5. CGMFARESGN 0F MASS memos:
WET?! ENEKED UNES AND LINE CROSSES
EN THE WTORY RAT

'fhasis {or tbs Degree of Ph. D.
MICHGAN STATE UNIVERSITY
fiavid Richard Pratt
3950

. ‘ L‘e

 

\
1
S
\.
\

3y
R
“we
1; 1%
ea,
’3:
-.-.

\
( -

A cmgmssm a; AAAss macaw
wm—e wage was Am: mas caessgs
554 THE AAAGAATQRY RAT

’S’hmés {at kha 939:» :75 Ph. D.
WHEGAR SEATE UNEVERSW
fiavié Richarfi 9:51“?
3960

 

This is to certify that the

thesis entitled

A COMPARISON OF MASS SELECTION
WITH INBRED LINES AND LINE CROSSES
IN THE LABORATORY RAT

presented by

David Richard Pratt

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Animal Husbandry

/ / ,"l I , . 1’ y , ’
(7’7/ /" ' I ‘1‘" ’
. I ’

.I/

I ‘1..— /' ) ,, , ’°
’ ’ I ' .-‘ , 1
1’ L V ./ “L V %7L «‘4 I ’/

*7

Major professor

 
  
     

  

LIBRARY

Michigan State
University

 

A COMPARISON OF MASS SELECTION
WITH INBRED LINES AND LINE CROSSES
IN (III-IE LABORATORY RAT

3?

David Richard Pratt

AN ABSTRACT

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

DOCTOR OF PHILOSOPHY

Department of Animal Husbandry
1960

J "/7 «‘1

Approved /C c: A. Jam 2/. J‘b/é (MA;

ABSTRACT

.A COMPARISON OF MASS SELECTION
WITH INBRED LINES AND LINE CROSSES
IN THE LABORATORY RAT

By David Richard Pratt

Inbred lines of rats were developed from each of
two heterozygous populations of the laboratory rat by
mating sons, double sons, and triple sons back to their
dams. Four lines were deve10ped from.a two-strain base
population and four lines from a three-strain base popula-
tion. .A.mass selected group was developed from each of
the base populations.

Selection was made on the basis of weight gain
from the twenty-first to fiftieth day of age. Crosses
were made between the inbred lines within each of the popu-
lations and comparisons of the inbred lines and the line
crosses were.made with the mass selected groups of each
population.

Inbred males gained an average of 40 grams.more
than the inbred females. In the mass selected groups,
males outgained females by 54 grams.

As inbreeding increased, average weight gain of
the inbred animals decreased, and the difference between
the mass selected and inbred groups increased from.4 to
34 grams in the three-strain population and from 12 grams

to 21 grams in the two—strain population.

David R. Pratt

Crosses between the inbred lines restored part of
the loss in weight gains which was incurred during
inbreeding; however, no hybrid vigor in excess of the
.mass selected group was observed in either the 2-1 or 3-1
line crosses (2-1 and 3-1 represent the inbred lines
developed from the two-strain and three-strain base popu-
lations), although there was less difference between the
average gain of the mass selected group and the line
crosses than between the mass selected groups and the
inbred groups the previous generations.

Average litter size at birth in the fourth genera-
tion was significantly higher in both the 2-3 and 3-8
groups (2-8 and 3—8 represent the mass selected groups
developed from the two-strain and three-strain base
populations) than in the corresponding inbred group and
in the 2-3 group in the fifth generation also. As inbred
dams were used in both these generations, it was concluded
that inbreeding of the dam had.more influence on the size
of the litter at birth than did the inbreeding of the
litter itself.

Litter size at weaning was significantly larger
in the 3-8 group than in the 3-I group in the fourth
generation, and it was also larger in the 2-8 group
than in the 2-1 group in the fifth generation. An analy-

sis of the average litter death loss mowed no difference

David R. Pratt

in birth to weaning livability between the mass selected
and inbred groups.

Estimates of the variance components for weight
gains between litters within lines were smaller than the
between litter estimates for the mass selected groups.
Variance of weight gain within inbred litters in the 2-1
group tended to be reduced as inbreeding progressed.

The amount of heterozygosity in the base popula-
tion apparently had no effect on the selection in either
the mass selected or the inbred groups. Average weight
gains, litter size at weaning and at birth, and livability
from.birth to weaning were similar for both mass selected
groups and for both inbred groups including the line
crosses.

One of the major drawbacks to the development of
superior inbred lines by the son-to-mother.matings system
used in this experiment was the reduced prolificacy of the
inbred dams. Reduced prolificacy, in addition to the
already restricted size of the inbred lines, increased the
necessity of developing.more lines if this system were to
be used to its best advantage.

.Another drawback to the development of inbred
lines by a strict method of son-to-dam matings was the
small number of animals available for maintenance of the
individual lines and for making line crosses. Expansion
of the lines by brother-sister matings earlier in the

development of the lines, perhaps even as early as the

David R. Pratt

first generation of inbreeding, along with continued son-
dam matings is recommended to allow sufficient numbers

to maintain the proposed lines in each group as well as
supply sufficient number of females for more complete

testing of the better lines in line crosses.

A COMPARISON OF MASS SELECTION
WITH INBED LINES AND LINE CROSSES
IN THE LABORATORY RAT

BY

David Richard Hatt

A 'JHESIS

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

DOCTOR OF PHILOSOPHY

Department of Animal Husbandry
1960

 

FFéJSMW
a/M it I

 

ACKNOWLEDGEMENTS

The author wishes to express his deepest appre-
ciation to Dr. William T. Magee, under whose guidance
and supervision this work was done, for his interest,
encouragement, and help throughout the course of study.
His patience, willingness to offer assistance, and
inspiration will always be remembered.

Sincere thanks are also expressed to the other
.members of the guidance committee - Dr. w. D. Baten, Dr.
Lois Calhoun, Dr..A. S. Fox, Dr. L. D. McGilliard, and
Dr. J. E. Nellor - for their willing assistance, advice,
and encouragement throughout the course of study. The
association with the entire committee and the opportunity
to have known each of them as a scientist, teacher, and
friend will be cherished always.

