Computer simulation model of swine production systems: II. Modeling body composition and weight of female pigs, fetal development, milk production, and growth of suckling pigs

Abstract

Theoretical concepts and relationships used to develop a deterministic computer simulation model of female pigs during their reproductive life are described. The model predicts, in a continuous form, body composition and weight of female pigs, fetal development, sow milk yield, and growth of suckling pigs according to genotype, diet, and management conditions. The model simulates growth of adult female pigs. Dietary nutrients are used first for maintenance and second for fetal growth or milk production. Any surplus is retained in the body. Energy and protein body reserves are mobilized when a nutrient deficit occurs during gestation or lactation. However, the rates of body protein and fat accretion as well as the fat/protein ratio are limited by boundaries, the values of which depend on the nutritional and physiological status of the pig. The model's ability to simulate sow body weight changes and composition, fetal growth, milk production, and suckling pig's growth is illustrated.

Full text

C. Pomar, D. L. Harris and F. Minvielle
growth of suckling pigs
composition and weight of female pigs, fetal development, milk production, and
Computer simulation model of swine production systems: II. Modeling body
1991, 69:1489-1502.J ANIM SCI
http://jas.fass.org/content/69/4/1489
the World Wide Web at:
The online version of this article, along with updated information and services, is located on
www.asas.org
by guest on July 21, 2011jas.fass.orgDownloaded from

COMPUTER SIMULATION MODEL
OF
SWINE PRODUCTION
SYSTEMS:
II.
MODELING BODY COMPOSITION AND
WEIGHT OF FEMALE PIGS, FETAL DEVELOPMENT,
MILK PRODUCTION, AND GROWTH
OF
SUCKLING PIGS'
Chdido Porn&, Dewey
L.
Harris3 and Francis Minvielle4
UniversitC Lad, Quebec, Canada,
GlK-7P4
and
US.
Department of Agriculture,
ARS,
Clay Center,
NE
68933-0166
ABSTRACT
Theoretical concepts and relationships used to develop a deterministic computer
simulation model of female pigs during their reproductive life are described. The model
predicts, in a continuous form, body composition and weight of female pigs, fetal
development, sow
milk
yield, and growth
of
suckling pigs according to genotype, diet, and
management conditions. The model simulates growth of adult female pigs. Dietary
nutrients
are
used
first
for maintenance and second for fetal
growth
or milk production.
Any surplus is retained in the body. Energy and protein
body
reserves are mobilized when
a nutrient deficit occurs during gestation or lactation. However, the rates of body protein
and fat accretion as well as the favprotein ratio are limited by boundaries, the values of
which depend on the nutritional and physiological
status
of the pig. The model's
ability
to
simulate sow body weight changes and composition, fetal
growth,
milk production, and
suckling pig's growth is illustrated.
Key Words: Reproduction, Models, Simulation, Sows, Pregnancy, Lactation
J.
Anim.
Sci.
1991.
693489-1502
Introduction
Swine production efficiency
is
the result
of
many distinct but interacting factors (environ-
he
authors
wish
to acknowledge the many
sugges-
tions
concerning model elements
from
numerous
discus-
sions with scientists at the Roman
L.
Hruska
U.S.
Meat
Animal
Research Center at Clay Center, NE. These
include G.
L.
Bennett, R.
K.
Christenson,
I.
J.
Ford,
K.
E.
Gregory,
K.
A. Leymaster,
h4.
D.
MacNeil,
R
R. Maurer,
W.
G. Pond and
L.
D.
Young,
and
also
P.
Savoie from
Agriculture Canada at Lennoxville (Quebec) Canada. The
efforts
of
I.
M.
Dzakuma
in
developing the
figures
is
also
appreciated,
as
well
as
those
of
Sherry Kluver for
typing
the
manuscript.
TEis
study
was
pdy supported
by
a
&rant
from CORPAQ-Agriculture.
Qucbec.
2Present address:
Station
de Recherche, Agricdme
Canada,
C.
P.
90,
Lennoxville,
Qu&ec,
Canada,
JIM-123.
SSDA-ARS, Roman
L.
Hruska
US
Meat
Anim.
Res.
Center, Clay Center, NE.
4Present address: INRA-CNRZ, Labomtoire de
Gh.6-
tique Factorielle,
78350,
Jouy-en-Josas,
Prance.
'Dept.
of
Zootechnie,
BAA.
Received November
20,
1989.
Accepted October
16,
1990.
ment, nutrition, genetics, economics, etc.) that
act during both the
growth
period and the
reproductive life of the pig. The complexity
within and among these factors cannot
be
fully
comprehended in a quantitative and dynamic
fashion by either the human mind or by
traditional research (Baldwin,
1976;
Koong et
ai.,
1976;
Whittemore,
1986).
Therefore, sys-
tems analysis techniques by the means
of
simulation modeling are proposed
as
an
essential part of the scientific method (Whitte-
more,
1986)
because they allow
an
analysis
of
the whole system and the interactions
of
its
components (Koong et
al.,
1976).
This
paper describes a deterministic life-
cycle model that includes and extends the
young pig growth model of Pomar et al.
(1991).
The combined model simulates the
overall individual pig's life. The objective of
this
study was to develop a swine model that
would
be
more mechanistic than previous ones
(Men and Stewart,
1983;
Tess et
al.,
1983;
Black et al.,
1986;
Pettigrew et al.,
1986;
Singh,
1986)
in order to ensure a reliable
1489
by guest on July 21, 2011jas.fass.orgDownloaded from

