintenance

The efﬁciency of ME use for maintenance (k_) and for gain (19) are related to the
source of ME. Studies in which steam volatile fatty acids, glucose and protein were
given as the sole energy source, suggest that k. is in the range of 80-85%, is constant
for widely different foods and is predicatable from physiological experiments in which
the end products of fermentation have been given as the sole source of energy
(Blaxter, 1961). These results supported the conclusion of Ritzman and Benedict
(1938) from earlier calorimetric studies. However, Blaxter and Wainman (1964)

observed that k. was not constant, but appeared to increase with the feed quality.

14
Van Es and Nijkamp (1969) reported similar results from 41 balance trials with

lactating cows consuming mixed diets of concentrate, silage, and variable amounts of
hay. More recently, Blaxter and Boyne (1978) reported that k_ is affected by types of

feed.

3.3.2.3 Estimation of maintenance energy

In animal calorimetry, heat production attributable to maintenance metabolism
can be distinguished from HI by measuring the fasting heat production (FHP). The
total heat produced less FHP would be considered as HI (Holter, 1974).

There are several ways of estimating k... Brody’s ( 1945) scaling of energy
maintenance to BW'73 subsequently rounded by Kleiber (1965) to a scaling of BW.",
as useful estimate of FHP has gained widespread use. According to NRC (1989), k.,
was calculated as .086 Meal/kg” of BW. Tablel, from the results of Van Karnpen
(1987), shows that in animals with the same metabolic level or with equal amounts of
heat produced per kg BW'”, there is a positive relationship between BW and HP
expressed per animal. However, expressing HP/ kg BW results in a negative

relationship.

15

 

TABLE 1. The effect of body weight on heat production.
Body Weight HP/animal HP/kg BW I-IP/m2 HP/kg BW'”

 

 

 

_ kg _ k]
.1 89 890 4134 500
1 500 500 5000 500
10 2810 281 6050 500
100 15810 158 7327 500
1000 88915 89 8871 500

 

Generally, k. is estimated by measuring fasting metabolism or by regressing EB
on ME (Moe and Tyrrell, 1973). In the latter case, the intercept is taken as k.I but is
expressed in production units (NEmJ. Summarizing 332 energy balance trials, Moe
et a1. (1972) reported a maintenance requirement of 73 kcal NEﬂ/kg BW”. National
Research Council (1989) uses 80 kcal NEmm/kg BW'” for maintenance taking into
account usual physical actitivity of cows.

With respect to fasting metabolism of dry cows, Holter (1976) reported
103.4j;2.8 kcang BW-7’ 98.6i3.5 kcal/kg BW” at l and 31 days after lactation,
respectively, while Flatt et al. (1965) observed a fasting metabolic rate of 73.5
kcal/kg BW-7‘ in dry cows.

There seems to be little agreement concerning the maintenance requirement of
lactating cows. It has been assumed (Moe and Tyrrell, 1973; Tyrrell and Moe, 1972)
that FHP in cattle numerically is greater than NE“, determined by regression

procedure.

16
3.3.2.4 Genetic aspects for maintenance energy

Taylor et a1. (1986) have discussed the role of genetics in inﬂuencing efﬁciency
of maintenance requirements per unit metabolic body weight (MBW); estimated by
BW'”. They computed a coefﬁcient of phenotypic variation in MNT of 6.4% and a
repeatability of .70 for 2-year periods. Van Es (1961) obtained an estimate of among-
cow coefﬁcient of variation of 4-8% (in dry cows) and 5-10% (dairy cows and steers)
in 237 energy balance trials that he reviewed.

The results of Andersen (1980) showed within-breed variation in maintenance
requirements for beef bulls. A heritability of .31 was calculated from these data.
Taylor et al. (1986) suggested genetic differences in maintenance requirement may be
due largely to genetic differences in HI. In agreement with this postulate, Vercoe
(1970) found that genetic differences in level of production require different k. to
convert NE.“ to ME for maintenance (MEﬂ). Davey et a1. (1983) concluded the
maintenance requirement of cows was not inﬂuenced by genetic merit for milk

production.

3.3.3 Energy use for milk production
The utilization of energy for milk yield has major economic implications. Energy
sources making up ME can inﬂuence the product of energy partitioning. Flatt et a1.
(1969) observed that MKE increased and tissue energy decreased, as alfalfa was
substituted for concentrate. Tyrrell et a1. (1973) noted a shiﬂ in NE from tissue

deposition to MKE as equal feed-energy increments were changed from corn to beet-

l7

pulp.

3.3.3.1 Sources of variation in milk energy

Bauman (1985) indicated, based on literature, that little variation exists among
animals in the efﬁciency with which MB is utilized for milk. A slight increase in
efﬁciency of ME use for milk production was attributable to metabolizability of the
diet (Van Es and Nijkamp, 1969). Extensive analyses by Moe (1981) of results from
energy balance trials performed by Flatt et a1. (1969) showed that 1) use of ME for
milk or body tissue gain was unaffected by milk yield, amount of body tissue gain (or
loss), and stage of lactation; 2) variation among cow was due to the amount of feed
consumed. However, an equal digestible DM produced more milk from white clover
than from ryegrass (Rogers et al., 1979). Other sources of variation in k. have been
observed.

Kirchgessner et a1. (1982) found that kl increased as frequency of fwding
increased. The inﬂuence of cold temperatures on the energy requirement of lactating
cows was minimal (NRC, 1989). This was attributed to the high HP at high feed

intakes. Hooven et a1. (1968) found that BW change increased as kI increased.

3.3.3.2 Efficiency of energy utilization for milk
Energy utilization for milk is primarily a function of digestibility (Waldo and
Jorgensen, 1981). Van Es and Nijkamp (1969) found no effects of percentage of

crude ﬁber or of crude protein on efﬁciency of milk production. They concluded that

18
ME was used for milk production (k.) with and efﬁciency of 54-58% . Walter and

Mao (1989) has summarized the reported partial efﬁciencies for lactation and
observed a range of 54-75%. Multiple regression analyses were used by Moe et al.
(1970) and Moe et a1. ( 1971) to derive partial efﬁciencies for milk production. Partial
efﬁciency of ME for milk were 61-64%. Hashizume et a1. (1965) found k, of 74 and
82% for low and high concentrate diets. Calorimetric studies (Armstrong et a1. , 1964;
Flatt et al., 1965; Moe et al., 1972) have shown that lactating cows, use DE or ME

for milk production with a similar degree of efﬁciency.

3.3.3.3 Estimation of milk energy
The NE requirement for milk (NE) is deﬁned as the energy contained in the milk

produced (NRC, 1989). The Gaines (1928) equation proposed the use of 4% fat-
corrected milk as a measure of NE. The inadequacies of this procedure became
evident with dietary regimens (Iaben, 1963; Van Soest, 1963) aimed at producing
lower fat and a higher solids-not-fat (SNF) concentration. The landmark analysis
(Tyrrell and Reid, 1965) of 21 different combinations of milk components
demonstrated that the most practical equation for the accurate prediction of energy in
milk was:

Energy (kcal/kg) = 41.84 (% fat) + 22.29 (% SNF) - 25.58.
In an effort to correct on the basis of SNF, solid-corrected milk (SCM) was
computed as:

SCM (kg) = 12.3 (kg fat) + 6.56 (kg SNF) - .0752 (kg Milk).

19
The SCM equation has been used widely to predict energy in milk (Walter and Mac,

1989; Ngwerume and Mac, 1992) with a high degree of accuracy. NRC (1989)
computed NE as:

NE, (Mcal/kg of milk) = .3512 + [.0962 (% fat)].

3.3.3.4 Genetic aspects of milk energy

There is a lack of studies in the literature on the genetic parameters of milk
energy due to the following reasons: 1) the number of animals in respiration chambers
and energy balance studies was too small, and equipment and labor was too expensive
to extend these studies; 2) the studies were carried out mainly by nutritionists who are
mostly interested in using uniform animals. In this regard, another caveat is that most
studies of genetic differences between animals have focused on the genetic
relationship between milk yield and feed efﬁciency (Custodio et al., 1983; Grieve et
al., 1976; Wilmink, 1987). For this reason, present knowledge of genetic parameters
for milk energy is limited.

There is a unanimous agreement, in the literature, that direct selection on gross
feed efﬁciency has no advantage because of the high correlation between gross feed
efﬁciency and milk yield (Buttazoni and Mao, 1989; Custodio et al., 1983; Grieve et
al., 1976; Freeman, 1967; Korver, 1988). Variation between animals in appetite,
digestion, nutrient absorption, maintenance requirement, utilization of ME for MKE,
nutrient partitioning and output composition makes gross feed efﬁciency an imprecise

measure of efﬁciency.

20
3.3.4 Energy use for live weight gain or loss

Because production of milk during lactation has a high priority in the dairy cow,
production of milk may continue to be high despite insufﬁcient DMI. In such
situations, the animal must mobilize body tissue to compensate for the energy deﬁcit.
On the other hand, excessive intake of energy during late lactation and the dry period
can cause BW gain (Morrow, 1976). The composition of the BW gain or loss is
important in determining kI (Garrett, 1980). Some extensive reviews discuss
manipulation of growth (Elsley, 1976), energy use for growth (Millward, 1976), and
‘ nutrition and genetic effects on body composition (Lister, 1976). Moe et a1. (1971)
caution that live weight change alone may not provide an accurate measure of EB.

Partial efﬁciency of ME for body gain (or loss) was 75% (Moe et a1. 1970; Moe
et a1. 1971). Thorbek (1970) found partial efﬁciencies for protein and fat deposition
of 43 and 77%. Protein-deﬁcient diets shifted energy deposition from protein to fat

(Black, 1974).

3.4 Canonical correlation analysis

It often happens that we make measurements on several variables. Collectively
these variables make up a multivariate system which may be divided a priori into two
sets, with each set relating to a particular component of the system and with some
idea required of the association between these components. For example, we may take
12 measurements relating to the yield of alfalfa (e. g.height, dry weight, number of

leaves) at each of 1: sites in a region, and, at the same time, we may have recorded q

21

variables relating to the weather conditions at each of these sites (e.g. average daily
rainfall, humidity, hours of sunshine). The whole system thus consists of n units on
each of which (p + q ) variables have been measured. The overall (p + q)x(p + q)
correlation matrix contains all information on associations between pairs of variables
in the system, but attempting to extract from this matrix some idea of the association
between the two sets of variables is a difﬁcult task. This is because the correlations
between the sets may not have a consistent pattern, and these between set correlations
must in any case be adjusted somehow for the within-set correlations. Hotelling
(1936) proposed the method of canonical correlation which derives a measure of
maximum correlation between linear combinations of the original sets of variables. A
rigorous derivation of the canonical correlation model may be found in Anderson
(1958). A derivation of computing procedures for canonical correlation used in this

project is outlined in Appendix V.

3.5 (Co)variance component estimation

The basis for estimating variance components was established by Fisher (1925);
that basis being: equate quadratic forms in the observation vector to their expected
values and thereby construct a set of equations with unknown parameters the vector of
variance components to be estimated. Whence, the method yields equations linear in
the variance components that can be solved and the solutions taken as the estimates
but this method was conﬁned to balanced data.

In genetic studies, the data are unbalanced. Henderson (1953) extended the

22

knowledge of estimation of components of variance to unbalanced data with his
Methods 1, II, and III. Method 1, which computes sums of squares in the standard
analysis of variance (ANOVA) with balanced data, equates the mean squares to their
expectations and solves for the components, has been used extensively but cannot be
used on mixed models. Method II is unbiased by ﬁxed effects. It adjusts for ﬁxed
effects (in models having no interaction between ﬁxed and random factors), estimated
as if random effects were ﬁxed, then estimates components as in Method I, using the
adjusted data. These two methods enabled substantial analyses to be performed.
Method III yields estimates of components of variance that are unbiased by ﬁxed
effects and their interactions and has contributed relatively more to animal breeding
applications. However, the order in which reductions in sums of squares are
computed is noteworthy. Reductions in sums of squares using a full model minus
reductions in sums of squares from reduced models are equated to their expectations
and solved for components. Computing these reductions and their expectations may be
difﬁcult for large data sets. Hence other approaches such as the method of maximum

likelihood (ML) have been preferred.

3.5.1 ML estimates of (co)variance components
The ML method was applied to the general mixed model by Hartley and Rao
(1967). The scope of ML estimation for the estimation of variance components has

been reviewed by Edwards (1961) and Harville (1977). In general, for a given

staitistical model, parameters 0 to be estimated, and assumed distribution of the data,

23
the likelihood function L(6) can be derived. The ML estimates are the numerical

values of the parameters for which L(6) attains its maximum. Maximizing the

likelihood leads to estimates that are functions of sufﬁcient statistics, universally most
powerful, consistent, asymptotically normal, and often efﬁcient. Large computational
requirements restrict the use of ML for estimating variance components. Inherent to
ML are some undesirable properties. The ﬁrst is that the distribution of the data,
usually a multivariate normal distribution, is assumed known. Secondly, ML

estimators are biased as ﬁxed effects in the model of analysis are treated as if they
were known. This bias can be reduced by considering only the part of [(0) which is

independent of the ﬁxed effects (Patterson and Thompson, 1971) and hence invariant
of the loeation parameter. The latter approach is referred to as restricted maximum
likelihood (REML). Under normality the REML estimation is equivalent to both
minimum norm quadratic unbiased estimation (MINQUE; Rao, 1971') and local -
minimum variance quadratic estimation (MIVQUE; Rao, 1971"). Other properties of

REML are discussed by Harville (1977).

3.5.2 REML estimates of (co)variance components
Interest in estimation of (co)variance components by REML procedures has
increased in recent years since REML: 1) yields estimates less affected by selection
bias than Henderson’s ( 1953) Methods 1, II and III (Schaeffer, 1979); 2) allows for
estimation of genetic parameters after consideration of information on all relatives

(Meyer, 1986) without knowledge of true variance covariance components; 3) is

24
computationally feasible. Several REML algorithms have been used (Meyer, 1990)

but most of these are iterative, often leading to repeated re-ordering of the mixed
model equations (MME). For instance, the expectation maximization (EM-REML)
algorithm requires inversion of the mixed model matrix (MMM), and utilizes

information on ﬁrst or second derivatives in order to obtain estimates that maximize
L(O) . An alternative algorithm which avoids explicit evaluation of ﬁrst derivatives is
the derivative free (DF) algorithm (Graser ct al., 1987; Meyer, 1986) generally

referred to as derivative-free restricted maximum likelihood (DP-REML).

3.5.2.1 Derivative-free type REML algorithms

The best linear unbiased prediction (BLUP) method (Henderson, 1973) has
rapidly become the method of choice for genetic evaluation of animals. The notion of
utilizing the numerator relationship matrix in the analysis of BLUP under an animal
model (AM) was presented by Henderson (1952). In the AM the order of the ME
often exceeds the number of records making inversion of the M impractical. Use
of DF-REML has become excwdingly attractive with the widespread use of the AM.

The application of BLUP to multiple traits was described by Henderson and
Quass (1976). The inclusion of maternal effects and presentation of the reduced
animal model were made by Quass and Pollak (1980). The approach of DFREML is

suitable for AM including additional random components (Meyer, 1991). The use of a
direct sparse matrix solver to obtain L(6) can reduce cental processing unit (CPU)

time per round of a DP algorithm (Boldman and Van Vleck, 1991). Recently, a

25

method to approximate sampling variances and conﬁdence intervals for individual

parameters in a multi-parameter analysis has been described (Meyer and Hill, 1991).

3.5.2.2 Limitations of DF-type algorithms
The maximum value of [,(e) in DF-type algorithms has less signiﬁcant digits than

the maximized function. This condition could lead to false maxima, especially for
multilple traits and when correlations are high (Misztal, 1992). Hence, the method of
DF-REML , like EM-REML, does not guarantee identiﬁcation of global maxima in
the presence of local maxima. Groenveld and Kovac (1990), using a small data set,
explored if multiple solutions could be generated for a multivariate mixed model
estimating six covariance components by a DF-type algorithm. Multiple solutions
from the DF algorithm suggested existence of local maxima. However, the DF-type
algorithm was superior to all other algoritrns considered in that study. In terms of
CPU time it was faster by a factor of 22 misidentifying only one solution instead of 2

as EM-REML did.

4 Sources of Variation in Partitioned Uses of Energy Intake in Pluriparous

Lactating Holstein Cows in Energy Chamber

4.1 ABSTRACT

Data were energy chamber measures on 34 multiparous Holstein cows collected
during wk 6, 10, and 14 postpartum. For each period, cows were placed in digestion
stalls for a six-day excreta collection followed by two consecutive ll-h methane and
heat production measurements. Energy partition averages coincided with conventional
values. Sources of variation among cows in gross energy consumed, feeal, urinary,
methane, heat, milk, and maintenance energy, and tissue energy balance were
analyzed. Also analyzed were heat production, energy balance, milk energy,
maintenance energy expressed in ratios to various energy intake measures. A within-
period model containing ﬁxed effects of treatment, parity, season, and covariates for
maintenance and milk energy was used. Neither diet, parity, nor season effects was
found to be a signiﬁcant source of variation in all the variables. Milk energy as a
covariate was highly signiﬁcant in all variables except energy for heat production and
that for maintenance. However, the covariate maintenance energy was found to be a
signiﬁeant effect in heat production at wk 10 and 14 postpartum. The effect of

maintenance interacted with periods in most energy partition and efﬁciency measures.