Grateful appreciation is extended to Dr. R. H.
Nelson, head of the Department of Animal Husbandry, and
to Michigan State University for provision of the facili-
ties and materials necessary for this study and for
financial aid in the form of an assistantship.

The author wishes to express special appreciation
to his wife, Billye, for her inspiration, patience, and
encouragement and for an excellent typing of the manu-

script. Sincere gratitude is expressed to all his
ii

family whose encouragement, sacrifices, and faith made

the study possible.

iii

David Richard Pratt
candidate for the degree of

Doctor of Philosophy

FINAL EXAMINATION: November 15, 1960, 3:00 p.m.

103 Anthony Hall

THESIS: A Comparison of mass Selection With Inbred Lines

and Line Crosses in the Laboratory Rat
OUTLINE OF STUDIES:

major subject: Animal Husbandry

Minor subjects: Statistics
Anatomy-Physiology

BIOGRAPHICAL ITEMS:

Born: October 20, 1929, Dallas, Texas
Undergraduate studies:
Sul Ross State College, Alpine” Texas, l9t7-48
Southwest Texas State College, San Marcos,
Texas, 19A8-5l
Graduate studies:
Southwest Texas State College, l953-5h
Michigan State University, 1956-58, 1960

EXPERIENCE:

Member United States Air Force, 1951-53
VOcational agriculture instructor, El Campo,
Texas, 1954-56
Graduate assistant, Michigan State University,
1 6-58
Associate pggfessor, Oklahoma Panhandle Agricul-
tural and.Mechanical College, Goodwell,
Oklahoma, 1958-60

MEMBER: .American Society of Animal Production

iv

TABLE OF CONTENTS

INTRODUCTION . . . . . . . . . . . . . . . . . . .
OBJECTIVES . . . . . . . . . . . . . . . . . . . .
REVIEW OF LITERATURE . . . . . . . . . . . . . . .
MATERIALS.AND PROCEDURES . . . . . . . . . . . . .

Method of Developing Inbred Lines Using
Outbred Dams . . . . . . . . . . . . .
The FOLLfldation Stock Rats o o e o e e e, o e 0
Management Procedures . . . . . . . . . . . .
Development of the Heterogenous Base Popula-
tions of Rats . . . . .
Development of the Inbred Lines of Rats . . .
Development of the Selected Groups of Rats .
Adjustment of Data for Sex . . . . . . . . .
making Inbred Line Crosses . . . . . . . . .

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

Total Gain as a Measure of Performance . . .
Selection for Gain in the Mass Selected
Groups . . . . . . . . . . .
Selection for Gain in the 3-1 Group
Selection for Gain in the 2-1 Group
Comparison of 3-1 with 2-1 . . . . .
Comparison of 3-I with 3-3 Group . .
Comparison of the 2-1 and 2-8 Groups
Litter Size as a Measure of Performance .
Litter Size at Birth . . . . . . . .
Litter Size at Weaning . . . . . . .
variance of Weight Gains as a measure of
Pbrformance . . . . . .
Variance Cmmponents of Weight Gains
Between Litters . . . .
variance of weight Gains Within Litters.

SUMMARY AND CONCLUSIONS . . . . . . . . . . . . .
WENDICES O O O O O O 0 O O O O O O O O O O O 0
LI EMTURE CI ED 0 O O 0 O O O O O O O O O O O O O

Page

LIST OF TABLES

TABLE Page

1. Average gain (in grams) of the groups each
generation and differences between the
groups . . . . . . . . . . . . . . . . . . . 2h

2. The expected and observed reach in weight
gains for males and females with the
resulting "gain" in 2-8 group . . . . . . . 27

3. The expected and observed reach in weight gains
for males and females with the resulting
"gain" in 3-8 group . . . . . . . . . . . . 28

4. Average coefficient of inbreeding (F) of the
inbred litters each generation . . . . . . 30

5. The observed reach in weight gains for males
and females with the resulting "gain" in the
high and low lines of the inbred groups . . 32

6. Average litter size at birth for each group
by generations . . . . . . . . . . . . . . . 37

7. Average litter size at weaning for each group
by generations . . . . . . . . . . . . . . . #1

8. Average litter death loss of the groups
by generations . . . . . . . . . . . . . . . #2

9. variance components of weight gains between
litters of the group by generations. . . . . A3

10. Analysis of variance tables for weight gains
of the 2-1 group in generations 3, A, and
5A, used for making variance component
eStimateSooeeeeoeeooeeotechs

ll. Variance of weight gains within litters of
the groups each generation . . . . . . . . A6

vi

LIST OF APPENDICES

APPENDIX

A.

B.

C.

F.

G.

I.

J.

Formula for rat feed prepared at Michigan
State University (1 ton) . . . . . . . .

Litter size and average litter gain of the

groups in generation

1 O O O C O O O O O

Litter size and average litter gain of the

groups in generation

2 O O O O O O O O O

Litter size and average litter gain of the

groups in generation

3 e o o o o o o o o

Litter size and average litter gain of the

groups in generation

Litter size at weaning
the 2-1 line crosses
line of sire . . . .

Litter size at weaning
the 3-1 line crosses
line of sire . . . .

Litter size at weaning
gain of the selected

Litter size at weaning

A O O O O O O O O O

and average gain of
in generation 5 by

and average gain of
in generation 5 by

and average litter

groups in generation 5.

and average litter

gain of the three-line crosses in

generation 6 by line of sire . . . . . . . .

Litter size at weaning

gain of the selected groups in generation 6.

and average litter

vii

Page

52

53

5h

55

56

57

58

59

60

61

INTRODUCTION

.A program has been in operation with swine at
Michigan State University to test the feasibility of
developing inbred lines on the basis of mating sons to
outbred dams. This system would delay any detrimental
effects of inbreeding in the dams but would allow a
rather rapid increase in the inbreeding of the offspring
in each generation.

