SELECTION METHODS FOR GENETIC IMPROVEMENT OF

INDONESIAN FOWL COMPARED BY SIMULATION

BY

Maria Astuti

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

DEPARTMENT OF DAIRY SCIENCE

1978

ABSTRACT

SELECTION METHODS FOR GENETIC IMPROVEMENT
OP INDONESIAN FOWL COMPARED BY SIMULATION

BY

Maria Astuti

Effectiveness of four methods of selection, mass selection,

selection index, restricted selection index and independent culling
levels, was studied for each of nine simulated native chicken popula-

The nine populations were chosen arbitrarily out of the possible
.3, .IL

tions.

<316QQ<Q_VCV

combinations of three levels of heritability of body weight (.4,
.2, .1), heritability of egg weight

heritability of egg production (.3,
.1), genetic correlation between body weight and egg production

(.5, 03'
(-.6, -.4, -.2) and genetic correlation between body weight and egg

weight (.5, .3, .2).
Nine generations of selection were performed by truncation,

where upper 1.25% males and 12.5% females were saved for breeding.

The economic value per unit of body weight was 25 times the

economic value per unit of egg production.
Mass selection showed that predictions of direct responses were
When negative

more accurate than those for correlated responses.
genetic correlation and heritabilities were low, the correlated

responses were predicted more accurately than when both were high.

Selection index showed that within the same combination of heri-
tabilities, mean genetic change of body weight was highest when genetic
correlation between body weight and egg production was lowest, but on

the contrary the smallest negative change in egg production occurred

Maria Astuti
in the population with the lowest negative genetic correlation. As in
selection index, restricted selection index and independent culling
levels also showed that the magnitude of negative genetic correlation
affected the mean genetic change of the two traits in opposite ways.

Mean body weight response was largest when mass selection was
used, and was about doubled at generation ten when heritability was .4,
but egg production was decreased 1.7% per generation.

Comparable responses on mean body weight were obtained by
selection index and restricted selection index, and independent culling
levels showed slightly smaller responses.

Selection index showed no decrease in egg production in popula-
tions with genetic correlation equal to -.2.

Restricted selection index produced decreases in egg production
40%-SO% less than mass selection when genetic correlation was strong
(-.6 or -.4) regardless of heritabilities.

When selection index was considered 100% efficient, mass
selection was 20% less efficient and independent culling levels were 10%
less efficient, especially when genetic correlation was strong.

Examination of genetic correlations showed that during selection
experiments in all populations and for all methods of selection the
initial genetic correlation was maintained.

Mass selection and independent culling levels produced genetic
correlation in the selected offspring somewhat weaker than in the
unselected offspring, but the difference was clear only when the genetic
correlation was strong. Selection index and restricted selection index

maintained genetic correlation in all cases.

ACKNOWLEDGMENTS

I am deeply indebted to Dr. John L. Gill, my major professor
for his patient assistance and invaluable suggestions in the preparation
of this manuscript. I would like to express my appreciation for his
guidance and encouragement in my graduate program.

I am grateful to Dr. C. E. Meadows, Dr. T. H. Coleman,
Dr. W. T. Magee and Dr. R. W. Erickson for their participation and
suggestions in my graduate program, to Dr. R. R. Neitzel for his
assistance in computer programming.

I am also grateful for the leave of absence provided by the
Gadjah Mada University, for the financial support provided by the
Midwest Universities Consortiums for International Activities (MUCIA,
Inc.) and for the facilities provided by the Department of Dairy

Science at Michigan State University.

ii

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . .
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . .

I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . .
II. LITERATURE REVIEW . . . . . . . . . . . . . . . . . . .

Improvement of native stock . . . . . . . . . . . . . .

Monte Carlo studies in genetics . . . . . . . . . . . .

III.

IV.

V.

Genetic correlation and correlated response . . . . . .
Methods of selection . . . . . . . . . . . . . . . .

METHODS AND PROCEDURES . . . . . . . . . . . . . . .

Parameters used in simulated populations . . . . . . .
Simulated mating system . . . . . . . . . . . . . . . .
Generation of base parental population . . . . . . .
Generation of offspring . . . . . . . . . . . . . . .
Methods of selection . . . . . . . . . . . . . . . .
The process of simulation . . . . . . . . . . . . . . .

RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . .

Mean genetic change under mass selection . . . . . . .
Mean genetic change when selection index was used in
females . . . . . . . . . . . . . . . . . . . . . . .
Mean genetic change when restricted selection index was
used in females . . . . . . . . . . . . . . . . . . .
Mean genetic change when two stage selection was used
in females . . . . . . . . . . . . . . . . . . . . .
Change of the phenotypic mean under different method of
selections . . . . . . . . . . . . . . . . . . . . .
Relative percent efficiencies of different selection
methods for total economic response . . . . . . . . .
Genetic correlations during selection study . . . . . .

IMPLICATIONS FOR THE NATIVE CHICKEN POPULATION IN

INDONESIA . . . . . . . . . . . . . . . . . . . . . . .

iii

Page

vii

.1).

00b

16
22
22
23
23
26
28
30
33
33
37
4O
42
44
58
80

96

SUMMARY AND CONCLUSIONS

LITERATURE CITED . . .

iv

Table

10.

11.

12.

LIST OF TABLES

Page

Parameter Combinations Selected for Simulation Study . . . 34
Expected and Observed Mean Genetic Change on Body Weight,

Egg Production and Egg Weight Per Generation under

Mass Selection . . . . . . . . . . . . . . . . . . . . . 35
Expected and Observed of Mean Genetic Change on Body Weight

and Egg Production per Generation when Selection Index

Was Used in Females . . . . . . . . . . . . . . . . . . 39
Expected and Observed of Mean Genetic Change on Body Weight

and Egg Production per Generation When Restricted

Selection Index was Used in Females . . . . . . . . . . 41
Mean Genetic Change on Egg Production in Females under

Restricted Selection Index . . . . . . . . . . . . . . . 41
Expected and Observed Genetic Change of Body Weight and

Egg Production per Generation when Two Stage Selection

was Used in Females . . . . . . . . . . . . . . . . . . 43
Means of Body Weight, Egg Production and Egg Weight in

Population I Under Different Methods of Selection . . . 46
Means Body Weight, Egg Production and Egg Weight in

Population II Under Different Methods of Selection . . . 47
Means Body Weight, Egg Production and Egg Weight in

Population III Under Different Methods of Selection . . 48
Means Body Weight, Egg Production and Egg Weight in

Population IV Under Different Methods of Selection . . . 49
Means Body Weight, Egg Production and Egg Weight in

Population V Under Different Methods of Selection . . . 50
Means Body Weight, Egg Production and Egg Weight in

Population VI Under Different Methods of Selection . . . 51

Table Page

13. Means Body Weight, Egg Production and Egg Weight in
Population VII Under Different Methods of Selection . . . 52

14. Means Body Weight, Egg Production and Egg Weight in
Population VIII Under Different Methods of Selection . . 53

15. Means Body Weight, Egg Production and Egg Weight in
Population IX Under Different Methods of Selection . . . 54

16. The Expected and Observed of Total Response Per Generation
in Economic Unit . . . . . . . . . . . . . . . . . . . . 77

17. Relative Percent Efficiency of Different Selection
Methods . . . . . . . . . . . . . . . . . . . . . . . . . 79

18. Range of Sample Values of Genetic Correlation of Body

Weight and Egg Production and Body Weight and Egg
Weight in Base Population . . . . . . . . . . . . . . . . 82

vi

LIST OF FIGURES

Figure

1.1

1.2

1.7

1.8

2.1

2.2

2.3

2.4

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and
Under Different Method

Means of Body Weight and

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Egg Production in
of Selections . .

Population I

Population II

Population III

Population IV

Population V

Population VI

Population VII

Population VIII

Population IX

Egg Weight in Population I Under
Different Method of Selections . . . . .

Means of Body Weight and Egg Weight in Population II Under
Different Method of Selections . . . . .

Means Body Weight and Egg Weight in Population III Under
Different Method of Selections . . . . .

Means of Body Weight and Egg Weight in Population IV Under
Different Method of Selections . . . . .

vii

O O O O O O O O

Page

59

6O

61

62

63

64

65

66

67

68

69

7O

71

Figure Page

2.5 Means of Body Weight and Egg Weight in Population V Under
Different Method of Selections . . . . . . . . . . . . . 72

2.6 Means of Body Weight and Egg Weight in Population VI Under
Different Method of Selections . . . . . . . . . . . . . 73

2.7 Means of Body Weight and Egg Weight in Population VII
Under Different Method of Selections . . . . . . . . . . 74

2.8 Means of Body Weight and Egg Weight in Population VIII
Under Different Method of Selections . . . . . . . . . . 75

2.9 Means of Body Weight and Egg Weight in Population IX
Under Different Method of Selections . . . . . . . . . . 76

3.1 rG12 and rGl3 in the Unselected and Selected Offsprings

Under Mass Selection . . . . . . . . . . . . . . . . . . 83

3.2 rG12 and rGl3 in the Unselected and Selected Offsprings

Under Mass Selection . . . . . . . . . . . . . . . . . . 84

3.3 r612 and rGl3 in the Unselected and Selected Offsprings

Under Mass Selection . . . . . . . . . . . . . . . . . . 85

rG12 and rG13 in the Unselected and Selected Offspring

Under Selection Index . . . . . . . . . . . . . . . . . 86

rG12 and rGl3 in the Unselected and Selected Offsprings

Under Selection Index . . . . . . . . . . . . . . . . . 87

4.3 rG12 and rG13 in the Unselected and Selected Offsprings

Under Selection Index . . . . . . . . . . . . . . . . . 88

5.1 rG12 and r613 in the Unselected and Selected Offsprings

Under Restricted Selection Index . . . . . . . . . . . . 89

5.2 rG12 and rGl3 in the Unselected and Selected Offspring

Under Restricted Selection Index . . . . . . . . . . . . 90

5.3 rGl2 and rG13 in the Unselected and Selected Offsprings

Under Restricted Selection Index . . . . . . . . . . . . 91

6.1 r612 and rG13 in the Unselected and Selected Offspring

Under Independent Culling Levels . . . . . . . . . . . . 92

viii

Figure Page

6.2 rG12 and rGl3 in the Unselected and Selected Offsprings

Under Independent Culling Levels . . . . . . . . . . . . 93

6.3 rG12 and IG13 in the Unselected and Selected Offsprings

Under Independent Culling Levels . . . . . . . . . . . . 94

ix

INTRODUCTION

Nature has provided that animals will reproduce, but if we
let them reproduce naturally, we arrive at the problem of lack of
nutritious food from animals. This problem we observe now in part of
the world (underdeveloped and developing countries) where men haven't
interfered much in the life pattern of animals. On the other hand, in
developed countries lack of production from animals is not a problem
since for many decades men with science and knowledge have been able to
improve and change animal production and reproduction to suit their
needs.

Improvement in animal production is brought about primarily
by the change in genetic composition of animals through careful
selection of breeding stock. So it's obvious if developing countries
want to solve their food problem, an effort toward the increase of
animal production should be put forward.

Ideas to increase production by crossbreeding, upgrading,
linebreeding and entire replacement with exotic breeds have been
suggested, but all of these can't be justified and wholly adapted
without first conducting a study of the genetic potential of the
existing populations.

The author is interested in planning and working toward the

improvement of the native chickens in Indonesia. The native chicken

population is a good example of producing and reproducing as nature
provided. They are raised widely in the traditional way, well adapted
to local conditions and characterized by low production of meat and
eggs. Yet they provide nearly 90% of chicken meat supply for the
nation.

An extensive and intensive study of the population of chickens
by survey and experiment is necessary for estimation of genetic
parameters. The author is interested in improving body weight of
chickens at market age (sexual maturity). Most studies with other
breeds of chickens have indicated that body weight has a moderate to
high heritability and the variability of body weight is due primarily
to additive genetic variance. Therefore the author believes that
selection within the native chicken population should give a good
response.

As the practical studies can't be conducted for the time being,
a Monte Carlo simulation process may be used to mimic the native
chicken population. In this study three different levels of heri-
tability of body weight, egg production and egg weight, and three
different levels of genetic correlation between body weight and egg
production, and between body weight and egg weight are used in
combinations to simulate potential progress.

The objectives of this study are:

1. To design an optimal breeding plan to improve body weight at
sexual maturity.
2. To compare effectiveness of four different methods of selection:

(a) Selection on body weight by truncation on phenotypic

value and monitoring the change of egg production and egg
weight as correlated response.

(b) Selection on body weight and on egg production
simultaneously.

(c) Selection on body weight with the imposed restriction of no
change in egg production.

(d) Selection at two stages, first on body weight then on egg
production.

To predict direct response to selection under different combi-

nations of parameter values and selection methods.

To examine correlated responses under different combinations of

parameter values and selection methods.

LITERATURE REVI EW

Improvement of native stock

 

Recently there has been considerable interest and effort in
improving animal production in developing countries.

From the genetic standpoint animal improvement can be achieved
through selection and breeding.

Although suggestions have been made for genetic improvement
via crossbreeding, upgrading, replacement of native stock by imported
animals or selection within the native populations, they do not lead
to an easy decision. However McDowell (1972) cited a recommendation
from an expert on animal breeding and climatology of F.A.O. who
recommended the following procedures:

1. Objectively evaluate presently available animals and the
environments to which they are exposed.
2. Experimental work should be conducted to assess the merits

of indigenous animals, and their ability to respond to improved

environments. When sufficient merit exists, then programs of

improvement through selection should be implemented.
3. When evaluation indicates the need for grading up, then
animals that will be likely to contribute the desired end

result should be chosen.

From this recommendation it is obvious that a careful study of
genetic potential of native animals is a prerequisite for a plan for
further improvement.

Most improvements of native stock have been done by cross-
breeding or upgrading, primarily in dairy cattle and in swine. However,
a number of native stocks that have been improved without infusion of
foreign blood are Arabian horses, Brahman and Boran beef cattle,
Karakul sheep, Angora goats, several Chinese breeds of pig and the
Awassi sheep in Israel.

The improvement of Awassi sheep as dairy animals is interesting.
The program was initiated 50 years ago with local stock. Epstein
(1977) reported that the improvement in the course of 30 years from
1943 to 1973 was reflected by the number of ewes with actual lactation
yields of more than 400 kg. In 1943 only 0.08% of all milk recorded
animals yielded more than 400 kg but by 1973 the percentage had
increased to 18.9%. Some important things to note from this report
are:

1. milk yield is an additive trait

2. selection was practiced within the native flock by selecting
the additive genetic variation

3. improvement was achieved without introducing outside

genetic material.

In chickens it has been reported that the variance of body
weight is mainly due to additive genetic variance and the heritability
is quite high (Martin et al., 1953; Brunson et al., 1956; Kan et al.,

1959 and Siegel, 1962).

Varieties of native chickens are found in developing countries,
and dominate the chicken populations in the rural areas.

Uddin (1962) and McArdle (1972) suggested the improvement of
native chickens for egg production by crossbreeding or upgrading.

El-Issawi (1973) reported that in Egypt, Baladi, Fayoumi and
Dandarawi were developed as layers from indigenous breeds.

No reports were noted on selecting native chickens for meat
production.

Genetically there is a similarity between body weight in
chickens and milk production in Awassi sheep. Although the heritability
of milk production in Awassi sheep was not reported, the author
believes that the heritability is moderate because of the long term
response to selection. If the improvement of native chickens is con-
centrated on body weight instead of egg production, then selection on
additive genetic variation of body weight within the native population

seems very promising.

Monte Carlo studies in genetics

 

Monte Carlo technique has been introduced in genetic and
selection studies to confirm genetic theories and to predict results.
Monte Carlo methods comprise that branch of experimental mathematics
which is concerned with problems using random numbers. As the
genetic systems obey the laws of probability, they can be investigated

using probabilistic Monte Carlo methods.

Fraser (1957a) introduced the first techniques for the simu-
lation of genetic systems by high speed electronic digital computers.

The genetic framework was presented as a binary system in which the

symbols 0 and l were regarded as identifying the two alleles of a
single locus. An array of alleles in the chromosome was generated by
adopting the binary representation of the generated random numbers.

Fraser (1957b) studied the effects of linkage on the rate of
advance under selection. His results indicated that linkage in large
populations caused a correlation of rate of advance with degree of
linkage when the recombination is less than .5%. However, in small
populations the effect of linkage is considerably exaggerated. In his
next three papers on the subject Fraser (1960, a, b, c) discussed
the effects of linkage, dominance and epistasis. The study on epistasis
was extended by considering the effects of reproductive rate and
intensity of selection. Fraser also simulated selection against pheno-
typic extremes.

Since the works by Fraser, many simulation studies have been
reported, which primarily were concerned with artificial selection
theory. These studies contributed to the understanding of some of the
operating genetic mechanisms.

A detailed review of these studies is not warranted, but
examples of types of studies may serve to summarize.

Barker (1958a, b) investigated selection between alleles
at an autosomal locus and at a sex linked locus.

Martin and Cockerham (1960) reported the effect of linkage
on the progress from mass selection in small populations.

Gill (1965a, b, c) studied the effects of population size,
linkage, selection intensity and environmental variation upon genetic

change in simulated populations. In other studies Gill and Clemmer

(1966) reported the effect of selection and linkage on degree of
inbreeding. And a similar study was reported by Bogyo and Ting (1968).

Young (1966) presented the results from simulation studies on
the rate of decay of the additive genetic variance due to selection
and the resulting change in heritability.