4.2 INTRODUCTION
The fundamental aspects of energy metabolism were described by results from

several complete energy balance trials in direct or indirect calorimeters (Knott, 1934;

26

27
Moe, 1966; Moe and Tyrrell, 1973; Van Es, 1961). An explanation for the causes

of variation in energy efﬁciency of animals fed different diets was provided by
Armstrong and Blaxter (1957'; 1957”; 1965). They observed that the heat increment
of VFA was controlled by the amount of acetate in fattening sheep but had less effect
in sheep at maintenance. Thus, it was demonstrated that the end-products of digestion
were more important than nutrients consumed in understanding metabolic efﬁciency in
ruminants. Variation in partial efﬁciency of use of energy of VFA for milk
production and maintenance also was shown to be of considerable signiﬁcance. In
addition, the type of ration (Flatt, 1966; Tyrrell et al., 1973), level of intake (Moe,
1966), level of production (Flatt, 1969), stage of lactation (Janicki, 1985),
environmental conditions (Young, 1976), and size of the cow (Tyrrell et al., 1973)
can affect the partition of the energy consumed. However, effects of variation in
amount and type of diet on energy efﬁciency and energy partition are better explained
by a knowledge of amounts and balance of the speciﬁc metabolites which are
absorbed from the digestive tract.

A plethora of literature exists on the inﬂuence of ration composition and level of
intake on digestive efﬁciency, but there is a paucity of data that describe the
associated magnitude and sources of variation. Much of the previous work has
focused on proportions in partitioned energy intake by a typical, or an average, cow.
However, an examination and understanding of variation among cows is much
needed. Such an understanding may indicate whether the efﬁciency of a cow’s ability

to convert energy intake for production would be a worthwhile criteria for genetic

28

selection. This study was undertaken to determine the amount of variation and
examine speciﬁc sources of variation in each of the partitioned energy uses and

energy efﬁciency measures.

4.3 MATERIAIS AND METHODS
4.3.1 Experimental Procedure and Data

Measurements were taken on 34 pluriparous cows during the course of three
periods, wk 6, 10, and 14 postpartum. Diets were protein supplements, low-protein
concentrate, com-silage treated with urea at ensiling, and wilted grass silage fed
individually for ad libitum intake. All ration components were blended and fed twice
daily to provide at least 2.3 kg of orts daily as indicated by Holter (1982).
Composites of low-protein grain and supplements were ground in a Wiley mill (1-
mm), mixed thoroughly, subsampled, and analyzed for proximate nutrients, ADP, and
combustible energy (Parr adiabatic oxygen bomb calorimeter). Milk samples were
collected from each milking, composited over the collection period, and analyzed for
combustible energy according to methods described by J anicki and co-workers (1985).

Feces and urine were collected using mechanical separators and weighed daily; a
1% aliquot was taken each day and composited over the 6—d collection period.
Following excreta collection, cows were placed in an open circuit, indirect respiration
calorimeter to measure heat and CH. production for two consecutive ll-h periods.
Samples of composite chamber air were analyzed for C0,, CH“ and 0,

concentrations. Thereafter, energy balance was determined by difference between

29
inputs and outputs. Energy partition variables were fecal energy (FE), urinary energy

(UE), digestible energy (DE), methane energy, metabolizable energy (ME), heat
production (HP), heat increment, net energy (NE), milk energy (MKE), maintenance
(MNT) and energy balance (EB). Efﬁciency values of energy conversion were
expressed as ratios of energy in product and the product from which it was formed,

namely, MNT/ME, MNT/NE, MKE/ME and MKE/NE.

4.3.2 Model and Analysis algorithm

Within each of the three postpartum periods, crude mean and SD were computed
for each of the energy expenditure and efﬁciency variables. Proportions of partitioned
energy intake ﬁrst were expressed using crude means. Variation among cows was
expressed in CV. Variation in energy intake and in each of the partitioned energy
measures was analyzed using a within period model:

yw=u +S,+1}+P,+b,xwﬂ+b,xm+ew

where: ytﬂd is an energy partition (kcal) or efﬁciency (%) trait; p is a constant
common to all observations; St is ﬁxed effect of season, i=1,2, and 3 where 1 is
November through March, 2 is June through August and 3 is other months; 1} is

ﬁxed effect of treatment, j=1,,,,,4; P1: is ﬁxed effect of parity, k=2,3, 24; xiv” is

maintenance (kcal) as a covariate; X201! is milk energy (kcal) as a covariate; and 5W

is random residual error distributed as, NIOJOI) where of is assumed to be

3O

homogenous across all groups. When yw was milk energy or a related trait, qu
was dropped from the model. Similarly, when ya“ was maintenance or an associated

measure, x“ was excluded from the model. All analyses were performed using

SAS" GLM (1985).

4.4 RESULTS AND DISCUSSION

The partitioning of energy intake was based on raw means of cows over the
duration of the energy metabolism study. Figure 3 illustrates the overall means and
SD of the energy intake variable expressed as a percentage of gross energy (GE)
consumed. The chart depicted the ﬂow of dietary energy through a dairy cow: GE
less FE yielded DE, which gave ME after subtracting UE and CH, and so on.
Utilization of energy for production was more than that for maintenance. The
proportion of energy left over as tissue energy balance portrays that the experimental
animals were in positive energy balance. These energy partitioning data are
consistent with those reported by Flatt (1966). Changes in the postpartum partitioning
of ingested energy are discussed in more detail elsewhere (Janicki, 1985).

The within period means and SD for the energy partition traits are given in Table
2. Intake of GE was comparable for the three stages of lactation. Fecal and CH.
energy, HP, and EB appeared to increase whereas the DE, ME, and MKE decreased
over time. These changes in the energy utilization could be attributed to changes in

diet and production (J anicki, 1985) as lactation progressed. There was more variation

31

 

100

 

 

 

 

 

 

 

 

 

 

33.5% (2.“) 100% GE
eo ——«»
ao —
70 _L
FECAL
2.495 (.490 URINARY “'5‘ (2.“) DE
‘° ﬂ 4.59s (.ax) METHANE
III T—
E 12.795 (2.2%) ME
0 so 59.3% (3.0%)
* HEAT mcneueur
15.9% (2.0%) NE
‘0 _ 4a.”;
(3.39s)
30 — MAINTENANCE
2a.“ (5.2%)
20 ——
MILK PRODUCTION
WEIGHT GAIN OR LOSS
CONCEPTUS
1° —‘ HAIR
o 2.1% (5.3%) EB v

 

 

 

 

 

 

 

 

Figure3. Overall mean percentages forpartitioningofenergy intake (n = 34).
ValuesinparenthesesareSDforeachenergypartitiontrait.

32

 

 

 

 

o: 83. Ge :8 8m 88 a8 32 3e 3% on. E 8.33 98m
2 8:. 4: Sta 8 nos. 48 was" a a!“ n3 Inca 388%:
o a; «2 ~82 e 3.» e: 3.3 a. 3m «2 83 38885835“:
2 ”new 3» 36: S 3:. 48 R23 2 «an «a 884 @8832
an 38 an ~32 3 3.2 «8 “$2 a cam 2h «an: egos .8:
2 8am «on «Sea 2 88 «.3 83s 2 :2 5 Eva 8332a .8:
2 8mm 82 2.34 : 38 :5 Sam 2 ﬁne 8: 85 .acoﬁaegsqc:
a 84 2. 3:. 2 so a: in on an» n: can 38525.82
8 :4 2 £8 a n3. 2. 88 on «an 8 SE .388 has:
: $8 $2 wanna : $8 8: ~38 2 88. 32 80% .908 spouse
2 32 3e 88m 2 a: 8” Sea 2 89. «2. «88 38518".
: 38 $2 33 S 8850: 8%» 2 FEES. «$9 385.35
a [.53 I a lasso. II a [383. .II
>O 8 mm 53: >O mm mm :3: >O om mm :8: :5.
E: 35, a.:: x? use;

 

.33 Estuaoma 3.003 3.028 3 A3“ I 5 3.532: 33:0 .8. >0 v:- .Qm .mm .52.: moron 55; .N mania.

33

in the energy used for production than for maintenance. Variation in CH. was higher
than that for UE. However, with the exception of EB, variation among cows was not
very high.

The efﬁciency of energy utilization is shown in Table 3. Methane loss as a
proportion of DE, and HP as a ratio of ME increased by wk 14 postpartum. The
increased forage in the diet could account for this trend. Urinary losses were
relatively similar in all three periods. Likewise the efﬁciency of ME and NE for
maintenance did not change. Therefore the decrease in efﬁciency in milk energy
from ME and NE over time could be associated with reduced milk yield with time.
On the other hand, the use of ME for MNT appeared to be constant across all
periods. Consequently, the conversion of ME into tissue energy balance, increased
from wk 6 to 14 and represented the highest variation among cows. Postpartum
ﬂuctuations in the magnitude of the variation in the efﬁciency traits are indicative of a
substantial amount of variation in the efﬁciency of GE partitioning during lactation.

The amount of variation in each of the energy conversion traits was partitioned
using the within period model. The levels of signiﬁcance for all ﬁxed effects and
partial regression coefﬁcients for the covariates are shown in Tables 4 and 5 for
selected traits. Energy in milk was an important source of variation in HP, ME, NE,
and EB during the postpartum period (P < .05). The results suggested that variation
in the efﬁciency with which ME or NE are used for HP, MKE, or MNT is

independent of diet, season, and parity.

34

:38:an u m2!“- .3§3hwuouo H mm .355
SEE u 5:2 £928 o_e§:3308 1 mi 68323::— 1 P22 66:03.2.— .8: I .5 $928 2.533“. I m0 .383 at: 1 m9.

 

 

 

 

 

 

 

 

 

an: E 3 am was 3 3 3 98a 5 a. 3 3.8 3. a. on $53
a: 92 5 2n 5; 3 3 as 2: 2: 3 one 3: a2 3 as mzamxz
2: S. as as. a: as a; 2:. no. 3 3 a: we. 3 a. «.8 szuz
3: 3. A. 4.3 2: 3. m. «.3 an an a. a: 9: a... m. 3a MET—.22
2 an m. a: 92 an e. EN ”.2 N... A. 4.8 a: an 4. as $1.52
as S. w. 4.8 2: an a. 9% as S. m. at. as 3 m. 9% 82;:
3 w. _. as 0.2 a. a. we is E N am 3: 3 _. E. maize
a: e. _. an we a. _. an a: o. a an 2 e. _. an 825
a
>0 mm mm :82 >0 mm mm 5.: >0 mm mm 98: >0 9. mm as: .8382
9:33 9.23.3 .533 .326
.ae 5.9.8.383 38:: six 2 avgasﬁégeﬁdm .mm as... 8.5.. SEE-.355 .m 35:.

35

TABLE 4 . Critical levels (P - value) for main effects and partial regression coefﬁcients for
covariates from within period model: energy partition measures.

 

 

 

 

P - value b(MNT) b(MKE)
y I Wk. Estimate P - Estimate P - E:
W postpartu TreatmentParity Season of b SE value of b SE value

111

HP 6 .27 .84 .55 .79 .52 .15 .17 .09 .08 .15
10 .19 .97 .48 1.20 .66 .003 .19 .14 .01 .41
14 .13 .42 .26 1.77 .41 .01 .20 .07 .17 .34

ME 6 .18 .87 .36 -1.09 1.24 .39 .84 .22 .001 .30
10 .89 .65 .23 1.75 .90 .06 .82 .16 .0001 .52
14 .03 .55 .18 2.71 .98 .01 .68 .21 .003 .54

NE 6 .28 .82 .26 -.98 .98 .33 .68 .18 .001 .30
10 .65 .69 .39 1.50 .82 .08 .61 .15 .0003 .42
14 .31 .78 .63 1.91 .10 .07 .47 .21 .03 .25

MKE 6 .48 .70 .08 1.50 1.03 .17 Dropped .10
10 .82 .68 .21 .55 1.12 .63 -.06
14 .69 .18 .03 .09 .96 .93 .12

MNT 6 .51 24 .70 Dropped .04 .03 .17 -.01
10 .36 .20 .83 .02 .04 .63 -.05
14 .27 09 .60 .004 .04 .93

BB 6 .28 .82 .26 -1.98 .98 .05 -.32 .18 .08 .24
10 .65 .69 .39 -.39 .82 .55 -.39 .15 .01 .02
14 .31 .78 .63 -.53 1 .37 -.53 .21 .02 . 17

 

'E3 = energy balance, HP = heat production, ME = metabolinble energy, MKE = milk
energy, MNT == maintenance, NE = net energy.

36

.3555: u m2 .888
u P22 .355? u mum: .355 0—39.89: I m2 .guoavemuoo: H mm .0233 355 I mm.

 

 

 

8. 8. 88. 8.- 3.. 8. 8.- 8. 8. 8. z
3. 8. 88. 8.- 8. 8. 8.- 8. 8. 8. S
8. 8. 88. 8.- 8. 8. 8.- on. 8. 3.. 8 85mm
2. 8. 88. 88.- 8. a. a. 3
S. 88. 88. 8.- 8. 8. 8. 2
8. 8. 88. 88.. 888: 8. 8. 8. e 52.52
8. 8. 88. 88.- :. 8. 8. z
8. 88. 8o. 8.. 8. 8. 8. 2
8. 8. 88. 88.- 888: s. 8. an. e 38.52
8. an. 8. 8.- 8. 8. s. z
2.- G. 8. 8.- R. 8. 8: 2
:. 3&8: 2. 8. 8. 8. 8. an. o 858.:
2. 8. 8. 8.- 8. 8. 8. z
2.- 8. 8. 8.- a. 2. 8. 2
a. 8&2: I. 8. 8. 8. 8. 8. o 83.8.:
:.- 8. 88. 88.- 8. 8. 8. 8. 8. 2. 3
a. 8. 88. 8.- an. 8. 8. 8. 8. 8. 2
8. n8. 8. 8.. 8. 8. 8. a. 8. 8. o :22:
3.: mm a 8 3? mm a 8 =83- bﬁm 5.58; 33:98.. a.
"m .m 33.8 .m 38 a; _ a
8.28 9228 83.:

 

.588 35850 355 :32:
35: £58 :5: 53:38 .5. 858508 58632 3.5: 5 88¢“. 8 .5» A023 - 5 205. 52.5 .n man‘s

37

Not shown are comparable results for water intake and body tissue balance
(Appendix 111).

Energy for maintenance accounted for a signiﬁcant proportion of the variation in
HP, ME and NE at wk 10 and 14 postpartum. As expected, the partial coefficients
for maintenance generally were higher than those for milk production. Replacing
MKE and MNT in the model by SCM and BW'”, respectively, yielded almost

identical results.

4.5 CONCLUSIONS

Dietary energy consumed in the postpartum period by lactating Holstein cows was
partitioned by indirect calorimetry. Average values in partitioning of energy intake
agreed with textbook estimates. Variation in energy utilization traits was found to be
high while variation among cows, for most of the traits except energy balance, was
low. It was observed that the utilization of ME for milk decreased as lactation
progressed. For these data, dietary source of energy, season, and parity were not
very important factors in explaining variation in the partitioning of DE or ME for HP,
MKE, or MNT. Clearly, much of the variation in energy traits can be attributed to
energy in the milk.

Variation exists in the efﬁciency with which energy is utilized; if genetics is a
major factor in that variation, then evaluation of individual genetic merit for energetic
efﬁciency traits should provide useful management information. This could be made

possible from dietary attributes observed on the animal in the ﬁeld or barn. However,

38
the validity of this approach is still questionable because previous studies have not

examined energy chamber and ﬁeld data on the same cows. Therefore, comparisons
of energy partition measures determined from calorimetry with those from ﬁeld or
barn data, on the same cows, are necessary in order to provide pertinence for ﬁeld

data.

5 Energy intake and Gross Efﬁciency Comparisons from Calorimetric and
Field data on the Same Lactating Cows

5.1 ABSTRACT

Estimates on gross efﬁciency were obtained from feed intake and production data
on 30 pluriparous Holsteins cows during wk 5, 7, 9, 11, 13 and 15 postpartum.
Energy intake and efﬁciency measures from energy chamber on the same cows were
taken during wk 6, 10, and 14 postpartum. Measures of gross efﬁciency were
expressed in terms of utilization of metabolizable energy or net energy for production
and maintenance. For corresponding postpartum periods, comparisons were made
between chamber measures and ﬁeld estimates by canonical correlation analysis. All
energy and gross efﬁciency estimates from ﬁeld data closely approximated measures
of the same traits from energy chamber data. Variation among cows in gross
efﬁciency for ﬁeld estimates was one half that for chamber measures. On the other
hand, variation among cows in energy partition traits was consistent for both ﬁeld
estimates and energy chamber measures. Correlations greater than .87 were observed
between ﬁeld estimates and chamber measures on maintenance energy and milk
energy. Field estimates and chamber measurements of metabolizable energy and net

energy had correlations of .58 and .37, respectively.

39

4o
5 .2 INTRODUCTION

Alloeating energy intake to energy for milk yield in a lactating cow is an
important aspect in energy metabolism. Exact measures of energy intake from dietary
sources can be determined by direct (Knott, 1943) or indirect calorimetry (Moe etal.,
1972) but this can be costly. 80, several methods have been developed to predict feed
(Brown et al.,1977; NRC, 1989; Moore and Mao, 1990; Van Soest, 1967) and energy
(Moe and Tyrrell, 1972; Moore and Mao, 1992; NRC, 1989; Tyrrell and Reid, 1965;
Walter and Mao, 1989) intake using variables such as BW, milk production, forage
type, ﬁber content, age, parity, and season. Genetic selection for energetic efﬁciency
is of increasing importance.