1490
POMAR
ET
AL.
response into a wide range of combined
nutritional, genetic, and managerial situations.
This
goal required a model that integrates
fundamental biological processes regulating
the accretion of body protein as well as energy
and protein metabolism during growth of
female pigs, gestation, and lactation based on
theoretical concepts for
the
independent estim-
ation of lipid and protein accretions or losses.
The model should also incorporate a large
number of factors affecting the efficiency
of
swine production systems. Among these fac-
tors, genotypic and nutritional effects and their
interactions with growth, body composition,
and the productivity of the sow were required.
Theoretical concepts and relationships pre-
sented here constitute the second part of the
framework for a complete life-cycle model of
pigs.
Model
Description
Strategy
The simulation model described here
predicts body composition and weight of
female pigs during the overall reproductive
life, along with fetal development,
milk
pro-
duction, and suckling pig growth. Genotype
parameters, diet composition, reproductive per-
formance, and management alternatives
are
model input variables. Rate variables are
expressed on a
daily
basis, energy is in
megajoules
(MJ)
and mass is in kilograms (kg)
when not specified in the text. The general
strategy used to model
growth
and composi-
tion of adult animals and nursing pigs is based
on the one previously reported (Pomar et al.,
1991)
with extensions to include the additional
physiological processes.
Body
Composition and Weight
of
Adult
Female Pigs
As
for young animals (Pomar et al.,
(1991),
empty BW of
an
adult sow
is
assumed to
be
95%
of its total body live weight and is
defined as the algebraic sum of its main
chemical components, which include the total
body
mass
of protein
0,
lipids
(LT),
water,
and ash. These and other abbreviations
are
presented in Table
1.
It
is
assumed in the
model that
all
diets
are
palatable
and
ade-
quately balanced in
all
known
nutrients (in-
cluding minerals, vitamins, and
all
amino
acids
TABLE
1.
ABBREVLATIONS
AND
ACRONYMS
USED
IN
THE TEXT
_______~~
Symbol
Meaning
A
B
C
DE
e
EM
IPI
m
LT
LTP
LTpr
LTr
M
PP
PPr
FT
PTP
m
PTr
t
TBW
TLW
constant
constant
constant
Digestible
energy
inWd
Natural
logarithm
base
Energy
requirements for maiatenance
Ideal protein intake
Intrinsic
potential
for protein accretion
Total body
mass
of lipids
Body fat
mass
of
the
fetus and
suckling
pigs
Fat accretion
rate
of the fetus
and
suckling
pigs
Body lipid accretion rate
constant
Total
body
protein precursor
mass:
Protein precursor accretion rate
Total body protein
mass
Body protein
mass
of the fetus
and
suckling
Protein accretion rate of the fetus and suckling
Body protein accretion rate
Time
Total litter weight at
birth
Total litter weight
approximation for DNA,
Pigs
Pigs
y(t) Average daily
milk
yield
except lysine). Total body protein precursor
mass (PP) used in
this
model
is
an approxima-
tion of the total empty body deoxyribonucleic
acid (Pomar et al.,
1991).
The intrhsic
potential for protein accretion
(IPTr)
is the
maximal amount of protein that an animal can
retain
in
a day when there is no external
(mainly nutritional) limitation. The terms PPr,
PTr, and LTr represent, respectively,
the
amounts of PP,
PT,
and LT mass retained in
each integration step.
The estimation of IPTr for adult animals is
done in the same way
as
for young pigs. Under
conditions of normal growth and when extrin-
sic factors, such as nutrition,
are
not limiting
the retention of body protein, adult pigs can
retain as much protein as IPTr allows. When
intrinsic
or
extrinsic factors limit protein
accretion under conditions of protein gain. PPr
is
restricted in the same proportion
as
PTr
(Pomar
et
al.,
1991).
However, at the end of
gestation sows generally lose weight
(if
the
products of conception are excluded). Weight
loss may also occur during lactation. Under
situations of
body
protein losses, that is, when
PTr is negative, PP mass decreases proportion-
ally to the losses of
IT.
This
mechanism
allows the simulation of body tissue losses
by guest on July 21, 2011jas.fass.orgDownloaded from

RJiPRODUCJNG
SOW
MODEL,
1491
Energy Protein
+."."...<
i
j
Maintenance
-u-
,...
I
....
+
i
i
fiestation
..........) ..".."""........
".:
i
First
priority
Last
:
Figure
1.
Protein
and
nonprotein
energy
flow.
Solid
lined
path represent metabolic processes
and
dotted
lines
represent catabolic
associated
processes.
during periods of heavy nutritional demands. It
also allows depleted sows to express slight
compensatory protein
growth
to recover from
gestation and lactation body protein losses.
Lysine concentration in the diet is
used
to
calculate the ideal protein intake
(IPI)
for
gestating and lactating sows according to the
ARC (1981) optimal lysine diet concentra-
tions. Protein and energy
(EM)
requirements
for maintenance, as well as the energy and
protein efficiencies for
growth
and the energy
and protein losses in urine are calculated for
adult animals as
*
previously discussed for
young pigs (F'omar et al., 1991).
After energy and protein requirements for
maintenance have been satisfied, the remaining
available
ME
and ideal protein intake
(PI)
may be used for either fetal
growth
or
milk
production (Figure 1). The surplus,
if
any, is
retained in the body. Open and gestating
females generally gain weight because they are
either
still
growing or are recovering from
lactation body losses or both. The composition
of
this
gain has not been adequately studied.
Nevertheless, it is assumed that the
amount
of
PTr and LTr retained depends on the final
balance between requirements and ingested
nutrients available for
growth.
However, some
limits are imposed to avoid situations that do
not seem to
occur
normally
or
are not
consistent with
our
understanding.
Thus,
changes in
body
weight are in relation to
changes of
PT
and(or) LT mass and, generally,
these changes
are
in the same direction at the
same time. However,
this
hypothesis will
probably not
be
valid in near steady-state
situations because numerous anabolic and
catabolic processes are occurring simulta-
neously (Close and Fowler, 1985).
In
fact,
Fuller et al. (1976) and Close et al. (1978)
observed that, even when BW gain
is
close to
zero an animal may deposit
as
much as
25%
of
its intrinsic potential for protein accretion
(IlpTr),
with
offsetting losses
in
other compo-
nents, mainly LT. Whittemore et al. (1981)
also observed that young pigs of
5
kg
live
BW
in steady-state condition can lose as much as
50
g of lipids daily and retain
an
equivalent
amount of water. Therefore, simulated PTr and
LTr are allowed to have opposite algebraic
signs only when both protein and fat retention
are close
to
zero (Figure
2).
It
is
difficult to determine from the litera-
ture a value for the maximal amount of
PT
that
a
sow can retain while losing lipids from
its
body. Also, the amount of LT that a
sow
can
retain while losing body protein is not well
understood. Because these
body
changes are
probably related to the
sow's
nutritional status,
we arbitrarily assumed that 10% of the
maximal
daily
fat losses, that
is,
.06%
of the
sow LT mass, can
be
retained or lost when PTr
is zero (see Figure
2
for a diagram of these
boundaries). The hypotheses involved
are
basically intuitive and more data are needed to
evaluate accurately the dynamics of body
composition changes when sows are near
steady-state weight condition and to ascertain
the presence and magnitude
of
limits
to
the
processes involved,
as
well as the degree of
genetic control on these limits.
As for growing pigs, the minimum LTr/PTr
boundary for sows and gestating gilts
is
expected to be determined by the genotype.
However,
this
assumed minimum LTrmr
ratio
during the
growth
period
is
taken
as
the
inverse
of
the
slope
of the upper
boundary
of
the upper right quadrant (FTr and LTr
2
0)
during the sow's reproductive life and 10% of
the maximal
daily
fat losses with
zero
PTr is
taken as the intercept Figure
2).
This
slope is
represented
as
.4
in
Figure
2.
The slight
adjustment to the intercept of the minimal LTr/
by guest on July 21, 2011jas.fass.orgDownloaded from