Hill and Robertson (1966) studied the effect of linkage on
limits to artificial selection.

Parker et a1. (1969, 1970a, b) examined selection response and
genetic correlation in simulated populations.

Bereskin (1969, 1972) reported the effect of inbreeding on
genetic and phenotypic trends and selection response.

Bulmer (1976) studied the effect of selection on genetic
variability.

Sather et a1. (1977) reported the effects of population size,
selection intensity, mating system and selection method in simulated
populations.

Data for those studies were generated on the assumption of
known initial gene frequencies and number of loci affecting the traits.

However in other studies Monte Carlo methods were used to
generate data for metric traits for which realistic specification of
gene frequencies and number of loci can't be made. The generation of
data was based on a specified biological model and statistical
. estimates of biological parameters. Several studies using this
procedure have been reported.

Magee (1965) studied the estimation of response to selection

when selection was applied for two traits.

Everett et a1. (1967) used simulated records of milk pro-
duction to estimate the genetic and environmental trends.

Strohbehn (1974) evaluated response to selection and mating
systems with different levels of heritability, selection intensity
and progeny testing.

Bruns and Harvey (1976) studied the effects of varying selection
intensity for two traits on estimation of realized genetic parameters.

These studies show that when data can't be made available from
a real experiment, then Monte Carlo methods for generation of biolo-
gical data make possible selection studies, which can serve as guides
for establishing long term experiments that will have optimal chance

of success.

Genetic correlation and correlated response

 

Genetic correlation between traits is important when dealing
with metric traits, as selection applied to one trait generally results
in correlated response in other traits not under selection.

The most important underlying cause of genetic correlation
appears to be pleiotropy, i.e., some genes affect two or more traits.
Other possible causes of genetic correlation are usually considered
to be minor or transient such as linkage (Falconer, 1960).

Two traits can be correlated genetically either positively or
negatively and theoretically in selection experiment the correlated
response can be predicted if the genetic correlation and the herita-
bilities of the two traits are known.

Many studies on genetic correlation and correlated response

have been reported, with considerable emphasis on

10

1. whether the theoretical treatment of correlated response to
selection in terms of the genetic correlation is adequate to
explain the responses realized in experiments
2. asymmetrical correlated response
3. the effect of selection on genetic correlation
4. negative genetic correlation between traits
Falconer (1954) experimented with mice to check the validity
of the theory of genetic correlation. Two way selection was made for
body weight in one pair of lines and for tail length in another pair.
The results showed an agreement between the two independent estimates
of the genetic correlation between weight and tail length, one from
each pair of lines. This agreement showed that the theoretical
treatment of the genetic correlation between two traits adequately
accounts for the observed correlated responses.

Reeve and Robertson (1953), selected for wing and thorax

length in Drosophila melanogaster for fifty generations. The genetic

 

correlation between traits is .7. A good agreement was found between
estimates of the genetic correlation between the two characters in
their base population and the correlated responses when either of the
two were selected separately.

Clayton et al. (1957),working with Drosophila melanogaster,

 

observed the correlated response on number of sternopleural bristles,
when selection was practiced on number of abdominal bristles. The
genetic correlation between abdominal bristles and sternopleural
bristles in the base population was small but slightly positive

(.05 to .10). They reported that a moderate agreement with predicted

11

correlated response was observed in earlier generations but the
correlated response became entirely unpredictable with further
selection in later generations. They suggested that the magnitude of
genetic correlation affects the accuracy of the expected response and
the effect of random drift in the correlated trait causes the response
somehow to deviate from the expectation when the genetic correlation
is weak.

Asymmetrical correlated response in selection experiments have
been reported, where discordance of the pattern of correlated response
with the expectation was observed.

Falconer and Latyszewski (1952) experimented with mice, where
mice were selected for growth rate on high and low planes of nutrition.
They considered the same measurement made under two different environ-
ments as two separate traits. Genetic correlations observed were equal
in early generations but were markedly different in later generations.
They attributed this asymmetry to changes in the basic parameters due
to selection applied.

Bell and McNary (1963) and Yamada and Bell (1963) also
observed asymmetry of genetic correlations when selection was applied

to Tribolium castaneum under two different environments.

 

Clayton et a1. (1957) reported asymmetry in correlated
response of sternopleural bristle number in Drosophila melanogaster
when selection was practiced to increase or decrease abdominal
bristle number.

Siegel (1962) reported results of a double selection
experiment which provided a comparison of the genetic correlations

obtained between body weight and breast angle at eight weeks of age

12

in white Plymouth Rock chickens. The results showed that selection for
body weight apparently resulted in greater changes in breast angle than
selection for breast angle did for body weight. The relative efficiency
of changing breast angle through selection on body weight as compared
to direct selection for breast angle was 71% in males and 94% in
females, but, the percentage efficiency for changing body weight
through selection for breast angle as compared to direct selection for
body weight was 33% for males and 29% for females. He concluded that
the asymmetrical response could be due to unequal heritabilities, where
the realized heritability of eight week body weight (.31 for males and
.28 for females) was approximately one third larger than the respective
value for breast angle (.24 for males and .21 for females).

Nordskog and Festing (l962),also experimenting with chickens,
reported an asymmetry of the realized genetic correlation between body
weight and egg weight when selection was applied for both high and low
directions for body weight and egg weight.

Bohren et a1. (1966) and Parker et al. (1970a) reported asym-
metrical correlated response from simulation selection experiments.

Bohren et a1. (1966) suggested that in experiments it is
unlikely and more surprising to find symmetrical than asymmetrical
correlated response.

Parker et al. (1970a) reported asymmetrical response in later
generations although the response was symmetrical in early genera-
tions.

The effect of selection on genetic correlation has been of

considerable interest to research workers. Falconer (1960) theorized

13

that if selection has been applied to two traits simultaneously the
genetic correlation between them should be expected eventually to
become negative, because those pleiotropic genes that affect both
traits in the desired direction will be strongly acted on by selection
and brought rapidly toward fixation. They will then contribute little
to the variances or covariances of the two characters. The pleiotropic
genes that affect the traits in opposite ways are less strongly
influenced by selection and will contribute most of remaining covari-
ances of the two traits, and the resulting genetic correlation will be
negative.

Friars et a1. (1962), Lentz and Gowe (1974), Emsley et a1.
(1977), from selection experiment with poultry, reported a decline over
time in estimates of genetic correlation for traits related to
egg production.

Parker et a1. (1969, 1970a, b) examined the effect of truncation
selection of a primary trait upon the genetic correlation and the
correlated response in a secondary trait in simulated populations.
Results from an additive model indicated that high selection intensity
resulted in a decline in genetic correlation while low selection inten-
sities maintained the genetic correlation. A model with complete
dominance showed that the change in genetic correlation where selection
was by upper truncation followed the same patterns as for the additive
model. Selection by lower truncation with high intensity of selection
showed a more rapid decrease in genetic correlation.

Cheung and Parker (1974) conducted an experiment with mice
over 14 generations. They reported that estimates of realized genetic

correlation between 6-week body weight and 6-week tail length showed

14

that more intense selection led to some decrease in genetic correlation
between the two traits.

Knowledge of the magnitude of the genetic correlation and the
direction of the change of the correlated response in a secondary trait
upon selecting the primary trait is important. Breeders are more con-
cerned with the existence of negative genetic correlation as simulta-
neous selection for two traits with negative genetic correlation may
produce little progress and selection to improve one trait will cause
a decrease in the other trait.

Evidence of negative genetic correlation among economic
traits in livestock and poultry has been reported by research workers.

In dairy cows a negative genetic correlation between milk
yield and butterfat percentage was reported by Johansson (1950),
Freeman and Dunbar (1955), Tabler and Touchberry (1955), Farthing and
Legates (1957) and Johnson (1957). The values reported were in the
range of -.58 to -.20.

Fredeen and Jonsson (1957), as cited by Falconer (1960),
reported a genetic correlation of -.47 between body length and backfat
thickness and -.96 between growth rate and feed efficiency (lbs feed/
1b gain) in pigs.

In chickens, evidence of negative genetic correlation between
body weight and egg production has been reported. This means that
joint selection for body size and egg production in a given flock
reduces progress in each trait as compared to that made with the same
selection differential provided no selection pressure is placed on the

other trait. This has special importance in the poultry industry for

15

broiler producers. The broiler strains have large body size but are
low in egg production, so it would be impossible for the broiler
breeders to improve both traits significantly by simultaneous selection.

A wide range of values of negative genetic correlation between
body weight and egg production has been reported.

When experimenting with New Hampshire chickens, Dillard et a1.
(1953) reported genetic correlations of -.17 and -.20, Jerome (1956)
obtained a correlation of -.59 and Shimizu et al. (1968) observed a
moderately negative genetic correlation.

Nordskog (1960) studied White Leghorns and reported a positive
correlation of +.08, not significantly different from zero. In the
heavy breeds he obtained a highly significant negative genetic
correlation of -.35.

Siegel (1963), using White Plymouth Rocks in his study,
reported a range of negative correlation from -.01 to -.71.

Gowe et a1. (1973), measuring the genetic correlation in
White Leghorns chickens, reported values in the range of —.17 to -.37.

Working with broiler chickens, Mehta et al. (1975) reported
a correlation of -.14.

These studies indicated that negative genetic correlation
between body weight and egg production is obvious and substantial.

To improve native chickens in developing countries for meat production,
attention should be given to the existence of the negative correlation,
especially because it is known that egg production in native chickens
is relatively low. Both traits should be considered when one formu-

lates selection schemes.

16

Methods of selection

 

In plant or animal breeding three methods of selection can be
used if breeders are interested in selecting more than a single
character.

1. Tandem method.

Selection is practiced for only one trait at a time until

satisfactory improvement has been made in this trait.

Selection efforts for this trait are then relaxed, and efforts

are directed toward the improvement of a second trait, then a

third, and so on.

2. Independent culling method.

A certain level of merit is established for each trait and all

individuals below that level are discarded regardless of their

superiority or inferiority in other traits.
3. Selection Index method.

Extra merit in one trait may offset defects in another. Each

trait is weighted by its relative economic value and a linear

function of such values gives an index value for net merit
based on all the traits. Then animals with the highest index
value will be kept for breeding purposes.

The selection index method has been the subject of number of
studies either theoretical or experimental.

Smith (1937) developed an index designed for the selection of
plant lines, using Fisher's concept of discriminant functions to
derive a linear equation based on observable characteristics as the

best available guide to the genetic value of each line.

l7

Hazel (1943) by use of Wright's path coefficients, developed
a selection index which maximized the correlation between the index
and the aggregate genetic value of the traits selected for. Hazel's
selection index is defined as:

+----+bx

= +
I b x b2x2 n n

l l
where the bi's are the derived optimum weighting coefficients for
the characters xi.

The aggregate genetic value of the index is defined as
H = algl + a292 + - ' ' + aigi + ' ' ' + angn
where the a's are economic values corresponding to one unit of the
xi. The problem is to obtain a set of values for the bi's which
maximizes the correlation between I and H. The desired solution to
the bi's is obtained from a set of simultaneous linear equations
which can be presented in matrix notation as:
are.
where b_is a column matrix of unknown bi's
§_is the phenotypic variance covariance matrix
G is the genotypic variance covariance matrix
“a is the column matrix of economic weighting factors
Hazel and Lush (1942), Young (1961) and Finney (1962) presented
theoretical papers comparing the three methods: tandem selection,
independent culling, and selection index. They theorized that
selection index is never inferior to tandem selection or independent
culling. However, Rasmuson (1964L found contrary to theoretical

expectations,that selection based on independent culling levels was

superior in two experiments to selection index for increasing the

18

total number of bristles in Drosophila melanogaster. On the other hand,

 

Sen and Robertson (1964) compared index, independent culling levels, and
tandem methods for increasing both abdominal and sternopleural bristles,
and concluded from their results that the index method was superior as
predicted by selection theory.

In practice there are situations when we want to change the mean
of one trait while holding the mean of a second trait constant. For
this purpose the concept of a restricted selection index has been
derived mathematically by Kempthorne and Nordskog (1959). The idea is
that the usual index equations be solved subject to the simultaneous
conditions that the covariance between the index and a linear function
of the genotypes involved be zero, thus preventing selection on the
index from causing any genetic change in this function.

In solving for the index weighting factors, Kempthorne and
Nordskog (1959) did not retain the equations carrying the reStriction
explicitly in the system.

The index weighting factor is obtained by solving the final
form of equation:

.12 = [i - 25199 (99195;) 499.] Bilge
where: b_is a column vector of index weighting factors
is identity matrix

is phenotypic variance-covariance matrix

IC)

is genotypic variance—covariance matrix

0 I Q Ith O I
is a column vector of zero's Wlth a one in the i pOSition,

If)

.t . . . . . .
where the i h trait is the trait subject to the restriction

is a column vector of economic values

|m

19

The effect of restriction is to multiply the unrestricted
selection index by a matrix, obtained after performing a sequence of
matrix operations.

Cunningham et a1. (1970) presented a paper on restricted
selection index and showed a different procedure for the solution of
index weighting factors. The procedure has proved to be very practical
and simple. A matrix equation taking the original form of unrestricted
selection index with a modification of rows and columns is used. In
this method the equation carrying the restriction is retained.

The equation is:

2 £3.- 2 .9
J = a. or
I
93-2 ed __
P*b* = G*a

where E is the phenotypic variance-covariance matrix
§_is the genotypic variance-covariance matrix
. .th . . .
G. is the 3 column of the genetic variance-covariance
*9
. .th . . . . .
matrix for the j trait, the trait subject to restriction
.b is a column vector of index weighting factors
2d is a column vector of dummy weighting factors
a_is column vector of economic values
Cunningham's method is equivalent to Kempthorne and Nordskog's as
the solution for b_in Cunningham's equation is:
-1 -l -l 1 -1
b=E|I-PG.G'.P G'. G'.]PGV
'- —' -' ‘9 (‘9 -‘ “9) ‘0 " "'
. .t . . . .
where 9a is the 3 h column of the genetic variance-covariance matrix,

expressed as §_g_ in Kempthorne and Nordskog's method.

20

Several experiments have been devoted to verification of the
theory of the restricted selection index. Abplanalp et a1. (1963)
selected Broad Breasted Bronze turkeys for seven generations using an
index calculated to produce maximum gains in 9-week weight while
holding 24-week body weight constant. The results were in fair agree-
ment with predicted values.

Scheinberg et a1. (1967) compared relative efficiency of
restricted index selection and unistage tandem selection on three

traits in Tribolium castaneum. They reported that the results in

 

general confirm the biological validity of the restricted selection
index theory.

Okada and Hardin (1967) experimenting also with Tribolium
castaneum, reported results from 7 generations of selection.
Restricted selection index was effective for increasing 14-day larva
weight while holding 30-day adult weight constant, and when selection
was continued into 13 generations, the results showed the same trend.
They suggested that the restricted selection index is biologically
valid (Okada and Hardin, 1970).

Eisen (1977) conducted a study on bidirectional restricted
index selection for 10 generations in mice. His goal was to maximize
genetic change in postweaning weight gain from 3 to 6 weeks of age,
while keeping genetic response in feed intake at zero. The results
indicated that the first four generations of selection appeared to
verify the biological validity of the restricted selection index since
genetic gain in feed intake was essentially zero. The overall

regressions of feed intake on generation number or on index units

21

were not significantly different from zero, but nevertheless subsequent
to generation four feed intake tended to change in the same direction
as post weaning gain in both index lines. He explained that the
failure of the intended complete genetic restriction on change in

feed intake may have been the result of a biological incompatibility
between the goals of the restricted index and the genetic correlation
between the traits in the index.

In summary, most studies have shown that, theoretically and
experimentally, index selection is superior to other methods of
selection, and that a restricted selection index is effective in
permitting genetic progress in one trait without substantial genetic
change in a correlated trait.

Most selection studies have involved two traits with positive
genetic correlation, but no studies have been noted in which the

genetic correlation between two traits under selection is negative.

METHODS AND PROCEDURES

Parameters used in simulatedgpgpulations

Values of genetic parameters used in this simulation process
are based on reviews of literatures concerned with other populations
of chickens. It's been assumed that the genetic variability of body
weight is only due to the additive genetic variance, i.e., dominance
and epistasis play a trivial role.

Three levels of heritability for each trait and three levels
of genetic correlation between body weight and egg production and
between body weight and egg weight were selected. The selected
heritability parameters for body weight were .1, .3, .4, for egg
production .1, .2, .3, and for egg weight .1, .3, .5. The selected
genetic correlations for body weight and egg production were -.2,
-.4, -.6, and for body weight and egg weight .2, .3, .5.

The parameters used for mean body weight, 12 month egg
production and egg weight were 1.6 kg, 104 eggs, 38 g, respectively
and the standard deviations were .2 kg, 5 eggs, 4 g respectively,
based on recent studies reported by Wartomo et al. (1976). For each
trait the environmental variances corresponding to different
heritability values were calculated easily, as the additive variance

was the only genetic variance.

22

23

Simulated mating system

The size of breeding flock was kept constant through all
generations: 12 males and 120 females were divided into three groups,
I, II, and III, with 4 males and 40 females in each group. Within each
group, males and females were assigned randomly and equally to four
pens; therefore 1 male and 10 females were in each pen.