Numerous studies have shown that selection for milk yield brings linear
increments of feed efﬁciency (Blake 1979; Freeman, 1975). Nonetheless, Blake and
Custodio (1984) concluded that efﬁciencies of nutrient utilization have not been
inﬂuenced by selection for milk production .

Despite the rising costs of feeding cattle, current dairy cattle evaluations do not
consider information on individual feed intake nor on efﬁciency of energy
partitioning. This is, in part, because the establishment of such a data base, by
installing calorimetric apparutus on farms, would be both expensive and impractical.
Therefore, accurate approximation of energy efﬁciency using data obtainable under
normal conditions would be highly desirable. Walter and Mao (1989) compared
estimates of energy intake from ﬁeld collected data with literature chamber results and

found that accurate approximation was plausible.

41

However, in order to determine the validity of efﬁciency estimates from ﬁeld
data, it would be necessary to compare estimates from ﬁeld data with exact measures

from energy chambers on the same cows which was the objective of this study.

5.3 MATERIALS AND METHODS
5.3.1 Experimental Design and Data

Data used in this study came from a study which examined the effects of
percentages of crude protein and nitrogen solubility in the diet and their interactions
on digestibility, energy and protein balances (Janicki, 1985). In that study
measurements were taken on 30 pluriparous cows, in energy chamber, during wk 6,
10, and 14 postpartum. Diets were as described by Holter et al. (1982) and Janicki
(1985) and energy balance was determined by methods described by Saama et a1
(1992‘). Energy intake variables were metabolizable energy (ME), net energy (NE),
milk energy (MKE), and maintenance energy (MNT). Gross efﬁciency (GREF)
measures from energy chamber were thus expressed as ratios of MNT/ME,
MNT/NE, MKE/ME and MKE/NE.

Recorded for each cow in wk 5, 7, 9, 11, 13 and 15 postpartum were DMI, milk
yield, milk fat, and body weight. Estimated ME (M) and estimated NE (eNE), and
estimated GREF (eGREF) were obtained from DMI:

eME (Mcal/d) = [ (1.57 x Grain) + (1.29 x C8) + (1.07 x HCS) - (1.31 x Orts)],
eNE (Mcal/d) = [ (.93 x Grain) + (.77 x CS) + (.65 x HCS) - (.78 x Orts)],

where Grain, CS and HCS are daily consumption, in kilogram per day, of grain

42

concentrate, corn silage and haycrop silage, respectively, and orts in kilogram per
day. Coefﬁcients for energy value of feeds are those reported by NRC (1989).
Estimated energy in milk (eMKE) was,

eMKE (Meal/kg of milk) = [(.3512 + (.0962 x 96 fat)].
Estimated MNT (eMNT) was (NRC, 1989),

eMNT (Mcang of BW'") = .086 x BW'75(kg).

5.3.2 Analysis Procedures

The variables for comparison were: 1) chamber measures of MB, NB, MKE,
and MNT in wk 6, 10, and 14 postpartum; 2) estimates from ﬁeld data on eME,
eNE, eMKE, and eMNT from averages of wk 5 and 7, 9 and 11, 13 and 15
postpartum; 3) chamber measures of GREF in MKE/ME, MNT/ME, MKE/NE, and
MNT/NE; and 4) eGREF from ﬁeld data as eMKE/eME, eMNT/eME, eMKE/eNE
and eMNT/eNE.

Paired comparisons of means, SD and CV’s, and computations of product-
moment and rank correlations were done: 1) between GREF and eGREF in ME; 2)
between GREF and eGREF in NE.

Canonical correlation analysis (CCA) was performed to compare estimates from
ﬁeld data to chamber measures. Canonical correlations refer to correlations that are
independent of each other (Hotelling, 1936). Its use is most suited to examining
correlations between a group of p X-variables and a group of q Y-variables, when one
wishes to test the null hypothesis that X, and Y, variables are independent. Various

linear combinations in X, and Y, are established in CCA. Then correlations between

43

the linear combinations from the sets of variables are computed. The highest
correlation would correlation between X, and Y,. Thus, the CCA model reduces the
dimensionality to a few linear functions of the measures under study.

The null hypothesis that a canonical correlation is 0 in the population was tested
by a likelihood ratio (Lawley, 1959). Redundancy analysis (Cooley and Lohnes,
1971), which measures the standardized proportion of total variation in a variable, X,
or Y” that is predictable from linear functions of X, or Y, also was performed. All

analyses were accomplished using SAS‘ (1985).

5.4 RESULTS AND DISCUSSION

Means of energy estimates from ﬁeld data were within the normal range and
approximated closely energy chamber measures (Table 6). Corresponding standard
errors also were similar. Variation among cows in energy use for MKE and MNT
was slightly lower for estimates from ﬁeld data. Averages for ME and MKE as
estimated from ﬁeld data were slightly higher than those for chamber measures while
NE and MNT means were slightly lower. Similar trends were observed for means
computed within periods (Table 7). Differences between ME and eME and between
MKE and eMKE decreased as lactation progressed while differences between MNT
and eMNT and between NE and eNE remained fairly constant. The large differences
in wk 6 postpartum may be due peak lactation as shown in Figure 4, and the state of
negative energy balance, during that period. Therefore, ﬁeld estimates at wk 14

postpartum perhaps were most representative of actual energetic efﬁciency of cows.

44

TABLE6. Overall mean, SE, SDandCVforenergymeasuree fromenergychamberand
from ﬁeld data on the same 30 pluriparous Holstein cows in early lactation.

 

 

Energy measure Energy chamber
Mean SE SD CV
--—- Meal/d —— %
Metabolinble energy 50.68 .64 6.03 11.9
Net energy 39.78 .52 4.98 12.51
Milk energy 24.74 .52 4.91 19.83
Maintenance energy 13.29 .08 .79 5.98

 

Field
Mean SE SD CV
-— Mcal/d —— %
61.19 .92 8.76 14.31
36.44 .55 5.21 14.29
26.33 .48 4.58 17.39
10. 11 .07 .63 6.20

 

TABLE7. Mean, SE, SDandCVofenergymeasuresfromenergychamberandestimates
from ﬁeld data on the same 30 pluriparous Holstein cows in early lactation.

 

 

Energy measure Energy chamber

Mean SE SD CV

— Meal/d —— %

WED
Metabolizable energy 51.84 1.17 6.39 12.32
Net energy 41.13 .94 5.13 12.47
Milk energy 27.21 .90 4.91 18.05
Maintenance energy 13.32 . 16 .86 6.45
ﬂk 19 mstpartug}
Metabolizable energy 50.58 1.07 5.85 11.56
Net energy 39.65 .91 4.96 12.51
Milk energy 24.72 .86 4.74 19.15
Maintenance energy 13.26 .15 .80 6.06
W
Metabolizable energy 49.61 1.06 5.83 11.75
Net energy 38.57 .85 4.66 12.08
Milk energy 22.27 .70 3.83 17.20
Maintenance energy 13.29 .14 .74 5.59

 

Field
Mean SE SD CV
— Mcal/d —— 96
64.57 1.56 8.56 13.26
38.39 .93 5.10 13.29
28.91 .80 4.39 15.20
10.18 .13 .70 6.82
61.06 1.68 9.20 15.06
36.36 1.00 5.48 15.08
26.46 .78 4.26 16.10
10.04 . 12 .64 6.40
57.93 1.35 7.39 12.75
34.56 .81 4.41 12.77
23.61 .64 3.52 14.89
10.09 .10 .54 5.40

 

45

However, results from paired t—tests indieated that means were signiﬁcantly different
(P < .0001) in all postpartum periods. Notwithstanding, CV for mean absolute value
of differences between ﬁeld estimates and chamber measures were as high as 70%.
This suggested differences between estimates and measures for energy traits were
quite erratic and misleading as evidence for correspondence between chamber

measures and ﬁeld estimates.

 

 

“MI.
1

j

 

 

 

 

 

 

ﬁgure 4. Lactation curve for 30 pluriparous Holstein cows
The ﬁeld estimates of efﬁciency in eMKE and eMNT from eME were

signiﬁcantly lower than measures from energy chamber (P < .0001) as shown in
Table 8. Approximation of NE utilization for MKE by eMKE / eNE were higher,
while that for eMNT was lower (P < .0001). This implies that the formulae for

estimating MKE was more precise than that for MNT.

46

TABLE 8. Overall mean, SE, SD and CV for efﬁciency measures from energy chamber
and estimates from ﬁeld data on the same 30 pluriparous Holstein cows in early lactation.

 

 

 

Energy chamber Field
Efﬁciency measure1 Mean SE SD CV Mean SE SD CV
% %
MKE/ME .48 .01 .08 16.21 .43 .01 .07 6.97
MKE/NE .62 .01 .10 16.51 .73 .01 .12 7.03
MNT/ME .27 .004 .04 13.62 . 17 .002 .03 6.91
MNT/NE .34 .01 .05 14.08 .28 .004 .05 6.91

 

‘ME = metabolizable energy, MKE = milk energy, MNT = maintenance energy, and
NE = net energy.

The within period means for GREF and eGREF are given in Table 9. In constrast
to results in Table 8, among cow variation in energy utilization from ﬁeld data was
not always lower than that for corresponding chamber measures; variation in GREF in
ME and NE for maintenance was higher at wk 10 and 14. From wk 6 to wk 10
postpartum, mean differences between GREF and eGREF remained consistent.

The product-moment and rank correlations between ﬁeld estimates and chamber
measures in energy are presented in Table 10. Correlations among MNT and eMNT
were the highest. Correlations between MKE and eMKE, between MNT and eMNT
were higher than correlations between either NE and eNE or ME and eME. The rank
correlations were moderate to high and consistent with the product-moment

correlations.

47

38.8.8.2." ulna-.Reosggegu ﬁzzéwgamﬁu g.§3£§§n HS.

 

 

 

 

$2 8. 8. 8. 2.: 3. 8. mm. 22.2:
3.2 8. 3o. 2. 3.: 8. 8. a. mick:
2.2 8. 8. S. 8.2 2. 8. %. mzauz
3.: 8. S. 2.. 8.2 S. 8. 3. mg:
933
:2 8. 8. ”N. NE: 8. 8. 3. szz:
2 .2 8. 3o. 2. 3.2 3. 8. 2. $5.22
an: 8. 8. s. 3.2 2. 8. 1.. mzauz
«62 S. 8. 3. 8.2 8. 8. 3.. mg:

a a c
2.: 3. 8. 2. 2 .2 8. 5. no szzz
8: 8. «.8. 2. 2 .2 3. 3. ea. $5.22
:2 :. 8. 8. 2.2 2. 8. 2. mzaxz
8.2 8. 8. an. 2 .2 8. 3. 2.. mg:
3

a 9..
>0 on mm :8: >0 em 8 as: .853... Egan
not 8520 38m

 

doc-8: been 5 .38 53:6: nap—gm on 058

95.5.3. :8 eégnaaﬁezogee. Baggage... 5238.368: 533...

48

TABLE 10. Product-moment correlations between ﬁeld
and energy chamber measures of energy intake from data
on the same 30 pluriparous Holstein cows in early
lactation; values in parentheses are rank correlations.

 

 

 

Field estimate2
Chamber measure‘ eME eNE eMKE eMNT
ME .71 .71 .58 .37
(.73) (.66) (.63) (.18)
NE .65 .65 .57 .32
(.73) (.66) (.62) (. 18)
MKE .58 .58 .87 .21
(.56) (.53) (.86) (.22)
MNT .23 .23 .24 .91

(.32) (.29) (. 16) (.92)

'ME = metabolizable energy, NE = net energy, MKE =
milk energy, MNT = maintenance energy.

2eME = estimatedME,eNE= estimatedNE,eMKE =
estimated MKE, and eMNT = estimated MNT.

 

Correlations between GREF and eGREF, in Table 11, were relatively high but
lower than those in energy traits. The rank correlations between GREF and eGREF
were, in most instances, higher than the product—moment correlations. Negative
correlations between utilization of either NE or ME for MKE and use of MB or NE
for MNT reﬂect that the lactating cow must sacriﬁce efﬁciency for production in

order to partition more energy for maintenance.

49

TABLE 11. Product-moment correlations between ﬁeld
and energy chamber measures of energy efﬁciency from
data on the same 30 pluriparous Holstein cows in early
lactation; values in parentheses are rank correlations.

 

 

 

Field estimate2
eMKE/ eMKE/ eMNT/ eMNT/

Efﬁciency eME eNE eME eNE
measure‘
MKE/ME .59 .59 -.25 -.24

(.59) (.56) (.04) (.03)
MKE/NE .55 .56 .24 -.24

(.59) (.56) (.04) (.03)
MNT/ME -.04 -.04 .59 .60

(-.26) (-.25) (.71) (.64)
MNT/NE -.06 -.06 .57 .57

(-.26) (-.25) (.71) (.64)

'ME - metabolizable energy, NE = netenergy, LIKE == milk
energy, MNT=maintenanceenergy.
M=estimatedME,eNE=estimaredNE,eMKE=
estimatedMKE,andeMNT =estirnatedMNT.

 

Linear combinations of the ﬁeld estimates and chamber energy measures were
examined by CCA. Tables 12 and 13 show CCA results for comparisons between
chamber measures and ﬁeld estimates of energy partitioning and efﬁciency,
respectively. Because the comparisons involved four energy partitioning or efﬁciency
variables at a time, we could have, at most, four orders or dimensions. As expected,
the eanonical root for ﬁrst order was the largest. The ﬁrst dimension also gave the
largest correlation among the linear combinations of the chamber and ﬁeld variables.
Within each period, summing all four canonical roots yielded the total variance. At
wk 6, 10, and 14 postpartum the ﬁrst squared canonical correlation was signiﬁcant (P

< .0001) and the ﬁrst two dimensions accounted for over 90% of the total variation

50

with the highest cumulative proportions occuring in wk 14. These dimensions depicted
convincing evidence for strong linear associations between the factors.

Results of the redundancy analysis showed slightly reduced cumulative
proportions of variation in GREF, at wk 6 and 14 postpartum, which was indicative
of lower precision in those estimates. This might also imply that some of the
negatively correlated variables could have been acting as suppressors.
Notwithstanding, for energy intake and GREF measures in the energy chamber, the
highest proportion explained by the ﬁeld variates was at wk 10.

A factor loading is a correlation between the underlying canonical variable and
the observed trait in question. The factor loadings for the energy intake variables
showed that all the measures from ﬁeld data contributed signiﬁcantly in the
relationships between the canonical variables and energy traits (Appendix V). The
chamber canonical variables had the highest loadings for ﬁeld MKE and MNT. On
the other hand, the mixture of signs on the factor loadings for the GREF measures
conﬁrmed the existence of suppression. The GREF in ME and NE for MKE acted to
suppress the relationships between the canonical variables and the GREF in ME and
NE for maintenance. This could be so because, in the ﬁrst vector, the contrast was
between efﬁciencies for milk energy and those for maintenance. It is worth noting
that the second canonical variable for the ﬁeld variables, at wk 10, had very strong
positive correlations with all the ﬁeld GREF variables. These data are in agreement

with the initial observations that the post-peak GREF estimates were more precise.

3. V .5. .3. V 5.8 .800. v ken...

 

3 . 84 . . . 3 *—
ha. a. 8. 8 can
mv. 50. no. “N ass
:3. an an. to N nae—o

33:65 39 a a
I .5

e332;

22325 33 .o a

0—;

33:50 39... e
o as

8. .v
«N. n
hm. N
3.5 ~
.03— 520

Eggnog 3 v5

2 .o a... a 58 «5.3 8 25. 2. .8 83.8... $8629 :2» 938.33 vs. sou 82.3 9% 33338 a: :8. .8286 .2 39:.

 

 

 

 

 

 

 

 

 

1
5
8. V .7 .5. v tee .88. v he:
9. 8; 2. 8; no. 84 v
305. a. 8. mm. 32.. a. no. a. «no. a. 3 an. n
nee—o. so. 3. vnd :53. 3. an. and «name. 8. en $6 N
sauna. no. no. :6 238. 8. 3. mad— esano. on. an wed a
95 ouSeoBoa count: .03— 95 cusses; can; .03— 95 ems-880a can; .03— 820
968.550 38.3 R 9513550 33.3 $ 255.550 38.3 I
.1 x3 2 S o M3

 

Esp—1&3.“

12.2 ease-asuéﬁaonoeasasgieoisgﬁuagaaaaag9533;389:8233.239;

52
5.5 CONCLUSIONS

Estimates of energy intake and gross efﬁciency estimates from data obtainable
from ﬁeld data closely approximate measures from energy chamber. Therefore,
establishing a database on energy partitioning and energetic efﬁciency of individual
cows from ﬁeld data may be worthwhile if such measures are desirable for
management and if genetic evaluation of animal’s energetic is desired. The use of
post-peak ﬁeld data to estimate energetic efﬁciency provided more reliable estimates
of energy partitioning than those obtained during peak lactation. This study examined
measures versus estimates, a closer examination would necessarily involve the
partitioning of phenotypic means and variation into genotypic and environmental

means and variation.