ZMore use of laboratory animals to check logical
statistical conclusions to learn the differences between
observation and expectation has been advocated (Craft
gtwgl., 1951). The use of laboratory animals has also
been suggested as an aid in solving theOretical problans
in livestock selection.

The purpose of this study with the laboratory rat
has been as a pilot study of the program.with swine. The
similarities of swine and the rat as to nutrition require-
ments, litter size, and growth pattern after weaning allow
valid comparisons. The reproductive habits of the rat
make it possible to observe many more numbers than with
swine in a.much shorter time. The size of the rat allows
these numbers to be kept for observation at a.much lower

expense than would be possible for swine.

1

The true value of an inbred line must be deter-
mined from its value in crossing (Craft gt 31., 1951).
A method of improving swine, based on a comparison of
performance of crossbreds within and between breeds with
the performance of selected, well-bred, and well-managed
non-inbreds, was advocated by Winters 33,21. (19L4).
Thus this project was designed to study the performance
of the crosses between inbred lines compared to that of

animals produced by mass selection.

(l)

(2)

(3)

OBJECTIVES

objectives of this study were:

Tb compare mass selection with selection
among inbred lines as a method of increasing
rate of gain;

To determine if the amount of heterozygosity
in the base population affects the results
of selection;

To test for general and/or specific com-
bining ability in the crossing of inbred

lines;

(A).And to indicate possible guides in the breeding

systems to be used in swine breeding.

REVIEW OF LITERATURE

From.the observations of 15 generations of brother-
sister.matings of the laboratory rat, King (1918) found
that the inbred rats, both male and female, were heavier
in body weight than stock albino rats. Also, she found
that the variability of the body weights decreased as the
inbred generations advanced.

In additional work King (1919) observed that in
10 more generations the inbred strains were lighter in
body weight than earlier generations. However, the inbred
strains were still much superior to the stock albino
strains reared under the same conditions. The results of
25 generations of brother-sister matings showed that no
deteriorationlof rate or extent of growth in body weight
had been produced.

Body weight in lB-week-old rats was found to
average significantly heavier in crossed progeny than in
the component inbred lines from.which the crosses were
derived (Craig and Chapman, 1953). This hybrid vigor
observed in cross progenies was apparently due to domi-
nance and/or overdcminance. The performance of inbred
lines was used to predict the relative value of the lines
for crossing with other inbred lines.

A

Jones (1918) recognized that the effect inbreeding
would have on any organism would depend on the hereditary
constitution of that organism at the time inbreeding
began. He further found that inbreeding generally led to
the elimination of undesirable characters which are con-
trolled by one pair of genes by recognition of these
characters as they were expressed through homozygous
recessives and the culling of the affected animals.
.However, inbreeding was usually accompanied by a loss of
size, vigor, and productiveness, which was generally
returned with the crossing of the inbred strains.

Inbreeding experiments in other species generally
have shown a decrease in productivity with an increase
in inbreeding.

Hybrid mice were found by Chai (1956) to be no
.more superior for bioassay than an inbred parental line.
One line of inbred.mice showed a greater response of the
seminal vesicles to androgen than did another inbred line
or the F1 hybrid of these two lines. The nature of the
stimulus and the genetic constitution of the test animal
was shown to have a great influence on the usefulness of a
line or line-cross used for bioassays.

Tantawy (1957) observed a decline in wing length
and thorax due to inbreeding by mating sibs and first
cousins in Drosophila melanogaster. He also found (1956)
the coefficient of variation declined in.brother-sister

(matings but increased over the controls mated by a sys-
tematic cyclic method of outbreeding, designed to keep
inbreeding at a minimum.

Brother-sister matings of guinea pigs have pro-
duced a decline in all.measures of vigor. Eaton (1932)
dhowed these changes took place uniformly rather than
suddenly.

.A decline in vigor in all characteristics studied,
particularly in frequency and size of litter, was observed
by wright (1922a) after 13 years of brother-sister matings
of guinea pigs. The most marked decline was in frequency
and size of litter. He also found a greater decrease in
gains after birth than in birth weight. Also, the
decrease in the per cent raised of the young born alive
was greater than the decrease in the number born alive.

Improvement over parental stocks of animals pro-
duced by crossing inbred families of guinea pigs was
reported by wright (1922b). This improvement was found
in each trait studied and apparently was manifested most
for adult weight and disease resistance in the progeny of
the first cross females mated to inbred males. mortality
at birth, mortality between birth and weaning, and size
and frequency of litters each were improved somewhat when
both sire and dam were crossbred. .Although the improve-
ment in each factor was small, the combined improvement
resulted in superiority greater than that of random bred

stocks over the inbreds.

Greenwood and Blyth (1951) showed that superiority
of inbred line crosses in poultry depended on the charac-
teristics measured. In crosses between inbred lines, egg
size and body weight were intermediate between parents;
but winter egg production was superior to that of the
parents. Viability in the crosses was similar to viability
in the best inbred lines.

Reproductive fitness in line crosses of domestic
fow1,.measured by the number and weight of November eggs,
was generally superior to contemporary inbreds (Shultz,
1953). lHowever, this superiority was not over the pro-
duction-bred flock from which the inbred lines were
developed. The superior crosses tended to be those
between lines selected differently rather than between
those selected similarly.

In a herd of Holstein-Friesian cattle, an average
decrease in birth weight of about one-eighth of a pound
was reported for each one per cent increase in inbreeding
of a calf itself (Nelson and Lush” 1950). They found a
slightly smaller effect due to the inbreeding of the dam.

Regan 33 31. (19h?) observed increased total mor-
tality in dairy calves with increased inbreeding. His
data included abortions, stillbirths, and postnatal
deaths to four months of age.