To simulate 4 rotations of mating in one year, within each
group the males were shifted from pen to pen, so each male mated with
40 females to make a sire's family. From each mating, each sire had 4
offspring, 2 males and 2 females, so the sire's family consisted of
80 males and 80 females. To continue the next generation 1 male and
10 females were selected from each sire's family, so that 1.25% males
and 12.5% females were saved for breeding. The next mating of the
selected parents in each generation was done by shifting the males in
group II to group I, selected males in group III to group II, and
selected males in group I to group III, while the females remained in
assigned groups.

With this mating system, inbreeding was assumed to be zero

or trivially small.

Generation of base parental population

 

A repetitive pseudo-random number generator was used to
simulate the probabilistic genetic mechanism. A library program,
RANF, was available in the MSU Computer system (CDC 6500) for the
generation of uniformly distributed random numbers in the range

0 < r. < +1.
_. 1 _.

24

For the uniform distribution, variance is equal to the square of the
range divided by 12.

Therefore, V(ri) = (1)2/12 = l/12

Twelve random uniform numbers (ri) were added together to form a random
sum in the range 0 to 12.

Then the variance of the sum equals 12 (1/12) = 1, $0 the standard
deviation of the sum also is 1.

By subtracting six from any sum of twelve ri, one obtains
random numbers in the range of -6 f.ej :_+ 6, and the ej are normally
distributed with a mean of zero and standard deviation of 1, providing
standard normal random deviates (RND). It is well known that the
deviates so produced do not conform well to the normal distribution in
the extreme tails, but that is of little consequence in studies of
this type.

The base populations were generated to provide genotypes of
the male and female parents from which to generate the progeny from
which selections were made. For convenience in notation, subscripts
l, 2 and 3 are used for body weight, egg production and egg weight
respectively on symbols for genetic and phenotypic parameters. Sub-
scripts on random normal deviates are consecutive as deviates are
introduced into simulation methodology. Similar numbering applies
to constants introduced.

Genetic deviation of body weight in an individual is
G = Gl(RND)l

additive standard deviation of body weight and

s
{3'
{D
H
(D
8
II

E

a random normal deviate

25

The genetic deviation of egg production in the same individual

must include a term OG1(RND)1 Cl' where Cl is a constant value and

2
C =
Cov GlGZ/O G

1 1°

Therefore the genetic covariance between body weight and egg

production is Cov G G

1 2, which can be calculated for the required

genetic correlation. But the genetic deviation of egg production

cannot be generated simply as: G = OGl(RND)

C
2 l l
. . 2 2 . 2 2
From this expreSSion 0 G2 = 0 G1, but in fact 0 G2 # 0 G1

. . 2 .
To restore the variance of egg production to 0 G2 another term is

then added to the expression so that

G = US (RND)1 C

+
2 1 (RND)2 C

1 2

The variance of this expression is

2
0 G2 = 02G1 (Cl)2 + (C2)2, and the value of C2 can be calculated.

The phenotypic value of body weight is:

= + + O
P pl Gl (RND)3 El

where U - mean genetic value of body weight (1.6 Kg), and
CE = environmental standard deviation of body weight
Phenotypic value of egg production is:

P = U + G + (RND)4 0E

2 2 2 2
where 02 = mean genetic value of egg production (104 eggs), and
0E2 = environmental standard deviation of egg production

With the same procedure the genetic deviation of egg weight

can be obtained: G = 0G1 (RND)l C3 + (RND)5 C

3 4

26

The value of C4 can be calculated from

O“2 (I? 2 2
= +

G3 G1 (C3) (C4)

The phenotypic value of egg weight is
= + +

P 1.13 G3 (RND)6 OE

3

mean genetic value of egg weight (38 grams), and

2
:3‘
(‘D
H
(D
“C
U)
ll

8

environmental standard deviation of egg weight
In this study the standard deviations used for body weight,

egg production and egg weight are .2 kg, 5 eggs, and 4 g, respectively.
Twelve males and 120 females were generated as the base

parental population.

Generation of offspring

 

The genotypic value of male and female parents was used to
generate the offspring.
The genetic deviation of body weight in an offspring is

generated as
601 = 1: (GMl + GFl) + (V .5) (061) (RND)7

where symbols M and F stand for male and female.

Genetic deviation of egg production of the same individual

= Cov G G /02G .

must include a term (V .5) «5G ) (RND) (C ) where C
1 7 5 5 l 2 l

The genetic covariance between body weight and egg production is

Cov GlGZ' and can be calculated for the required genetic correlation.

27

But the generation of genetic deviation for egg production can't be
expressed simply as:

= / o
coz MGM2 + GFZ) + ( .5) ( Gl) (RND)7 (C5)

. . . 2 2 ,
The variance of this expreSSion gives 0G2 = GG1, but in fact
2 2 . 2
0G2 # 0G1, so another term is then added to restore 0G2, and the

genetic deviation for egg production is generated as:

602 = MGM2 + ch) + (V .5) (0G1) (RND)7 (C5) + (RND)8 (C6)

2

2 2 2
Further O'G2 = B O'G2 + 5 02Gl (C ) + (C6) , so the value of C can be

5 6

calculated.
The phenotypic values for body weight of male and female

offspring and egg production of female offspring were generated as

+ +
PO U GOl (El (RND)9 and

1 1

+ + '
PO u GO2 0E2 (RND)lo

2 2

With the same procedure the genetic deviations of egg weight for male
and female offspring were generated as

GO3 = 5(GM3 + GF3) + (V .5) (061) (RND)ll (C7) + (RND)12 (C8)

Then the phenotypic deviation of egg weight for female offspring is:

= + +
P03 U3 G03 OE (RND)13

Based on the breeding scheme that was presented before, each
male was mated to 40 females during the four rotations of mating in
one year and each mating gave 2 male and 2 female offspring. The

contribution of each male parent to the offspring population is then,

28

80 males and 80 females. Twelve males were used as parents, therefore
1920 offspring, 960 males and 960 females, were generated in each

generation.

Methods of selection

 

In every generation the size of breeding population was kept
constant at 12 males and 120 females. For this purpose 1.25% and
12.5% of males and females respectively were saved for breeding.
Four methods of selection were used and compared:

1. Within each sire family individuals were ranked. Selection
of males and females was based on phenotypic value of body
weight. Then one top male and 10 top females were selected.
Therefore, 12 males and 120 females were selected from 12
sire families to be used as parents for the next generation.

2. Selection on body weight and egg production simultaneously,
where males were selected based on their own phenotypic value
and females based on Index value. For each parameter combi-
nation, an index equation was developed:

I = b x + b x

l l 2 2
where x1 = phenotypic value of body weight
x2 = phenotypic value of egg production
b1 = weighting factor for body weight
b2 = weighting factor for egg production

The values for b1 and b2 were obtained by solving the normal

equations Pb’= Ga

where P phenotypic variance-covariance matrix

G genotypic variance-covariance matrix

29

a.= vector of economic values, where economic values for
body weight and egg production were set equal to 25
and 1 respectively.
In setting up the index an assumption was made that environ-
mental correlation equals zero. The top male based on own
phenotype and the top ten females based on index values were
selected from each sire family.
Selection on body weight where no change on egg production is
permitted. Males were selected on own phenotype and the top
male was selected from each sire family. Selection on females
was based on restricted index value. The restricted index was
adapted from Cunningham et a1. (1970).
The index used for selection is: I = b x + b x

1 1 2 2

The index was calculated for each of the parameter

combinations. x phenotypic value of body weight

1
x2 = phenotypic value of egg production
bl = weighting factor for body weight
b2 = weighting factor for egg production
The values for b1 and b2 were obtained by solving the equation:
3*? = 9*3
where P* = is a matrix obtained by adding a row and column to

the §_(phenotypic variance-covariance matrix), the

(n + 1) column and row of the Pf consists of the
column and row vector of genetic covariances between
the restricted trait (egg production) and body weight

and zero in the final position.

30

G* = is a matrix obtained by adding a row of zeros to
§_(genotypic variance-covariance matrix)
b* = is a column vector of b's and bd (dummy variate)

is a vector of economic values where economic

In:
ll

values for body weight and for egg production were
25 and 1 respectively
Using the index values, the best 10 females were then selected
from each sire family for breeding.

4. Selection at two stages, the first stage is selection on body
weight then at the second stage selection is on egg production.
Selection for males was based on own phenotypic value and the
top male was selected from each sire family. Selection for
females was done in two stages. Females were selected based
on their phenotypic value for body weight and from each sire
family 20 out of 80 females (25%) were saved at the first
stage, then at the second stage these females were selected
based on their egg production and 10 out of 20 females (50%)
were saved for breeding.

In all methods of selection egg weights were monitored
as the correlated response due to the positive genetic

correlation between body weight and egg weight.

The process of simulation

 

For all methods of selections, a FORTRAN program was made
general in generating the base populations and offspring populations,
and a variable "input" made it possible to simulate different parameter

combinations.

31

The process of simulations can be summarized as follows:

The FORTRAN program was composed of the main program and
several subroutines.

Each run of the program started by reading the additive genetic
standard deviations, and environmental standard deviations
selected for study.

Generating males and females in the base population: The
library program RANF was called to generate a Random Normal
Deviate. Then 12 males and 120 females were generated in the
main program according to the formula as it was explained
previously.

Generating offspring: This generation program took place in the
subroutine. The parents were mated as described before, then
the offspring populations were generated according to the
formula described previously.

The next step was selection of offspring, leaving 12 males and
120 females to be parents for the next generation. Two
subroutines were used, one for selected males and the other
for selected females. From this point the program was a
little different for each method of selection. Males were
selected in the same manner based on own phenotypic value

of body weight, but selection of females was dependent on

the method of selection, explained in the previous part.

To obtain ten generations of offspring a "do loop" was made in

the main program following generation of the base population.

32

Results were stored on magnetic tape to avoid excessive
printing and for statistical analysis.
The Stat-package program available in the MSU Computer system

was used for statistical analysis.

RESULTS AND DISCUSS I ON

Three levels of heritability of body weight, egg production and
egg weight and three levels of genetic correlation between body weight
and egg production and between body weight and egg weight make 81
possible parameter combinations. In this study, nine combinations were
selected arbitrarily to create nine different populations (Table 1).
Four different methods of selection were applied in each population.
For each method of selection, each population was permitted to pass
through 9 generations of selection and mating.

The printout of data was not obtained for each generation, but
data were stored on magnetic tape for the statistical analyses. Results
from the statistical analyses were printed out as means, variances-
covariances, and correlations of phenotype and genotype of body weight,
egg production and egg weight for the offspring population and for the
selected parents in each generation.

Replicated runs were considered unnecessary, as checking on
several replicated runs showed essentially the same results, because of
the large number of birds per run.

Meanpgenetic change under mass
selection (Method I)

 

 

The expected and observed genetic changes of body weight, egg
production and egg weight are presented in Table 2. The expected
genetic change on body weight was calculated by using the formula

33

34

Table l.--Parameter Combinations Selected for Simulation Study.

 

 

 

 

I
. ‘ 2 2 2 r r
Po lat . .
pu ion h BW h EP h EW. GBW EP GBW EW
I -.6 5
II .4 .3 .5 -.4 .3
III -.2 .2
IV -.6 .5
V .3 .2 .3 —.4 .3
VI -.2 .2
VII -.6 .5
VIII .1 .l .l -.4 .3
IX -.2 .2

 

 

 

 

presented by Falconer (1960): A81 = i hi 0P1, where i is selection

2 . , . . . . .
intensity, h is heritability for body weight and UP is phenotypic

1 1

standard deviation of body weight.

Genetic change on egg production:

AG2 = i h h rG 0P , where h:

l 2 12 2 is the heritability for egg

production, rG is the genetic correlation between body weight and

12

egg production and 0P2 is the standard deviation of egg production.

Genetic change on egg weight:
A=‘ hrG P, 2
G3 1 hl 3 13 o 3 where h3

is the genetic correlation between body weight and egg

is the heritability for egg
weight, rGl3
weight and 0P3 is the standard deviation of egg weight. From this
point subscripts l, 2 and 3 will be used to indicate body weight, egg
production and egg weight respectively.

The selection intensity was obtained from the table presented

by Becker (1975). Selection intensity for males and females were

35

 

 

 

 

 

 

 

 

50mm. omoa. oona.u mHHN.I mmvo. vmvo. xH
boam. Nvmm. wvmv.u ommv.u mmvo. vmvo. HHH>
thm. ommv. moah.l vmmo.| move. vmvo. HH>
memo. mmom. owmo.| mmam.l Nova. HBNH. H>
mmhh. mmoh. wmmo.an memo.an nmma. HBNH. >
avov.a momN.H hemm.an vomm.an moma. Anna. >H
Nmoo. numb. hamm.l hmm5.a mmma. vmoa. HHH
mmma.a owma.a nmmm.an mhov.al mama. vmoa. HH
bumm.a vvmm.a nmmw.au HHON.NI moma. vmma. H
om>ummno pwuommxm Um>uwmno owuommxm om>ummno pwuommxm
cofiumasmom
N
8V moq 3mm $23 3 85 H3

 

 

 

.cofiuowamm mmmz noon: soapmumcww Mom Davao: mom

 

pom :ofluosooum mom .unmflwz xoom co mmcmau oeuwcmw cmmz ow>ummno pom Umuommxmnl.m magma

36

2.589 and 1.647 respectively, where 1.25% males and 12.5% females were
selected for breeding.

The prediction of mean genetic change on body weight as a direct
response to selection was more precise than the prediction for egg
production and egg weight, both as correlated responses. This pro-
bably happened because the correlated responses were affected more by
random chance than the direct response.

The prediction of mean genetic change for egg weight as
correlated response due to positive genetic correlation was more
accurate compared to the prediction of egg production as correlated
response due to negative genetic correlation.

Despite a larger random error that seemed to affect data of
population IV, when the genetic correlation between body weight and
egg weight was low (.2) the accuracy of prediction was somewhat less,
a result also indicated by Clayton et a1. (1957).

The mean genetic change in egg production was predicted
accurately when the heritabilities were low and even more so when
genetic correlation was also low. Under high heritabilities the
prediction was poor, especially in combination with high negative
genetic correlation.

As expected, the mean genetic change in body weight increased
as the heritability increased from .1, .3 to .4. Within the same
levels of heritabilities of body weight, egg production and egg weight,
the observed correlated responses on egg production and egg weight
were increased as the magnitude of genetic correlation increased.

Under the same level of genetic correlation of body weight and egg

37

production and body weight and egg weight, the correlated responses on
egg production and egg weight were increased as the heritability
increased.

The effect of the magnitude of the genetic correlation on the
correlated responses depends on the level of heritability of the two
traits. The effect is greater in combination with higher heritability.
Further, combination between the highest heritability and the highest
genetic correlation gave the highest mean genetic change on correlated
traits as was expected. However, the observed value of mean genetic
change on egg production in population I was somewhat deviate, probably
due to random chance.

Mean genetic change when selection index
was used in females (Method II)

 

 

The expected and observed mean genetic change on body weight
and egg production are presented in Table 3.

In the offspring population the expected mean genetic change
on body weight and egg production was calculated as the average genetic

change contributed by male and female parents.
+

= AGM AGF
2

AGO

Where subscripts 0, M and F stand for offspring, male and female
respectively, and AG is the expected genetic change.

Males were selected based on their own phenotypic value of
body weight, where 1.25% of males were saved for parents. The expected
genetic change of body weight in males was calculated as

. 2.
AGMl - i hl 0P1

38

and the mean genetic change of egg production in males was calculated
as the correlated response, so

AGM2 = l hl hz 1612 6P2

Selection on female parents was based on the non restricted
index value, which was calculated based on the method presented by
Hazel (1943). The expected genetic change on body weight and egg

production was then calculated according to VanVleck (1976).
Cov(G1;I) i
OI

AGFl =

 

__ Cov(G ,I) i
AGFZ— 2

 

CI

2
b O G + b CovG

where, Cov(G1,I) 1 l 2 12

2
+
b2 G2 blCovG12

I
COV (G2 I )

and b1 and b2 are the index weighting factors. The variance of index
2 2 2 2
is calculated as: o = b o p + b 02p2 + 2 b1 b2 Cov P

I 1 1 2 12

Table 3 shows close agreement between the expected and observed
values of genetic change on body weight and egg production, due to the
fact that the genetic and phenotypic variances and covariances indicated
no trends over times. However, random chance seems to affect more the
genetic change in egg production as a trait correlated to body weight,
and the prediction for body weight was more accurate.

Within the same heritabilities, the highest mean genetic change
on body weight was when the negative genetic correlation was the lowest.

But genetic change of egg production was highest but negative when the

39

negative genetic correlation was the highest. And when the genetic
correlation was low (-.2), a positive genetic change on egg production
was observed. Apparently when a small negative correlation was the
case, the selection index was effective in preventing a negative genetic
change on one trait but also tended to decrease a positive change on

the other trait.

Table 3.--Expected and Observed of Mean Genetic Change on Body Weight
and Egg Production per Generation when Selection Index
Was Used in Females.

 

 

 

 

 

I
3 AGl (Kg) A62 (unit egg)
Popula-
tion
Expected Observed Expected Observed
I .1469 .1721 -l.7083 -l.1053
II .1362 .1503 —.3875 -.7506
III .1516 .1766 .0683 .0627
IV .1141 .1223 -l.1571 -.8502
V .1140 .1311 -.5710 -.4621
VI .1160 .1379 -.0408 .1830
VII .0306 .0312 -.2658 -.2337
VIII .0330 .0393 -.0790 -.0734
IX .0353 .0373 .1066 .1979

 

 

 

 

 

40

Mean genetic change when restricted
selection index was used in females
(Method III)

 

 

 

Restricted selection index was constructed to impose no
change in egg production while selecting for body weight.