6 Comparisons of Genetic Parameters for Energy Intake Estimated from
Energy Chamber and from Field Collected data on the Same Lactating Cows

6.1 ABSTRACT

Measures of energy intake from energy chamber ean be approximated closely by
estimates from ﬁeld collected data according to a study using the same data as in this
one. This study estimated genetic parameters of these measures and of partial energy
requirements from energy chamber and ﬁeld collected data. Data from 67 primiparous
Holstein cows collected at peak and post peak lactation consisted of measures of DMI,
milk yield, BW, metabolizable energy, net energy, and maintenance energy. From
DMI and milk yield, energy partitioning was estimated. Univariate and multivariate
animal models were used to estimate genetic parameters for these energy traits. .
Partial energy requirements were estimated using an animal model which included
covariates of age at calving, milk energy, maintenance energy, and weight change.
(Co)variance components were estimated by a derivative-free REML algorithm.
Genetic and phenotypic variations and heritability estimates in energy intake variables,
at postpeak lactation, were similar from chamber and from ﬁeld data. This was not
always the case at peak lactation. There was little difference in solutions for
covariates with and without animal effects. However only solutions for maintenance

energy from animal models matched literature values.

53

for i

dais

U168!

 

 

 

54
6.2 INTRODUCTION

The potential for increasing milk production through feeding is well appreciated.
Efﬁciency is usually deﬁned as the ratio of output over input or its inverse. Selection
for improved efﬁciency may replace selection for total outputs such as milk yield in
dairy enterprize today and future. Feed consumption data is required in order to
measure efﬁciency. Good knowledge about partial energetic requirements is
fundamental to establishing energy efﬁciency criteria.

Freeman (1967) showed that the direct measure of efﬁciency under commercial
conditions does not seem to be economically feasible. He concluded that, "Selection
for higher milk yield automatically improves feed efﬁciency”. Notwithstanding,
Grieve et al. (1976) and Custodio et al. (1983) examined the relationship between
estimated transmitting ability for milk production and digestibility of dietary
components in Holstein cows. Both studies concluded that digestive ability of a cow
was independent of predicted transmitting ability. Buttazoni and Mao (1989) found
that the genetic correlation between net efﬁciency and production was only 60% . We
can attribute this lack of association to the low variability among cows in digestive
ability.

Van Es (1961), Wagner (1965) and others (Andersen et al., 1959; Saama et al.,
1992', Taylor et al. , 1986) demonstrated that little variability exists among cows in
their ability to digest a given diet, particularly when intakes are standardized.
However, considerable variation exists in maintenance energy requirement (Bauman,

1985; Taylor et al. , 1986) and energy requirement for producing milk (Saama et al. ,

55

1992') in cattle. Korver (1988) reviewed the importance of different components of
efﬁciency in selection programs.

Genetic aspects of feed and energy intake have been studied (Blake and Custodio,
1984; Freeman, 1967; Korver, 1988). Genetic parameters for feed intake (Korver,
1988; Stone et al., 1960) and fwd efﬁciency (Blake and Custodio, 1984; Buttazoni
and Mao, 1989; Hooven et al., 1968), energy intake (Taylor et al., 1981) and energy
efﬁciency (Buttazoni and Mao, 1989) traits of lactating cows also are documented.
These studies indicated that feed and energy efﬁciency are moderately heritable traits.
But genetic estimates can be valid only in data collected from a large number of
animals.

In view of the high cost of calorimetric determinations of energy partitioning,
generating similar information from ﬁeld collected data is highly desirable. Walter
and Mao (1989) estimated net efﬁciency of energy conversion from on-farm data and
found them to be in close agreement with published chamber results. They indicated
that in order to verify these results, similar comparisons involving ﬁeld and chamber
data on the same cows would be desirable. Saama et al. (1992”), using ﬁeld estimates
and chamber measures of energy utilization on the same cows at peak and postpeak
lactation, showing that ﬁeld estimates approximated energy chamber measures
closely, hence supported the validity of using ﬁeld data to approximate chamber
energy measures. However, the efﬁcacy of using ﬁeld data to estimate genetic
parameters for energy efﬁciency needs to be examined. At peak and postpeak

lactation, using energy chamber measures and ﬁeld estimates of energy partitioning on

56

the same cows, the objectives of this study were to make comparisons between
chamber and ﬁeld: 1) genetic parameter estimates for energy utilization traits; 2)
partial energetic efﬁciency and weight change requirements; 3) partial energetic
efﬁciency and weight change requirements with and without animal effects in the
model.

6.3 MATERIAIS AND METHODS

6.3.1 Experimental Procedure and Data

Field collected and energy chamber data on 28 pluriparous Holstein cows were
available from a study which examined the effects of percentages of crude protein and
nitrogen solubility on digestibility, energy and protein balances (Janicki, 1985); herein
referred to as study A. Chamber measures were collected at wk 6, 10, and 14
postpartum. Barn DMI, BW, and milk yield were recorded at wk 5, 7, 9, 11, 13, and
15 . From separate energy balance trials (Holter et al., 1992), study B, ﬁeld and
energy chamber data on 39 primiparous Holstein cows were available. In study B
cows were in the energy chamber at wk 7 and 16 postpartum. Barn DMI, BW, and
production data was recorded at wk 6, 8, 15, and 17 after calving. Diets were as
speciﬁed in (Janicki, 1985; Holter, 1992), energy balance was determined by methods
described in (Saama et al. , 1992‘), and ﬁeld estimates of energy partitioning were
obtained using formula outlined in (Saama, 1992‘).

The variables for analysis were metabolizable energy (ME), net energy (NE),

milk energy (MKE) and maintenance energy (MNT). Estimated from the ﬁeld

57
collected data were estimated ME (eME), estimated NE (eNE), estimated MKE

(eMKE) and estimated MNT (eMNT). Two periods of measurement, peak (PL) and
postpeak (PPL) lactation, were established. From study A, PL data were wk 6
chamber data and averages of wk 5 and 7 ﬁeld data; PPL were wk 14 and averages of
wk 13 and 15 postpartum ﬁeld data. From study B, PL data were wk 7 chamber
measurements and averages of wk 6 and 8 ﬁeld estimates; PPL were wk 16 chamber
data and averages of wk 15 and 17 postpartum ﬁeld data. At PL and PPL, weight
change (WC) was computed as the difference between BW (kg) at wk 5 and 7, and
between wk 13 and 15 from study A. Similarly WC was calculated as the difference
between BW (kg) at wk 6 and 8, and between wk 13 and 15 postpartum from study B

measurements.

6.3.2 Estimation of Genetic Parameters

Using estimates from ﬁeld or measures from energy chamber, within PL and PPL
periods, the ab trait, i = MB, NB, MKE, orMNT‘wasentered oneatatimeinan
animal model (AM) [1]:

Y: = a,+e, [1]
where y, is a vector of 67 observations for the ah trait; a, is a vector of unknown
random effects of 40 sires, 10 dams with records, 49 dams without records and 57

animals without offspring on the an trait which was assumed to be distributed as

N(0,Ao:) where a: is the additive genetic variance of the ith trait and A is the

58

additive genetic relationship matrix between the total of 156 animals; e, is a vector of
67 random residuals for the ab trait corresponding to y and was assumed to be
distributed as,N(o, 103) where a: is the residual variance with

80) - E(a) = Etc) = 0-

Using ﬁeld estimates or chamber measures, MB, NB, MKE and MNT were

entered two at a time, within PL and PPL, in a multivariate AM [2]:
y = Z.a + e [2]

where y, a, and e are as deﬁned in [1]. For a pair of energy intake traits, the

y 0 V R 2.011
random elements in [2] had distribution: e ~ N o R R 0
a 0. 6.21 0 G.

 

 

where,v = R + szz.’ R =1° ® 110,6A = A ® 60 with 2. being an incidence
matrix for the animal effects, A is the numerator relationship matrix of order 156,
Re is residual covariance matrix among measurements or estimates on the same
animal, 00 is covariance matrix for additive genetic effect among measurements or

estimates on the same animal, and ® denotes Kronecker (direct) product.

Within PL and PPL, the estimated genetic and phenotypic covariance matrices

59
from ﬁeld data were compared with those from chamber data using a generalization

of Bartlett’s likelihood ratio test by Box (1949). The variance ratio test was used to

make speciﬁc comparisons between individual variances.

6.3.3 Estimation of Partial Fmergetic Efficiency
Within PL or PPL, partial requirements for MKE, MNT, WC were computed
from an AM [3], analogous to that ﬁtted by Ngwerume and Mao (1992),

y, = b,(Age) + b2(MKB) + b3(MNT') + b,(WC) + a, + e, [3]
where y‘ is NE; 1,” b2, b3 and b4 are partial regression coefﬁcients for age at
calving (months), MKE (Mcal), MNT (Meal), and WC (kg), respectively, with a,
and ‘1 are as deﬁned in [1]. Age at calving is included in [3] beeause of its effect on
nutrient partitioning (Bauman and Currie, 1980; Bauman et al., 1985). (Co)variance
component estimation in [1] and [2] and solutions for b, in [3] were obtained using a
derivative-free REML algorithm described by Meyer (1991). For each run,
convergence was declared when the variance of the log-likelihood function was less
than 10". Sampling errors for individual parameters were estimated using univariate
approximation techniques outlined by Meyer and Hill (1991).

Omitting a, from [3] gave a multiple regression model (MRM), [4] , which was

used to estimate partial energy efﬁciencies ignoring animal effects. Analyses for the

MRM were performed using SAS‘ (1985).

60
6.4 RESULTS AND DISCUSSION

Genetic parameter estimates from [1] and [2] and solutions from [3] and [4] were
used for the purpose of making comparisons between chamber measures and ﬁeld
estimates of ME, NE, MNT, and NIKE. The direct use of these etimates may not be

appropriate due to the very small sample size.

6.4.1 Genetic and Phenotypic Variation

In general, convergence was reached after approximately 30 evaluations of a
mixed model equations. The genetic and phenotypic standard deviations, heritability
estimates and associated standard errors for energy intake traits, at PL and PPL are
shown in Table 14. At PL, genetic variation in ME, MKE, and MNT were not
different from that in eME, eMKE, and eMNT. However, genetic variation in NE
and in eNE was signiﬁcantly different (P < .05). During PPL, genetic variation in
all intake traits estimated from chamber and those estimated from ﬁeld data was very
similar. With the exception of phenotypic variation in MNT and eMNT, all chamber
and ﬁeld energy intake characteristics were comparable (P < .05), at PL.
Notwithstanding, at PPL, phenotypic variation in chamber energy utilization traits and
corresponding ﬁeld traits was not different. Although the heritability estimates for ME
and eME, MNT and eMNT were alike at PL, the heritability estimate for eMKE was
higher than the heritability estimate for MKE. Furthermore, the heritability estimate
for eNE was twice as high as that for NE, at PL. Also, standard errors for heritability

estimates, at PL, tended to be high. Yet at PPL, heritability estimates for all

61

TABLE 14. Additive genetic and phenotypic SD and heritability for energy measure traits at peak
and postpeak lactation.

 

 

 

 

 

 

 

Chamber Field
Additive Additive
genetic SD Pbenotypic Heritability genetic SD Pbenotypic Heritability
Trait' (Meal) SD (Meal) estimate SE (Meal) SD (Meal) estimate SE
peak
ME 1.08 8.50 02 .03 1.05 10.98 01 .02
NE 588' 7.25 .66 .34 3.85 6.42 .36 .41
MKE 4.10 5.38 58 .50 4.84 5.41 .80 .38
MNT .70 1.07‘ .43 .25 .59 .80 54 .25
postpeak
ME 1.02 7.80 .02 .03 .98 9.64 .01 .01
NE 1.09 6.49 .03 .02 .95 5.79 .03 .04
MKE .95 4.33 .05 .06 .92 4.44 .04 .06
MNT .73 1.02 .51 .29 .59 .77 .58 .27

 

1ME = metabolizable energy, MKE - milk energy, MNT =- maintenance energy, and NE = net
energy.
‘Corresponding variance components signiﬁcantly different (P < .05).

chamber traits considered were, in some instances, identical to heritability estimates
for corresponding ﬁeld traits. Buttazoni and Mac (1989) found comparable heritability
estimates of .05;t.37 and .13i.34 for NE and NE for maintenance from single trait
sire models, respectively. No prior heritability estimates for ME and MKE could be

found in the literature.

6.4.2 Genetic and Pbenotypic Correlations
An average of around 240 evaluations of the mixed model equations was required
before reaching convergence. Estimates of genetic and phenotypic correlations for
energy usage traits are given in Table 15 . The estimates were generally consistent

within data source but disparagingly divergent when compared between data sources.

62

.385 .0: u m2 2: $3.28 33:18 n 22 £905 um:— u. 22 £908 233808 a N2.

 

 

 

 

 

 

 

 

ca. 3. 3. we: Nb. no. 2.. on. S. 2.. on. 8. E2
ow. 8.2 ea. 9. 3. 3... mm. 8. ea: we. 3.- 3. $2
No. we. 2. me. R. $6 2.. on. 3. oo. 2. 2.... m2
mm. N— .- 0N. mm. mm. Nm. mm. 2 .- NN. ~o. :. 8. N2
22 ”5.2 m2 m2 22 a2 m2 m2 22 ”5.2 m2 m2 P22 m¥2 m2 m2 28,—.
32% eon—Beau Eer— BASS—U
imam Meek

 

.325 833:3: .388 303.3 32595 36.08 aged—9:8 0303» v:- anwamv 25.3 «none—oboe 03308...— .«e «8:83 .2 mam—<9

63
This may be attributable to the small sample size and its effect on the log-likelihood

surface. This could have led to the possibility of local maxima at the point of
convergence. Groeneveld and Kovac (1990) observed that, for small datasets, multiple
solutions can exist from multivariate derivative-free REML algorithms. The space
around the converged solutions was not investigated.

The genetic and phenotypic (co)variance matrices for the chamber and ﬁeld traits,
at PL and PPL, were signiﬁcantly different (P < .05). Regardless, at PL and PPL,
estimates of genetic and phenotypic correlations between MNT and ME, and MNT
and NE from ﬁeld estimates were in reasonable agreement with those estimated from
chamber measures. The estimate of genetic correlation between NE and MNT at PPL
was much higher than the value of -.3 reported by (Buttazoni, 1989) but gave the

most accurate portrayal of the biological relationship between those two traits.

6.4.3 Partial Energy Requirements

The partial regression coefﬁcients for covariates in AM at PL and PPL are shown
in Table 16 for chamber and ﬁeld data. Although the coefﬁcients for age at calving
from chamber and ﬁeld data were generally in close proximity, they were much
closer at PPL than at PL. While maintenance requirements would consist of the
energy required to maintain and conduct activities related to homeostasis, milk energy
and weight change requirements are usually associated with homeorhesis (Bauman and
Currie, 1980). The requirements for MKE and eMKE, at PL and PPL, and WT and

eMNT, at PL were proximate and within the range of values reported by Walter and

64
Mao (1989) and others (NRC, 1989; Ngwerume and Mao, 1992). At PPL, the

requirements for MNT were higher than requirements for eMNT (P < .05). A
similar trend was observed for WC. In addition, the R2 values were remarkably
higher for PPL analyses. In general, these trends for MKE and age at calving were
not altered by exclusion of animal effects from the underlying statistical model
(Appendix VI).

The partial regression coefﬁcients in Table 17 are from MRM of [4] which
ommitted the animal effects. Visual appraisal of results at PL, reveals only trivial
differences between MRM and AM. The closeness between coefﬁcients for age at
calving and MKE was greater with MRM. Nevertheless, trends for MNT and WC
were reversed by using MRM but magnitude of differences between coefﬁcents from
ﬁeld estimates and chamber measures was consistent, at PL and PPL. Whereas
estimated requirements for MKE from chamber measures using MRM, at PL, and
those from ﬁeld data, at PL, coincided with values published by NRC (1989), it is
worth noting that the estimate for maintenance requirements, at PL, from chamber

measures using AM was the only one that agreed with values reported in the literature
(Walter and Mao, 1989). The theoretical expectation of y under [3] and [4] is the

same but inferences from parameter estimates were not the same.

65

—. V keg-.8.V Ki... .8. V kit... 48. V $21....

 

 

 

 

 

 

 

m8. 1.8m. 2m. 8...- «a. .3... am. am. a... as... 2...?
«S. :33... .8. :2”. a... 2.3.... as. SN. :82. 858.52
m2. tings. n1. ensign. 0.2. 1.3.3. «3. nainmvb. :82. .358 3:2
n8. .8.- .8. .8.- «8. 8o. «8. «8. .38.... M.......8 a on...
mm .680 mm .680 mm .680 mm 680 .86..
5. u .6 2...... .8. r. .6 8.5.6 5. u .6 so... .8... .6 .255
.898. as...

 

.838. .8988 .6- 69. .a 238. 8.8632 2......5. :. 8......>8 .5. .680. 38.2.68“. 8.82m... 1...... .2 .53....

_. V mova- .mo.V k: .5. V $1.... .28. V Kai:

 

 

 

 

 

 

 

.8. NS. NS. 258. SN. .58. z... 3... a... as... 2...;
a8. :33 no... :38... NS. 258.. 9.... .8. :82. 8:833:
00—. anaram... N3. 2.3.1.». as. nice—no. N2. sienna. 232.. 3.58 5.2
“8. §. .8. v8. .8. «8. n8. 8... .38.... 9.3.8 .- on...
.5 .680 mm 680 mm .680 mm .680 .86..
.3. u .6 2...... 5.. n .6 8.5.6 a... n .6 so... .8 u .6 .355
.898. 48..