Tyler 35 31. (l9t7) reported an average decrease

in birth weight of 0.28 pounds for each one per cent

increase in the inbreeding of calves. However, the inbred
calves of the sires transmitting genes for heavy birth
weights tended to be heavier than outbred calves. This
was due to their having more of their sires desirable
genes for birth weight which cancelled part of the reduc-
tion in birth weight due to homozygosity of undesired
recessive genes for birth weight.

The result of inbreeding of 362 litters of Chester
White swine was a downward trend in litter size at birth,
28 days of age, and at 70 days of age (Hetzer 32 31.,
19h0). Litter size seemed to be more affected by the
differences of inbreeding of the litters themselves than
by the inbreeding of the sires and dams.

An increase of 10 per cent in the inbreeding of
dams resulted in a decrease of about 0.6 pig per litter
at farrowing time (Stewart, l9h5). However, the litter
size at birth was apparently not affected by the inbreed-
ing of the litter itself.

Godbey and Starkey (1932) reported a negative
correlation between degree of inbreeding and weaning
weight of Berkshire swine.

.A decrease in fertility and very high mortality
in the second generation of brother-sister mating of
Poland China swine caused McPhee 95 31. (1931) to discon-
tinue their experiment with swine sib matings.

Half-sib swine matings resulted in a decrease in
the number of pigs farrowed as reported by Willham.and
Craft (1939). A gradual decrease in the average number
of pigs weaned per litter as well as the per cent of
pigs raised to weaning occurred as inbreeding increased.
The coefficient of inbreeding in the eighth generation
was A5.6 per cent.

Craft (1953) observed that for each 10 per cent
increase in inbreeding there was a decrease in number of
pigs farrowed of approximately one-third of a pig per
litter and about one-half pig per litter for number weaned.
He also found an apparent reduction in strength and liveli-
ness of pigs at birth as inbreeding was increased. Litter
size was much more difficult to maintain in inbred lines
than was growth rate.

Previously Craft (l9h3) had reported that,
although there might be slight decreases in fertility,
vitality and growth rate with inbreeding, the rate of
inbreeding in swine could be increased five to ten times
as fast as that rate commonly practiced by pure bred
breeders without loss of individual merit.

maintenance of litter size in inbred lines of
swine was found by Comstock and Winters (l9h4) to be
‘much.mere difficult than maintenance of growth rate.

They therefore emphasized the necessity for maximum atten-
tion to selection for fertility in the development of
inbred lines.

10

Single crosses between inbred (F=.42) lines of
swine showed that inbreeding had a greater effect on
viability than on rate of growth (Dickerson gt_gl., 1946).

‘Winters 23,31, (19h4) obtained a degree of
increased vigor from crossing inbred lines of swine in
keeping with the decrease of inbreeding of the crossbreds.
Also, the increased vigor was greater in line crosses
between breeds than within breeds; thus, the increase in
vigor was apparently affected by the genetic diversity
of the parental stocks. Vigor was based on measurements
of fertility, litter size, survival, rate and economy of
gain, and score for body conformation. They also found
a tendency for the superior lines to produce the superior
crossbreds.

Sierk (19h8) reported an increase in vigor of
approximately 15 per cent for the best crosses between
inbred lines compared with non-inbred crosses and non-
inbred swine. The best line crosses were between lines
having the greatest genetic diversity in the foundation
animals. Thus, this indicated the importance of genetic
diversity for heterosis (Sierk and Winters, 1951).

Inbred line crosses of Duroc swine resulted in an
increase in both number of pigs per litter and litter
birth weights (Chambers and Whatley, 1951). Hybrid vigor
as expressed in the increase in viability of the pigs and
productivity of the two-line cross gilts was greater than

 

ll

hybrid vigor expressed as increased growth rate in indi-
vidual pigs. Two-line cross litters raised by inbred
dams performed equally as well as outbred litters for
the same reason. Two-line cross gilts, mated to an
inbred boar of a third line, produced three-line cross
litters which were superior to both two-line cross and
outbred Duroc litters for most of the characteristics
studied. In most cases the increased number of pigs per
litter accounted for a large percentage of increase in
total litter weight.

{An inbred line of Chester White swine with an
average inbreeding coefficient of .a in the sixth genera-
tion, which had a satisfactory reproductive rate and no
important defects, was found by warwick and Wiley (1950)
to perform well in line crosses. When the inbred Chester
gilts were mated to inbred boars of Landrace x Duroc
foundation stock, the resulting cross line pigs were
heavier at birth; had superior gaining ability, longer
carcasses, and a larger average number of pigs per litter;
and required less feed per unit of gain than the conven-
tionally bred purebred and crossbred swine. .

Garwood (1956) analyzed data collected from.Duroc
swine at experiment stations in Ohio, Oklahoma, and
Nebraska. He found that hybrid vigor in general was more
evident when two-line crosses were compared with inbreds

than when two-line and three-line crosses were compared.

l2

Heterosis in the two-line crosses over the inbreds was
.measured as the increase in performance of the reciprocal
crosses over the average of the parental lines. Heterosis
was determined by the mean estimations of inbreds and
two-line crosses which were unadjusted for differences

in effects of litter inbreeding. Heterosis was deter-
.mined in the Ohio two- and three-line crosses on the
basis of the mean estimation of the crosses which were
adjusted for difference in the inbreeding of the dam.

In the two- and three-line crosses at Nebraska and
Oklahoma, general means were used to calculate heterosis.
The heterotic effect was taken as the increase in produc-
tion of three-line crosses over the two-line crosses.
Specific combining ability was found in two of the four
Nebraska lines.

Garwood also reported no differences between

.outbreds and three-line crosses at Oklahoma or between
the Ohio outbreds, rotational crosses, or three-line

crosses e

MATERIALS AND PROCEDURES

Wei—Web 1 marching
Outbred gay;

Inbred lines can be developed by the mating of
females to their full brothers. A male offspring of
this mating can be mated back to his dam. A son in this
double son litter can be mated back to his dam to pro-
duce triple sons. These triple sons can be mated to
their dams and also to full sisters to produce enough

females available for the crossing between lines.