The expected and observed mean genetic change in body weight
and egg production are presented in Table 4. To obtain the expected
value of genetic change the same formula as was applied when selection
index was used. The b value was the index weighting factor in the
restricted selection index equation. The observed and expected
values showed a close agreement. Only females were selected based on
the restricted index values. Selection on males were based on their
own phenotypic value of body weight which caused a correlated response
on egg production.

The theoretical expectation showed that the restricted
selection index was effective in holding the mean genetic change on
egg production in females close to zero (Table 5). An exact zero
value was not obtained, although a zero covariance between genetic
egg production and index was imposed, probably because of rounding
errors. Mean genetic change of egg production contributed by females
was trivial but males were selected in a different way. Males contri-
buted half of the genetic value of egg production and this caused the
change. The genetic contribution from males was large as the
selection intensity was high especially when the heritabilities of
two traits and the genetic correlation were high.

Data from the phenotypic mean in Tables (l3, 14, 15) showed

that the restricted selection index was quite effective when the

41

Table 4.—-Expected and Observed of Mean Genetic Change on Body

Weight and Egg Production per Generation When Restricted

Selection Index was Used in Females.

 

 

 

 

 

 

AG (Kg) AG (unit egg)
1 2
Popula-
tion
Expected Observed Expected Observed

I .1315 .1538 -l.3452 -.8821
II .1652 .1811 -l.2176 -1.0017
III .1661 .1858 -.4479 -.6032
IV .1051 .1292 -.9501 -1.0060
V .1165 .1298 -.6339 -.4979
VI .1243 .1506 -.3213 -.5210
VII .0352 .0374 -.3869 -.0459
VIII .0383 .0388 -.2579 -.4164
IX .0415 .0395 -.1278 -.1586

 

 

 

 

 

Table 5.--Mean Genetic Change on Egg Production in Females under

Restricted Selection Index.

 

 

Population A621
I .0000
II .0000
III .0010
IV .0023
V .0006
VI .0006
VII .0028
VIII .0021
IX .0034

 

AG

1This expected value of AG

,1 '
= Cov(G2 )i

 

2

O
I

2

was calculated by the formula

42

heritabilities of both traits were .1, especially when the genetic
correlation was -.2.

The values in Table 4 showed that within the same heritabi-
lities, the genetic change on body weight was the highest when combined
with the lowest negative genetic correlation. But the highest negative
genetic change in egg production was observed with the highest genetic
correlation.

Mean genetic change when two stage

selection was used in females
(Method IV)

 

 

 

The method of two stage selection or independent culling
level was used to select females for parents.

The expected and observed values of mean genetic change are
presented in Table 6. The observed values were in close agreement
with the expected values. This table shows that the genetic change
in body weight can be predicted more precisely than the genetic
change on egg production. The latter was affected more by random
chance as explained before.

The expected genetic changes in females were calculated

following the approach of Harvey and Bearden (1962).

2
AGFl ‘ (0‘1 hi + 9‘2 JEG12 hi hzmpi
AG *(oc h2+orG hh)dP and
F2- 2 2 1 12 1 2 2

= ll-rPlZlZ iZ-rPlzil
a1 ____ET___— a = —__—3T——__
_- 2 I
l r P12 1 r P12

43

Table 6.--Expected and Observed Genetic Change of Body Weight and
Egg Production per Generation when Two Stage Selection
was Used in Females.

 

 

 

 

 

 

 

 

 

 

K AG 't e
Popula— AGl ( g) 2 (uni 99)
tion
Expected Observed Expected Observed
I .1555 .1565 -1.7735 -1.0439
II .1461 .1474 -.6383 -.6015
III .1507 .1674 -.0099 -.0384
IV .0955 .1148 -.9529 -.7984
V .1156 .1204 -.5010 -.6750
VI .1128 .1241 -.0456 -.2791
VII .0338 .0405 -.3669 -.4019
VIII .0355 .0423 -.l729 -.2008
IX .0371 .0389 -.0126 -.0878
Where, i1 and i2 are the selection intensities for body weight and egg

production, which are 1.27 and .80 respectively (Becker, 1975). And

rP12 is the phenotypic correlation between the two traits, calculated

according to the formula (Falconer, 1960)

r = + E
P12 hl h2 rG12 e1 e2 r 12
Assumption was made that the environmental correlation (rElz) was
equal to zero. So rP12 = h1 h2 rG12

The genetic change of males was calculated with the same

procedure as already explained.

44

The values in Table 6 showed that within the same levels of
heritabilities the mean genetic changes of body weight were more when
the genetic correlation was lowest. The negative change on egg
production was small when the correlation was low {-.2).

Harvey and Bearden (1962) also presented the formula to calcu-
late the expected correlated response for the unselected trait.

The correlated response on egg weight can be calculated as follows:

= +
AG3 d1 rG h1 h3 a3 rG23 h2 h3

However as in this study, the value of rG , correlation of egg

23
production and egg weight, was not known in the population so the
expected values of the correlated response can't be obtained,

Change of thepphenotypic mean under
different method of selections

 

 

The changes of the phenotypic mean in different populations
under different method of selections are presented in Table 7 to
Table 15. From this point method I, II, III, or IV designates
how females were selected: mass selection (method I), selection
index (method II), restricted selection index (method III), independent
culling level (method IV).

Mass selection in all populations was very effective in
increasing the mean of body weight.obviously because selection was on
the additive genetic variation. The percent of improvement was nearly
double when the heritability of body weight was .4, and was about 75%
and 25% when heritability was .3 and .1 respectively. The correlated
responses in egg production and egg weight were as expected. Egg pro-

duction was declining due to negative genetic correlation between body

45

weight and egg production. And the positive genetic correlation
between body weight and egg weight caused a positive correlated
response on egg weight.

The percent decline of egg production at generation ten was
16.9%, 16.9%, 8.4%, 15.5%, 8.6%, 4.5%, 6.1%, 4.3% and 1.3% in
populations 1 to 9 respectively. The corresponding percent increase
of egg weight was 45.5%, 26.4%, 15.3%, 33.8%, 21.7%, 15.0%, 8.4%,

7.6% and 7.4%.

In some populations the percent decline in egg production
was noticeable when the heritability and the genetic correlation were
high, but the response to selection on body weight was also high.

For example in population I the average decline of egg production per
generation was about 1.7%. At generation 10, the mean body weight
was 3.26 kg and it seems that a good marketable size of livebird is
around 3 kg. If this method is going to be applied then, should
selection be continued after 10 generations on the same direction

if egg production continues to decline while body weight continues to
increase? Or, should selection be changed by keeping body weight
constant while improving egg production? To answer this question, one
needs to identify which of these nine simulated populations the native
chicken population is most likely to resemble.

Selection using selection index showed that mean body weight
was increased and egg production declined. Though egg weight was not
subjected to selection, mean egg weight was also increased as a cor-
related response due to positive genetic correlation with body weight.

The purpose of using selection index is to select body weight

and egg production simultaneously to get a maximum total economic

46

me>mH mcwaaoo ucwocwmmpcfl
woocw coauomamm owUOMHummH
Moose cowuomawm

”>H coruwz
"HHH wonumz
.HH mocha:

 

 

 

 

 

 

 

 

 

 

 

 

cofluomawm mmmE "H condo:
HNMN.nm vonQ.mm mhvv.mm damn.wm menv.nm meow.om vena.nm nnoo.om hemo.n nnoo.n mm-.n Hmm~.n 0H
mmoo.~m ammu.vm msum.mm mmoo.mm Nooo.vm nnmm.mm Hmmm.ma mmmw.oo vmmm.~ mmnm.N ~0h0.n «mmu.n m
monm.mv mvvw.um m~0v.~m mmm~.nm wmwn.oo Hmow.hm ha~w.mm vnmm.mm mwnb.~ mmvo.m omwm.~ vndm.n a
«omn.mv mmHn.mv o-~.av owov.mm hnv~.hm mo~o.ma mvvo.om momm.nm Hooo.N omm¢.~ monh.~ ~N¢h.~ s
Nach.hv nmwo.hv nonm.hv menu.om mono.ho mono.om wuow.ho chom.vm ovmv.N HHOM.~ canm.~ nomm.w o
hoNa.vv mooo.ov Nmmm.mv On~m.hv omhv.mm boon.ooH ~m~m.om mth.om onwm.~ ova.~ omen.~ owhn.m m
moom.nv mmnu.vv nnmm.nv ommo.mv HoH~.ooH ammo.oo~ amow.oo~ somv.mm w~v~.~ ne~o.N w~w~.~ Hov~.~ v
000v.av wuuh.~v mom~.~v omuw.nv aomh.dou ov~o.~om msqv.ood mam~.oo~ bnwm.n wnom.H mmuo.~ hmmo.n n
ammm.mn ~mom.ov anom.mm rwvo.ov wmqv.~oa mmmm.no~ mooa.~o~ Onmo.~oH ooom.u wmvh.~ mamm.~ wawm.u N
ohmv.mn hmww.mn mwow.mn mm~h.mm ouvm.n0u moHH.voH mmoo.no~ mem.no~ seem." “moo.u N~h©.~ bhvo.q H
>H HHH HH H >H HHH HH H >H HHH HH H cowuu
numcuo
Rome unmawz seem can.

Ame aromas mom

 

Ammo uwcov cowuosfioum mom

 

 

 

.COwuooawm no avenue: ucmuouuuo moot: H newumHomom :« unawoz mom can noduosooum vow .unoflws atom uo mcoozlu.h manna

47

mHo>mH OGHHHso Homocmmwocw
xwocH coHuomHom omuowuuwmu
xmocH :oHuomem

">H corpus
"HHH conuwz
"HH porno:

 

 

 

 

 

 

 

 

 

 

 

 

:oHuomem mmmE ”H conuoz
omnm.om omHh.hv amvo.mv vown.wv ~mnr.no HHon.mm mHmv.>m NnHo.om onmm.~ vmvw.n movo.n hmm~.n oH
ommH.av ammo.wv onn.ov wo>~.hv mem.mm Homn.om onwm.>o mmn~.wm mmmn.~ ammo.n hHmm.~ omnH.n a
vaw.hv voHv.mv nmmm.nv ommn.mv nH~m.wo noom.no o~0m.mm Hmom.om NHNm.~ momm.~ Howm.~ enmm.m m
vomo.mv Hoow.nv «www.mv mno~.vv Hnow.mm ONoH.mm nmno.ooH vowm.~m Hhov.~ nnmo.~ wmnm.~ nono.n s
Nvmm.ve hmmH.~v o~mm.ve Hnn~.nv mnho.ooH oth.om wme.HoH «hmo.vo omen.~ wham.~ oonn.~ v-0.~ o
mem5.nv moon.Hv Fva.~v vwmm.Hv vvnh.ooH mor~.ooH mamm.HoH mono.wo sohH.~ VHw~.~ mmVH.~ hmnv.~ m
omeo.~v eoHo.ov mnmn.~v omho.ov mmvm.HoH o~mm.HoH wHHw.~0H -o>.ho ammo.~ ~mmo.~ HhNo.~ woh~.~ v
somH.ov moo~.mn nhv~.Hv anm.oq nmo~.HoH vmhn.~oH o~m~.~oH o~vo.mm tho.H nowm.H «Hmm.H mwwo.~ n
memo.mn oHnm.wn smoo.mn mmom.mm ooHH.~0H cosm.noH MHm~.moH m~mm.HoH mohb.H h~m5.H mhmh.H wmwm.H w
HmHm.mm hmcm.mn men.mm mmH~.mn mnmm.moH mmwh.voH momH.voH wovv.noH nmwo.H omHo.H mme.H wowo.H H

>H HHH HH H >H HHH HH H >H HHH HH H
Ame pamflmz mom imam “Haas :oauosooum new logo uamamz zoom

 

 

 

 

 

.coHuooHom Ho mooguw: acmHmHHHn Hops: HH :oHumHomom :H uonws mom can coHuoavOum mum .uanws >000 mcnvzll.m anda

48

 

 

 

 

 

 

 

 

 

 

 

 

mHm>mH mcHHHDO unwocwmmpcw ">H vogue:
meCH :OHuomHmm omHOHHummH “HHH venom:
xoUCH :oHuomem “HH vogue:
coHuomem mmmfi "H vogue:
noho.mc comm.hv mnmo.vv QMHm.mv momm.voH HHmm.mm Nomm.m0H NoOH.vm omMH.n omvm.m ~mo~.n mmvm.n oH
tho.nv hmmh.hv mmmv.vv mwom.~v mm~w.voH non.mm om~o.voH omen.mm nmmm.N HHo~.m memo.n vowH.n m
vmmm.Hv omno.>v hOMH.nv mmmh.~v mwm~.moH mvvo.ma ~0~m.v0H mmvm.mo thh.~ m0mo.n HHom.N vaa.~ m
been.Hv Hme.mv NHoH.nv moHv.Hv vmmm.mOH vmmm.mm mHoh.noH oven.oo Nmoo.u noon.~ mvon.~ ath.N n
comm.ov bmnm.vv ovnH.mv w>~o.ov mmmH.mOH vomo.om mHmm.voH Nwmo.no eonq.~ nmoo.~ mmwm.m mhmm.~ o
mmoo.ov oHHm.nv Hnmm.He m~o~.ov mMHH.moH omo~.HoH mHMH.voH Hhmn.>m nHo~.~ moov.~ mHmm.~ mevAN m
memv.mm mmm0.nv nmm~.ov mmnH.ov mono.voH n~vm.~oH ovom.voH Hom~.mm movH.~ wo-.~ HHmH.~ vm-.~ v
noH~.mm mbw~.~v Hth.mm o~mm.mm momv.voH HoH~.moH HMHm.voH oan.HOH omvm.H mmho.~ vmmo.H vao.~ m
coho.om omoo.ov Nov~.mm onmm.mm mmHv.voH vvmm.moH noo~.c0H ~h0~.HoH ome.H omnm.H ommh.H mmmw.H N
MHom.hn ms~m.mm bhHh.hm mmmm.hm now>.voH oomH.voH mvmm.voH cmmh.~oH mmnw.H nvoo.H noHo.H Hovo.H H
>H HHH HH H >H HHH HH H >H HHH HH H :oHuo
luocwo
13 £13? mom Bow $5; 839.603 mom 85 Emma; zoom 85....

 

 

 

 

.coHuomHmm Ho moonuoz ucmuoHuHo hops: HHH coHumH5dom :H unoHo3 mom can :oHuoaooum mom .u:OH03 zoom m:mo:uu.o oHnaa

449

mHm>mH mcHHHso unmocmmmGCH
xoocH :oHuomem owuoHHummu

XOQCfl GOMHUQHOM

u>H compo:
“HHH monumz
"HH mocha:

 

 

 

 

 

 

 

 

 

 

 

 

coHvoonm mme “H oosuwz
osmv.av mmao.Hm omHo.om mmH~.om moom.mm omme.mm cmam.mm mmmm.~m mmmm.~ sass.~ mmma.~ ovmm.~ oH
Hmmm.me ommH.om mmom.mv san.mo Hmmm.am nmam.om ommH.mm anv.om mmmm.~ nose.“ enmm.~ n-~.~ m
aomm.mv sonn.mv mmHm.ae oaam.mv move.mm HNmo.mm Henm.am anmm.~o vamq.~ mva.~ mmHm.~ mmmm.~ m
mavo.mv mmma.me comm.mv mmom.me mamm.mm nnmm.mm HHHm.nm mao~.vm mmmn.~ mmmn.~ mmmm.~ m~me.~ a
mm~o.vv mHhs.ve NHHn.mv mnn~.ev amwm.ooH vaH.ooH mva.mm «HHm.mm mmm~.~ mma~.~ amm~.~ coon.~ m
mHmn.~v ovHv.mv AHma.ov mHm~.nv mHvH.~oH mamv.HoH v~m6.mm mvmm.sm non.~ HHnH.~ AHHH.~ vmmH.~ m
onma.ov ~sm~.~v omsp.mv mmmm.~v moma.~oH mHmm.HoH a~m~.mm noom.mm mmvm.H mo~o.~ amvo.~ ~H~o.~ e
ommH.ov amHH.Hv mam~.~c moHH.Hv man.noH emma.~oH swam.ooH somm.ooH mmmm.H mmom.H Hemm.H mmmm.H m
mem~.on Hmoa.on mena.ov ven~.mn Namo.moH vmam.noH amen.HoH mama.HoH momm.H mvma.H HmHm.H mmom.H ~
Hooa.sm mmmm.mn mmon.mm ohmm.nn Homm.moH mon~.voH amHm.~oH mHmm.noH nHNm.H muHm.H Hmmm.H momm.H H
>H HHH HH H >H HHH HH H >H HHH HH H coHHu
Inocoo
Amy uanoz mom Ammo HHCDV :OHuosooum mom Home uanwB hoom 693w

 

 

 

 

.coHuumHom Ho wpozum: ucoHoHHHa Hops: >H coHumH3Q0d :H uanoz mom pom coHuooooum wow .unuHoz Hoom acmwxna.oH anma

50

mHo>mH mcHHHso unoocomoocH
xmocH coHuowHom oouoHHummH

n>H oomuoz
”HHH oomuoz

 

 

 

 

 

 

 

 

 

 

 