1492
FQMAR
ET
AL.
PTr
boundaq
relative to the young animal
model (Pomar et al., 1991) is implemented to
allow small changes in body composition near
the steady-state weight conditions as observed
in real situations. When there
is
violation of
this
boundary (upper left quadrant of Figure
2),
there is a deficit of
ME
available for LTr.
An
equivalent situation is reached when there
is
a
surplus
of
ideal protein intake
(PI)
for
PT
retention
(PI
>
PTr), and, in both cases, a
fraction of the
PI
is
deaminated as proposed
for growing pigs.
The boundary at the top in the upper right
quadrant (positive PTr), (Figure
2)
is
the
PTr,
or
genetic potential for protein accretion. The
amount of
LT
retainWday
by
a sow (right
boundary in the upper right quadrant of Figure
2)
is
function of the energy surplus
it
receives.
All these boundaries
are
dynamic and they
are
:
Total body mass of lipid
PTr
:
Body protein accretion rate
LTr
:
Body lipid accretion rate
evaluated at each integration step. Thus, as the
sow gets older, PTr decreases because body
protein synthesis approaches a plateau and
protein degradation increases with the in-
creases in body protein mass. At the limit, old
sows allowed ad libitum access to
feed
will
reach their adult protein mass and steady-state
body weight.
In
these circumstances, energy
and protein intake
are
only used to satisfy
maintenance requirements.
At the end of gestation or during lactation,
sows generally lose weight. Under these
circumstances, body fat and protein stores
are
used to overcome the nutritional deficit. The
composition of the body weight loss is difficult
to predict because of many different factors
such as feed intake, body composition of the
sow, body composition of the fetuses, and
so
on. However, protein or fat reserves are
PTr
.....
.-
..
__.
.....
.-
---
...........
...____
.--._
t
.:-
.
--...
Genetic potential
for
protein accretion
r
=
.4*PTr
-
a
Maximal Daily
fat losses
(
.6%of LT)
--.
...
LTr
=
20
*PTr
-
a
8
=
10%
of maximal
daily fat
losses
.....
Maximal Daily
...
+
protein losses
...
(
.6%ofPT)
...........................................
:;qKB:--
-
..
.-
LTr
Figure
2.
Rotein
(IT)
and
fat (LTr) accretion
and
loss
during
gestation
and
lactation,
by guest on July 21, 2011jas.fass.orgDownloaded from

REPRODUCING
SOW
MODEL
1493
seldom catabolized alone. Bowland
(1967)
and
Tess et
al.
(1983)
assumed that the protein/fat
ratio was constant at
1/20
for lactation weight
losses.
A
more realistic approach seems
feasible, which assumes that the amount of
lipids and protein catabolized
is
related to
needs associated with the energy and protein
deficit.
As
for body protein gain, some
boundaries
are
assumed
so
as
to avoid
seemingly unrealistic situations (lower left
quadrant of Figure
2).
The nature of these
boundaries is to represent
1)
the
metabolic
capacity of the sow to catabolize protein and
fat tissues (maximal daily fat and protein
losses),
2)
the
minimal
protein that should
be
catabolized from the body per unit of catabo-
lized fat (upper boundary in lower left
quadrant with a slope of
1/20},
and
3)
the
minimal fat that should
be
catabolized per
unit
of catabolized protein (lowest boundary in
lower left quadrant). Little information is
available for the maximal amounts of
PT
and
LT mass that can
be
catabolized,
as
well
as
for
the slope of these boundaries. Black et al.
(1986),
quoting Greenhalgh
et
al.
(1980),
King
(1982),
and
King
and
Dunkin
(1986),
quoting
Greenhalgh et
al.
(1980),
King
(1982),
and
King
and
Dunkin
(1986)
assumed that
.6
to
.8%
of the total body protein could
be
lost
daily without reduction of
milk
production.
Our
model assumes that
.6%
of LT and
PT
can
be
lost daily before fetal growth or
milk
yield
is reduced. It
is
also assumed that the slope of
the upper boundary for weight losses is
20,
accepting the value of Bowland
(1967)
and
Tess et
al.
(1983).
Because
no
estimates were
found
in
the
literature for the slope
of
the
lower boundary,
this
is determined in the
model by connecting, by a straight line, the
point of maximal PTr daily losses
(-.6%
of
PT
mass)
and maximal daily LTr losses
(-.6%
of
LT mass) with the point LTr
=
.1
LT and pTr
=
0
(lower left quadrant
of
Figure
2).
When
energy deficit is
too
high in relation to these
boundaries, extra PI
is
deaminated,
as
described for
similar
situations in growing pigs
and adult females retaining
IT.
Under some
nutrient requirements and diet composition,
required
FT
losses may
be
too
large
in
relation
to LT losses. Under these circumstances,
control
of
feed intake and(or) total sow
production would possibly
be
implicated in the
stabilization of the amount of protein and
fat
catabolized. However, these mechanisms
are
uncertain,
and
the model assumes that
addi-
tional LT is degraded and lost as heat. Only
when maximal protein or fat losses are reached
will
milk
production
be
reduced.
Fetal Growth
Although other approaches are also avail-
able (Verstegen et
al.,
1987),
the factorial
method is used
to
estimate total energy
requirements during pregnancy because
this
method has
been
proven adequate (Van-
schoubroek and Van Spaendonck,
1973).
To
represent the increase of maintenance energy
requirements
(EM)
with the progress of gesta-
tion (Verstegen et al.,
1971),
ARC
(1981)
proposed to increase the basic requirements of
the sow by
1
H/kg
of metabolic weight, per
day,
and for each day after the
40th
d of
gestation.
In
our
model,
EM
is calculated
separately for the sow and each fetus with the
same relationship
as
used for growing pigs
(Pomar
et al.,
1991).
These values are then
added together to estimate the total
EM
requirements at each specific integration step.
Although the higher heat production observed
during late pregnancy may not
justify
the use
of the same
EM
predictor throughout
the
sow
reproductive cycle (Verstegen
et
al.,
1987).
other
results do not show differences in
EM
requirements between pregnant and nonpreg-
nant sows (Noblet and Close,
1980;
Walach-
Janiak
et al.,
1986). A
similar approach is used
in
the model to evaluate the total protein
requirements for maintenance during pregnan-
cy.
Total birth weight
(TBW)
is
calculated as
follows:
TBW
=
1.1
(litter
size
born
alive)
(fetal birth weight),
where the additional
10%
accounts for
still-
births, reabsorbed fetuses (Tess et
al.,
1983),
and conception products. Litter size and fetal
birth weight
are
model input variables that
depend
on
genotype, parity, estrus at breeding,
and preceding lactation length. The following
equation of Pomeroy
(1960)
allows the predic-
tion of the total weight of the litter
(TLW
throughout pregnancy:
TLW
=
.1
(.2447t
-4.06)3 (TBW/1396),
where t is the day
of
gestation. The first
by guest on July 21, 2011jas.fass.orgDownloaded from