The Foundation mg; 3.91%

Three inbred strains of rats were used: a strain
of hooded rats, an albino strain from the Michigan State
University Chemistry Department referred to as the Hoppert
strain, and an albino strain from the University Isotope
Laboratory. The average inbreeding coefficient of each
of the original inbred strains was estimated to be .80.

Management Pracedures
Matings were made by placing the female(s) and a
male in the breeding cages for ll. days, a period long
enough to include at least two estrous cycles for each

female. When it was evident that reproductive failures

13

lb

might limit numbers too greatly, the females were left in
the breeding cages an additional three to five days.

This additional period would permit time for three estrous
periods.

‘Upon removal of the females from the breeding
cages, they were placed in individual littering cages
until the litters were born and weaned. During the
littering period close daily observations of the females
were made, and birth dates and litter size at birth were
recorded.

Until the birth of the first three-way inbred
litters, the dams were allowed to keep all the rats in
their litters and to raise as many as possible. However,
beginning with the first inbred litters, all litters of
over eight rats were reduced to eight at five days of
age. The heaviest four females and four males were saved
as far as was possible. If there were not four of one
sex in litters of eight or more, five or more of the
majority sex were retained so that there would still be
eight in the litter. It was considered that maintenance
of litters not larger than eight would tend to reduce any
effect that larger litter numbers.might have on growth
rate. Since there is a limit to the amount of milk pro-
duced by a dam.and this factor would naturally affect
growth rate in the litter, it was thought that by limiting

15

litter size to a moderate number each rat would have ade-
quate nutrition for maximum.growth before weaning.

Where there was opportunity to do so, the males
and females saved were also selected for color pattern.
males and females representing each color pattern -
hooded, self, or albino - in the litter were saved if
possible. The primary interest, nevertheless, was to
save as equal a number of both sexes as possible and
still keep only eight. This reduction in litter size to
eight was done in all litters of both groups after it
was started.

During the suckling period the dam was self-fed,
and no attempt was made to prevent the young rats from
free access to the dam's feed. The rats generally began
to consume feed between sixteen and eighteen days of age.
No attempt was.made to estimate the effect of the feed on
weaning weights, but it was thought that these few days
of feed helped to prevent any setback at weaning and thus
gave a clearer indication of the growing ability.

"At 21 days of age the litters were weaned, ear-
notched, and weighed. They were placed in feeding cages
and fed as a litter. Depending on size of litter and
availability of cages, the litters were separated by
sex either at weaning or within the next 21 days to pre-
vent overcrowding and to insure virginity in the females.

.At 50 days of age they were again weighed and the gain in

16

the 29-day feeding period after weaning was used as the
basis for selection.

The ration fed the first generation was a standard
commercial laboratory animal ration. After the first
generation the rats were fed a similar ration which was
formulated and prepared at Michigan State University. The
formula for this latter ration appears in the Appendix
on page 52.

To give all rats equal opportunity to utilize the
ration and to avert possible errors in feeding, the same
ration was fed to all groups of rats ad libitum; and no
attempt was made to adjust the ration to compensate for
either age or weight of the rats.

'Weaning weights for the first four generations
were recorded to the nearest .2 gram. The 50-day weights
were recorded to the nearest gram. Between the weaning of
the fourth and fifth generations, the scales normally
used were broken; and the scales substituted for the fifth
generation weights were accurate only to the nearest .5
gram. Therefore, the weaning weights of the fifth and
succeeding generations were recorded only to the nearest
.5 gram.

Development gf'thngeterogenous Eggs
Populations‘gf‘figgs
In January 1957 two inbred hooded male rats were

each mated with two inbred albino female rats. The female

 

1?

offspring of the resulting four litters were divided into
two groups so that each litter was represented in each
group. One group of females was mated back to the hooded
males, care being taken that no female was.mated with her
sire. This group was then closed to outside matings and
was called the two-way group.

[An additional selected mating was made prior to
the start of the inbred lines. The fastest gaining male
and female of each litter were selected, and matings were
.made with the restriction that no sibs be mated together.
This additional mating provided a base population which
had a family structure. Without the additional mating of
the selected group, the relationship between sibs would
have been virtually no higher than that between non-sibs
because of the high degree of inbreeding (F-.8) in the
original lines.

The other group of females was mated to males from
a second inbred albino strain. This group was then closed
to outside matings and was called the three-way group.

Both the two-way and the three-way groups were

subsequently divided into an inbred and a selected group.

Essentially the same pattern was used in develop-
ing the inbred line of both the two-way and three-way
groups. From.the heterogenous base population, two

females and a male from each litter were selected on the

 

 

18

basis of gain during the feeding period. One female was
mated to her full brother to begin an inbred line; the
other female was used in the development of a selected
group. It was not expected that the high gaining females
in a litter would have the same rate of gain; therefore,
the selection of one for the inbred group and one for the
selected group was made at random. This random selection
was to minimize any bias which might influence later com-
parisons of the inbred group with the selected group.

The fastest gaining male in the litter produced
by the brother-sister mating was mated with his dam. The
procedure previously mentioned or.mating sons, double sons,
and triple sons back to the dam was followed in developing
the inbred line of rats. The use of the same females,
which were mated to their youngest son each generation,
allowed the inbreeding of a line to be increased as
rapidly as possible without the influence or variation
which.more than one dam might cause.

In the two-way group full sib matings to begin
the inbred lines were made among the offspring of the
generation in which the covariance coefficient between
litters was equal. At a particular season the two-way
inbred group (designated 2-I) was a generation behind the
three-way inbred group (3-1), which was begun with the
full sib matings among the offspring of the first three-

strain cross.