 

xoUCH COHuoonm HHH oozuoz

HHnXMUthHmwm .mm2msH “H“ mXUvfiVm£H
nvwm.mv m~m0.~e mwmv.vv MNMH.bv mvmm.mm Homo.oo mmvH.om H0-.mm omen.~ evob.~ mmoh.~ mnom.m oH
mmhm.mv ~m¢0.~v hmhc.vv wam.mv mvmm.mm memo.ooH oHHh.mm nmm~.om vao.~ oovo.~ hHho.~ mmvh.~ m
Hmmm.mv wmvm.Hv Hoom.nv mon.oc Nonv.ooH HHov.ooH mmHm.wm omom.nm HmHm.~ mmwm.~ Nmmm.~ momm.~ m
Nan.mv Hmo~.Hv vvmv.~v mch.cv Hnmm.mo mobm.HoH o-o.¢o «Hhh.mm w~o«.~ mmwm.w mhmm.~ mmwv.n h
mNmo.mv mth.H¢ mooo.Hv NoH~.ev memo.ooH mHom.HOH HHH>.mo vmhh.mm hvm~.~ Nmo~.N ~vm~.~ Hmom.~ m
vmwm.mv ammo.ov man.Hv oth.nv wmmH.HoH mmmH.~0H mmnn.ooH mvmm.mo mmmH.N mmMH.~ hmmH.~ ohmm.m m
ean.mv wvmm.mn mmho.mm vmm~.~v Hvov.HoH hm~h.~0H mNnN.~oH ONmH.HoH homo.~ oHHo.N mmmm.H oomo.m v
hwhm.~v hHHm.hn hoHH.ov mnm0.He nHmm.HoH oohm.~0H HHom.~oH mHmn.HoH hono.H Nnom.H «mom.H Humm.H m
~m50.Hv hvmv.hn thv.mn m~mv.mn omom.~oH Hmom.noH ~v~H.moH vvon.~oH Onoo.H vmmz.H mmmz.H Hobh.H N
hnmo.mn Hvow.hn ommm.mn ohmh.mn w~m~.voH FMHH.VOH mmbn.n0H oo-.v0H mmoo.H hmmm.H mmow.H Howm.H H

>H HHH HH H >H HHH HH H >H HHH HH H
Hoe umoHos mom Homo nHoov :oHooooon mom Home ummHoz zoom

 

 

 

 

 

.coHuooHom mo moosooz HcouomHHo soon: > ooHuoHoooo cH Honoz mom oco coHoooooHo mom .oaoHo: zoom mooozu-.HH «Home

51

mHo>mH mGHHHDU unoccomeCH
xmocH :oHuoonm oouoHHummH

woocH coHuoonm

">H oomuoz
uHHH oomuoz
uHH oonuoz

 

 

 

 

 

 

 

 

 

 

 

 

COHuomHom mmmE "H conumz
mwmh.He Homo.~v mmmm.mn mamm.vv ~0m~.HoH mwow.mm Onom.m0H ooHo.mo whwh.~ ~mmm.~ mmww.~ wv>m.~ oH
omno.ov mmHm.Nv -~n.am mnom.mv oqu.HoH hHmm.mm mvoN.eoH mHmo.ooH mnvm.m mhmo.~ mowh.~ wmvm.m m
NmHQ.mm moh~.nv nvoH.mn mmm0.~v momh.HoH moov.ooH oHov.wOH hHHo.HoH MHmm.N anh.~ hmoo.m oomo.~ m
mohm.mm envo.~v mmmm.mn mmH~.~v mmHH.~oH mHoH.HoH enhv.oOH nvh~.HoH omHv.N mnnm.~ vmom.~ ohmm.~ h
mnnn.mn Hzom.Hc wnmo.mn whHm.ov momm.moH boom.HOH voHH.moH QFVH.~OH vvm~.~ behm.N mvwm.~ omHv.N o
wnnH.mm ommH.Hv HmOm.mn mono.ov ovom.~0H oHom.HoH wmom.on moom.HoH mva.~ mHv~.N ~wo~.~ omo~.~ m
nnho.mn oOmH.ov hv~>.hn mm~o.ov memm.noH wwvo.NOH nowh.moH ~m>m.~oH mmm0.N ammo.~ ~mno.~ vnHH.N v
Norw.mn ammo.mn Hmv~.mn moon.ov echm.noH mono.voH mumm.moH homo.noH thm.H vmmm.H mvwm.H Nwmm.H m
Hm0m.wm mmwm.mn vavv.mm oomh.mm mmHh.moH Ommm.moH NhHH.moH omnm.moH ooom.H Hons.H enoh.H omHm.H N
Hmom.hn vmzH.wm NmVH.wn mmhm.mn mnvv.noH HmNH.voH vHHv.voH mNHh.moH mmmw.H v~H0.H hhmw.H Hmmm.H H

>H HHH HH H >H HHH HH H >H HHH HH H IHMMWM
Hoe ummHms mom Homo uHcsv :oHuosooum mom Home ummHos zoom Hoes:

 

 

 

 

coHuuoHom Ho moonuo: acououuHo Hops: H> goHuoHDQOA :H usuHoz mom poo coHu0500um mom .ucmHmz >vom mcooz|I.~H oHnme

mHoewH mcHHHoo usoocomoocH
woocH :oHuooHom oouoHHumoH
woocH coHuoonm

”>H oomuoz
"HHH oomooz
HHH ooauoz

52

 

 

 

 

 

 

 

 

 

 

 

 

:oHuoonm wme "H vogue:
mmmm.Nv havo.Hv mHoo.Hv hmHm.Hc NmmN.ooH memo.ooH NHnm.HoH wooH.mm vao.H movo.H mon.H mmoo.N oH
mmhm.Hv voom.ov comm.ov Nva.ov hNHo.ooH mmoH.o0H Nmmm.HoH MOHN.mm vam.H omom.H mmnm.H momm.H m
meno.Hv nmNh.mm mvmm.ov ano.ov mnmm.ooH mth.ooH mmHv.NoH Nnom.oo ommm.H VHmw.H mmvm.H Hham.H m
NmoH.Hv cenn.mn mmHv.ov Hoom.om oon.HOH ovmo.HoH zoom.HOH tho.ooH oon.H von.H NNNm.H oovm.H z
Nmnh.ov NmNN.mm NnHo.ov ommv.mm hmom.NOH mmOH.NOH ova.NoH Hoom.HoH vnmb.H morn.H ommh.H mmHm.H o
nNmo.ov os0m.mm vomz.mn NNmo.mn man.NoH zmHm.NoH mon.NOH «Nun.NoH anh.H oomm.H mnmh.H oooh.H m
momm.an mnHm.hn thn.on omNm.mm mmHm.n0H mmoH.moH Hohn.NoH comm.NoH nvoo.H choo.H momh.H won>.H v
omNo.on mmzo.nn ovom.mn momm.mn Nwmo.moH MHHN.moH HNNN.noH mnvm.noH onmo.H vao.H momm.H tho.H m
mNon.mm hmmn.mn ween.mn mmv~.mm hnoN.moH mHhv.moH NmNm.moH OnNo.voH Non.H nNmo.H Nomo.H ono.H N
Homo.mm Hmmm.hn NNNh.mn oVMH.mn ommm.moH whom.noH hmom.m0H mmmv.voH nmmm.H muco.H mme.H HHmm.H H

>H HHH HH H >H HHH HH H >H HHH HH H
Ame ummHoz.oom Homo oHooo ooHooooon mom Home umoHoz zoom

 

 

 

 

 

.coHuuoHom mo moozuoz ucwuwHHHo Hoot: HH> coHuoHsaoa cH unonz mom can coHuosoOum mom .unuHoz >oom mcmwzul.nH oHnue

53

wHw>mH mCHHHoo ucwpcomwch
xoUCH :oHuoonm omHOHHummH

xwwcw COHuoonm

u>H oomuoz
”HHH oonuoz
uHH oomuoz

 

 

 

 

 

 

 

 

 

 

 

 

:oHuoonm mmmE ”H wonuoz
vmzH.Ho ovoo.ov zoom.ov ammo.He omvm.HoH mmmm.ooH vam.noH «HHm.mm -oo.H mmom.H oozo.H zooo.~ oH
nvwm.ov ovop.mn ovom.oo Hmov.ov oomH.~oH ~mom.ooH ozzm.moH oHom.mm mHmm.H mmHm.H mmmm.H mmzo.H o
somm.ov m-o.mn mmzm.on mVHo.oo zooo.moH ovzo.HoH «HH~.voH Homo.ooH ozzm.H omzm.H anm.H mon.H m
m~vo.oo zHoo.mm Homz.on nmmm.on nomm.~oH mHnm.HoH mmmz.voH H~m~.HoH momm.H Hmnm.H mon.H nmzm.H z
owHo.ov ooom.mm ommm.mm HonH.mn ommH.noH mamm.HoH zHHv.ooH nooo.~oH mmmz.H ooom.H mnmz.H zmmm.H o
mo~o.mn n~n~.mn HHoH.om comm.on oozo.noH msz.~oH ovmn.ooH ooHn.~oH omoH.H mooz.H mmmz.H momz.H m
memo.om vomv.mn m~om.mm zooo.om mmmm.~oH mmom.~oH zozH.ooH ommm.~oH msz.H mmmz.H oon.H nooz.H o
mmmm.mn ammo.mn mmnm.mm omoo.mm Hmo~.noH wzzm.~oH somz.moH oHoo.noH mozm.H Nazm.H nozm.H whom.H n
zomo.mm Homn.mn onv.mn zom~.mm mzom.moH moan.moH ~H~0.ooH emmH.noH noHo.H Hmmo.H H-0.H comm.H m
-H~.mn oo~m.zn Han.mn mmNH.mn momz.noH Homz.moH mmmm.noH mmmo.noH voom.H ozHo.H anm.H memo.H H

>H HHH HH H >H HHH HH H >H HHH HH H :oHoo
Inocou
Amy uano3.mmm Ammo UHGDV :oHuooooum mom Amxv uanmB zoom H655:

 

 

 

 

.coHuooHom Ho mpocuoz ucououuHo Hoot: HHH> coHumHodom CH uzmHoz mom can coHuozooum mum .unoHoz >vom acmo::I.VH oHnda

mHm>oH mCHHHso unmocmmoocH
xooCH coHuomem oouoHHumoH
xwocH :oHuomHom

">H oomooz
"HHH oomuoz
uHH oomuo:

54

 

 

 

 

 

 

 

 

 

 

 

 

 

 

COfluUmem mmME "H @OSUTZ
hmmv.mn HmNm.mm mmmv.mn mHHN.Hv monw.noH oNNo.NoH ommm.mOH Nmom.NoH Nwmm.H oNom.H mvz0.H ovNo.N 0H
omom.on mvmn.mm vam.mn omnH.Hv ommo.noH MHmm.NoH onvo.moH NNmm.NoH whoo.H ano.H movo.H Nmom.H m
mmHH.mm ommm.mn oOMN.om Nomw.ov mvmn.moH vam.NoH oNom.voH m00m.moH mNmm.H moom.H voom.H NNvo.H m
memo.mm cmmo.mm ohmo.mm zmmH.ov mNVN.noH Nmoo.NoH VNHm.voH mHhH.moH vmvm.H wmmm.H mwnm.H mHom.H n
momm.mm thv.mn voNH.om mHHH.ov omvm.moH vav.moH mmmo.voH ommm.moH mmmn.H ommN.H mvvm.H omvw.H w
mmnm.mm ozNH.mm cmmh.mm hHHm.om Nmmm.moH NoHo.voH moNN.voH Noom.moH ommn.H momz.H Noom.H mHom.H m
oovm.mm hva.mm Nozm.mm NHmm.om mumm.moH oon.moH mmmm.moH nmvv.moH MHNn.H meN.H mmnz.H «HwN.H v
oNom.mn ooo~.mm oHHv.mm mmmN.mn mmNm.moH Hmom.moH vHNo.voH vmwN.moH ome.H Nzoz.H mmH>.H Hva.H m
waN.mn vMHH.mm mNoH.mn omNm.wn oovm.moH thm.voH hzvN.v0H mmmm.noH mmNo.H omvo.H mmmo.H ammo.H N
homo.mn hmom.hm mem.hn Nth.mn oooz.moH ovvm.ooH Homo.voH onH.voH moom.H HNHo.H Hmmw.H omHo.H H

>H HHH HH H >H HHH HH H >H HHH HH H
loo omoHos mom Homo oHooo ooHHosoon mom Ammo umoHoz zoom

 

 

 

.coHuoonm Ho mcosuoz ucououuHQ Hope: xH coHumHoooo cH uchoz mom can coHuuscoum mom .uonoz >oom mcowzln.mH ”Hams

55

response, and it was theorized that this method is never inferior
compared to other methods (Lush, 1942; Young, 1961; Finney, 1962).
Discussion on total economic response will be presented in later
part of this section.

At generation ten, the means body weight was improved by 92.9%,
87.8%, 98.5%, 65.4%, 73.6%, 76.0%, 17.3%, 22.0%, 20.5% in populations
1 to 10 respectively, and the corresponding decreased on egg production
in the same population was 9.6%, 6.4%, 0.0%, 6.8%, 4.0%, 0.0%, 1.8%,
0.0%, 0.0%. The mean egg weight was increased by 43.3%, 26.6%,
18.5%, 27.2%, 15.2%, 4.6%, 6.0%, 5.3%, 4.2% in populations 1 to 9
respectively.

In all populations the response on body weight was greater
under mass selection than selection index. The decline of mean egg
production and the increase of mean egg weight also were greater for
mass selection. These results were really obvious, as mass
selection was concentrated only on body weight but in
selection index method egg production was accounted for, which in
turn caused less decrease in egg production.

Further the results also indicated that the percent response
was dependent on the level of the heritabilities, the higher the
heritability the higher the response. The percent decrease and the
percent increase in egg production and egg weight,respectively,
showed a consistent pattern that within the same levels of herita-
bility the magnitude of percent response was ranked directly with the
magnitude of genetic correlation. However this pattern was not

clearly observed on percent response on body weight. In populations

56

where the genetic correlation of body weight and egg production was
low (-.2) the percent decrease of egg production was equal to zero,
although the observed phenotypic means showed slight increase. It
seems that the effect of low negative genetic correlation in this
study is trivial and even selection using selection index eliminates
the effect to some extent.

The results from selection using restricted selection index
showed that mean body weight and egg weight were increased
although egg weight was not selected directly. Previously it was
shown that when the genetic change on egg production was calculated
using the theoretical expectation, this method should be successful
in preventing genetic change in egg production. However, the observed
phenotype showed that mean egg production was declining slightly.

The reason for this decline can be explained. Although restricted
selection index was successful in imposing no change on the mean
genetic egg production in female parents the index was not used in
selecting male parents. Thus it may be concluded that the genetic
change was contributed by male parents.

The percent decline of mean egg production at generation ten
was 7.2%, 8.6%, 5.4%, 8.4%, 4.2%, 4.3%, 3.4%, 2.6%, 1.8% in popu-
lations l to 9 respectively. The question whether the restricted
index selection in this study has been successful can't be answered by
experimental evidence.

In restricted selection index studies that have been reported
both males and females were selected the same way, as the traits were

expressed in both sexes. To improve this study males should also be

57

selected using restricted selection index, where the necessary informa-
tion may come from family or progeny. All the information used to
construct the selection index for females come from individual data

but as the information in males did not, another restricted selection
index is needed.

The mean of body weight at generation ten was increased by
86.6%, 100.8%, 100.9%, 72.0%, 73.3%, 83.9%, 21.1%, 21.4%, 21.7% in
populations 1 to 9 respectively. The percent increase of mean egg
weight at generation ten was 44.9%, 29.4%, 23.5%, 32.0%, 13.4%, 11.8%,
9.3%, 6.7%, 2.5% in populations 1 to 9 respectively. Populations with
the higher heritability combinations showed more percent increase
on either body weight or egg weight. The magnitude of genetic corre-
lation between body and egg weight played a role in determining the
percent increase of egg weight. In populations with the same heri-
tability combinations, the higher the genetic correlation the higher
the percent increase.

For selection using independent culling levels, the results
showed that mean body weight and egg weight were increased, accom-
panied by a decrease in mean egg production. Mean body weight at
generation ten was increased by 85.7%, 81.7%, 92.3%, 63.9%, 64.8%,
67.5%, 23.2%, 22.3%, 21.6% in populations 1 to 9 respectively. The
mean egg production was decreased by 10.1%, 4.9%, 0.2%, 6.7%, 5.1%,
2.1%, 3.2%, 1.8%, 0.1% in populations 1 to 9 respectively. The
percent increase of mean egg weight was 38.4%, 30.2%, 14.7%, 25.8%,
17.2%, 10.4%, 11.7%, 7.8%, 3.6% in populations 1 to 9 respectively.

The method seems good in reducing the decline of egg pro-

duction to some extent, as egg production was also subjected to

58

selection. The percent of decline was somewhat less than the

percent decline under mass selection. When genetic correlation between

body weight and egg production was low the decline was trivial,

especially in populations where low genetic correlation was combined

with low heritabilities. Obviously the percent increase in mean body

weight was a little less compared to what was obtained by mass

selection. Mean egg weight was increased due to correlated response.
Figures 1.1 to 1.9 show the changes of mean body weight and

egg production in populations 1 to 9 respectively. Figures 2.1 to

2.9 show the changes of egg weight and body weight.

Relative percent efficiencies of

different selection methods for
total economic response

 

 

 

The relative percent efficiency was calculated followed the
approach presented by Young (1961). For each method the total
economic response was calculated as: Hi = a1 AG1 + a2 AG2
Where subscript i can be I, II, III, or IV indicated which method of
selections is considered. Here al and a2 are the economic value of
body weight and egg production respectively and AGl and AG2 are
genetic change of body weight and egg production.