1494
POMAR
derivative of
this
equation is used in
our
model
to estimate the basic daily fetal growth.
Moustgaard (1962) showed that fetal chemi-
cal composition is not constant throughout
pregnancy. However, fetal mass is
small
compared with the whole sow
mass
until late
pregnancy. Therefore, for simplicity, basic
fetal
mass is assumed to contain
11
%
protein
and 1% fat, which is within the range of values
observed at birth by Pomeroy (1960), Wood
and Groves (1965),
Curtis
et al., (1967), and
Okai
et
al.
(1977). From results of
Curtis
et al.
(1967) and
Okai
et al. (1977), fetal ash content
is taken as 3.9% of the input litter weight.
Feed intake level during gestation slightly
affects the size and chemical composition of
the litter (Vanschoubroek and Van Spaen-
donck, 1973; Walach-Janiak et al., 1986).
Basic protein retention
in
the fetus
(PTPr
=
.11
x
fetal growth rate) represents the maximal
potential for protein accretion. It is only under
severe protein restriction that
FTPr
should
be
decreased. Nevertheless,
this
situation is un-
likely
to
be
reached because
it
would be the
result of very poor protein ingestion and very
poor sow body condition.
In
fact, results of
Pond et
al.
(1968) suggest that severe protein
restriction during pregnancy affects the dam’s
body condition rather than the offspring
development. Therefore, we assumed that the
small but real effect of the sow intake on fetal
weight (Vanschoubroek and Van Spaendonck,
1973) will focus on
body
fat rather than on
protein mass. Thus, the
1%
fat assumed for
fetal weight gain represents the minimum fat
retained on a daily basis. Despite the limited
information available, we assume
also
that
extra fat can be retained
by
the fetuses only
when the sow is also gaining fat. Thus,
5%
of
the energy used for the
sow
lipid retention is
directed toward fetal
fat
deposition
(LTPr).
Results of Vanschoubroek and Van Spaen-
donck (1973) indicate that
an
increase in pig
birth weight reaches a plateau with high
intakes. This plateau is simulated in the model,
limiting the extra fat retained by the fetus to
five times the basic fat retention.
On
the other
hand, situations in which energy and protein
available for fetal retention are under the
minimal potential are not included in the
model. Such conditions
are
seldom observed in
well-managed swine production units.
Very few estimates exist for the energetic
efficiency of fetal
growth.
The values range
from .2 (Hovell et al., 1977) to
.8
(ARC,
ET
AL.
1981).
As
proposed by Close et al. (1985) and
adopted by Black et al. (1986), the efficiency
of
ME
utilization for fetal
growth
is assumed
to
be
.6.
This
efficiency is increased to
.8
when
the energy used for fetal
growth
comes from
the sow body reserves. Values for IPI effi-
ciency for fetal growth are scarce
in
the
literature, but the real value is probably
between
.8
and
1.0.
If
we assume that
this
coefficient reflects the fetal protein accretion
as
a propodon of the protein available, and
that some protein goes into other conception
products, a value of .8 seems reasonable.
During the fetal protein retention process,
losses of
body
protein
also
can take place
in
the sow. When the protein used for fetal
protein retention comes from the sow body
protein reserves, protein efficiency is assumed
to
be
.95.
Milk
Production
Few
data
are available to accurately predict
milk
yield curves for the sow throughout
lactation.
This
results from the difficulty of
obtaining a satisfactory and easy estimate of
the sow’s
milk
yield and from the differences
between methods.
For
these reasons, cow milk
yield information was used
in
our model to
characterize the shape of the sow
milk
yield
curve.
This
was done even though these two
curves are not necessarily similar and different
sows possibly have different
milk
curve
patterns (Salmon-Legagneur, 1958).
One of the most popular models to describe
dairy
cow
milk
yield
is
the one proposed by
Wood
(1969):
y(t)
=
AtBe(-Ct),
where y(t) is the average daily
milk
yield
in
the week t, e is the base of the natural
logarithms, and
A,
B, and
C
are positive
parameters defining the lactation curve.
Dhanoa (1981) showed that the reparameteri-
zation of
Wood’s
model by making B
=
M
x
C
(where
M
is the time between calving and the
peak
milk
yield) has better mathematical
properties because the correlations between
nonlinear parameters are reduced. Both forms
of
Wood‘s
model can
be
used in a continuous
form, but
both
are inaccurate when
t
is small.
Black et al. (1986) suggested the transforma-
tion
t’
=
t
+
10
in
Wood’s
model to allow for
an intercept different from zero.
In
the present
model, Dhanoa’s equation was
fitted
to sow
by guest on July 21, 2011jas.fass.orgDownloaded from

REPRODUCING
SOW
MODEL
1495
milk
yield
data
presented by Elsley (1971).
The model was standardized for a peak
milk
yield of one unit. The transformation t’
=
t
+
25 was adopted because it decreases the
residuals from regression and allows y(t
=
0)
to be greater than
zero.
A
nonlinear least
squares minimization procedure
(SAS,
1985)
was used for
this
purpose and solutions were
A
=
33.26
x
106,
M
=
2.46, and
C
=
66.45
x
lW3. Initial
daily
milk
yield is 70% of the
maximal yield and the
peak
milk
yield is
reached at the end
of
the 4th wk of lactation.
Potential
milk
production for gilts is as-
sumed to
be
74% of the sow
milk
yield
potential (Elsley, 1971).
As
observed by
Salmon-Legegneur (1965), no difference
be-
tween second and subsequent lactations
is
included in the model. Breed differences in
milk
yield have been observed in sows (Allen
and Lasley. 1960). However, because insuffi-
cient data
are
available to characterize the
milk
yield curves of
all
the breeds, daily
milk
production is calculated in the model as the
product of the standardized
milk
yield (max-
imal value of 1) and the peak
milk
yield. Peak
milk
yield is an input variable supplied by the
user according to the genotype of the simu-
lated sow. However, this value is termed udder
potential because it represents the sow mi&
yield only when there is no nutritional or other
factor limiting the synthesis of
milk.
Colostrum composition is considerably
dif-
ferent
from
that
of
milk
(Penin,
1955;
Bowland, 1967; Fahmy, 1972; Brent et
aL,
1973) and gradual changes in
milk
composi-
tion
occur
during the first days of lactation
(Penin, 1954, 1955; Brent et al., 1973). The
energy content in sow
milk
is relatively
constant throughout lactation (Brent et al.,
1973). Therefore,
it
is taken to be
5.0
M.l/kg
according to the data of Brent et al. (1973) and
Klaver
et
al.
(1981).
Milk
protein concentra-
tion is generally high at parturition but
declines rapidly during the first 24 h. The
decline then slows and a minimum occurs at
about the 16th d after farrowing (Penin, 1954,
1955; Brent et al., 1973). However, changes in
milk
protein content between the 2nd and 5th
wk after parturition are relatively small (Perrin,
1954; Pond et al., 1962; Brent et
al.,
1973).
Based on the data of Pond
et
al. (1962),
Bowland (1967), Elsley (1971), Fahmy (1972),
Brent et al. (1973), and Klaver et al. (1981),
milk
protein concentration is assumed to
be
5.6%.
Sow feeding level does not have a large
effect on
milk
yield in early lactation
(O’Grady et al., 1973; Klaver et
al.,
1981).
However, O’Grady et al. (1973) observed that
sows with low energy intakes had lower
milk
yields in late lactation,
this
effect increasing
with the number of parities. They
also
argued
that sows with low energy intakes could not
maintain later
milk
yields by using their body
reserves
as
at the beginning of the fist
lactation. These results agree with those of
Klaver
et
al.
(1981). who noted that body
condition of the sow seems to
be
the primary
factor influencing
milk
production. Therefore,
we assume that the sow energy or protein
intake only
affects
milk
yield in relation
to
the
body reserve condition of the sow. Similarly,
milk
composition
is
not greatly affected by
energy intake (O’Grady et
al.,
1973). but small
effects have been observed on
milk
protein
concentration (O’Grady et
al.,
1973; Green-
halgh et
al.,
1980). Here,
milk
composition
is
also assumed to
be
independent of the amount
of
nutrients ingested.
Klaver et al. (1981) observed that sows with
poor body condition, but fed at high levels,
restricted the use of their body tissues for milk
production.
On
the other hand, sows that
generally gain more weight during gestation
tend to use more
of
their body reserves during
lactation (Greenhalgh et
al.,
1980). The most
important factor associated with weight losses
during lactation
is
probably the body condition
of the sow. Thus, sows in good condition can
rely on their reserves to replace the nutrient
deficit
occunring
during the
periods
of high
milk
production.
In
contrast,
thin
sows depend
mainly on their feed intake. These mechanisms
are simulated in the model because potential
losses of lean and
fat
tissues are expressed as a
function of their
own
masses. Indeed, fat sows
will tend to lose more weight during lactation
because they will
be
able to supply body
nutrients more easily than
thin
sows. Influence
of preceding lactations
on
the subsequent ones
is also simulated through
this
mechanism.
Simulated daily
milk
yield is less than the
calculated udder potential when 1) pig
milk
requirements
are
lower than
this
potential or
when
2)
dietary nutrients and body reserves
cannot satisfy all the sow nutrient require-
ments. Therefore, the effect
of
litter size on
milk
production observed by Salmon-Legag-
neur (1965) and Elsley (1971)
is
simulated in
this
way.
Small
litters will usually
nor
require
by guest on July 21, 2011jas.fass.orgDownloaded from