 

9.58. «8&8. 3. .19. a .33 1......- ... 83.2.8 8. .680. 38.8698 8.828.. 1...... .o. 39:.

66
6.5 CONCLUSIONS

Genetic parameter estimates and partial requirements for energy intake traits from
ﬁeld collected and energy chamber data were quite similar. This similarity was
greater with data collected during postpeak lactation. Accurate measurement of
individual intake and production data is not limited to experimental herds.

Milk recording and management programs can provide individual concentrate-
intake data, especially those systems with automated individual feeders. Forage intake
and testing data can be obtained on a herd basis. In practice, cows are fed according
to milk yield. This may cause a high correlation between feed intake and feed
efﬁciency. Korver (1988) suggested that considering only the ﬁrst 60 days of
lactation, during which cows have a negative energy balance and are fed less
adequately according to production requirements might alleviate this problem. But
direct selection on gross efﬁciency has little advantage (Buttazoni and Mao, 1989;
Korver, 1988). So, for purposes of estimating genetic parameters for net efﬁciency,
the authors suggest using intake and production during 60 to 150 days as these data
provided a closer approximation.

Several formula for estimating energy intake from ﬁeld collected data are
available from the literature. Standards nwds to be established with regard to which
formula to use for prediction. Wide acceptance of such formula can be anticipated if
the statistical properties of these formula are well elucidated. This is a matter that has

received little attention in the literature.

67

There was trivial evidence to suggest estimates of partial energy requirements
from animal models and multiple regression models differed. Including animal effects
in the model reduced the error sums of squares but did not necessary increase
accuracy of estimates. Omitting animal effects led to discrepant estimates of energy
requirements for maintenance. Research is needed to examine the biological and
statistical merits and demerits of using animal models versus multiple regression

models to estimate energy requirements for maintenance.

7 SUMMARY

Energy balance trials involving 34 pluriparous Holstein cows were conducted at
the University of New Hampshire, Durham, during wk 6, 10, and 14 postpartum.
Dietary energy was partitioned by indirect calorimetry. Average percentages in
partitioning of energy intake were in agreement with classical values. With the
exception of energy balance, within cow variation in energy intake traits was low.
The utilization of metabolizable energy for milk energy decreased as lactation
progressed. Evidence from a within period model indicated that milk energy
accounted for a highly signiﬁcant pr0portion of the variation in energy intake and
efﬁciency traits.

Field estimates of energy utilization measures were computed from dry matter
intake, consumed by the 34 Holstein cows, at peak and post peak lactation. Both
energy intake and gross efﬁciency estimates from ﬁeld collected data approximately
closely corresponding measures from the energy chamber. The precision of ﬁeld
estimates was higher at postpeak lactation.

From a separate energy study, energy chamber measures and ﬁeld estimates of
energy intake on 37 primiparous Holstein cows were obtained. Data from the two
studies were merged and genetic parameters for metabolizable energy, net energy,
milk energy, and maintenance energy were computed. Partial energetic requirements
were then estimated from animal models and multiple regression models. Excluding
animal effects from the underlying statistical model did not lead to a change in

estimates for energy requirements. It was veriﬁed that genetic parameter estimates for

68

69

energy intake traits estimated from data obtainable from barns were in close

agreement with those estimated from energy chambers.

8 APPENDICES

Appendix 1: Frequency distributions for study A

TABLE 1.1. Frequency distribution of 34 pluriparous Holstein cows by treatment
and Wk of measurement in early lactation.

 

 

 

Wk postpartum
Treatment‘ Wk 6 Wk 10 Wk 14 Total
High CP - high N 9 9 9 27
Low CP - 9 9 9 27
low N
High CP - low N 8 8 8 24
Low CP - high N 8 8 8 24
Total 34 34 34 102

 

‘ CP = crude protein, N = nitrogen.

70

71

TABLE 1.2. Frequency distribution of 34 pluriparous Holstein cows by parity
group and Wk of measurement in early lactation.

 

 

 

 

Wk postpartum
Parity Wk 6 Wk 10 Wk 14 Total
Lactation = 2 14 41 14 42
lactation = 3 10 10 10 3O
lactation = 4 5 5 15
Lactation = 5 4 4 4 12
Lactation = 7 1 l 1 3
Total 34 34 34 102

 

72

TABLE 1.3. Frequency distribution of 34 pluriparous Holstein cows by month and
Wk of measurement in early lactation.

 

Wk postpartum
Month Wk 6 Wk 10 Wk 14 Total

 

 

 

 

January

February
March

April

May

June

July
August
September
October
November
December
Total

NWNNNMUJWNMt-‘rﬁ
NUJUDNw-hWNUtt—ww
KMMMNNWtBMt—WWN

8f
5’:

 

73

TABLE 1.4. Frequency distribution of 34 pluriparous Holstein cows by season and
Wk of measurement in early lactation.

 

 

 

 

 

 

Wk postpartum
Seasonl Wk 6 Wk 10 Wk 14 Total
Cold 15 12 14 41
Mild 9 12 1 1 32
Warm 10 10 9 29
Total 34 34 34 102

 

1Cold = November to March, Mild = April, May, September, and October, and
Warm = June, July, and August.

TABLE 1.5. Frequency distribution of 34 pluriparous Holstein cows by energy
balance status and Wk of measurement in early lactation.

 

 

 

Wk postpartum
EBl status Wk 6 Wk 10 Wk 14 Total
Negative balance 13 11 8 32
Positive balance 21 23 26 70
Total 34 34 34 102

 

‘EB = energy balance

Appendix 11: Frequency distributions for study B.

TABLE 11.1. Frequency distribution of 51 primiparous Holstein cows by
treatment and Wk of measurement in early lactation.

 

 

 

Wk postpartum
Treatment‘ Wk 7 Wk 16 Total
WCS + Ca-LCFA 19 19 38
WCS 18 18 36
Control 14 14 28
Total 51 51 102

 

lCa-LCFA = calcium salts of long-chain fatty acids and WCS = whole cotton seed

TABLE 11.2. Frequency distribution of 51 pluriparous Holstein cows by parity
group and Wk of measurement in early lactation.

 

 

 

 

 

 

Wk postpartum
Parity Wk 7 Wk 16 Total
Lactation = l 18 18 36
Lactation = 2 11 11 22
lactation = 3 10 10 20
lactation = 4 6 6 12
lactation = 5 2
lactation = 6 3
lactation = 8 l 1
Total 59 27 102

 

74

75

TABLE 11.3. Frequency distribution of 51 primiparous Holstein cows by month
and Wk of measurement in early lactation.

 

 

 

Wk postpartum
Month Wk 7 Wk 16 Total
January 8 8 16
February 3 6 9
March 5 8 13
April 2 2 4
May 1 6 7
June 3 2 5
July 1 3 4
August 6 l 7
September 3 l 4
October 6 6 12
November 7 2 9
December 6 6 12
Total 51 51 102

 

76

TABLE 11.4. Frequency distribution of 51 primiparous Holstein cows by season
and Wk of measurement in early lactation.

 

 

 

 

Wk postpartum
Season1 Wk 7 Wk 16 Total
Cold 29 3O 59
Mild 12 15 27
Warm 10 6 16
Total 51 51 102

 

lCold = November to Match, Mild = April, May, September, and October, and
Warm = June, July, and August.

TABLE 11.5. Frequency distribution of 51 primiparous Holstein cows by energy
balance status and Wk of measurement in early lactation.

 

 

 

 

Wk postpartum
EBl status Wk 7 Wk 16 Total
Negative balance 39 37 76
Positive balance 12 14 26
Total 51 51 102

 

lEB = energy balance

Appendix 111: Critical levels and regression coefficients for effects in within
period model to partition variation in dietary and energy intake traits

TABLE 111.1. Critical levels (P - value) for main effects and partial regression coefﬁcients for
covariates from within period model: selected energy partition measures.

 

 

 

 

P . value b(MNT) b(MKE)
Trait' Wk. Estimate P- Estimate P-
POWmTreatment ParitySeason of b SE value of b SE value
GE 6 .61 .64 .71 -1.30 2.21 .56 1.27 .40 .004 .18
10 .88 .57 .11 2.92 1.69 .10 1.41 .30 .0001 .48
14 .13 .63 .22 3.52 1.58 .04 1.26 .33 .0008 .50
W1 6 .74 .29 .003 .002 .003 .57 .001 .N1 .14 .26
10 .41 .14 .21 .003 .003 .43 . .001 .96 .11
14 .33 .78 .01 .001 .002 .59 .001 .0004 .03 .23
FE 6 .87 .40 .92 -.19 1.05 .86 .39 .19 .05 -.01
10 .37 .37 .11 .94 .93 .32 .46 . 16 .01 .26
14 .85 .48 .48 .71 .85 .41 .52 .18 .01 .16
UE 6 .12 .93 .47 -.03 .08 .66 .04 .01 .02 .20
10 .003 .37 .07 .16 .06 .01 .05 .01 .MI .68
14 .05 .07 .84 .09 .08 .26 .(B .02 09 38
CH‘ 6 .75 .53 .04 .11 . 17 .54 .96 .03 .96 14
10 .93 .79 .64 .11 .12 .37 .001 .02 .WI 23
14 .61 .64 68 .03 .10 .71 .04 .02 .04 03
DE 6 .16 .85 .65 -l.l2 1.38 .43 .88 .25 .002 .28
10 .94 .77 .u 1.97 1.00 .06 .95 .18 .(XJOI .54
14 .02 .55 .21 2.81 1.05 .01 .74 .22 .002 .55

 

'CH. 8 methane energy, DE 8 digestible energy, GE =- gross energy, FE = fecal energy, MKE =
milk energy, MNT = maintenance energy, UE = urinary energy, W1 = water intake.

77

78

888888an "E26888? amazegaoaﬁauiaﬁee88 «...-3.48“. mm.

 

 

 

 

a..- 8. .... 8... e4. ..4. .h. 4.
o... 8. .... 8... 8. 8. 44. ...
8. .885 4N. .... 4.... 8. ..4. 8. e a... ..2m
8. 8. .... 88. 8... 8. ..4. 4.
8. 4n. .... 8. 8. .... 8. o.
8. .88... .... .... .... 8. 8. 8. e ...... ....
o..- 2. ... 8.- 8. 8. .... 4.
8.- .8. ~.. .... 8. 8. on. o.
8.- .88... 8. 4.. .... 8. 8. .4. e .6... ....2
.... 8. .... 8.- 8. 8. 8.- 8 8. 8. 4.
4..- 8. 8. . 8. 8. 4...- 8. 44. .n. ...
8.. 8. .... N8. 8. 4... 8.- 8. on. 8. o ..m
.... 8. 8. 8.- 4~. 8. ... 8. 8. ...... 4.
8. .... .... 4...- nm. 8. 8. 8. 4n. .n. ...
8. 4... 8. 4...- 8. 8. e. .- 2. .... 8. c ....
.~. 4.. 4. .N. 4.. 8. 8. 8. m4 .... 4.
.N. .... 8. 8. 8. an. 8. ..4. 8. .a. ...
8. 8. 8. t. 8. en. ...- 84. 4... on. o ...
8.: mm .. ... 3.: mm .. ... 8.8... be... .888; 83.888
-.. 38...... ... 38...... ...? .....s
.88.... 2.22... 3.2.-..

 

.8568... 8.5.!— 388 3.8.8 ".038 veto.—
..EE: BE... 83.38 .0. 38.0508 8.82m». 3...!— 65 Beebe 5!: he. .31., - k. 326. 102.5 .N.E mums.

79

ggumb

.8838 8883818 u ...zs. .888 ....a u 8.3. .888 388 n mu .8......» .48. u 8. .888 8.888... u .5.

 

 

 

 

4.. 8. ... 88. 88.- ... 48. 8. 4.
8.- 84. . . 8. 88. . 88. 8. 8. ...
8. 4... .8... .88.- 8. 88... 88. .... 84. 8. 8 8.536
4...- 8. . . 88. .8... . 8. 8.. 8. 4.
84. 8. . . 8.. .8... .8... .... 8.. 8... ...
8.- 8. . . 8. ...8. . 8. 88. 84. 8 men-5
48. 88... 8.. 8.8 .... 2.. 8. 4.
44. .8... 8... ...n 8. 8. 8. ...
... 8. 8.... 48.4 .885 8. 8. 88. 8 382...:
88. 8... .8. 88. 8. 8. 84. 4.
..4. 48... 88. 8... 8.. 8. 8. ...
8. 8. ...... 8... .885 ..8. 88. 8. 8 382825
8.- 84. .8... .8... 88. ...... .8.- 8. 8... 8.. 4.
8. 8... ...8. . 8... .8. .8...- 8. 8. 8. ...
8.- 8... .8... . 8. .8. 8....- .8. 8. .... e mom...
8.: m8 8 .8 3.: mm 8 .8 8888 8.88.8585 9.8.8
... 9.68m -.. 3.58m ...? ...e...
.85... ...-22... 8.27.8

 

.80-5888 828.050 33.8 "—3.08 veto.—

..85 82. 83:88 a. 38.8888 8.88.82 .58 ...... 588 8...: 8. .8.: - m. ..oB. .8...... .8.... .39;

Appendix IV: Product moment and rank correlations between energy and
efﬁciency measures during early lactation

80

.§3§§ @3358 u “22093.?me 33:58 u $20.52? n 940 .5
3333:3508 Beauooumzo .355 8535.33. u 822 .355 Mac. u a: .355 .8 I m2 .355 33356 a mi

 

 

 

 

 

 

 

 

 

 

 

 

a. 3. 3. 3. 2.. 3.- 3.- a. 3.- 3.- 3.- 3. 3.- 3:3.- $5.22»
8. 3. 3. 3. 3. 3.- 3.- n. 3.- 3.- 3.- 3. 3.- 3.33.3202.on
S. 3. a. 8. 3. on. an. 32. 3. 2.- 2.- 8. 8. 8. 8. mZoBu—zo
3. 3. a. 8. 3. on. an. 2. 3. 2.- 2.- 8. 8.. 8. 8. mzoauzo 2mm—
3. S. 8.- 8.- 3. 8.- 3. 3.- 3.- 3.- 3. 8.- .8.-F.- Mia-z:
3. on. 3.- 3.- 3. 8.- 8.- 2. 3.- 3.- 3.- 3. G.- 2..-3.- Mia-z:
3.- 3. on. 3. 8.- : 3. 8.- 3. 3. 3. a. 8. 2. mzaxz
3.- 3.- an. on. :- 8: 8. 8.- 3. 3. 3. 8.- 5. _. 8. mix—g
2. 2.- 2. S. 8. 8 8. 8. 3. 3. 3. 3. 2. 3.3. .52»
9..- 3.- 3. 3. 8. 3.. 3. 3. 3. 3. 3. 3. 8. 3. on. 8.2»
3.- 3.- 3.- 3.- mm. on.- 3. 3. S. 3. a. 2. 3. 3. 3. m2»
3.- S: 3.- 3.- 33.- on 3. 3. .n. 3. 8. 2. 3. 3.3. m2» elnﬁ
2. 2. _. _. 3. 3. 8. 8. a. 3. 3. 3. 32. 3. 3. .322
a... m- at 3. 3 33.- a. 3. 3. 3. an. 3. 2. an. 3. m2:
2.. a; 8. 8. 8.- a. 8. 8. 3. 3. 3. 3. 3. 3. 3. m2
3.- 3.- 8. 8. a S. _. S. c... 8. 2.. 3. 3. 3. 3. ngﬂu

 

 

 

 

NZo 929 NZo mic
2.220 \FZZoxmv—Zo 320

m2 m2 m2 m:

2.22 2.22 a: a:

F220 mu—Eo m2» m2»

P22 mg: m2 m2

 

 

20E

.8620

Eur—

.3520

 

.Sn I 5 cacao: >15 cm 2.60 083 05 no 83 82.. 505350 .33 .8.—o5
..o 5.35:. 595:0 R55 .33 Soc 52.5.. 33558 92.530: :5.— vs Gig-5&3 gas-8:2 :55 4.2 339

81

chm-30500553582035 u 9220.05.55M—m6v85 u $20.52? a m20£m5
0333:3803: 355nm20£w5500c§ u 22 .3552... u Maw—2.3.5530: I m2..350_§355 u m2

 

a. in. in.
a. 3”. 3.”.
an. on. a.

an. on. 8.

mm. mm. 5.. ho.-
wm. mm. no. 3.-
8.- 8.- mm. mm.
8.- 8... an. an.
mm. mm. n—. 3.
2n..- «n.. we. 8.
an: 5.- 3.- Z”.-
8: 3.- Zn.- 2”.-
mn. mm. —. a.
on; on; n. n.
3.- 3..- 3. no.
nn.- nnf S.- 5.-

3. 33. 3.- 8. 3. 3.- 3.- 3.- 3. 33.- 3:3.- 3.22.322...