Table 16 shows the expected and observed total response per
generation in economic units, where 1 unit equal to $.10. There was
close agreement between the expected and observed values in most cases.
Both the expected and observed value showed that combination of

heritabilities and magnitude of genetic correlation affected the

total economic response. This was true for all method of selection.

S9

 

D
u:_.
"’ 11 Method I
- V Method II
D U Method III
"3 - 0 Method IV
05 Solid Symbol (body wt.)
—- Open Symbol (egg prod.)
O
fl -
M
O
C” H-
N /
O
H¢~ _.
o3.
éfiw
. )—
igc:
.U7 —- _‘
“o ' [15'
”N
2 - .. ,,
do 2*
°‘ N
i— _ g
‘3 o
'1 - Io53
N 5L
:2
0: - F
H
b
O
"T t- as
.4
O
”?-- ..,5
«1 l J ll .J 11 J 4] J J l

 

 

 

1.bo 2.00 3.00 4.00 5.00

5.00 7.00 8-00 9.00 1010

Generation

Figure 1.1. Means of Body Weight and Egg Production in Population I

"' o ' rc = -06, rG

UBder Different Method of
h3 - 12 13

Selections. (hi = .4, h2 = .3,
= .5)

 

60

 

 

‘7’
f3
- 7' A Method I
m V Method II
" a Method III
0 0 Method IV
“2 - Solid Symbol (body wt.)
00 Open Symbol (egg prod.)
T’
C)
v4,_.
«7
O
m {—
N
O
a". P
5N
J
30
."3 — H
:geq
:0 "' —
C'.
“0
gr? F- ,/ -.
N
b / ‘
O
H r- a. “ / u—
N afiga "g-I' -
‘ I: _~\;
— W "§._§; a —
c: /' "Illii; “*‘F
q, )— ‘ - —I
"' L
/ ‘ _
'9 _ ./ I —1
H
o F
U1..— —
'1 J J J J J J, J, J .lJ .J
1.00 2.00 3.00 4.00 5.00 5.00 7.00 9.00 0.00 1010
Generation
Figure 1.2. Means of Body Weight and Egg Production in Population II

Under Different Method of Selections.
h} - .5, r612 a -.4, rG13 = .3)

2 _ 2 _
(hl " 04' hz - -31

E

#5

E

fir

333 sues”

'POJJ

3-50
I

3.30
1

3-10

2.90
1’

3363 82.515052“)
1

 

 

61

Method I

Method II

Method III

Method IV

Solid Symbol (body wt.)
Open Symbol (egg prod.)

C)El<lt>

 

 

HI}!
t .. d
0:
a
w -— 115'”
. “ to
EN g
‘- -o
O 8
'1 "--'—“"Illall""’ “‘= ‘i’b5p'
N D
o i — — d J
C? a”
v1
0
'7 ~43:
*4
C)
e—— 4;:
v-o l l J J J J J 1 J l
1130 2.00 3.00 4.00 5.00 5.00 7.00 8.00 9.00 10'.0
Generation
Figure 1.3. Means of Body Weight and Egg Production i3 Population III

0 der Different Method of Selections. (h1 - .4, h = .3,

2
h3 = .5, rGlz g -02, 2613 3 01)

62

 

 

K

V

s

 

D
“3 —- A Method I
m V Method II
_' 0 Method III
C 0 Method IV
"3 _. Solid Symbol (body wt.)
6) Open Symbol (egg prod.)
CD
H >—
F?
CD
C? —.
N
c: "”1'
F‘ H.
EN
5 L_ /
.30
zuut. .—
so: .
ID
Go /
m
"7
g - F‘ -
CM
— d
C)
".' t- .. -
N T V‘té/‘I—z
1 /4g -\- ‘ _
S: J”" . L ? ‘.“‘==§!!-I-nfl
r; / ‘1
o .1
r~ __ -
fl
0
'C'a- n
1"! _L 1 J J J J J J l l
1110 2.00 3.00 4.00 5.00 6-00 7.00 9.00 9.00 10'.0
Generation

Figure 1.4. Means body Weight and Egg Production in Papulation Iv
Under Different Method of Selections. (h1 = .3, h2 = .2,

2
h3 = -3. r612 = -.6, rGl3 = .5)

E

3 suean

%paid 33

63

O
“3 w- A Method I
M V Method II
-— 0 Method III
C: (3 Method IV
«3.. Solid Symbol (body wt.)
"; Open Symbol (egg prod.)
r—
O
‘1 un—
M
P
O
0) L- r
N
D
r~ __ 1”],
’8‘:
55 H. ///,l
.33, ’
3 . — - IL!
-c4
:1
E ._ /’ ._ g
28 / 8
CU h—
o - //’ “llfw
ZN 0c:
* / - 3
C: o.
'1 - -.._ _ / - 105"
CH ’ ‘ “‘—-? 4;7::n-.m.____. 1
‘ “‘-s::==—-2"IIIINI-—H . ___“
- J ‘- “v—S‘ "‘
(3 n"" “‘:;IIIII* ’
<p —- ‘ -i$f
" r/
O /
r:,_, / ‘185’
'4 ./
1... —
:5
III—.— — F
:4 I .J J J] J .1 J _l l

 

 

 

J
1Tb0 2.00 3.00 4.00 5.00 $.00 7.00 8.00 9.00 1010

Generation

Figure 1.5. Means Body Weight and Egg Production in ngulationzv

Uader Different Method of Selections. (h1 = .3, h2 .2,

h = .3. r6 3 -04, I'G = .3)

3 12 13

64

 

 

 

 

= a
01d Bfigegueaw

i

43'

fir

 

-—I_-
o
m .. A Method I
,3 V Method II
_ 0 Method III
0 Method IV
S, _ Solid Symbol (body wt.)
:6 Open Symbol (egg prod.)
o
H .—
m
o /
03 ._
N
L.
o
'T L /'/
1301
£5 _.
Do /
Sm _
'N
:3 .J
2o
3"? -
EN
2 H-
N 9
‘7‘ d
o
03 .4
fl
0
(Is —4
H
o
“F- '1
v1 J JJ, J Jr 1 J l l J J
1700 2.00 3.00 4.00 5.00 6.00 7.00 0.00 9.00 1010

Generation

Figure 1.6. Means Body Weight and Egg Production in ngulation VI

Uader Different Method of Selections. (h1 = .3, h2 = .2,

= =—.' =.2
h3 .3, rG12 2 r613 )

 

65

 

 

 

CD
“2 +—
_. V Method II
D 0 Method III
m _ 0 Method IV
v; Solid Symbol (body wt.)
__ Open Symbol (egg prod.)
C3
"" r—
m
p.
CD
01 .—
N
F.
D
f‘ _.
130i
:5" .—
°CD
an.» -- e
-N
'o
m '— I—
CD
0
5M _ _
q; .
ZN
CD
9‘ L— r. \ c—
N ‘.——— —_ \ ¥ - —
_ A\:=E; ‘ .—
O
m / —
O r- /
‘4
h— —
j/
0
h _. , —-
o ' /
fl 9"
__ 1:!!" ._
O
".L— -
‘4 l J. #1 1 4L 1 Jo eJ lJ wL
1.b0 2.00 3.00 4.00 5.00 5.00 7.00 8.00 9.00 1010
Generation
Figure 1.7. Means Body Weight and Egg Production in Papulation VII

Uader Different Method of Selections. (h1 = .1, h2 = .1,

h3 = .1, r612 = -.6, r613 = .5)

E

05'

833 suean

\
9.
'POJd

2a

75’

66

3-ISO l

(3 Method I

V Method II

D Method III

F— 0 Method IV

Solid Symbol (body wt.)
Open Symbol (egg prod.)

3-10 3.30
l

2.90
I

“(5)- 70

Means Bd. Wt.
2.50

I l

l |

2-30
I
l
3 suean

>AM7

33

'POld

2.10
i?

1 .90

-70

1

 

1 .PO
r
l
W

L J J J J J l J J J
1.b0 2.00 3.00 4.00 5.00 5.00 7.00 8.00 9.00 1010

Generation

 

 

 

Figure 1.8. Means of Body Weight and Egg Production i9 Population VIII
U der Different Method of Selections. (h1 = .1, 112 c .1,

h3 = .l, r612 8 -.4, r613 = .3)

67

 

 

 

 

CD
It.) .—
fi) 13 Method I
-— V Method II
o 0 Method III
(I? _. 0 Method Iv
fl, Solid Symbol (body wt.)
__ Open Symbol (egg prod.)
C)
“ —
m
D
O, _-
o3
r’
C)
h —
ZEN
é —
13
EU? L. —-IZ$'
,6¢M
00 {-— ... g
2:: g
00— ox”
2N o9?
__ 00
-' °v
3t 3
O; —- --~___:- -—~.... ..___ — _;..._-;-g— ___1____'\ ‘ (’59
o /
a: r— ] f _ 9:.
I. i
L. .—
2
. - "W"
'1
O
o..- - 75'
v0 _1 J JJ, till 4] J J JJ J J
1.130 2.00 3.00 4.00 5-00 5.00 7.00 8.00 9.00 10'.0
Generation

Figure 1.9. Means of Body Weight and Egg Production i3 Population IX
U der Different Method of Selections. (h1 = .1, h2 = .1,

h3 = .1, r6 = -.2, r6 = .2)

12 13

68

3.50
1

Method I

Method II

Method III

Method IV

Solid Symbol (body wt.)
Open Symbol (egg wt.)

3.30
I
o 04 1,

Means Bd. Wt. (k
2-10 2.30 250 $40 2.90 3-10
I

1.90

 

1-70

P —

J .JJ J J J JJ J .J Jr JL,

 

LFO

 

 

1.b0 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 iOLO

Generation

Figure 2.1. Means of Body Weight and Egg Weight in Population 5
U der Different Method of Selections. (h1 = .4, h = .3,

_ 2
h3 - c5. r612 -06, r613 - .5)

it :9 £9
'3“ 383 sueau

3‘.
‘3)

69

 

S __ A Method I
a; V Method II
0 Method III
_ 0 Method 1v
2, Solid Symbol (body wt.)
or; '- Open Symbol (egg wt.)
:3
V? _.
m
P
8
(q. «>58
‘52
o P _ $7
00
g I—
.338 _. .
' - SD
0“ g 7
:3 «h- ;« ' ; - 5%
as:
(a 3
o - " ‘
EN 9 g?
-‘ s:
c: 5’
H
.1 ‘ “as:
8
o — a
q...
53
o b — 37
'4
P —
8..
- _ 130
'1 J J I J J J J J JL

 

 

 

1.b0 2.00 3.00 4.00 5.00 5.00 7.00 8.00 9.00 now

Generation

Figure 2.2. Means of Body Weight and Egg Weight in Population II

Uader Different Method of Selections. (h1 = .4, h
h = . B - =
3 5, r612 .4, rGl3 .3)

2=‘3'

3.50
I

3-30
I

3.10
IT

2.90
I

“‘9. 70

'CD
U
3“,—

Bd.
2.

8588

M
2.

1-90 ZHIO

1-70

 

l-PO
I

 

70

Method I

Method II

Method I II

Method IV

Solid Symbol (body wt.)
Open Symbol (egg wt.)

(30‘60

J J J J J J J J

J I l
S 6.
'JM 883‘ Queen 3 *

L
‘3
8)

"JD

 

J

 

1.130 2.00 3.00 4.00 5.00 5.00 7.00 8.00 9.00 10'-0

Figure 2.3.

Generation

Means of Body Weight and Egg Weight in PoBulation EII

Under Different Method of Selections. (h = .4, h
h2 _ 1 2
3 " .5, r612 8 -02, r613 g 01)

= .3,

71

 

 

 

 

—II-—
0
ID —
N) AsMethod I
I- VMethod II.
o a Method III
"3 _. 0 Method IV
«a Solid Symbol (body wt.)
__ Open Symbol (egg wt.)
c:
v0 .—
w)
c:
U, h- - 53
N
P d
(D
'T " -4 57/
'BEN
0O 3
SW -- 1 ‘ _ m
'93“ _ g
_. , m
° 5
cc: m
30, F’ "l9‘rn
z:' 00
°* 00
— _' S
f?
'c": — - 91;;
N V
c: L
m e c - 3!
v4
0
r~ _. ..3q
F1
O
In...)— ‘30
rd J J JJJ l J, J J 1 JJ_ {
130 2.00 3.00 4-00 5.00 6.00 7.00 8-00 0.00 10.0
Generation

Figure 2.4. Means of Body Weight and Egg Weight in Pogulation 5V
UBder Different Method of Selections. (h1 = .3, h2 = .2,
113 = .3, rGlZ : -06, r613 8 OS)

72

I
I

- AbMethod I
V Method II

_ 0 Method III
(3 Method IV

Solid Symbol (body wt.)
Open Symbol (egg wt.)

3.10 3.30 3.50
I

2.90
I
l
a

(k)
.70
I I
I J.
-%

 

 

 

 

Jo
3m .—
. - so
N
f _ / _ §
: U
58 _ / a a
z ' // ’ z z '— 9151
‘V :E.IIIW’ g
3 , :25
d -— 2 ’ ' “2’5
3 fl
.4 " “
o -
7‘
_ /
8. -
O _ H3O
v4 J I l l J I J JJ J J
1.I)o 2.00 3.00 4.00 5.00 5.00 7.00 8-00 9.00 10'.0

Generation

Figure 2.5. Means of Body Weight and Egg Weight in Pogulation g
UBder Different Method of Selections. (h1 - .3, h2 = .2,

h3 8 .3, r612 ‘ -.4, r613 = .3)

73

 

 

:3
un__ .A.Method I
00 V Method II
D Method III
0 " 0 Method IV
fl, Solid Symbol (body wt.)
«:7 ' Open Symbol (egg wt.)
0
fl -
F) P
c:
‘2‘ .. “ 53
0%
IL— c-I
c:
h II— - 57
»(V
00
ii w— ..
.c:
EM "- <- w
ON 2
:3 .L. ._4 g
35’. "’
o ’I_ -4I¢§g
ZN oo
<f- . _ 3 - ft?
0 O
"‘ _ - no?
04 9 ‘ "
.- ,..s '.+
c:
"2 ._ a 30
fl
c>
'7 .L. - 3‘]
H
4|— -—I
:3
“%~- -ao
"“ J J J J J J 4 J I I

 

 

ljbo Juno 3.00 q.oo*:§loo

6-00 7.00 8.00 9.00 10in

Generation

Figure 2.6. Means of Body Weight and Egg Weight in Population g1

U der Different Method of Selections. (bl = .3, h
h3 = 03' IQ = -02; IG

12

-2 2:.2'
13")

74

 

 

 

 

D
m _
t6 .A.Method I
_ V Method II
a Method III
a _ 0 Method IV
“a Solid Symbol (body wt.)
_- Open Symbol (egg wt.)
0
'jIp-
0")
CD
03 — —-I
cw
C
'7- 4
1504
5 P —
e313 - 0
'N
E _ _.
D
58 _
0 - --1
SN
(3
‘I‘
N
O
0)
v4
0
h
.4
O
vh- —
vfl J J JJ J1 J J _J .1 JJ JJ_I
1130 2.00 3.00 4.00 5.00 5.00 7.00 8-00 3.00 .1010

Generation

Figure 2.7. Means of Body Weight and Egg Weight in Population XII
Dader Different Method of Selections. (h1 = .1, h2

h3 = .1, r612 = -.6, r613 = .5)

= .1,

5’8

3 suean g

32

~¢
(if). ‘JM 38

Is

~30

75

 

 

 

O
Lo _ A Method I
(:3 V Method II
_ 0 Method III
0 Method IV
5; Solid Symbol (body wt.)
- b' Open Symbol (egg wt.)
to
(3
H .—
M
D
a) L. " 53
N
O
'7 '— -I :7
3N
é — -
.o
‘5": _ - 6'0
~04
::
a — - :
Ibo a
g"? " " “(m
Seq g
—- -" s
H
3 9.;
C; I) 03’
-— - - 3,227 .-
.. /é__s" /
m I—- "f w ‘-/J ; -38
H /
O / /
/
v1 <‘__.4p"'
. -30
‘4 J I JJ, J I JJ 1 J I I
1.b0 2.00 3.00 4.00 5-00 6-00 7.00 3.00 9.00 10.0
Generation
Figure 2.8. Means of Body Weight and Egg Weight in Pogulation ¥III
U der Different Method of Selections. (h1 = .l, h2 = .l,
h3 = .1, r612 = -.4, r613 = .3)

76

 

 

 

 

:3... 4A»Method I
0‘; 7 Method II
_ :1 Method III
0 Method IV
E: Solid Symbol (body wt.)
or; " Open Symbol (egg wt.)
C)
'4 .—
M
8
. " .q 58
C‘
2
A ‘ _ q if
.x
v P —
do
an? ._ d 5'0
WN
'u
m __ _ g
go a
£6? P- - 946:,
cu m
— '_ 2
n
c: .
'1 o - 9’2:
N . L’
h— ~ . --i
- - a ’7 ’k.
g _ , _. -_ _ _:—'——-=.r;—"—-3='—:_:? 52'; 38
’ t "V
‘* , -
L. __
O
h ,_ - 3‘]
v.
P -
O
“zap ‘30
vi l J l] l] __L, J J Ill J l
1.bo 2.00 3.00 4.00 5.00 6.00 7.00 3.00 9.00 101.0
Generation

Figure 2.9. Means of Body Weight and Egg Weight in PoBulation 5X
U der Different Method of Selections. (h1 = .1, h2 = .1,
h3 = .1, r612 = -02, I'G13 g .2)

77

Table 16.--The Expected and Observed of Total Response Per Generation

 

 

 

 

 

 

 

in Economic Unit. (1 unit = $.10)
ethod E
' I II III IV
Population
I 2.0339 1.9642 1.9423 2.114
0 2.8768 3.1972 2.9629 2.8686
II E 2.7677 3.0175 2.1874 3.0142
0 2.6363 3.0069 3.5258 3.0835
III 3.5013 3.7217 3.5493 3.7576
0 3.7558 4.3523 4.0418 4.1466
IV 1.6211 1.6954 1.6774 1.4346
0 1.5923 2.2073 2.2240 2.0716
V 2.1399 2.2790 2.2786 2.3890
0 2.2901 2.8154 2.7471 2.3350
VI E 2.6587 2.8592 2.7862 2.7744
0 2.9984 2.8592 3.4437 2.8234
VII .4246 .4992 .4931 .4781
O .4382 .5463 .8891 .6106
VIII E .6364 .7460 .6996 .7146
.5877 .9091 .5536 .8567
IX E .8482 .9891 .9097 .9401
O .9600 1.1304 .8289 .8847
METHOD I Mass Selection
METHOD II Selection Index
METHOD III Restricted Selection Index
METHOD IV Independent Culling Levels

78

Within the same levels of heritabilities, the larger the genetic cor-
relation, the less the total economic response.