 

19

Development 2; Egg Selected Groups 93 Rgtg

The same males used in the brother-sister matings
to develop the inbred lines were used to develop the mass
selected groups. Each male was mated to a high gaining
female non-sib from.a litter of the same generation and
group, with the restriction that inbreeding be kept at a
minimum. In a population closed to outside breeding, an
increase in inbreeding is inevitable; thus, it was neces-
sary to impose this restriction on all matings of this
group to minimize the inbreeding. TWelve litters were
represented in this initial generation.

In the succeeding generation of the mass selected
groups, the eight fastest gaining females from.a group
representing the fastest gaining female of each litter was
chosen to be dams of the next generation. From the eight
litters thus represented, the next four fast gaining
females were chosen as dams for reserve litters in case of
reproductive failures, with the restriction that no more
than two of the 12 females would be from the same litter.

In the same manner the four high gaining males
from a group representing the fastest gaining male of
each litter were selected as sires of the next generation.
Each.male was mated to three females, one of which was
the reserve mating. The four reserve females were mated
with the additional restriction that no two litter mates

would be mated to the same male.

20

.Adjustment 2; Data for Sex

Because of a large difference between the weight
gains of the males and females in the first generation,
an adjustment was made in the data. The average differ-
ence between the male and female weight gains on a within
litter basis was found for the combined selected groups
and for the inbred group in the first generation. The
average difference of 5A grams in the selected groups
was added to the observed weight gain of each female in
the selected groups to remove differences due to sex.

The average difference of A0 grams in the inbred group
was added to the observed weight gain of each female in
the inbred groups.

An analysis of variance showed no significant
differences between these adjustment factors and actual
differences between sexes in the other generations of
inbreeding; thus, it was concluded that the factors of
5A grams for females in the selected groups and #0 grams
for females in the inbred groups were valid factors for
all generations of inbreeding. Reference for statistical
procedure was made to Snedecor (1956).

In the base population a difference of 37 grams
was found between male and female weight gains on a within
litter basis. As there was no inbreeding that generation.
and no real advantage in using an adjustment factor con-

stant for other generations, a 37 gram factor was added

 

 

21

to the observed female weight gains to equalize sex

influence.

Making Inbred Ling Crosses

When the 3-1 dams were about a year old and had
produced three litters, the third generation males were
.mated to two sisters as well as their dams. The lines
were thus increased in numbers, and possible loss of
lines through reproductive failure was averted. The addi-
tional litters provided a larger number of females for
making the line crosses.

In order to make the 2-I line crosses at the
same season as the 3-1 crosses, the 2-I males in genera-
tion 3 were mated to two sisters as well as their dams.
Generation 3 was the third generation of inbreeding in
the 3-1 group but only the second generation of inbreeding
in the 2-I group.

From among the expanded inbred lines of the
fourth generation, four lines were used in both the 2-I
and 3-I groups for making line crosses. The original
intent of the experiment was to select the four superior
lines in each group on the basis of gain for making line
crosses, but the lack of sufficient numbers in all the
lines prohibited selection for gain among the eight
original lines in each group. The four lines used in

each group were used because of the numbers in the line

22

and not because they had been selected for their superior
gains.

The fastest gaining male in each line of the
resulting fourth generation was mated to females from
each of the other lines and, where it was possible, to
a female from each litter in that line. The same pro-
cedure which was used for the 3-1 line crosses was used
for.making the line crosses among the litters of the

expanded 2-I lines.

RESULTS.AND DISCUSSION

Selection for Gain in the Mass Selected Groups

A comparison of the group generation averages for
total gain in weight from.the twenty-first to the fif-
tieth day in the two-strain and three-strain selected
groups (2-8 and 3-3) revealed similar trends for both
groups. Little increase in total gain was made in either
group until the fourth generation, at which time the
average gain in both groups increased. The increases in
average gain in the first generation of selection over
that of the base population can best be explained as a
result of a change in management procedure. It was in
this generation that litter numbers were restricted to
eight, apparently causing a.more desirable pro-weaning
Period that resulted in a more favorable post-weaning

period.
In the fifth generation the average gain of both

groups decreased to a level below the previous generation

levels. Table 1 shows the group average gain for each

generation.
Apparently several generations of selection were

needed to accumulate enough genes for gain to change gene

23

 

 

2h

.nonmoneoom nannn one an 0Hsm no
commons on» no owonops on» one 4 one .m .N donpswodow non IN non ewsnops omao

.donpsnodem nunnn on» an QHum sang
no nomnono no owonobo one one muonpsnodow noon woman Hon oHIm non ewsnowo snap

.msonw cannon anonpmuoonmu can no wunndnwon
on» nouns consensus» s anus: names pom one macaw unsnumuosp on» no monocouean

'LII
III?

 

em and can a- ems man 0
a as can can one on on can and ems m
as am and can can o~_ 4m and new ass 4
~ on men can «on mg as and man and m
om- «H can and «on c on men men men a
a «on on a and “on men . s
04H ooa ooom
on-m awnmv H-~ an-m swam H-m
means 23: Hrs HA or... 33: 332 H- Hum mA 3:288
:.m-~ m-~ swam m-m m-m mans

 

 

undonw on» essence
neocononnnv one donpsnonom nose museum can no Awasnw on. snow owonoh<_ .H mgmda

 

 

25

frequency enough to increase the mean gain. It has been
shown (Lush, 1958) that a slow rate of increase in the
.mean under selection could be due to either a low herita-
bility or a small selection differential or both.

The decrease in the average gain in the fifth
generation of selection was not expected if previous
increases had been due to an increase in frequency of
genes with additive effects. If the decrease in gain in
the fifth generation were due primarily to environmental
factors, it is probable that a large part of the increase
in the mean gain made in the previous generations was
environmental. The fact that all groups did poorly in
the fifth generation indicated an extremely poor average
environment during that time. Average environment is used
to denote the gross environmental conditions which affect
the complete population that generation in contrast to
environment which is used in the classical animal breeding
concept to denote the non-genetic factors which affect
the individual animals within the population. The group
averages were actually averages of gains incurred in two
seasons, the second season being a repeat mating of the
first.