3. 33. 8.- 3. 3.- 3.- 3.- 3. 3.- 3.33.- mzonzo

2. 3. 3. 3. 3. 3. 2.- 37 8. 3n. 3. 2.- mzoaazo

2. 3. 3. 3. _. 3. 32.- 32.- 8. 3n. 3. 2.- 3.2.33.2» Hula
8. 3. 3. n. 33.- 3.- 3.- 8. 3.- 8.33.- szz:

3.... 3. 3. 3. 3.- 3.- 3.- 3. 3.- 3,-3.- 3.22.22

2. 3_. R. :. 3. 3. 3. :. F. 2:8; mzaxz

32. 2. 8. _. 3. 3. 3. 32. 3. 2.2.- mgzg

3. 3. 2. 2. 33. 3. 8. 3. 3. 3_. 32. ES»

3. 3. 33. 3. 3. 3. 3. 3. 3. 3. 3. 8.2»

3.- 3.- 2. 8. 33. 3. 3. 8. 3. 3. 33. m2»

3.- 3.- 2. 8. 33. n. 3. 8. 3. 3. 33. m2» 33.-m

33. 3. 2. t. 33. 33. 3. 3. 3. 8. 8. .322

3..- 3..- 3. 33. 3. 3. 3m. 3m. 33. 3. 3. 3.2

3.- 3.- 2.- 3.- a. 3. 3. 3. 8. 3. 8. 32

8.- 3.- 7 2.- 33. 33. 3. 3. 8. 3. 8. 3238453

 

m20 M20 m20 N20
3.220 2.220520 320

m2 m2 m2 m2 P220 No.20 m20 N20 222 E2 m2 m2
2.22 2.2252 52

 

20E

83.520 new. .3520

 

.Aonu 58%.50535308305586602223080

vs .3305 .3 82.58 59820 3.305 vs 303-. 5233 3:339:00 SST—030: x5.— vca 2.3—3730353 «5.3.9.3092 .~.>— mAmSh

82

.§§§I2§§.§é§§ua20.§g§um%.§
8§§§IN§.B§§§ I 22 Saga—«E I 52.330608: I N2.§0_§3806 I N2

 

a. «m. m.
a. um. m.

hm. hm. a.

hm. hm. a.

he. he. No. 8.
mh. mh. no. 8.

m.- m.- «n. E.
oNr oNr hm. hm.

3. 3. N. N.

wnr NW. 0*. 0*.
o6 6.- «N: ”N.-
a.- a: wN.- «N.-

2. N. NN. NN.
mh.- mm: mm. mm.
on; On; no. we.

*0.- 3: No. No.

an.
an.
m—.
Z.

a.
NNI
mm.-
2.
Nu.-
on.-
on.-
I.
he.-
an.-
mm.-

mh.
mh.
N".
—.
ma.

Nm.-
hm.-
m2.
mm.-
no.-
no.-
3.
Nh.-
mm.-
hm.-

mm.-
mm.-
m.
n.
8.-
hN.-

hm.
mo.-
3.
NN.
”N.
3.-
m.
2 .
MN.

—m.-
um.-
Nn.
Nn.
N.-
Nm.-
hm.

he.

8.3

3.
hN.
Nm.

QN.
0N.
mm.
cm.
wN.
_m.

I .-

N— .-

Nm.
VN.
«N.
no.
N_.

m.

Nm.

3.-
3.-
at.
at.
mm.-
3.-
o.
5.
_m.

~h.
~h.
Nm.
va.
3.
h.

an...
an.-
2.-
3.-
av.-
$.-
~m.
«N.
2.
mh.

nu.-
ma.-
3.-
2.-
3.-
e.-
«N.
«N.
3 .
Nh.
3.

"N.
am.
we.
mh.

“N.
~N.
Nm.

3.
«N.
um.

5833.332n38

3n- 33.-3.- 329322»
33.- 3...- 3; 320322»
3. 3.- 33. 3205.2»
3. 3o. 3. 32035—2» 23%

mm: hh.- h.-
no.- mh.- ah.-
ah. no. mN.
Na. 3. hN.
ah. mm. 3.
an. 0. uh.
an. o. mh.
ho. 1N. hN.

No. Nh.
2h. am.
ah. 3.

322.22
322.22
323.2
323.2 Emma
.22»
mu2o
32»
m2... Elem
.322

N2
N2 383

 

N20 N20 N20 N20
32202-220220 520

N2 N2 N2 N2
2.22 2.2252 N22

220 N220 N20 N20

22 N22 N2 N2

 

30E

.3530

30mm

.2820

 

..oml cvgaﬁuao—xiﬁgoﬁlﬂzgsgag0
gguogggﬁqﬁggggaoﬁ—DE8Eggciavgoﬁuﬁag .m.>—NAN<.—.

83

.338333323 "220%.?32530880 ﬂ $20.?5ﬁca8258 I m20£uuoﬁ
ﬁgégum§.§§g I #22 SERRA—ME u 32.906030: ﬂ mz.§£§3808 ﬂ m2

 

 

8. 8. 8. 8. 8. 2.- 2.- 8. 8.- 8.- 8.- 2. 8.- 8.- 8.- mzonzo
8. 8. 8. n. 8. 2.- 2.- 8. 8.- 8.- 8.- :. 8.- 3-8.8252...»
8. 8. 8. 8. 8.- 8. 8. 8. 8. 8.- 8.- 8. 8. 3. mzoauzo
8. 8. 8. 8.- 8. 8. 8. o. 8.- 8.- 8. 8. 8. 8. mzoauzo vlvﬂ
8. 8. 8.- 8.- 8. :. 8.- 8.- 8.- 8.- 8.- 2. 3..- 8.- z..- szz:
8. 8. 8.- 8.- 8. 8.- 8.- 2.- 3.- 8.- 8.- 8. 8.- 2..- 8: 8,522
2.- 2. 8. n. :. 2.- 8. 8.- 8. :. :. 2.- 8. 2.- 2. mama:
2.- 2.- 8. 8. 8.- 3.- 8. 2.- 3. 8. 8. 8.- P. 8. 3.- mixgﬂd
3.- 3.- 8.- 8.- :.- 8.- 2.- 8.- 8. 8. 8. 8. 2. n. 8. .52»
8.- 8.- 3. 3. 8.- 8.- 8. 8. 8. 8. 8. 8. 8. 8. 8. 8.2»
8.- 8.- 8.- 2..- 8.- 8.- 8. 8. 8. 8. 8. 8. 8. 8. m2»
8.- 8.- 8.- 8.- 8.- 8.- 8. 8. v. 8. 8. 8. 8. 8. m2... ﬂu
8.- 8.- 2. :. 8.- 2.- 8. 8. 8. 8. 8. 8. 8. ...z:
8.- 8.- 3. 8. 8.- 8.- R. F. N. 8. 8. 8. z. 8. 8. 8.2
v.- 2..- 8.- 8.- S.- 8: 2.- 8.- 8. n. 8. 8. 8. 8. 8. m2
...- ? 8. 8. 8.- 8.- 8. 8.- 8. R. 2.. 2.. 8. 8. 8. m2 weds-Eu
m2» m2» m2» m2» m2 m2 m2 m2 .rzzo 5.2» m2» m2» 9228.2 m2 m2

2.220 2.220520 320

$.22 $22 a2 a2

 

30E

Esau

30mm

8536

 

.22 I 53-inch: «38.38%0538353 52.3080
98 .8.—0:02 .05-86 .8955 .385 98 20¢ .8233 2.31058 Gig-.833 A5,. 98 cﬁgv 50608-892 .m.>— Emir

Appendix V: Canonical Correlation Analysis

Canonical correlation analysis was used to relate energy chamber values to ﬁeld
estimates on the same lactating cows. The objectives of CCA, in this project were to
ﬁnd linear combinations that produce maximum correlation between linear
combinations of the energy chamber and ﬁeld variables; and to look at the pattern of

association between the two sets of variables.

Pmceduros for Canonical Correlation Analysis

Assuming that Xl, ..., X, and Y,, ..., Ya are two sets of random variables. Let X

be a set of ﬁeld estimates. Deﬁne X as [X,, ..., X,], the predictor variables and X ~
MVN(U,, 2,). Let Y be a set of energy chamber measures. Deﬁne Y as [Yb ..., Yq],

the outcome variables and Y ~ MVN(U,, 2,). After X and Y are partitioned into the

energy intake and gross efﬁciency variables, let the Pearson correlation matrix,,lg7 of

all these intake or gross efﬁciency variables be

 

 

X Ru I R12 p
ny = R _ a ’
Y 32: l R22 ‘1
P q

where Rn contains intercorrelations among the ﬁeld variables, R22 is the
intercorrelations among the calorimetn'c variables, and R21 = Rl2 cross-correlations

between the chamber and ﬁeld variables. If X and Y are of full rank, then deﬁne the

34

85

pxp matrix,G of rank,k and a qxq matrix D as,

G = RIB Rake-21 12;, and

D ‘ Rz-lenRt-llkz’r
Because both G and D are non-symmetric matrices of the form, E“ H to decompose
either G or,D deﬁne F as the upper triangular Cholesky decomposition of E“.
Let,p = FHF’ then obtain the A eigenvalues and W eigenvectors of P. Let

V = F’W. The diagonal elements of, A lip-«1;: are the nonzero latent roots of

5“]! and the columns of V are the orthonomal latent vectors of E411. Note that the

eigenvalues of G and D are equal.
Let A contain the latent vectors of G. Similarly, let B contain the
latent vectors of D . Hence, A are canonical coefﬁcients or weights for the chamber

variables, 3 are the canonical coefﬁcients for the ﬁeld variables, and the diagonals of A

are the squared canonical correlations (R3) between the two sets of variables.

k
Observe that, "(10‘ 2 N3 x! >)‘2>"'>)‘v gives the total variance.

t-l

Form, U, = x4, a linear combination of the ﬁeld, and, V] = YB, a linear

combination of the calorimetric variables, such that the correlation between (J, and VI

86

is maximized. These linear combinations are the canonical variates. We are

interested in the correlations between these canonical variates. It follows immediately

that because A and B are orthomomal, the correlation matrix of U‘ and V1 is,
U '1, | A W
R _ x

I
V LA I I‘ 4

 

 

 

Therefore, U are the canonical variates of the ﬁeld variables and V are the canonical

variates of the energy chamber variables. Thus, ”1 is the ﬁrst canonical variate of the

ﬁeld estimates, and V1 is the ﬁrst canonical variable of the chamber measures; U2

will be the second, and so on.

If al and V1 have the maximum canonical correlation of all linear
combinations, then (up V1) are the ﬁrst pair of canonical variates, which are
independent, i.e., the correlation between (”9 V) and (UP V1) is zero. The
correlation between (U1, V1) would be the ﬁrst canonical correlation and is given by
A, . The variable-variate correlations between [0,; (x1, ..., X,“ and
[V,; (Y,, ..., Y0] are the canonical factors or factor loadings. Thus, the entire

relationship between F ﬁeld variables and q calorimetric variables is expressed only

in terms of k parameters 11, 12, ..., 1,. Hence the name canonical correlations.

87

TABLE v.1. Cumulative proportions of standardized variance of the chamber energy
intake measures explained by the chamber and ﬁeld canonical variables at wk 6, 10

andl4postpartum.

 

Wk 6 Wk 10 Wk 14
Proportion explained Proportion explained Proportion explained
by by by
Order Chamber Field Chamber Field Chamber Field
1 .46 .42 .26 .24 .52 .44
2 .67 .57 .78 .66 .79 .65
3 .84 .67 .89 .72 .96 .69
4 1 .67 l .72 1 .69

 

TABLE v.2. Cumulative proportions of standardized variance of the chamber gross
efﬁciency measures explained by the chamber and ﬁeld canonical variables at wk 6,

10 and 14 postpartum.

 

Wk 6 Wk 10 Wk 14
Proportion explained Proportion explained Proportion explained
by by by
Order Chamber Field Chamber Field Chamber Field

88

Suuoao 8: u m2?! 33.360 35.3.33— " P22 Saga—ME u mm: $328 £33808 u m2.

 

 

 

 

 

 

 

mm. mm. :- 8. n. :- nmr :- 3: mm. m2 \ PZE

3. w. :- 3. 3. 2.. am: 3.. am.- Nn. m2 \ p.22

3.. 3.- Na. 9: v. a: _N.- a: X”.- 5: m2 \ mum}

R. E..- N». 3: an. an: _N.- n».- um: 3.. m2 \ mum:
_ _ a _ N _ N _ N _ 2E9 \ EEO

23m Baa-AU 20E Beau—U 20E 8:830
and? $833 @9539..—
2 a? 2 M? e a?

 

 

date-8 3 e5 2 .e
a; a 85.88 3865.. sea a: Co see .933: as 8. music. 20¢ a: .265 .v.> mam-:-

SMuoco B: I m2 v5 .335 3:353: u P22 chm-8:0 USE I a: 33.85 33380:. I m2.

 

 

:. - S. S. 2.- .- n. 3. 8. mu. 8. 8. 6. t. a. 3. 2.. ...z:
a. 8. n. 2. .- t 2.- S. 2. 8. a. 8.- 3. mh.-R. n. 3.8. 8.2
8.- n. m». 8.- an. S. F- 8. 2. :- 8. N. 2..- a; =- 3: 8.4“. mz
8.- n no. 3.- an. E. t..- G. S. 2.. me. «a. ﬂ.- 3- u. no.- 3- an. us.

a a _ n N _ ... N a n N _ n u _ n N a 2:?qu
as". .2520 so: .3620 as". .3530

$52.3 3533 3533
:5 2 as, :3

 

 

559.8 z a: 2 .e as. a .238... sis 908 as no :8. «Reg. 2.. 8.. .9533 :3 a: .255 ...-5 Sai-

Appendix VI: Mean Regression Coefficients in multiple regression models to
estimate partial energetic coefﬁcients.

Multiple regression models for metabolizable and net energy intake were similar

to those analyzed by Walter and Mao (1989).

NE], = D0 + b,SCM, + e,

MEI, = b0 + b,SCM, + e,

NE], = b0 + b,SCM, + szgtChng, + e,

NEI/MBW, = b0 + b,(SCM,/MBW,) + e,

NEIIMBW, = b0 + b,(SCM,/MBW,) + b2(WgtChng/MBW,) + e,
N5], = bo + b,NESC,,, + e,

NEI, = b0 + b,NE,C,,, + 1’2“”?me + e,

MEI/MBW, = b0 + b,(NE,C,,jMBW,) + e,

MEI/MBW, = b0 + b,(NE,C,, J(M13 W) + awe,“ 1MB 1",) + e,

89

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

Two-stage multiple regression for net energy intake were:

Stage 1: NEI/MBW, -.- b0 + “”153qu + NEWp/MBW, + e,

Stage 2: (NEI, - boMB ivy-SCM, = b, + b,(WgtChng/SCM,) + e,’ [10]

Stage 1: Nat/SCM, = b, + “NEH," + NEWMJSCMQ + a,

Stage 2: (NEI, - b,SCM,)/MBW, = b0 + b2(WgtChng/MBW,) + e,’ [11]

91

dogggggmboﬁﬁuﬁuSSEBm ounce: u gm?"
.ﬁEeBBCoa-«ES u 209

 

 

 

 

23.82
a? a .m «E. 88: BE
88.: 53 A83 38¢ 3:3 9:3 53 3.3 69¢ 88:82
83. an. 83 own. as. «R. mom: «3.8 ted. 8331?
3.3 53 53 an; 33 A23 and Ans-mi 53¢
Se. 5.- xz. 8m. 8m. 9.... 892 «3.2 33a 2255
33 an: $3 5.3 53 803 $2.9 $8.6 88d
n2.- 8? 8m.- Ge. as. «an. 8&2 «32 8.2a 88303
53 53 A83 and 83.8 6:5
”a. so. «we. 8de 8%: En? 32m "E
$3 33 Ex: Gov-.9 5nd 62.5
«2. 5. m8. ”3.8 v8.8 :28 8553
6:3 2:3 6:: 823 A83 Gene 83:82
a. can. «8. 3.68 Sosa BEN 2.8.13
53 53 Ex: 2%.: 63.9 83.9
2m. 8m. «3.. 22.: 83: «8.8 2258
an: at; an: ass 8&6 28.:
one. a8. 3... 28.2 «8.: 33m Saga:
- + e - + a - + m is:
"35033? $63... 3

 

.8853 3 56% 8. mm Amy venom «58838:. 8:3 05 888. v:- EBB A+v 2660a v:- A-vozuewoa we aver—om 85

.888 03$? .8 928 .0: 3 @2388 035....» 33393 55 3038 E 836503 86838 :82 ._ .~> ant-b

92

ﬁggggggugﬁﬁguoBEBmowgzuugﬁ?

.8? 38:33.88 .... 28.

 

 

 

23882
83. 83 32.. 88. 38: BE
38.: 38¢ 33¢ 38¢ 38¢ 38¢ 38¢ 38¢ 38¢ 38: 82
83. 38.- 83 88. «3. 8t 8:. 5. E. vs. .313
38¢ 32¢ 32¢ 32¢ 32¢ 333 38¢ 38¢ 38¢
v8. 88. 8m. 8m. 8n. :3. 3:. 8:. 8". 3oz "5
38¢ 32¢ 38¢ 38¢ 38¢ 38¢ 38¢ 38¢
88.- 83.- 88. 38. 8... 8:. 9:. 88. 58.6 "E
38¢ 38¢ 38¢ 38¢ 38¢ 38¢ 38: 82
En. 8v. 8... 3:. «2. E. a... .935
32 ¢ 32¢ 32¢ 38¢ 38¢ 38¢
8... 8m. :1 2:. 8:. 3:. 32m "E
38¢ 38¢ 38¢ 38¢ 38¢ 38¢
88. 38. 5.. 3. E. .8. .255 ”E
- + m - + a - + m 85:
”398323 .5035 SEE?