Table 17 shows the expected and observed relative percent
efficiency for total economic response for different selection methods.
Each value in the table was calculated as: (Hi/H11) x 100%

Where subscript i can be either I, II, III, or IV. Thus, selection
index was considered to be 100% efficient.

On the average the expected values and observed values indi-
cated that selection index method was better than the other three
methods. When the results from mass selection and selection index
were compared, the observed values showed that except in population VI
mass selection was about 20% less efficient. Here, in comparing
percent efficiency there was no indication of the effect of different
parameter combinations. Random error might be the cause for observed
value in population VI a little over 100%.

Is 20% less efficiency for mass selection too much to pay
for the simplicity and conveniences of the method? Selection index
requires genetic parameters such as heritabilities, genetic and
phenotypic correlations, and economic value of traits which have to be
made available before the index can be constructed. Further, experts
suggested that the index should be recalculated each cycle of
selection as the required genetic parameters might change due to
selection. Another thing to note is characteristic of selection
index, i.e., it has its greatest value only for a particular
population with particular economic values of traits which the index

is intended for.

Table l7.--Relative Percent Efficiency of Different Selection Methods.

79

 

 

 

 

 

 

 

 

Method
I II III Iv
Population
I E* 103.55 100 98.88 107.63
0* ; 89.98 100 92.67 89.72
E
II ‘ 91.72 100 72.49 99.89
0 ; 87.68 100 117.25 102.55
III 3 94.08 100 95.37 100.96
1
1
; 86‘29 100 92.86 95.27
z
IV 1 95.62 100 98.94 84.62
0 i 72.14 100 100.75 93.85
v E 93.89 100 99.98 104.83
0 81.34 100 97.57 82.93
VI 92.99 100 97.45 97.03
0 ' 104.87 100 120.44 98.75
VII . 85.06 100 98.78 95.77
I
I 80.21 100 162.75 111.77
I
I
VIII E 2 85.31 100 93.78 95.79
0 ! 64.65 100 60.89 94.24
1x I 85.75 100 91.97 95.05
i 84.92 100 73.33 78.26
*E = Expected
*O = Observed
METHOD I : Mass Selection
METHOD II : Selection Index
METHOD III : Restricted Selection Index
METHOD IV - Independent Culling Levels

80

The observed percent efficiency of restricted selection index
showed that this method was almost as efficient as when selection index
was used. However, in the populations where genetic correlation was
low the observed value seems subjected to larger random error which
made further interpretation difficult.

Independent culling method was somewhat less efficient than
selection index method. On the average this method was about 10% less
efficient. Using independent culling method the population size was
reduced after the first stage of selection where females were selected
for body weight at sexual maturity. At this stage only 25% females
were saved until the second stage where females were then selected for
egg production. From the number of females that entered the second
stage selection, 50% were saved for parents producing the same size
of breeding flock as when other methods were used. This means that
with this method the cost to maintain the flock was reduced 75%
compared to when selection index was used. The reduction of cost and
avoidance of the known complexity of constructing a selection index
seems more than enough to pay for the 10% loss in efficiency from
use of this method. In practice independent culling is as simple as
mass selection but the results in this study also indicated it's more
efficient than mass selection.

Genetic correlations during
selection study

 

 

The required genetic correlations, IG and rG (subscript

12 13

l, 2 and 3 is for body weight, egg production and egg weight,

respectively) were first introduced when the base populations were

81

generated, according to the simulation method presented in a previous
section. A separate run was made to check the sample values of

rG12 and rGl3 for different parametric values of rG12 and rGl3 required.
Table 18 shows the range of rG and rG for 30 sample populations with

12 13
sample size equal 132 individuals. These values indicated that the
simulation technique used to build up the required correlation was
successful, as the parametric values fall within the range of values
observed.

The second time the required genetic correlations were forced
in when the offspring populations were generated, with the procedure
already explained.

The genetic correlations in unselected offspring and selected
offspring in each generation were calculated as the product moment
correlation of genotypic values. For different selection methods the
genetic correlations in the unselected and selected offspring were
plotted together for every two generations and are presented in
figures 3.1 to 6.3. In all populations the genetic correlation was
maintained regardless of the method of selection, because the geno-
typic and phenotypic variances and covariances did not change
substantially.

Results from mass selection indicated that in all levels of
heritability used in this study truncation selection decreased the gene-
tic correlations in the selected offspring, except when the genetic
correlations were low (rG12 = -.2 and r613 = .2). Thus both the
positive or negative genetic correlation showed the same results.

The decrease of IS and IS in the selected offspring were observed

12 13

more clearly when heritability of body weight = .4, 1612 = -.6 and

82

Table 18.--Range of Sample Values of Genetic Correlation of Body
Weight and Egg Production and Body Weight and Egg Weight
in Base Population.*

 

 

 

 

 

h2 h2 h2 Parametric Values Range of Sample Values
1 2 3
r
G12 rG13 rG12 rG13
-.6 .5 -.67 to -.41 .27 to .53
.4 .3 .5 —.4 .3 -.55 to -.25 .16 to .43
-.2 .2 -.31 to .01 .04 to .40
- 6 .5 — 69 to -.42 .31 to 66
3 2 3 - 4 .3 - 58 to - 26 12 to 44
- 2 2 - 39 to -.08 03 to 38
-.6 5 - 71 to -.52 34 to 64
1 1 l - 4 3 - 49 to -.23 16 to 43
- 2 2 - 36 to -.02 08 to 37

 

 

 

 

 

 

 

132 individuals

* Range of 30 samples and sample size

83

 
   
 

 

 

 

Unselected Offsprings
---------- Selected OffSprings
.7..
.5 1
’-—-——-- .—
\,r
.3 "
rGl3-.5
01 T
O r: - 4—4 .1; 1
..3...
15.. ,r~~-- A
/\ \\\
fig),
rGlZ=—.6

rG

 

 

 

 

 

.50
’/~\\
.3 o //’ \\
1” \\fi\
=.2 -‘~v" \
13 .1V ‘\
° 1 I, z 7* 7.

=- W A . /
12 '2..3‘\ ’/4_;>-—-" ’

“51.

 

 

Figure 3.1. rGlz and rGl3 in the Unselected and Selecteg Offsprings

= 03' h = 05)

Under Mass Selection. (hl 8 .4, h 3

2

84

Unselected Offsprings

.7.. --------- Selected Offsprings

 

-03 4»

Isl-23‘. 6.”; 0

 

 

 

 

 

 

03 p ’I’ \\
‘ \‘
\ / ‘
I.613— 2 '1} \\‘v”
o————3=r—-7— T - _
-4 0 l”’ ‘ ‘ ~ 1 ”’7 ’0
/ -~ _--- -”
rG --.2'3“
12
-.5.“-

 

Figure 3.2. rG12 and r613 in tbs Unselecsed and Salected Offsprings Under
Mass Selection. = .3 h2 2, h3 =.3)

85

 

 

Unselected Offsprings
'7" ---------- Selected Offsprings
5, ~ I. ___———:’
V‘“~ ’x’
rG =.5 v
13 -3 1L
.1.,
O 4 ——‘ <7 44 -
.J1., l '7 6 8 IO
-.3 If

 

 

 

 

 

 

Figure 3.3. rG and r6 3 in She Unselscted andZSelected Offsprings Under

Mass Selection (h1 - .1, h2 - .1, h3 = .l)

86

 

 

 

 

 

Unselected Offsprings
7 ----------- Selected Offsprings
‘9'-“ N.

05 W _‘

.3 r

'1 0

01 [ 7- ;] Z 3 IO
..3.
...5'.

—-——-” ‘\\
v-’-‘\
rG =-.6'“7+
05"“
,A I
. /
r613: 3 ‘sv ’ \v/
'1‘»
O : fir 4 i j
.7.

-44L 7' 6 48 lo

 

 

 

 

Figure 4.1.

 

r6 2 and rGl in the Unselectfid and Selected Offsprings Under
Selection Index. (h1 = .4, h2 = .3, h3 = .5)

Figure 4.2.

87

Unselected Offsprings

 

 

 

---------- Selected Offsprings
’f---- --——_
./
~.‘. 1’
s§v/

i ; ' ‘
7 4 8 lo

[I

 

 

 

 

 

 

 

rG and r6
1
Selgction Index.

in the Unselectad and Selected Offsprings Under
2

I
H

-.3.

”5‘

r6 =-.6 r'24

 

13 ' ’1‘

D

 

88

Unselected Offsprings

---------- Selected Offsprings

’
”

’ Q,

»4
'Q
0‘.
up
8

 

IG12=".2 “‘3 1

 

-.51

 

Figure 4.3. rG 2 and rGl in the Unselectsd and Selected Offsprings Under
Se1ec sex.

tion In (h1 = .1, h = .1, h = .1)

2 3

89

Unselected Offsprings

 

 

 

 

 

 

 

J ---------- Selected Offsprings
a? I
I
I
05 ~‘ /”’F~‘.\~///
r6 = 5 “~/
13 .3 i
.1 1b
0 11¢, ‘ i’ A A
2. V 6 3 10
‘01 1
rG - 6
12
r‘ ..
r513 3
rGlz-— 4
rG13= 2
rG12=- 2
-054»

 

Figure 5.; r612 and rG in the Unselected and Selested OffsBrings Under

, 3
Restricted Selection Index. (h1 = .4, h2 - .3, h3 = .5)

90

 

 

 

 

 

 

Unselected Offsprings
0? U ---------- Selected Offsprings
.5 M
\V’-
rGl3=.5 ,3 _
.1 .
0 7 = < - .
-.1 3 y 5 3 10
-430
rGlZ=-.6
IGJ3= 3
O. 1 7 6 8 IO
-01
-03 d) »
~---¥’ 5‘“-.-
=-.4 '"5 u
rG12
.5
r613=.2
rGlZ=-.2

 

 

Figure 5.2. rG and r6 in the Unselected and Selested Offsprings Under

3
Restricted Selection Index. (hl - .3, h2 = .2, h3 = .3)

13’
03 '

01p

91

Unselected Offsprings

---------- Selected Offsprings

A
\ \~.‘

‘ \
~ ~Q‘
5-

 

-'1 1

 

 

”.1
‘Q
0\ .
no
3

p.
.1.
9x
00
a

 

 

 

Figure 5.3.

rG
6.42..

 

in the Unselected and Selested Offsprings Under
(h1 = .1, h2 = .1, h3 = .1)

and rG 3
icted Selection Index.

92

Unselected Offsprings

 

 

 

---------- Selected Offsprings
.’1
05¢ [A
\ - -._. \\
r613= 5 ‘
03"
01 1
o 12 t i ‘ 4
-.11 q' ‘8 ’0
-03 1
-.5 ‘ / —— \ ” ~~
/ ‘M
l
_- -.7.
r612 6

 

 

l3
0 A 1 v 4. 2
-n1. 7 ‘ 8 1%
/F--‘-.~‘ ”
- 3 /” “\--———r"/

 

 

 

 

 

 

Figure 6.1. r6 2 and rGl in the Unselectedzand Selested Offsprings Under
Independent gulling Levels. (h1 I .4, h2 = .3, h3 = .5)

Figure 6.2.

 

93

Unselected Offsprings
---------- Selected Offsprings

 

 

 

 

 

 

rG 2 and rG1 in the Unselectedzand Selested Offsprings Under
Independent gulling Levels. (hl - .3, h2 = .2, h3 8 .3)

94

Unselected Offsprings

 

 

 

 

 

---------- Selected Offsprings
o7 ‘
”‘~‘ ”
051 ‘ ’ V’
rGl3=.5
I3 ‘
01 ‘
0 A. #‘ : 4 t
”1‘ z 7 3 lo
-034
5 f”’ \‘\
- ‘ ” \\\~‘
-.71
c =- 6
r 12
05‘
-u-n-A
r6 = 3 ‘\‘\/” VI
13
011
O 4 j i ‘ f
”1‘ z y ‘ m
-3. ”’
’l’P~-_ ’
-0 1 ‘N-- I
r6 =- 4 5 ‘ ”
12
05.1
0.31
I“

----~

 

 

 

Figure 6.3. rG 2 and rG1 in the Unselected and Selested Offsprings Under
Independent gulling Levels. (h1 = .l, h2 = .1, h3 = .l)

95

r613 = .5. Further, examination on the genetic variances and covari-

ances indicated that the covariances were decreased slightly more than
the variances. The same evidence was reported by Parker et al. (1969)
and Cheung and Parker (1974).

When either selection index or restricted selection index was
used, the results didn't indicate that the genetic correlations in
the selected offspring were decreased but it appeared that the genetic
correlations in the unselected offspring were maintained after trun-
cation (Figures 4.1 to 4.3 and figures 5.1 to 5.3). This evidence
was seen more clearly when the genetic correlations were high.
Probably this can be explained by the fact that when the index was
used, the truncation was applied not directly on the phenotypic
distribution of traits but rather on the index values.

Results from independent culling seem to indicate that the
genetic correlations in the selected offspring were reduced only when
the genetic correlations were high (rG12 = -.6 to -.4 and r613 = .5 or
.3), and this applied for all levels of heritabilities used in this
study (Figures 6.1 to 6.3). When the genetic correlations were low,
the interpretation of results became difficult.

Possibly if more generations were added, the effect of trun-
cation on the genetic correlations in the selected offspring could

be examined more clearly as was done by Parker et a1. (1969) who

carried the selection for 30 generations.

IMPLICATIONS FOR THE NATIVE CHICKEN

POPULATION IN INDONESIA

The resulUsfrom this simulation study indicate that improve-
ment of body weight in native chicken populations is possible by
selection within the populations. Although the exact genetic para-
meters are unknown, it seems reasonable to assume that the parameters
are within the range of parameter values used in this study.

Selection index showed the best result but the method
can't be recommended at this time because of its complexity and
impracticality.

Mass selection and independent culling levels are simple and
practical and have proved effective in improving body weight in other
populations. The applications of these two methods are straightforward,
understandable, and little difficulty should be encountered in
beginning and continuing such selection for several generations.
However, egg production may be decreased as much as 1.7% per gener-
ation by mass selection and somewhat less by independent culling levels.
If one may not worry for a while about the decrease in egg production
and if mass selection should be the first choice, then the opera-
tional cost should be reduced even more than when independent culling
levels are used. Although in this study mass selection was simulated

for a designed breeding experiment, extensive application of this

96

97

method in the native chicken populations in the rural areas is not
impractical. The method may be combined with random mating, then
progress per generation will depend primarily on the selection inten-
sity. The selection intensity to be practiced should be based on
judgments about each local area regarding acceptable culling
percentages.

Some have suggested improving the native chicken populations by
crossbreeding because of the promising results from heterosis in the
first generation following the crossing. But, after the first
generation, there is confusion about planning for further improvement.
It is not justifiable to continue to breed the first cross as the
effect of heterosis will disappear. Furthermore in a crossbreeding
program there is uncertainty about which breeds will be appropriate
for crossing with the native chickens. Also, it is questionable if
the chosen breed will adapt well to the local conditions.

Others may think to replace the native chicken populations with
imported broiler strains as they see the results in developed
countries. However, one should not forget the conditions of the rural
areas and how those differ from environments that the broiler strains
have been developed in and intended for.

The broiler industry is growing in Indonesian urban areas.
Urbanization might make possible changes in existing conditions for
poultry husbandry, leading to a decision to discard the native
chicken populations, which at present still provide nearly 90% of

chicken meat for the nation. But, again urbanization is a relatively

98

slow process, affecting only minor portions of the land, and decisions
concerning ways to improve poultry production are urgently needed.
Given the results of this study concerning the near optimal
efficiency of mass selection and its practicality compared to other
selection schemes, it appears to be the method of choice for short-
term progress in native chickens. Modifications may be required in the

future if egg production declines to unacceptable levels.