The average number of rats weaned per litter in
the fifth generation (Table 7, page Al) decreased as did
the average gain. Some factor, possibly a viral disease,

which reduced the litter size may also have inhibited

26

growth rate or gain. Identification of the factor(s) was
not feasible. The small number of rats in the sixth
generation with essentially no increase in gain further
indicated some relationship between small litter size and
decreased growth rate.

The reach of the selected rats and generation
gains in averages which were actually obtained for the
selected groups are shown in Tables 2 and 3. The reach
ranged from 11 grams to 2A grams in the 2-8 group and from
5 grams to 20 grams in.the 3-S group. Reach was defined
as the difference between average gain of those males and
females which had offspring that lived to 50 days of age
and the average gain of all animals in the generation from
which the parents were selected.

.A comparison of the observed reaches with the
expected reaches computed by the method outlined by
Lush (19A8) revealed a rather close agreement between
the two values. Generally the reach obtained was less
than the estimated reach. The selected animals in the
base population were divided into the inbred group
foundation and the selected group foundation. This
division was taken into account,and the large differences
between the observed reach and the esthmated reach in
the base generation and in the first generation of
the 2-S group can best be explained as sampling errors.

The reason for the observed reaches in the fifth

on eadoammonomonm

m

o onoqow poo

.owsnobo nonponeoow nsonbonm nonna.omono>o.Wowwoaon headacnv moron oomMIIItllillill

.epnso finance no a one no oooaon (I11
.dOHuonpoo canvases do oonon on Adonaqononnno n lmwmwm owonoh<

 

 

27

 

as. E

 

 

 

and
. m
is a.mm Hm.- ma- s.s - n.am .mu man s
. .ma a.ma e.oa m
on n.4n mm. a a m 0 an a.mm “on m
an o.o~ oo. o m.oa a.m a.ma a m
. .mm a.ma o.am “on
on 0.5a oo. o p.4n m m a m.sm «on a
ma Hzom «5.4.. 0H 04 of: o.NH o 3.. or: warm
as «.mm o.m - o.en o.o n.5a a mm mm.
w. V.
w. mms no no r. x a
w a a . an is m in I...“ 3 w
90 .:s w u he he Te we no e I
re. s .4 t. Joe I u... o a be e
99 s U w 9 0... IL A In... NO 19 0+
9.9 a.m. u 7: use see was n.e 9.9 o .&
e.n // .u mxunv “a n.b. may T. n o
A non. no u T:L m. T. . and saw. u
rue e.u e a. .us.mu H. and T;d e rt
ens n.b. e :.h n.e s.e e.u n.e u
w m m an an a an .
u. use u. "H n. u.
sqsonw mum on endow: wunaasmon one man:
noaoaon one noaoa.non owns» newnos_on moron oobnoono was vopoeqno any .N mamga

 

 

28

.omsnobo nonpsnonow moonponq means.owsno>s nonuonosom cocoon

e maosvo

 

 

 

 

 

nofloaun one means non endow pawnos on meson cobnemno one oopoecwo sea

sqsonw mum on sansm: wnnpasnon on» non:

Ended:
1.. a
.opndo Hmanou no a one .oopooaom moauacn was no
ma
I! .donponboo onsonspm do venue on Adonpaononnno nonpooaomv nooonaoopwmmmmwunoqonn
o All
om. s.“Mflmmm owmnohd
mH.Hu cm: ‘11
ems o
om m.na m¢.au has o.oaa o.Hm Hm o.ma a.ma and m
an m.sa as. on a.ma o.~a mm o.oa m.ws Has 4
SH n.5H mm. m 0.5H o.m Nu ¢.ma m.Hm aha m
NH 0.5H be. a m.wa m.~ Hm ~.4H H.Hm mod m
«H H.H~ om.a c a.mm 0.4 cm 4.0H h.nm and H
m b.4a 0.4 o.m h a.ma N.wa cod boom
N" no nu .m as v. 11.
v v . . mm In a In a a m
A A m. n.e u.- 7:9 uxu Tue .0.» e
e 91 w r or o no In r
so as I law In A as c. .s a
en. en- U I. e e so We as We Ce 1
a.“ fium. A" m_ mmwns p. n.b. was In. m "w
e.n nu C. MAM msd flfid .qu end “um. U
“an H24 a snow" o.e see e.e 8.. .d.b
99 ”M. W 1T. 89 W 8W W U
mus m. a. Mr” m. u. "a .u

.n gamma

 

 

29

and sixth generations being larger than the estimated
reaches was also best explained as sampling errors.

The negative changes in average gain in the 2-S
group in the fifth generation and in the 3-S group in the
fifth and sixth generations indicated that average
environmental influence was so large and poor that an
increase in gene frequency for desired genes was not

expressed. Generation heritability estimates calculated

as Egggfifééghgggg ranged from numbers greater

than one to negative numbers, with an average of .22 for
the 2-8 group and .26 for the 3-8 group. Estimates of
heritability of gain for each generation, based on reach
and change in generation means, are also found in Tables
2 and 3.

The differences in total gain of the 2-S and 3-8
groups were relatively small but were consistently in
favor of the 3-S group in the early generations. In the
later generations the advantages of the 3-8 group became

smaller and even reversed.

Selection for Gain in the 3-I Group
The average gain of the 3-I group increased in the
first generation of inbreeding; however, this increase was
probably due to a large extent, if not entirely, to the
previously.mentioned change in management procedure during

this generation. Thereafter there was a general decline

 

 

30

in performance in the inbred group as a whole. The large
decline in the second generation indicates that the slight
recovery.made in the third generation was primarily an
average environmental recovery.

Inbreeding progressed at a rapid rate as shown in
Table A, and selection in the group was not sufficient to
maintain the average performance of the inbred group.
Performance of the high producing line declined in keeping
with the group average and was only slightly superior to
the average of the group. .As there was only one litter
per line in each generation from which to select, a
decline in performance was expected.