 

.885g E .555 08 mm AB v2.89 «588338 235 05 8030- a:- 355 T5 95809 an. A-¢o>uamou .«o «vow—om .85 3m:
\ 358 .8 a 3533 2.35 .8883. as, :3 3m: :8 .2. . s 33.3. 288. a 388808 88882 532 3.3 .59;

93

gtagaggugﬁgsmeBGomgzugz
.8... 38:38:93 u 283.3883 a m2.

 

 

 

 

833.3 8%.— 2388:
3838.2
38¢ 32¢ 38¢ 38¢ 38¢ 88¢ 3.8.: 85 38¢ 38:82
83.3 E. 38. 3:. 38. 32.. 38.3 5.83 38.8 3.8.13
38¢ 88¢ 38¢ 38¢ 38¢ 38¢ 38.8 38.8 38.8
38. 38.- 88. 83.. mg. 8.. 83.3 «2.8 «8.8 22". E
38¢ 38¢ 38¢ 32¢ 32¢ 33¢ 38.3 38.9 38.3
_8. 28.- 88.. 3.8.— v8. 3:.— 838 38.8 818 888.2003.
38¢ 88¢ 38¢ 88.: 88¢ 33¢ 38:82
83.. :3. 83.. 83.8 388... 3883 88.335
32¢ 82¢ 32¢ 88.8 38.9 33.8
38.. 3.8.3 383 88.: 83.8 «8.8 225.3
88¢ 83¢ 88¢ 38.8 88.8 38.:
8:.— 883 88.. 8a.: 83.: 8.8 h.33....88.
- + m - + m - + .— .082
"16...,sz .18sz 3

 

.ggﬁggmwécvoi .ggeuaoagooauagA+vo>u§§—u

33388 ..o 3612— 326 809 333808 an 395888 0383: 2819301 5; £388 3 886508 88858 502 .n.~> SEE

94

 

datum 855983 05 mat—6 «3 N boa... 3 Eu» .8 30— 3m own—o: u ”an?
.38 3253338 u 28 as. 388 .2 u m2.

 

 

 

on? 83.3 83.3 29:32
33...: SE
33¢ 32 ¢ 3: ¢ 38¢ 38¢ 38¢ 32 ¢ 38¢ 38¢ 33...: 82
333 .8.- 34. £3. 38. 38. :13 83.3 8.« B. .335
38¢ 38¢ 33¢ 3«5 :33 33 38¢ 33¢ 38¢
83. 38. 38. 33 83. 333 33.3 awn 38.3 203 "3.
38¢ 38¢ 38¢ 38¢ 38¢ 33¢ 38¢ 38¢ 38¢
38.- _«o. 38.- a: 38.. 82 «3: 33. 83.3 .3836 "3.
38¢ 38¢ 38¢ 3.2 ¢ 33¢ 38¢ 38: 82
E. «8. :3. «2.3 38.3 38.3 B. 323.»
333 38¢ 38¢ 38¢ 38¢ 38¢
23. «.3. 83.3 38... «33 83.3 22.3 "E
32¢ 33¢ 38¢ 38¢ 38¢ 38¢
«3.. :3 83.. 3: «8.3. «$3 3830 "3.
- + m - + a. - + .3 Es:
3.5.32? gamma... ...

 

3805383 5 3305 8: mm .3: votom .393ng 0.5.8 05 898. v:- 3328 A+¢ 9569* 65 3 25:96 me «var—um ~25 3m: \ 38:0

2.33338. 8 32:82. 032.: .8389... as, 5:. 3m: :9. .2. . 2 33.3. 238. a 283508 838932 :82 63> 39:.

95

.Bigaoﬁgu—BN 533:8».583 3m 0953 u man?"
2.4::— vsgmvas u 20m.

 

 

 

84. a.“ 83.. o8. 29582
88: SE
2.8.: 88¢ 88¢ 88¢ 88¢ 88¢ 88¢ 88¢ 88¢ 88: 8:
82 .2. v8.2 888 88. 88. Eu .2. 8.. m8. 3.. 5235
88¢ 88¢ 88¢ 82¢ 22¢ 82¢ 88¢ 28¢ 28¢
88.- 8t- 88.- 88.2 8: 85 3.8. 88.. :3 20E 8 2
82¢ 22¢ 8§¢ 82¢ 82¢ 88¢ 88¢ 88¢ 83¢
88.- $8.- 88.- 8...— 83 ~82 :8. 82 28.. .3830 H2:
82.: 88¢ 88¢ 88¢ 88¢ 88¢ 88¢ 28¢ 88¢ 88: 82
e8.— 88. e8.— 8... 3.. 88. 82. 5. 5. B. .235
82¢ 88¢ 88¢ 88¢ 88¢ 88¢ 88¢ 88¢ 88¢
82 .- v9..- 88.- 98. v82 88.— 8a. :8. 8a. 22... 8:
22¢ 82¢ 88¢ 88¢ 83¢ 28¢ 88¢ 28¢ 88¢
«8.- 88.- 8w. «8.2 .22 8: 88. 88. 88. .3820 H8:
. + m - + .2 - + m .082
"856.83? .8635 ....

 

88528.2. 5 .565 2: mm An: cocoa .5838:— 825 05 8803 was 3.85 A+v 03:89 was 3033?:

he «Bron .85 .358 03833308 3 .355 .0: 3 3.2638 033...; .59-2.3 55 £358 E 886508 €286qu use: .m.—> maﬁa.

9 REFERENCES

Andersen, P. E., J. T. Reid, M. J. Anderson, and J. W. Stroud. 1959. Inﬂuence of
level of intake upon the apparent digestibility of forages and mixed diets by
ruminants. J. Anim. Sci. 18:1299.

Anderson, M. J., D. C. Adams, R. C. Lamb, and J. L. Walters. 1979. Feeding
whole cottonseed to lactating dairy cows. J. Dairy Sci. 62:1098.

Anderson, T. W. 1958. An introduction to multivariate statistical analysis. New
York, Wiley & Sons, Inc. Chapter 12.

Andrew, S. M., H. F. Tyrrell, C. K. Reynolds and R. A. Erdman. 1991. Net
energy for lactation of calcium salts of long-chain fatty acids for cows fed silage-
based diets. J. Dairy Sci. 74:2588.

Andrew, S. M., R. A. Erdman, and D. R. Waldo. 1990. Extent and site of tissue
protein and lipid mobilization during the lactation cycle in dairy cattle. Page 82 in
Proc. Maryland Nutr. Conf., Baltimore.

Armsby, H. P. 1904. The respiration calorimeter at the Pennsylvania Experiment
Station. Expt. St. Record 15:1037.

Armstrong, D.G. and K.L. Blaxter. 1957‘. The heat increment of steam volatile fatty
acids in fasting sheep. Brit. J. Nutr. 11:247.

Armstrong, D. G., and K. L. Blaxter. 1957". The utilization of acetic, propionic and
butyric acids by fattening sheep. Br. J. Nutr. 11:413.

Armstrong, D.G., and K. L. Blaxter. 1961. The utilization of energy of
carbohydrate by ruminants. Proc. 2nd Symp. Energy Metab. EAAP Publ.
10:187.

Armstrong, D.G., K.L. Blaxter and N.M Graham. 1957. The heat increments of
mixtures of steam-volatile fatty acids in fasting sheep. Brit. J. Nutr. 11:392.

Armstrong, D. G., K. L. Blaxter, and R. Waite. 1964. The evaluation of artiﬁcially
dried grass as a source of energy for sheep. III. The prediction of nutritive value
from chemical and biological measurements. J. Agric. Sci. , Camb. 62:417.

Armstrong, D.G., K.L. Blaxter, N.M Graham, and F. W. Wainman. 1961. The

utilization of the energy of two mixtures of steam-volatile fatty acids by fattening
sheep. Brit. J. Nutr. 12:177.

96

97

Atwal, A. 8., and J. D. Erﬂe. 1992. Effects of feeding ﬁsh meal to cows on
digestibility, milk production, and milk composition. J. Dairy Sci. 75 :502-507 .

Bauman, D. E., and W. B. Currie. 1980. Partitioning of nutrients during pregnancy
and lactation: a review of mechanisms involving homeostatis and homeorhesis. J.
Dairy Sci. 63:1514.

Bauman, D. E., S. N. McCutcheon, W. D. Steinhour, P. J. Eppard, and S. J.
Sechen. 1985 . Sources of variation and prospects for improvement of productive
efﬁciency in the dairy cow: A review. J. Anim. Sci. 60:583.

Black, J. L. 1974. Manipulation of body composition through nutrition. Proc.
Australian Soc. Anim. Prod. 10:211.

Blake, R. W. 1979. Research needs to supply milk protein in the human diet. J.
Dairy Sci. 62:1963.

Blake, R. W., and A. A. Custodio. 1984. Feed efﬁciency: a composite trait of dairy
cattle. J. Dairy Sci. 6722075.

Blaxter, K. L. 1961. The utilization of the energy of food by ruminants. Proc. 2nd
Symp. Energy Metab. E.A.A.P. Pub1.10:211.

Blaxter, K. L. 1989. Energy metabolism of ruminants and man. Cambridge Univ.
Press.

Blaxter, K. L., and A. W. Boyne. 1978. The estimation of the nutritive value of
feeds as energy sources for ruminants and the derivation of feeding systems. J.
Agric. Sci., Camb. 90:47.

Blaxter, K. L., and F. W. Wainman. 1964. Use of energy of different rations by
sheep and cattle. J. Agric. Sci., Camb. 63:113.

Boldman, K. G., and L. D. Van Vleck. 1991. Derivative-free restricted maximum
likelihood estimation in animal models with a sparse matrix solver. J. Dairy Sci.
74:4337.

Box, G. E. P. 1949. A general distribution theory for a class of likelihood criteria.
Biometrika 36:317.

Brody, S. 1945 . Bioenergetics and growth. Reinhold Publ. Co. , New York, NY.

98

Brouwer, E., A. J. H. Van Es, and H. J. Nijkamp. 1961. Nutritive value of early
and late cut hay for maintenance of cows. Proc. 2nd Symp. Energy Metab.
E.A.A.P. Publ. 10:153

Brown, C. A., P. T. Chandler, and J. B. Holter. 1977. Development of predictive
equations for milk yield and dry matter intake in lactating cows. J. Dairy Sci.
60: 1739.

Bursztein, S., D. H. Elwyn, J. Askanazi, and J. M. Kinney. 1989. Energy
metabolism, indirect calorumetry, and nutrition. Williams & Wilkins, Baltimore.

Buttazoni, L., and I. L. Mao. 1989. Genetic parameters of estimated net energy
efﬁciencies for milk production, maintenance and body weight change in dairy
cows. J. Dairy Sci. 72:671.

Cardoniga, C. P. , Satter, L. D. 1988. Energy versus protein supplementation of high
alfalfa silage diets for early lactation cows. J. Dairy Sci. 72 (Suppl.
1):298.(Abstr.)

Chalupa, W., B. Vecchiarelli, A. E. Elser, D. S. Kronfeld, D. Sklan, and D. L.
Palmquist. 1986. Ruminal fermentation in vivo as inﬂuenced by long-chain fatty
acids. J. Dairy Sci. 69:1293.

Church, D. C., and W. G. Pond. 1982. Basic Animal Nutrition and Feeding, ed.
John Wiley & Sons. Pages 84-95.

Cooley, W.W., and Lohnes, PR. 1971. Multivariate data analysis. John Wiley &
Sons, Inc., New York, New York.

Coppock, C. E. 1985. Energy nutrition and metabolism of the lactating dairy cow. J.
Dairy Sci. 68:3403.

Custodio, A. A, R. W. Blake, P. F. Dahm, T. C. Cartwright, G. T. Schelling, and
C. E. Coppock. 1983. Relationships between measures of feed efﬁciency and
transmitting ability for milk of Holstein cows. J. Dairy Sci. 66: 1937.

Davey, A. W. F., C. Grainger, D. D. S. Mackenzie, D. S. Flux, G. F. Wilson, I.
M. Brookes and C. W. Holmes. 1983. Nutritional and physiological studies of
differences between Friesian cows of high or low genetic merit. Proc. New
Zealand Soc. Anim. Prod. 43:67.

DePeters, E. 1., S. J. Taylor, A. A. Franke, and A. Aguirre. 1985. Effects of
feeding whole cottonseed on composition of milk. J. Dairy Sci. 68:897.

 

99

DePeters, E. J., S. J. Taylor, C. M. Finley, and T. R. Farnula. 1987. Dietary fat and
nitrogen composition of milk from lactating cows. J. Dairy Sci. 70:1192.

Donker, J. D., and F. A. MacClure. 1982. Responses of milking cows to amounts of
concentrate in rations. J. Dairy Sci. 65:1189.

Edwards, A. W. F. 1961. Likelihood. Camb. Univ. Press.

Elsley, F. W. R. 1976. Limitations to manipulation of growth. Proc. Nutr. Soc.
53:323.

Emmans, G. C., and J. D. Oldham. 1988. Modelling of growth and nutrition in
different species. Pages 13-21 in: Modelling of livestock production systems. S.
Kover and J. A. M. van Arrendonk, ed. Kluwer Academic Publ. Dordrecht,
Netherlands

Ferguson, J. D., J. Torralba, P. L. Schneider, B. Vecchiarelli, D. Sklan, D. S.
Kronfeld, and W. Chalupa. 1988. Response of lactating cows in commercial
dairies to calcium salts of long chain fatty acids. J. Dairy Sci. 71 (Suppl. 1):254
(Abstr.)

Ferretti, L., P. Leone, and V. Sgaramella. 1990. Long range restriction analysis of
the bovine casein genes. Nucleic Acids Res. 18:6829.

Fisher, R. A. 1925. Statistical methods for research workers. Oliver and Boyd,
Edinburgh.

Flatt, W. P. 1966. Energy metabolism results with lactating cows. J. Dairy Sci.
49:230.

Flatt, W. P. , and K. A. Tablet. 1961 . Formulae for computation of open circuit
indirect calorimeter data with electronic data processing equipment. Proc. 2nd
Symp. Energy Metab. E.A.A.P. Publ. 10:39.

Flatt, W. P., C. E. Coppock, and L. A. Moore. 1965. Energy balance studies with
dry, non-pregnant dairy cows consuming pelleted forages. Proc. 3rd Symp.
Energy Metab. E.A.A.P. Publ. 11:131.

Flatt, W. P., P. J. Van Soest, J. F. Sykes, and L. A. Moore. 1958. A description of
the energy metabolism laboratory at the U. S. Department of Agriculture,
Agricultural Research Center in Beltsville, Maryland. Proc. 1st Symp. Energy
Metab. E.A.A.P. Publ. 8:53.

100

Flatt, W. P., P. W. Moe, A. W. Munson, and T. Cooper. 1969. Energy utilization
by high producing dairy cows. 11. Summary of energy balance experiments with
lactating holstein cows. Proc. 4th Symp. Energy Metab. EAAP Publ. 12:35,
Oriel Press, New Castle-Upon-Tyne.

Freeman, A. E. 1975. Genetic variation in nutrition of dairy cattle. Page 19 in The
effect of genetic variance on nutritional requirements of animals. Natl. Acad. Sci.
Washington, DC.

Freeman, A. E. 1967. Genetic aspects of the efﬁciency of nutrient utilization for milk
production. J. Anim. Sci. 26:976.

Gaines, W. L. 1928. The energy basis of measuring milk yield in dairy cows. Illinois
Agr. Expt. Sta., Bull. 308.

Garrett, W. N. 1980. Factors inﬂuencing energetic efﬁciency of beef production. J.
Anim. Sci. 51:1434.

Garrett, W. N., and D. E. Johnson. 1983. Nutritional energetics of ruminants. J.
Anim. Sci. 57 (Suppl. 2):478.

Graser, H. U., S. P. Smith, and B. Tier. 1987. A derivative-free approach for
estimating variance components in animal models by restricted maximum
likelihood. J. Anim. Sci. 64:1362.

Grieve, D. G., G. K. MacLeod, T. R. Batra, E. B. Burnside, and J. B. Stone. 1976.
Relationship of feed intake and ration digestibility to estimated transmitting
ability, body weight and efﬁciency in ﬁrst lactation. J. Dairy Sci. 59:1312.

Grings, E. E., R. E. Rofﬂer, and D. P. Dietelhoff. 1991. Response of dairy cows
in early lactation to additions of cottonseed meal in alfalfa—based diets. J. Dairy
Sci. 74:2580.

Groeneveld, E. , and M. Kovac. 1990. A note on multiple solutions in multivariate
restricted maximum likelihood covariance component estimation. J. Dairy Sci.
73:2221.

Hartley, H. 0., and J. N. K. Rao. 1967. Maximum likelihood estimation for the
mixed model analysis of variance model. Biometrika 54:93.

Harville, D. A. 1977. Maximum likelihood approaches to variance component
estimation and related problems. J. Am. Stat. Assoc. 72:320.

101

Hashizume, T., H. Morimoto, T. Masubuchi, M. Abe, and T. Hamada. 1965. Some
aspects of energy utilization by cows during gestation and lactation. Proc. 3rd
Symp. Energy Metab. E.A.A.P. Publ. 11:111.

Henderson, C. R. 1952. Speciﬁc and general combining ability in heterosis. John. W.
Garver, editor. Iowa State Univ. Press.

Henderson, C. R. 1953. Estimation of variance and covariance components.
Biometrics 9:226.

Henderson, C. R. 1973. Sire evaluation and genetic trends. Page 10 in: Proc. Anim.
Breeding Genet. Symp. in honor of Dr. J. L. Lush. Am. Soc. Anim. Sci. and
Am. Dairy Sci. Assoc., Champlaign, IL.