1496
mMAR
all the sow udder potential, whereas large
litters
will
demand
it
all earlier and during
longer periods. Efficiency of
ME
utilization for
milk
production
is
assumed to be
.70
accord-
ing to De Lange et al.
(1980),
whose estimate
is close to the recommended value of
ARC
(1981).
Because
the composition of weight
changes of the sow during lactation cannot
be
measured accurately, body energy efficiency
for
milk
production
is
not well
known.
However, as observed for other productions,
body energy reserves are probably used more
efficiently for
milk
energy production than
dietary
energy. Thus, an efficiency
of
.SO
seems reasonable, which is close to the value
presented by
ARC
(1981)
and Tess et al.
(1983).
Also,
as proposed for growth, ideal
protein available for production cannot
be
utilized totally in the
milk.
Few values have
been published, and the model assumes that
80%
of the available ideal protein intake can
be
utilized in
milk.
This
value is increased to
90%
when the synthesized
milk
protein is
produced from the body protein reserves.
ET
AL.
Pig
Growth During Lactation
As
observed by Fahmy and Bernard
(1970),
no weight differences
are
assumed between
male and female pigs during fetal development
and preweaning growth.
A
similar
assumption
is
made for the body composition of prewean-
ing pigs.
Based
on
the preweaning growth curve
developed by Robison
(1976)
and on body
composition data of Manners and McCrea
(1963)
and Wood and Groves
(1965),
Tess et
al.
(1983)
proposed to predict protein
(PTP)
and fat
(LTP)
preweaning pig mass from birth
to
56
d of age
as
follows:
F'TP
=
.1595
+
.019Ot
+
.00032t2
LTP
=
.0145
+
.0225t
=
.00045t2
where t is the age of the pigs in days. The first
derivatives
of
these equations are used in the
model
to
predict the maximal
PTP
(PTPr)
and
LTP
(LTPr)
accretion rates during lactation.
We assume
that
suckling pigs consume
milk
to satisfy maintenance energy and protein
requirements and for protein and lipid growth.
When the sow cannot supply all the required
milk,
suckling pigs will tend to
eat
creep feed
to compensate for the deficit in energy
(Greenhalgh et al.,
1980).
The user of the
model needs to
specify
the age at which creep
feed
will be available
in
the
pen.
However,
under the age of
14
d, intake of solids by pigs
is assumed to
be
negligible
(NRC,
1987).
Large litters and depleted sows will lead to
higher creep feed consumption as a result of
higher nutrient requirements and lower
milk
production, respectively. Under these circum-
stances, maximal
daily
creep
feed
intake is
limited in
the
model to
tbree
times
the
proposed average creep feed intake relation-
ship for suckling pigs
(NRC,
1987).
This
equation predicts the average digestible energy
intake (DE) between
14
and
35
d of age and
has
the following form:
DE (MJ/kg)
=
.0469t-
.6347,
where t is the age of the pig
in
days.
Milk
GE
is
assumed to
be
97%
digestible
(Klaver et al.,
1981;
Tess et al.,
1983).
In
the
same way,
milk
protein digestibility
is
taken as
95%.
Total DE and protein are the sum of
milk
and creep feed digestible nutrients. The
PI
in
creep feed is calculated as for growing pigs
and total
PI
available is the result
of
adding
PI
in creep feed and in
milk.
This
latter
calculation assumes that
95%
of the digestible
protein of
milk
is retained in the pig's body.
Next,
PTFV
and
LTPr
are
calculated
as
was
done for body protein (PTr) and fat (LTr)
accretion rates in growing pigs. During periods
of restricted intake, the model allows
LTPr
to
decrease down to .25 times
PTFV.
This
value is
lower than any observed LTr/PTr ratio for
growing pigs.
It
is generally accepted that the
proportion of fat
in
the suckling pig gain is
low and it increases as the pig gets older.
Ad
libitum
Feed
Intake
of
Lactating
Sows
Lactating sows
are
generally allowed
ad
libitum access to feed, whereas gilt and sow
feeding is restricted when females
are
open
and during gestation. Feed intake during
lactation
is
low immediately after farrowing
but rapidly increases as lactation proceeds.
in
NRC
(1987),
DE intake for
a
lactation length
of 28 d was predicted
by:
DE (MJ/d)
=
56.07
+
2.49t
-
.072t2,
where t
is
the lactation day.
This
relationship
by guest on July 21, 2011jas.fass.orgDownloaded from