SUMMARY AND CONCLUSIONS

A simulation study was performed to examine the effectiveness
of four different methods of selection: mass selection, selection index,
restricted selection index, and independent culling levels (two stage
selection). These methods were compared for each of nine simulated
native chicken populations. The populations were distinguished by
different genetic parameter values of heritabilities of body weight, egg
production, and egg weight, and genetic correlations of body weight with
egg production and egg weight. Respective values for the nine popula-

tions are: I: .4, .3, .5, -.6, .5

VI: .3, .2, .3, -.2, .2
VII: .1, .1, .l, -.6, .5
VIII: .1, .1, .1, -.4, .3
IX: .1, .1, .1, -.2, .2
Simulation continued for ten generations with selection
beginning at generation one. Size of breeding flock was kept constant
through generations (12 males and 120 females) by saving the upper

1.25% males and 12.5% females to become parents for the next

99

lOO

generation. The economic value per unit of body weight was 25 times
the economic value per unit of egg production.

Method of selection differed for females but the same selection
was applied to all males, phenotypic mass selection for high body
weight.

Examination of mean genotypic values showed that in all
methods, selection theory was relatively accurate in predicting the
response. For mass selection,predictions of direct response were more
precise than those for correlated responses, and prediction was more
accurate for positively than for negatively correlated responses.

When the negative genetic correlation and heritabilities were low,

the correlated responses were predicted more precisely than when both
were high. Direct responses increased as heritability increased and
the amount of correlated response depended on the magnitude of genetic
correlation, as expected.

For the selection index method, within the same combina-
tion of heritabilities, mean genetic change of body weight
was highest when genetic correlation between body weight and egg
production was lowest, but on the contrary the smallest negative
genetic change in egg production occurred. In populations with
negative genetic correlation equal -.2, this method prevented negative
genetic change on one trait but also tended to decrease positive
'change on the other trait.

Restricted selection index showed much the same results as for
selection index: that the magnitude of negative genetic correlation

affected the mean genetic change of the two traits in opposite ways.

101

The same evidence also was found when independent culling levels
were used.

Mean body weight response was largest when mass selection was
used. At generation ten,mean body weight had almost doubled, or
increased 75% and 25% when heritability was .4, .3 and .1
respectively. This was accompanied by decrease of egg production and
increase of egg weight, the magnitudes depending on the heritabilities
and genetic correlations. Populations with the combination of highest
heritabilities and genetic correlation showed that egg production was
decreased 1.7% per generation.

In most populations selection index and restricted selection
index showed comparable responses of mean body weight. Independent
culling levels produced slightly smaller responses. Selection index
showed no decrease in egg production in populations with genetic
correlation equal to -.2.

Restricted selection index produced decreases in egg
production 40-50% less than mass selection when the genetic corre-
lation between body weight and egg production was either -.6 or
-.4, regardless of the heritabilities.

The relative efficiency for total economic response showed
that in all populations when the selection index was considered 100%
efficient then, mass selection was about 20% less efficient, inde-
pendent culling levels was 10% less efficient and restricted
selection index was almost as efficient,especially when genetic
correlation was high.

During the selection process, in all populations and for all

methods of selection, the initial genetic correlations were maintained

102

and the genetic and phenotypic variances and covariances did not
change more than trivially. Mass selection and independent culling
levels reduced genetic correlations in selected offspring but those
differences were clear only when the genetic correlation was high.
Selection index and restricted selection index maintained genetic
correlations in all cases.

Selection index method showed higher total economic response
than either mass selection or independent culling levels, but
because of the complexity and impracticality of the index for
Indonesian conditions, it shouldn't be the first choice. Genetic
parameters needed to construct an index should be made available
from the population where the selection is going to be applied, but
values of the genetic parameters are still unknown for the native
chickens. Furthermore selection index requires identification of
individuals along with the record for each measurement used in the
index. Present conditions would not permit such identification without
major changes in current practices. However, mass selection and
independent culling levels are simple and could easily be practiced
under present conditions.

If selection is for the improvement of body weight per se then
mass selection gives the highest response,but when total economic
response is considered, then independent culling levels provide
better results. But when independent culling levels are used, after
the first stage of selection,at sexual maturity,25% of the offspring
are saved to enter the second stage of selection and later only half

of them will be selected for breeding. In mass selection at sexual

103

maturity only 12.5% of individuals are saved for breeding. It is
questionable if the additional cost of keeping 125% more individuals
would be paid for by the additional economic return from independent
culling. The decrease in egg production is somewhat smaller for
independent culling than for mass selection. If the decrease in egg
production is considered to be an important problem then, it is
questionable if as much as 1.7% decrease per generation by mass
selection could be accepted.

When a negative genetic correlation exists and effort is made
to prevent negative change of a secondary trait (but not to obtain
positive change), then results of this study indicate that application
of restricted index selection is not effective when practiced on only

one SEX .

LITERATURE CITED

LI TERATURE C ITED

Abplanalp, B., F. X. Ogasawara, V. S. Asmundson. 1963. Influence of
selection for body weight at different ages on growth of
turkeys. British Poultry Sci. 4:71.

Barker, J. S. F. 1958a. Simulation of genetic systems by automatic
digital computers. III. Selection between alleles at an
autosomal locus. Aust. J. Biol. Sci. 11:603.

Barker, J. S. F. 1958b. Simulation of genetic systems by automatic
digital computers. IV. Selection between alleles at a
sex-linked locus. Aust. J. Biol. Sci. 11:613.

Becker, W. A. 1975. Manual of quantitative genetics. Washington
State University. Pullman, Washington 99163.

Bell, A. E. and H. W. McNary. 1963. Genetic correlation and asymmetry
of the correlated response from selection for increased body
weight of Tribolium in two environments. Proc. XI. Int. Cong.
Gen. p. 256.

Bereskin, B., C. E. Shelby, and L. N. Hazel. 1969. Monte Carlo
studies of selection and inbreeding in swine. I. Genetic
and phenotypic trends. J. Animal Sci. 29:678.

Bereskin, B. 1972. Monte Carlo studies of selection and inbreeding in
swine. IV. Selection response. J. Animal Sci. 34:726.

Bogyo, T. P. and S. W. Ting. 1968. Effect of selection and linkage
on inbreeding. Aust. J. Biol. Sci. 21:45.

Bohren, B. B., W. G. Hill, A. Robertson. 1966. Some observations on
asymmetrical correlated responses to selection. Genetical
Research 7:44.

Bruns, E. and W. R. Harvey. 1976. Effects of varying selection
intensity for two traits on estimation of realized genetic
parameters. J. Anim. Sci. 42:291.

Brunson, C. C., G. F. Godfrey, and B. L. Goodman. 1956. Types of gene

action in the inheritance of ten-week body weight and breast
angle in broiler. Poultry Sci. 35:524.

104

105

Bulmer, M. G. 1976. The effect of selection on genetic variability:
a simulation study. Genetical Research 28:101.

Cheung, T. K. and R. J. Parker. 1974. Effect of selection on heri-
tability and genetic correlation of two quantitative traits in
mice. Can. J. Genet. Cytol. 16:599.

Clayton, G. A., G. R. Knight, J. A. Morris and A. Robertson. 1957. An
experimental check on quantitative genetical theory. III.
Correlated responses. J. Genetics. 55:171.

Cunningham, E. P., R. A. Moen, and T. Gjedrem. 1970. Restriction of
selection indexes. Biometrics 26:67.

Dillard, E. U., G. E. Dickerson and W. F. Lamoreux. 1953. Heritabi-
lities of egg and meat production qualities and their genetic
and environmental relationships in New Hampshire Pullets.
Poultry Sci. 32:897.

Eisen, E. J. 1977. Restricted selection index: an approach to
selecting for feed efficiency. J. Anim. Sci. 44:958.

El-Issawi, H. F. 1975. Prospects for increasing world poultry
production. World. Animal Review 16:42.

Emsley, A., G. E. Dickerson and T. S. Kashyap. 1977. Genetic Para-
meters in progeny test selection for field performance of
strain-cross layers. Poultry Sci. 56:121.

Epstein, H. 1977. The Awassi Sheep in Israel. World Review of
Anim. Prod. 13:19.

Everett, R. W., C. E. Meadows and J. L. Gill. 1967. Estimation of
genetic trends in simulation data. J. Dairy Sci. 50:550.

Falconer, D. S. and M. Latyszewski. 1952. The environment in relation
to selection for size in mice. J. Genet. 51:67.

Falconer, D. S. 1954. Validity of the theory of genetic correlation.
An experimental test with mice. J. Heredity 45:42.

Falconer, D. S. 1960. Introduction to quantitative genetics. The
Ronald Press Company. New York.

Farthing, B. R. and J. E. Legates. 1957. Genetic covariation between
milk yield and fat percentage in dairy cattle. J. Dairy Sci.
40:639.

Finney, D. J. 1962. Genetic gains under three methods of selection.
Genetical Research. 3:417.

Fraser, A. S. 1957a. Simulation of genetic systems by automatic digi-
tal computers. I. Introduction. Aust. J. Biol. Sci. 10:484.

106

Fraser, A. S. 1957b. Simulation of genetic systems by automatic
digital computers. II. Effects of linkage on rates of advance
under selection. Aust. J. Biol. Sci. 10:492.

Fraser, A. S. 1960a. Simulation of genetic systems by automatic
digital computer. V. Linkage, Dominance and Epistasis. In
Biometrical Genetics. Kempthorne, 0., Ed. Pergammon Press,
New York, p. 70.

Fraser, A. S. 1960b. Simulation of genetic systems by automatic
digital computers. VI. Epistasis. Aust. J. Biol. Sci. 13:150.

Fraser, A. S. 1960c. Simulation of genetic systems by automatic
digital computers. VII. Effects of reproductive rate and
intensity of selection on genetic structure. Aust. J. Biol.
Sci. 13:344.

Fredeen, H. T., and P. Jonsson. 1957. Genic Variance and Covariance
in Danish Landrace Swine as evaluated under a system of
individual feeding of progeny test groups. 2. Tiers.
Zuchtbiol., 70:348.

Freeman, A. E., and R. S. Dunbar, Jr. 1955. Genetic analysis of the
components of type, conformation and production in Ayreshire
cows. J. Dairy Sci. 38:428.

Friars, G. W., B. B. Bohren and H. E. McKean. 1962. Time trends in
estimates of genetic parameters in a p0pulation of chickens

subjected to multiple objective selection. Poultry Sci. 41:
1773.

Gill, J. L. 1965a. Effects of finite size on selection advance in
simulated genetic populations. Aust. J. Biol. Sci. 18:599.

Gill, J. L. 1965b. A Monte Carlo evaluation of predicted selection
response. Aust. J. Biol. Sci. 18:999.

Gill, J. L. 1965c. Selection and linkage in simulated genetic
populations. Aust. J. Biol. Sci. 18:1171.

Gill, J. L. and B. A. Clemmer. 1966. Effects of selection on degree
of inbreeding. Aust. J. Biol. Sci. 19:307.

Gowe, R. S., W. E. Lentz, J. H. Strain. 1973. Long term selection for
egg production in several strains of White Leghorns:
Performance of selected and control strains including genetic
parameters of two control strains. 4th Europ. Poult. Conf.,
London, p. 225.

Harvey, W. R. and G. D. Bearden. 1962. Tables of expected genetic
progress in each of two traits. USDA Agricultural Research
Service Publication No. 20-12.

107

Hazel, L. N. and J. L. Lush. 1942. The efficiency of three methods
of selection. J. Hered. 33:393.

Hazel, L. N. 1943. The genetic basis for constructing selection
indexes. Genetics 28:476.

Hill, W. G., and A. Robertson. 1966. The effect of Linkage on Limits
to artificial selection. Genet. Res. 8:269.

Jerome, F. N., C. R. Henderson and S. C. King. 1956. Heritabilities,
genes interactions, and correlations associated with certain
traits in the domestic fowl. Poultry Sci. 35:995.

Johansson, 1. 1950. The heritability of milk and butterfat yield.
Anim. Breed. Abstr. 18:1.

Johnson, K. R. 1957. Heritability, genetics and phenotypic corre-
lations of certain constituents of cow milk. J. Dairy Sci.
40:723.

Kan, J., W. F. Krueger, J. H. Quisenberry. 1959. Non-additive gene
effects of six broiler traits as studied from a series of
diallel matings. Poultry Sci. 38:972.

Kempthorne, O. and W. A. Nordskog. 1959. Restricted selection
indices. Biometrics 15:10.

Lentz, W. E. and R. S. Gowe. 1974. Change in genetic parameters of
White Leghorn Strains under Long term Selection. Poultry
Sci. 53:1947 (Abstr.).

Magee, W. T. 1965. Estimating response to selection. J. Anim. Sci.
24:242.

Martin, F. C. Jr. and C. C. Cockerham. 1960. High speed selection
studies. In Biometrical Genetics, Kempthorne, 0., Ed.
Pergammon Press, New York, p. 35.

Martin, G. A., E. W. Glazener, W. L. Blow. 1953. Efficiency of
selection for broiler growth at various ages. Poultry Sci.
32:716.

McArdle, A. A. 1972. Methods of poultry production in a developing
area. World Animal Review 2:28.

McDowell, R. E. 1972. Improvement of Livestock Production in Warm
Climates. W. H. Freeman and Company. San Francisco.

Mehta, N. T., S. C. Mohapatra, S. D. Abuja, P. N. Sharma, S. K. Pati,
R. P. Sharma. 1975. Studies on the construction of selec-
tion indexes for broiler dams with body weight, egg production
and egg weight. Indian J. Poultry Sci. 10:175 (Abstr.).

108

Nordskog, A. W. 1960. Importance of egg size and other factors in
determining net income in random sample tests. Poultry Sci.
39:327.

Nordskog, A. W., M. Festing. 1962. Selection and correlated responses
in the fowl. 12th WOrld's Poultry Congress. p. 25.

Okada, I., and R. T. Hardin. 1967. An experimental examination of

restricted selection index using Triboleum castaneum.
Genetics 57:227.

Okada, I., and R. T. Hardin. 1970. An experimental examination of
restricted selection index, using Triboleum castaneum. II.
The results of Long term one-way selection. Genetics 64:533.

Parker, R. J., L. C. McGilliard and J. L. Gill. 1969. Genetic
correlation and response to selection in simulated populations.
1. The Additive model. Theoret. and Apply. Genet. 39:365.

Parker, R. J., L. C. McGilliard and J. L. Gill. 19703. Genetic
correlation and response to selection in simulated papulations.

11. Model of complete dominance. Theoret. Appl. Genet.
40:106.

Parker, R. J., L. C. McGilliard and J. L. Gill. 1970b. Genetic
correlation and response to selection in simulated populations.

III. Correlated response to selection. Theoret. Appl. Genet.
40:157.

Rasmuson, M. 1964. Combined selection for two bristle characters in
Drosophila. Hereditas 51:231.

Reeve, E. C. R., and F. W. Robertson. 1953. Studies in Quantitative
inheritance. II. Analysis of a strain of Drosophila
melanogaster selected for Long Wings. J. Genet. 51:276.

Sather, A. P., L. A. Swiger, W. R. Harvey. 1977. Genetic drift and
the response to selection in simulated populations: the

simulation model and gene and genotype responses. J. Anim.
Sci. 44:343.

Scheinberg, E., A. E. Bell, V. L. Anderson. 1967. Genetic gain in
populations of Triboleum castaneum under uni-stage tandem
selection and under restricted selection indices. Genetics

55:69.

Sen, B. K. and A. Robertson. 1964. An experimental examination of
methods for the simultaneous selection of two characters, using
Drosophila melanogaster. Genetics 5:199.

109

Shimizu, B., Y. Hachinohe, H. Kawana. 1968. A study on genetic
improvement of New Hampshire chickens. I. An investigation
of selection methods. Japan. Poultry Sci. 5:126.

Siegel, P. B. 1962. A double selection experiment for body weight and
breast angle at eight weeks of age in chickens. Genetics
47:1313.

Siegel, P. B. 1963. Selection for breast angle at eight weeks of age.
11. Correlated responses of feathering, body weights and
reproductive characteristics. Poultry Sci. 42:437.

Smith, F. H. 1937. A discriminant function for plant selection.
Ann. Eugen. 7:240.

Strohbehn, D. R. 1974. Evaluation of response to selection and mating
systems with different levels of heritability. Selection
intensity and progeny testing through the use of simulation.
Ph.D. Thesis, Michigan State University, East Lansing, Michigan.

Tabler, K. A., and R. W. Touchberry. 1955. Selection indexes based on
milk and fat yield, fat percent, and type classification. J.
Dairy Sci. 38:1155.

Uddin, Q. 1962. Improvement of indigenous fowls for eggs in eastern
hot countries. 12th World's Poultry Congress, p. 92.

Van Vleck, L. D. 1976. Notes on the theory and application of selec-
tion principles for the genetic improvement of animals.
Department of Animal Science, Cornell University, Ithaca, N.Y.
Mimeograph.

Wartomo, H., Supijono, P. A. 1976. The performance of Kampung and
Kedu chickens. Unpublished report. Faculty of Animal
Husbandry, Gadjah Mada University, Yogyakarta, Indonesia.

Yamada, Y. and A. E. Bell. 1963. Selection for l3-day Larval growth
in Triboleum under two nutritional levels. Proc. XI Int.
Cong. Genet., p. 256.

Young, S. S. Y. 1961. A further examination of the relative effi-
ciency of three methods of selection for genetic gains under
less restricted conditions. Genetical Research 2:106.

Young, S. S. Y. 1966. Computer simulation of directional selection
in large populations. 1. The Programme, The additive and
the dominance Models. Genetics 53:189.