Henderson, C. R. , and R. L. Quass. 1976. Multiple trait evaluation using relatives’
records. J. Anim. Sci. 43:1188.

Hoffman, P. C., R. R. Grummer, R. D. Shaver, G. A. Broderick and T. R.
Drendel. 1991 . Feeding supplemental fat and undegraded intake protein to early
lactation Dairy Cows. J. Dairy Sci. 74:3468.

Holmes, F. L. 1985. Lavoisier and the chemistry of life. Madison. The University of
Wisconsin Press.

Holter, J. B. 1976. Fasting heat production in “lactating” versus dry dairy cows. J.
Dairy Sci. 59:755

Holter, J. B., H. H. Hayes, W. E. Urban, Jr., and A. H. Duthie. 1992. Energy
balance and lactation response in Holstein cows supplemented with cotton seed
with or without calcium soap. J. Dairy Sci. 75:

Holter, J. B., W. E. Hylton, C. B. Smith, and W. E. Urban, Jr. 1982. Reducing
concentrate feeding for lactating dairy cows. J. Dairy Sci. 65:37.

Hooven, N. W., Miller, R. H., and R. D. Plowman. 1968. Genetic and
environmental relationships among efﬁciency, yield, consumption and weight of
Holstein cows. J. Dairy Sci. 51:1409

Hotelling, H. 1936. Relations between two sets of variates. Biometrika 28:321.

Huffman, C. F. 1961. High-level grain feeding for dairy cows. J. Dairy Sci.
44:2113.

102

Janicki, F. J., J. B. Holter, and H. H. Hayes. 1985. Varying protein content and
nitrogen solubility for pluriparous, lactating Holstein cows: Digestive
performance during early lactation. J. Dairy Sci. 68:1995.

Jenkins, T. C., and D. L. Palmquist. 1982. Effect of added fat and calcium on in
vitro formation of insoluble fatty acid soaps and cell wall digestibility. J. Anm.
Sci. 55:957.

Johnson, R. E. 1980. Techniques for measurement of gas exchange. Pages 32-36 in:
Kinney. J. M., ed. Assessment of energy metabolism in health and disease.
Columbus, Ross laboratories.

Kennedy, G. C. 1967. Ontogeny of mechanisms controlling food and water intake.
Page 337 in: C. F. Code, ed., Handbook of Physiology. Sect. 6 - Alimentary
canal. Vol. 1. Control of food and water intake. Anim. Physiol. Soc.,
Washington, D. C.

Kent, B. A., and M. J. Arambel. 1988. Effect of calcium salts of long chain fatty
acids on dairy cows in early lactation. J. Dairy Sci. 71:2412.

Kesler, E. M., and S. L. Spahr. 1964. Physiological effects of high level concentrate
feeding. J. Dairy Sci 47:1122.

Kirchgessner, M., H. L. Mueller, F. J. Schwarz. 1982. Experimental studies on the
feeding frequencies of dairy cows. Proc. 9th Symp. Energy Metab. E.A.A.P.
Publ. 29:30.

Kleiber, M. 1935. The California apparatus for respiration trials with large animals.
Hilgardia 9:1.

Kleiber, M. 1965. Metabolic body size. Proc. 3rd Symp. Energy Metab. E.A.A.P.
Publ. 11:427.

Knott, J. C., R. E. Hodgson, and E. V. Ellington. 1934. Methods of measuring
pasture yields with dairy eattle. Washington Agric. Exp. Stn. Bull. 295, Pullman.

Korver, S. 1988. Genetic aspects of feed intake and feed efﬁciency in dairy cattle: a
review. Livest. Prod. Sci.: 20:1.

Laben, R. C. 1963. Factors responsible for variation in milk composition. J. Dairy
Sci. 46:1293.

Lawley, D. N. 1959. Tests of signiﬁeance in canonical correlation analysis.
Biometrika 46:59.

103

Lister, D. 1976. Effects of nutrition and genetics on composition of the body. Proc.
Nutr. Soc. 35:351.

Macleod, G. K., D. G. Grieve, and I. McMillan. 1983. Performance of ﬁrst lactation
dairy cows fed complete rations of several ratios of forage to concentrate. J.
Dairy Sci. 66:1668.

Mertz, S. H. M., and S. G. Van den Bergh. 1977. Regulation of fat mobilization in
adipose tissue of dairy cows in the period around parturition. Netherlands J.
Agric. Sci. 55:198.

Meyer, K. 1986. Restricted maximum likelihood to estimate genetic parameters-in
practice. Proc. 3rd World Congr. Genet. Appl. Livest. Prod. Lincoln, NE
12:454.

Meyer, K. 1991. DFREML. Programs to estimate variance components for individual
animal models by restricted maximum likelihood. User notes, Edinburgh Univ.,
Scotland.

Meyer, K. 1990. Present status of knowledge about statistical procedures and
algorithms to estimate variance and covariance components. Proc. 4th World
Congr. Genet. Appl. Livest. Prod. Edinburgh, Scotland XIII:407.

Meyer, K., and W. G. Hill. 1991. Approximation of sampling variances and
conﬁdence intervals for maximum likelihood estimates of variance components (in
press).

Millward, D. J., P. J. Gorlick, and P. J. Reeds. 1976. The energy cost of growth.
Proc. Nutr. Soc. 35:339.

Misztal, I. 1992. Derivative free vs. expectation maximization restricted maximum
likelihood algorithms. J. Dairy Sci. 75 (Suppl. 1):250. (Abstr.)

Mitchell, H. H., T. S. Hamilton, F. J. McClure, W. T. Haines, J. R. Beadles, and
H. P. Morris. The effect of the amount of feed consumed by cattle on the
utilization of its energy contents. J. agr. Res. 45:163.

Moe, P. W. 1981. Energy metabolism. J. Dairy Sci. 64:1121.

Moe, P. W., and H. F. Tyrrell. 1972. Net energy value for lactation of high- and
low-protein diets containing corn silage. J. Dairy Sci. 55:318.

Moe, P. W., and H. F. Tyrrell. 1973. The rationale of various energy systems for
ruminants. J. Dairy Sci. 37:183.

104

Moe, P. W., H. F. Tyrrell, and W. P. Flatt. 1970. Partial efﬁciency of energy use
for maintenance, lactation, body gain and gestation in the dairy cattle. Proc. 4th
Symp. Energy Metab. E.A.A.P. Publ. 13:65.

Moe, P. W., H. F. Tyrrell, and W. P. Flatt. 1971. Energetics of body tissue
mobilization. J. Dairy Sci. 54:548.

Moe, P. W., W. P. Flatt, and H. F. Tyrrell. 1972. Net energy value of feeds for
lactation. J. Dairy Sci. 55:945.

Moe, P. W., W. P. Flatt, and L. A . Moore. 1966. Effect of level of intake on
energy losses by dairy cattle. J. Dairy Sci. 49:714 (Abstr.)

Moore, T. L., and I. L. Mao. 1990. Prediction of total intake of dry matter and net
energy in a lactation. J. Dairy Sci. 73:1255.

Morrow, D. A. 1976. Fat cow syndrome. J. Dairy Sci. 59:1625

National Research Council (NRC). 1989. Nutrient Requirements of Dairy Cattle.
Sixth Rev. Ed. Natl. Update 1989. Acad. Sci., Washington, DC.

Ngwerume, F., and I. L. Mao. 1992. Estimation of residual energy intake for
lactating cows using an animal model. J. Dairy Sci. 75: (submitted: MS #
1354E).

Ohajuruka, O. A., Zhiguo Wu, and D.L. Palmquist. 1991. Ruminal metabolism,
ﬁber, and protein digestion by lactating cows fed calcium soap or animal-
vegetable fat. J. Dairy Sci. 74:2601.

Palmquist, D. L. and Conrad. 1978. High fat rations for dairy cows. Effects on feed
intake, milk and fat production and plasma metabolites. J. Dairy Sci. 61:890.

Palmquist, D. L., and T.C. Jenkins. 1980. Fat in lactation rations: review. J. Dairy
Sci. 63:1.

Patterson, L. . D. , and R. Thompson. 1971 . Recovery of the inter-block information
when block sizes are unequal. Biometrika 58:545.

Rao, C. R. 1971'. Estimation of variance and covariance components - MlNQUE
theory. J. Multivariate Analysis. 1:257-275.

Rao, C. R. 1971". Minimum variance quadratic unbiased estimation of variance
components. J. Multivariate Analysis. 1:445.

105

Reid, J. T. 1974. Energy metabolism in the whole animal. Pages 113-151 in: J. D.
Sink, ed. The Control of metabolism. The Pennsylvania State Univ. Press, Univ.
Park, PA.

Reid, J. T., O. D. White, R. Antique, and A. Fortin. 1980. Nutritional energetics of
livestock: Some present boundaries of knowledge and future research needs. J.
Anim. Sci. 51:1393.

Ritzman, E. G. and F. G. Benedict. 1929. A simpliﬁed technique and apparatus for
measuring energy requirements of cattle. N. H. Agr. Exp. Sta. Bull. 240.

Ritzman, E. G. and F. G. Benedict. 1938. Nutritional physiology of the adult
ruminant, Carnegie Inst., Washington. Publ. 494.

Ritzman, E. G. and N. F. Colovos. 1932. An automatic method for collecting solid
and liquid excreta from cows in digestion experiments. N. H. Agr. Exp. Sta.
Tech. Bull. 52.

Quass, R. L., and E. J. Pollak. 1980. Mixed model methodology for farm and ranch
beef cattle testing programs. J. Anim. Sci. 51:1277.

Robinson, P. H., R. E. McQueen, and P. L. Burgess. 1991. Inﬂuence of rumen
undegradable protein levels on feed intake and milk production of dairy cows. J.
Dairy Sci. 74:1623.

Rogers, G., R. Porter, and 1. Robinson. 1979. The utilization of perennial ryegrass
and white clover by lactating dairy cows. Page 59 in: Research report. Dairy
Res. Inst., Victoria, Australia.

Rogers, J. A., U. Krishnamoorthy, and C. J. Sniffen. 1987. Plasma amino acids
and milk protein production by cows fed rumen-protected methionine and lysine.
J. Dairy Sci. 70:789.

Saama, P. M., I. L. Mao, and J. B. Holter. 1992‘. Sources of variaton in partitioned

uses of energy intake in pluriparous lactating Holstein cows in energy chamber. J.
Dairy Sci. 75: (submitted: MS # 2095B).

Saama, P. M., I. L. Mao, and J. B. Holter. 1992". Comparisons of energy intake and
gross efﬁciency measures from calorimetric and ﬁeld data on the same lactating
cows. J. Dairy Sci. 75: (Submitted MS # ).

SAS‘ISTAT User’s guide. 1985. SAS Institute, Inc., Cary, NC.

106

Satter, L. D., 1986. Protein supply from undegraded dietary protein. J. Dairy Sci.
69:2734.

Schaeffer, L. R., and H. Soong. 1979. Selection bias and REML variance—covariance
component estimation. J. Dairy Sci. 61 (Suppl. l):91. (Abstr.)

Schauff, D. J., and J. H. Clark. 1989. Effects of prilled fatty acids and calcium salts
of fatty acids on rumen fermentation, nutrient digestibilities, milk production,
and milk composition. J. Dairy Sci. 72:917.

Schneider, P., D. Sklan, W. Chalupa, and D. S. Kronfeld. 1988. Feed calcium salts
of fatty acids to lactating cows. J. Dairy Sci. 71:2143.

Smith, N. E., L. S. Collar, D. L. Bath, W. L. Dunkley, and A. A. Franke. 1981.
Digestibility and effects if whole cottonseed fed to lactating cows. J. Dairy Sci.
64:2209.

Sniffen, C. J., and P. H. Robinson. 1987. Microbial growth and ﬂow as inﬂuenced
by dietary manipulations. J. Dairy Sci. 70:425.

Susmel, P., B. Stefanon, C. R. Mills, and M. Candido. 1989. Change in amino acid
composition of different protein sources after rumen incubation. Anim. Prod.
49:375.

Stone, J. B., G. W. Trimberger, C. R. Henderson, J. T. Reid, K. L. Turk, and J. K.
Loosli. 1960. Forage intake and efﬁciency of feed utilization in diary cattle. J.
Dairy Sci. 43:1275.

Susmel, P., B. Stefanon, C.R. Mills, and M. Candido. 1989. Change in amino acid
composition of different protein sources after rumen incubation. Anim. Prod.
49:375.

Taylor, C. S., H. G. Turner, and C. B. Young. 1981. Genetic control of equillibrium
maintenance efﬁciency in cattle. Anim Prod. 33:179.

Taylor, C. S., R. B. Thiessen, and J. Murray. 1986. Inter-breed relationship of
maintenance efﬁciency to milk yield in cattle. Anim. Prod.: 43:37.

Tepperman, J. , and H. M. Tepperrrran. 1970. Gluconeogenesis, lipogenesis and he
sherringtonian metaphor. Fed. proc. 29:1284.

Thorbek, G. 1970. The utilization of metabolizable energy for protein and fat gain in
growing pigs. Proc. 5th Symp. Energy Metab. E.A.A.P. Publ. 12:129.

107

Tyrrell, H. F., and P. W. Moe. 1972. Net energy value for lactation of a high and
low concentrate ration containing corn silage. J. Dairy Sci. 55:1106.

Tyrrell, H. F., and J. T. Reid. 1965. Prediction of the energy value of cow’s milk. J.
Dairy Sci. 48:1215.

Tyrrell, H. F., P. W. Moe, and L. S. Bull. 1973. Energy value of cracked corn and
dried beet pulp. J. Dairy Sci. 5621384.

Tyrrell, H. F., P. W. Moe, and R. R. Oltjen. 1974. Energetics of growth and
fattening compared to lactation in cattle. Proc. 6th Symp. Energy Metab. EAAP
Publ. 14:205.

Van Es, A. J. H. 1961. Between-animal variation in the amount of energy required
for the maintenance of dairy cows. Ph. D. Dissertation. State Agricultural Univ.,
Wageningen, The Netherlands.

Van Es, A. J. H., and H. J. Nijkamp. 1969. Energy, carbon, and nitrogen balance
experiments with lactating cows. Proc. 4th Symp. Energy Metab. E.A.A.P. Publ.
12:209.

Van Karnpen, M. 1987. Climatic conditions and energy metabolism of laying hens.
Page 199 in: Verstegen, M W. A., and A. M. Henken. Energy metabolism in
farm animals. Kluwer Acad. Publ., Boston.

Van Soest, P. J. 1963. Ruminant fat metabolism with particular reference to factors
affecting low fat and feed efﬁciency. A review. J. Dairy Sci. 46:204.

Van Soest, P. J. 1967. Symposium on factors inﬂuencing the voluntary intake of
herbage by ruminants: voluntary intake in relation to chemical composition and
digestibility. J. Anim. Sci. 24:834.

Vercoe, J. E. 1970. Fasting metabolism and heat increment of feeding in Brahman x
British and British cattle. Pages 85-88 in: Schurch, and C. Wenk, ed. Energy
Metabolism of Farm Animals. Juris Druck and Verlag, Zurich.

Wagner, D. G. 1965. Studies on the energy requirements of high producing dairy
cows. Ph.D. Dissertation. Cornell Univ. Ithaca, NY.

Wainman, F. W., and K. L. Blaxter. 1958. Closed circuit apparutus for the cow and
steer. Proc. lst Symp. Energy Metab. E.A.A.P.Publ. 8:80

Waldo, D. R., and N. A. Jorgensen. 1981. Forages for high animal production:
nutritional factors and effects of conservation. J. Dairy Sci. 64: 1207

108

Walter, J. P., and I. L. Mao. 1989. Modelling net energy efﬁciencies as quantitative
characteristics in lactating cows. J. Dairy Sci. 72:2362.

Wilks, D. L., C. E. Coppock, and K.N. Brooks. 1991. Effects of differences in
starch content of diets with whole cottonmd or rice bran on milk casein. J.
Dairy Sci. 74:1314.

Wilmink, J. B. M. 1987. Studies on test-day and lactation milk, fat, and protein yield
of dairy cows. Ph. D Thesis, Agric. University, Wageningen, The Netherlands,

Page 123.

Wohlt, J. B., and J. H. Clark. 1978‘. Nutritional value of urea versus performed
protein for ruminants. 1. Lactation of dairy cows fed corn based diets containing
supplemental nitrogen from urea and/or soybean meal. J. Dairy Sci. 61:902.

Wohlt, J. B., J. H. Clark, and F. S. Blaisdell. 1978”. Nutritional value of urea versus
performed protein for ruminants. II. Nitrogen utilization by dairy cows fed corn
based diets containing supplemental nitrogen from urea and/or soybean meal. J.
Dairy Sci. 61:916.

Wohlt, J. E., S. L. Chmiel, P. K. Zajac, L. Backer, D. B. Blethen, and J. L. Evans.
1991. Dry matter intake, milk yield and composition, and nitrogen use in
Holstein cows fed soybean, ﬁsh, or corn gluten meals. J. Dairy Sci. 74:1609.

Young, B. A. 1976. Effect of cold environments on nutrient requirements of
ruminants. lst Int. Symp. Feed Composition, Anim. Nutrient Requirements, and
Computerization of Diets, Logan, UT.

Yousef, I. M., J. T. Huber, and R. S. Emery. 1970. Milk protein synthesis as
affected by high grain ﬁber rations. J. Dairy Sci. 53:734.