REPRODUCING
SOW
MODEL
1497
does not consider the effect of sow parity on
lactation DE intake. It is assumed in the model
that DE intake during lactation is directly
related to the sow protein mass. Assuming a
sow
mean
body protein mass
(IT)
of approxi-
mately 27 kg, DE
intake
can be calculated by
multiplying the prediction
from
the above
equation by
pT127.
Reproduction-Cycle Model
Evaluation
Model verification was performed by
checking for both mathematical and logical
consistency throughout the whole model de-
velopment process as previously described for
the growing individual animal model (Pomar
et al., 1991).
Model
validation might
be
achieved by
comparing performance of the model (predict-
ed) to real system performance measurements
(actual). However, comparisons between pre-
dicted and actual composition of
body
weight
or weight gain cannot be fully made without
the precise information that will allow charac-
terization
of
the biological material and the
reproduction of the experimental conditions.
For example, genetic characterization of simu-
lated
animals
requires body chemical composi-
tion data at several points of the
growth
period
and life cycle.
Because
such data
are
often not
available, these comparisons are not feasible at
this
point for the reproduction-cycle model
evaluation. Simulated results in the reproduc-
tion cycle are strongly dependent on the
genetic characterization
of
the simulated
animals and on many environmental and
management conditions. Appropriately
designed experiments seem necessary in many
cases for full validation of detailed portions of
the model. Nevertheless, the model can
be
evaluated
in
several ways to judge its robust-
ness and suitability. First, because the model is
based
on
fundamental concepts, description of
these provides a first step toward model
verification.
Second, model results can be evaluated for
their reasonableness and ability to represent
specific production situations. For this evalua-
tion. protein mass
at
maturity for the simulated
sows was assumed to
be
equivalent to the
basic genotype
(35.5
kg; Pomar et al., 1991).
Temperature
and
seasonal
effects are
not
accounted for
in
this
model and protein
digestibility is assumed to be
75%.
To
represent a realistic scenario, the experimental
procedure described
by
Whittemore et
al.
(1980) was replicated with the model. Howev-
er, because of lack of adequate information
(observed feed intake and many other experi-
mental conditions) required to effectively
represent the
trial
genotype, comparisons
be-
tween
experimental and simulated results
are
inadequate and therefore they
are
not present-
ed.
Gilts were allowed ad libitum access to feed
with standard diets
until
93 kg live
BW,
at
which point the evaluation exercise starts. Gilts
were fed 1.8
kg
daily between 93 kg
BW
and
first parturition. Sows were simulated during
their first
two
reproductive cycles.
All
sows
were given a common nutritional regimen
containing 12.8 MJkg of DE, 154 g/kg of
CP
and
7
g/kg of lysine. During the second
pregnancy,
2.3
kg/d of
feed
was given to the
sows. Simulated sows were fed based on
an
ascending scale starting at 1.4 kg. The ration
was increased by
.45
kg daily up to a
maximum of 2 kg, plus
.5
kg/(d-pig) in the
litter. Thus, a sow with
a
litter of 10 pigs
would
be
fed 1.4 kg on the 1st d (parturition),
with an increase of .45 kg/d until total fed
would reach
7.0
kg/d (13 d later).
First mating
is
simulated at 116 kg
BW
and
at 12 d after weaning
in
the second mating.
Simulated litter
size
and weight
are
those
obtained in the experiment. Pig mortality
during lactation is simulated
as
proposed by
Tess et
aL
(1983) from the overall experi-
mental pig viability, which was estimated as
88.5%. Simulated pigs
are
weaned
at
35 d of
age.
Detailed
results
(Figures
3
to
5)
show that,
during half of the first and most of the second
gestation, dietary protein intake satisfies re-
quirements for both the fetus and the sow.
Therefore, only small amounts of body protein
reserves
of
the sow
are
mobilized to satisfy
fetal requirements (Figure
3B).
Because body
protein losses during the simulated lactations
are significantly lower than body protein
increases during both gestations, the sow
shows
a
net gain in body protein for the total
period. However, this is not the case
for
fat
reserves. Early in gestation, energy intake
fulfills
the overall energy requirements for
only a few days. Afterward, the sow’s fat
reserves
are
mobilized.
This
fat mobilization is
less in the second gestation. Thus, simulated
body
fat
mass decreases over the experiment.
During both gestations, part
of
the protein
by guest on July 21, 2011jas.fass.orgDownloaded from

1498
POMAR
ET
AL.
ingested
by
the sow
is
deaminated to increase
(or
decrease)
fat
accretion
(or
losses).
This
Occurrence is controlbed
by
the boundaries
imposed
by
the PTr/LTr slope (Figure
2).
For
this
reason, the slopes for protein and
fat
accretion shown in Figures
3B
and
3C
change
when retention becomes negative
as
the
accretion rate reaches the upper boundary.
Also,
the rapid increase in nutrient
require
iA
1
ments of the fetuses
as
the end of gestation
approaches seem to be responsible for the
deceleration of gain of
body
components
observed in the
sow.
At the beginning of both simulated lacta-
tions, energy and protein reserves rapidly
decrease. Then, feed intake rapidly increases
with the specified feeding program and,
therefore, body losses decrease and the sow
n
I1
c1
nm
4,
cl
9.001
I'"i;
1;
1%
,
1;
I
160
260
360
460
560
1.00
!
1
I
I
180
260
360
..
j:
;!
i!
I!
ti
60
580
Simulated
Age
of
Sow
(days)
Rgure
3.
Simulated
feed
intake
(A),
body
protein
(B)
and
fat
(C)
accretion
rates
of
the
sow
(lower
curve)
plus
fetuses
(shaded
areafromC1
to
Fl
or
from
C2
to
F2)
and
nursing
pigs
(shaded
area
fromF1 to
W1
or
from
F2
to
W)
with
the
curve
at
the
top
of
shaded
area
representing
the
combiued
intake
or
accretion
for
the
production
unit
of
sow
plus
fetuses
or
nursiug
pigs
during
the
first
two
paritits
(C
=
conception,
F
=
Farrowing
and
W
=
weaning).
by guest on July 21, 2011jas.fass.orgDownloaded from

REPRODUCING
SOW
MODEL.
1499
200
-I
g
180-
By”
140-
t=
100-
0
2
160-
h
m-
-
120-
U
80
t
can even gain weight for a short
period
of
time
(Figures
3
and
4).
However,
milk
energy and
protein requirements soon increase and the
sow has to once again mobilize her body
reserves
to
satisfy
the
increasing
milk
require-
ments of the growing litter
(Figure
6).
Milk
yield reductions at the end
of
both lactations
lead to the decrease in the sow’s body
losses.
For
only
2
d at the beginning
of
the first
gestation and approximately
5
wk during the
I
I
second gestation, the protein accretion rate
of
the sow reaches its maximal potential.
Discussion
This
evaluation exercise demonstrates the
ability
of
the model to simulate sow
body
weight gains
and
losses, body composition,
fetal
growth,
milk
production, and suckling pig
growth.
The
main
difficulty
encountered
in
n
n
s
a
0
0
m
40-
n
a-
C.
n
r”
36-
as
w-
03
LL-
32-
0
30-
m
28
-
28
I
I
I I
1
a
U
160
260
360
460
560
Simuloted
Age
of
Sow
(days)
Figure
4.
Simulated
sow
and
fetal weight
(A)
and
body
protein
(B)
and
fat
(C)
mass
of
the
sow
(lower
curve)
plus
fetuses
(shaded
areas
from
C1
to
P1
or
from
C2
to
F2)
and
nursing
pigs
(shaded
area
from
F1
to
W1
or
from
F2
to
W2)
with the curve
at
the
top
of
the
shaded
area
representiug
the
combined
weight
or
mass
for
the
production
unit
of sow
plus
fetuses
or
nursing pigs
duriug
the
first
two
parities
(C
=
conception,
F
=
farrowing
and
W
=
weaning).
by guest on July 21, 2011jas.fass.orgDownloaded from

1500
POMAR
ET
AL.
8.0
-
0
6.0-
Y
v
I!
4.0-
e
2.0-
f
0.
?
Day
of
Gestation
Figure
5.
Simulated conceptus weight and protein mass .-
40
P
during
the
first
reproduction
cycle.
representing real situations is to properly
characterize the simulated genotype. The basis
for modeling growth responses
in
this
model
was the experiment
of
Walstra
(1980),
based
on pigs of Dutch Landrace breeding in a
minimal disease condition between
1970
and
1974.
To
be
more reliable, the model must
be
calibrated to represent other breeds and
crossbred genotypes under various environ-
mental conditions.
In
fact, part of the differ-
ences observed between simulated and real
animals are certainly due to the lack of genetic
calibration.
Also,
body protein mass plays an
indirect, but important, role in the determina-
tion of energy and protein requirements for
maintenance. Maintenance requirements
in-
crease
as
pigs get heavier and older, and they
represent the most important energetic need in
adult females. Research is needed to estimate
with precision maintenance requirements
in
adult animals and the main factors that may
affect them. More precise estimates of these
requirements should lead to more accurate
simulation of the energetic metabolism of adult
females.
The simulation model proposed herein is
generally more mechanistic than others previ-
ously presented and
is
the first to use
fundamental concepts to simulate protein
growth potential of adult females. The pro-
posed model is also the first
to
separately
estimate energy and protein gain
or
loss during
the reproductive cycle, with boundaries de-
pending on the nutritional and physiological
status of the animal. Because of the incorpora-
---
Yllk
Yield
-
Milk
potential
I
I
I
0
10
20
30
Day
of
Lactation
Figure
6.
Simulated
milk
potential and
yield
and
growth
of
suckling
pigs
during
the
fmst parity.
tion of these mechanisms,
this
model can give
a more accurate representation of real systems.
Therefore, the model is expected to respond
more reliably to
a
wide range of combined
nutritional, genetic, and managerial situations,
if
these situations are adequately described in
quantitative terms.
Implications
This
computer simulation model extends
the basis for predicting the growth and body
composition of pigs to include the gains and
losses in protein and lipid components of open,
gestating, and lactating sows. The predictions
incorporate the interaction of performance
potentials, due to genetics and other factors,
with the amount and nutritional characteristics
of diets fed during these periods. This model
facilitates planning the feeding program to
maintain adequate body reserves to support
gestation and lactation needs without excessive
fatness. More precise control of feeding
programs can allow maximum reproduction
and lactation without excess costs.
by guest on July 21, 2011jas.fass.orgDownloaded from

REPRODUCING SOW MODEL
1501
Llterature
Clted
Allen,A.D.andJ.F.Lasley.
1960.
Milkproductionof sows.
J.
Anim.
Sci.
19:150.
Allen, M. A. and
T.
S.
Stewart.
1983.
A
simulation
model for
a swine breeding
unit
producing feeder pigs. Agric.
Sys.
10193.
ARC.
1981.
The Nutrient Requirements of Pigs. Agricul-
tural
Research Council, Commonwealth
Agricultural
Bureaux,
Slough,
UK.
Baldwin, R. L.
1976.
Principles of modem
animal
systems.
Proc. N.
2.
Soc.
Anim.
Rod.
36128.
Black,
J. L.,
R.
G. Campbell,
I.
H.
Williams,
K. J. James
and
G.
T.
Davies.
1986.
Simulation
of energy
and
amino
acid
utilization
in
the
pig. Res. Dev. Agric.
3(3):121.
Bowlaud,
J.
P.
1967.
Energetic efficiency
of
the sow.
J.
Anim.
Sci.
26533.
Brent,
B.
E.,
E.
R.
Miller,
D. E. Ullrey and K. E. Kemp.
1973.
Postpartum changes
in
nitrogenous constituents
of sow
milk.
J.
Anim.
Sci.
36:73.
Close, W. H., L. E. Mount and D.
Brown.
1978.
The
effects
of
uh
of
nutrition
and environmental temmture
on
the' energy metabolism of
he
growing pig. Br.
I.
~utr.
40:423.
Close, W. H. and V. R. Fowler.
1985.
Energy requirements
of pigs.
In:
D.J.A. Cole and
W.
Haresign
(E&.)
Recent
Developments
in
Pig Nutrition. p
1.
Butterworths,
London.
Close, W. H., J. Noblet
and
R
P. Heavens.
1983.
Studies
on
the energy metabolism of
the
pregnant sow.
2.
The
partition and
utilization
of metabolizable energy
intake
in
pregnant and non-pregnant
animals.
Br.
I.
Nutr.
53:
267.
Curtis,
S.
E., C.
J.
Heidenreich and
T.
G.
Martin
1%7.
Relationship
between
bodyweight
and
chemical com-
position
of pigs at
birth.
J.
Anim.
Sci.
26:749.
de Lange, P.G.B.,
G.
J.
M.
van Kempen, J. Klaver and
UWA. Verstegen.
1980.
Effect of condition of sows
on
energy balances during
7
days
before
and
7
days
after parturition. 3.
Anim.
Sci.
50886.
Dhanoa, M.
S.
1981.
A
note
on
an alternative
form
of the
lactation
model of Wood.
Anim.
Prod.
32349.
Elsley,F.WH.
1971.Nutritionandlactationinthesow.
hI.
R
Falconer (Ed.) Lactation. p
393.
Butterworths,
London,
UK.
Fahmy,
M.
H.
1972.
Comparative study of colostrum and
milk
composition of seven breeds of
swine.
Can. J.
Anim.
Sci.
52:621.
Fahmy,
M.
H. and C. Bernard.
1970.
Genetic and phenotypic
study of
pre
and post-weaning weights and
gains
in
swine.
Can.
J.
Anim.
Sci.
50593.
Fuller,
M.
F.,
AJF. Webster. R.
M.
MacPherson
and J.
S.
Smith.
1976.
Comparative
aspects
of the energy
metabolism
of Pietrain and Large White
x
Landrace
pigs during
growth.
In:
M.
Vermorel
(Ed.)
Energy
Metabolism. G. de
Bussac
Clermont-Perrand, Publ.
Eur.
Assoc.
him.
Prod. No.
19.
p
177.
Greenhalgh,
JFD.,
B.
Baird,
D.
A. Grubb,
S.
Done, A.
L.
Lightfoot, P. Smith, P. Toplis, N. Walker, D.
WW
and
M.
L.
Yeo.
1980.
Coordinated
trials
on
the protein
requirements of sows.
2.
A comparison
of
two
levels of
dietaryproteiningestationandfoarinlactation.
Anim.
Prod.
3m395.
Hovell,
P.
D.
DeB.,
R
M.
MacPherson,
R.MJ. Crofts and R.
I.
Smart.
1977.
The effect of pregnancy, energy intake
and
mating
weight
on
protein deposition
and
energy
retention of female pigs.
Anim.
Prod.
25:281.
King, R. H.
1982.
The
Effect
of Nutrition
on
the
Reproductive Performance of Fit-Litter Sows.
PhD.
Thesis. Univ. W. Aust., Nedlands, Aust.
King,
R
H.
and A. C.
Dunkin.
1986.
The effect
of
nutrition
on
the
reproductive performance of
mt-litter
sows.
3.
The response
to
graded
increases
in
food intake during
lactation.
Anim.
Pmd.
42119.
Klaver, J., G.J.M. van Kempen, P.GB. de Lange, M.W.A.
Verstegen and
H.
Boer.
1981.
Milk
composition and
dairy
yield of
different
milk
components
as
affected by
sow condition and lactatiodfeeding
regimen.
J.
Anim.
Sci.
52:1091.
Koong, L. J., R.
L.
Baldwin and M.
J.
Ulyatt.
1976.
The
application of
mathematical
modeling techniques
to
animal
science research. Roc. Symp.
on
Use
of
the
Computer
in
Animal
Science Teaching Res. and Ext.
Texas A&M Univ., College
Station.
Manners, M. J. and
M.
R.
McCrea.
1963.
Changes
in
the
chemical composition of sow-reared piglets during the
1st
month of lief.
Br.
J. Nutr.
17:495.
Moustgaard, J.
1962.
Foetal
nutrition
in
th pig.
In:
J.
T.
Morgan
and
D.
Lewis
(Eds.)
Nutrition of Pigs and
Poultry. p
189.
Butterworths, London.
Noblet, J. and
W.
H. Close.
1980.
The energy costs of
pregnancy
in
the sow.
In:
L. E. Mount
(Ed.).
Energy
Metabolism. p
335.
EAAP.
Publ.
26.,
Buttenvorths.
London.
Animals.
National Academy Press, Washington,
DC.
O'Grady, J.
F..
F.WSI. Elsley, R. M. MacPherson and
I.
McDonald.
1973.
The
response of lactating
sows
and
their litters
to
different dietary energy allowances.
1.
Milk
yield and composition, reproductive performance
of sows and
growth
rate of
litters.
Anim.
Rod.
17:65.
Okai,
D.
B.,
F.
X.
Aherne
and
R
T.
Hardin.
1977.
Effects
of
sow
nutrition
in
late
gestation
on
the body composition
and survival of the
neonatal
pig.
Can.
J.
Anim.
Sci.
57:
439.
Penin,
D.
R.
1954.
The composition of sow's
milk
during
the come of lactation.
J.
Dairy Res.
2155.
Perrin, D. R.
1955.
The
chemical
composition of
the
colostrum and
milk
of the sow. J. Dairy Res.
22:103.
Pettigrew,
I.
E.,
S.
G. Cornelius, V. R. Eidman, and
R
L.
Moser.
1986.
Integrations of factors affecting sow
efficiency: A modeling approach. J.
Anim.
Sci.
63
1314.
Porn,
C.,
D.
L.
Harris and
F.
Minvielle.
1991.
Computer
Simulation
Model of
Swine
Production Systems:
I.
Modeling the
growth
of young pigs.
J.
Anim.
Sci.
69
1468.
Pomeroy, R. W.
1%0.
Infertility and
neonatal
mortality
in
the sow.
IU.
Neonatal mortality
and
foetal develop-
ment. J. Agric. Sci.
5431.
Pond, W. G., J. A.
Dunn,
G. H. Wellington, J.
R.
Stouffer and
L.
D.
Van
Vleck,
1968.
Weight
gain
and carcass
measurements
of
pigs from
gilts
fed adequate vs.
protein-free
diets
during gestation.
J.
Anim.
Sci.,
22
1583.
Pond,
W. G., L. D. Van
Weck
and D. A.
Hartmaa
1962.
E'arameters for
milk
yield
and
for percents of
ash,
dry
matter,
fat and protein
in
sows.
J.
Anim.
Sci.,
21:293.
Robison,
0.
W.
1976.
Growth
patterns
in
swine.
J.
Anim.
Sci.,
421024.
Salmon-Legagneur, E.
1958.
Observations
sur
la
production
laitii?re des
trUies.
Ann.
Zootech. (Paris)
2143.
Salmon-Legagneur,
E.
1%5.
Quelques
aspects
des relations
NRC.
1987.
PrdiCting Feed
Intake
Of Food-Producing
by guest on July 21, 2011jas.fass.orgDownloaded from

1502
POMAR
ET
AL.
nutritionnelles entre la gestation et la lactation chez la
truie.
AM.
Zootech.
(Paris)
141.
SAS.
1985.
SAS User’s Guide: Statistics. SAS Inst., Inc.,
Gary,
NC.
Singh,
D.
1986.
Simulation
of
swine herd population
dynamics. Agric. Sys.
22157.
Tess, M. W.,
G.
L.
Bennett
and
G.
E.
Dickerson.
1983.
Simulation
of
genetic changes
in
life cycle efficiency
of
pork
production.
1.
A bioeconomic model.
J.
Anim.
Sci
56:336.
Vanschoubroek,
F.
and
R.
Van Spaendonck.
1973.
Fak-
torieller aufbau des energiebedds tragender zucht-
sauen.
2.
Tierphysiol. Tieremaehr. Futtermittekd.
31:
1.
Verstegen,M.W.A.,A.J.H.VanEsandH.J.Nijkamp.
1971.
Some
aspects
of energy metabolism of the sow during
pregnancy.
Anim.
Prod.
13:677.
Verstegen,M.W.A., J.M.F.
VerhagenandL.A.DenHartog.
1987.
Energy requirements
of
pigs during pregnancy: a
review. Livest.
Prod.
Sci.
16:75.
Walach-Janiak, M., St. Raj and
H.
Fandrejewski.
1986.
The
effect
of
pregnancy
on
protein, water
and
fat
deposition
in
the body of
gilts.
Livest.
Prod.
Sci.
15:
261.
Walstra, P.
1980.
Growth and
Carcass
Composition
from
Birth
to
Maturity
in
Relation to Feeding Level and
Sex
in
Dutch Laudrace
Pigs.
W.D. Thesis. Agric. Univ.
Wagenhgen, The Netherlands.
Whittemore,
C.
T.
1986.
An
approach
to
pig
growth
modeling.
I.
Anim.
Sci.
63:615.
Whittemore,
C.
T.. M.
F.
Franklin
and B.
S.
Pearce.
1980.
Fat changes
in
breeding
sows.
Anim.
Rod.
31:183.
Whittemore,
C.
T.,
H.
M. Taylor,
R.
Henderson,
J.
D. Wood
and
D.
C.
Brock.
1981.
Chemical
and dissected
composition changes
in
weaned piglets.
Anim.
Prod.
32203.
Wood,
A. J. and T.D.D. Groves.
1965.
Body composition
studies
on
the
suckling
pig.
1.
Moishue, chemical fat,
total protein, and total
ash
in
relation to age and body
weight.
Can.
J.
Anim.
Sci.
45%.
Wood,
P.D.P.
1969.
Factors
affecting
the
shape
of the
lactation curve
in
cattle.
Anim.
Prod.
11:307.
by guest on July 21, 2011jas.fass.orgDownloaded from

Citations http://jas.fass.org/content/69/4/1489#otherarticles
This article has been cited by 1 HighWire-hosted articles:
by guest on July 21, 2011jas.fass.orgDownloaded from