TABLE A. Average coefficient of inbreeding (F) of the
inbred litters each generation

 

 

 

Group Generation
Base 1 2 3 A
3-1 0 e25 e38 e“ e50
2-1 0 .00 .25 .38 .A5

 

There was an average decline in total gain of
2.53 grams for each 10 per cent increase in inbreeding.
The average inbreeding coefficient, which was calculated
by Wright's method (1922c), was .50 at the time line
crosses were made.

The reach expressed in the high producing line of
the 3-I group averaged 7.1 grams per generation. The

decline in gain of the high producing line averaged

 

 

31

8 grams per generation. The low producing line had an
average reach of 8.7 grams per generation, and the average
change in gain per generation was 0 grams. The level of
performance of the low line was very low and variable,
ranging from 125 grams to 156 grams. The level of per-
formance of the high performing line ranged from 155

grams to 168 grams. Table 5 shows the weight gains and
reach for the high and low inbred lines.

Selection for Gain in the 2-I Group

The average total gain of the 2-I group declined
fairly consistently each generation. The slight increase
which occurred in the fourth generation was considered to
be.more a result of average environmental nature or of
sampling error than a result of genic improvement.
.Actually the increase occurred in the third generation of
inbreeding of this group, as inbreeding in the 2-I group
began a generation later than in the 3-I group.

The average decrease in gain, as shown by the
regression of gain on inbreeding coefficient, was A.A
grams for each 10 per cent increase in inbreeding. The
average inbreeding coefficient of this group was .A5 at
the time line crosses were made.

The actual reach which was practiced in Line B of
the 2-I group averaged 7.7 grams per generation. The
line had an average net loss in performance of 10 grams.

Line B was consistently the high performing line. The

 

 

32

TABLE 5. The observed reach in weight gains for males
and females with the resulting "gain" in the high and
low lines of the inbred groups

 

 

.—

 

 

 

 

Average a
Generation Gain Reach Reach "Gain"
of Line in Males in.Females
2-I High Line (2-IB)
Base (1) 165 38 -3
2 , 185 9 0 20
3 163 l l -22
h 155 - 8

 

2-I Low Line (2-IE)

 

 

 

 

 

Base (1) 1A5 52 7
2 153 2 0 8
3 118 #1 0 “35
h ' lht 26
3-I High Line (3-IC)
Base 163 25 l
1 159 ll 0 - A
2 168 15 O 9
3 156 0 5 -12
h 155 #A - l
3-I Low Line (3-IA)
Base 137 0 2
l 156 O 0. 19
2 125 29 0 -31
3 1A8 16 5 23
A 137 -ll

 

aGain equals present generation average minus
previous generation average.
consistently low performing line (Line E) averaged 17
grams per generation for the reach, and the average per-
formance decreased only .3 gram per generation. ‘However,
the level of performance in.Line E was quite low, ranging

from a low average of 118 grams to a high of 153 grams.

 

 

33

Average performance for Line B averaged from 155 grams to
185 grams. The reach and weight gains of the high and

low lines each generation are shown in Table 5.

Comparison of 3-I with 2-I

Certain similarities in performance of gain were
noticeable between the two inbred groups. There was a
decline in performance in both groups as inbreeding
increased, and this decline existed in each line. When
compared on the basis of equal or near equal amounts of
inbreeding, the shmilarities were even more striking than
when compared on a seasonal basis. Both groups had a
slight recovery in the second generation of inbreeding.
One explanation for this recovery was that the dams had
an increased milk production or some other increased
mothering ability during their second lactations and thus
afforded their offspring a better pro-weaning environment.
In addition to this, there had to be a positive correla-
tion between the pro-weaning period and the post-weaning
period which caused a higher post-weaning gain due to
better pre-weaning environment. If, in fact, there were
factors which caused better post-weaning performance in
second litters because of better pro-weaning conditions,
there should have been an increase in the performance of
the cross-line rats from the repeat matings over that of
first matings. However, a decrease rather than an

increase in performance of the cross-line repeat matings

 

 

31+

occurred, thus tending to repudiate any tendency for

second litters to perform differently from first litters.

Comparison of 3-I Group with 3-S Group

The average gain of the 3-S group was greater
than that of the 3-I group in each generation except the
sixth, in which both groups did extremely poorly. There
was a trend for the difference between the average gain of
the two groups to become larger as inbreeding progressed.
Table 1 shows a comparison of the two groups. In the
first generation of inbreeding there was only an insig-
nificant difference of four grams in the averages of the
groups. However, an analysis of variance showed the dif-
ferences of 19, 18, and 3A grams in the second, third, and
fourth generations, respectively, to be highly significant
(P (.01).

The high producing inbred line, 3-IC, had an
average gain below the 3-1 group average in the first
generation and, consequently, a larger difference (10
grams) in the average gain from.the 3-S group. In subse-
quent generations, the 3-IC line average gain had smaller
differences from the average gain of the 3-8 group than
did the 3-I group as a whole. However, in only one
generation (the second) did the average of the high pro-
ducing inbred line equal that of the selected group.

Line crosses in the 3-I group recovered part of

the loss in productivity which occurred in the inbred

 

 

35

group. .Although there was no cross which surpassed the
selected group in performance, the inbred group as a whole
increased in performance when compared to the selected
'group. The difference in average gain between the
selected group and the line crosses was reduced to 16
grams, less than one-half the difference between the
selected group and the 3-I group the previous generation.
The difference between the average of the selected group
and the highest performing line cross (Line D 1 Line C)
was 9 grams.

In the sixth generation the average of the 3-I
group was one gram.more than the average of the 3-3 group.
.Although the performance of both groups was greatly
reduced, this insignificant difference indicated that per-
formance of the three-line crosses equalled that of the
selected group. There was only a four gram difference
between the high and low producing 3-I three-line cross.
As only three of the three-line crosses were available,
and only one litter per cross, no one cross was considered

better than the others.

Comparison of the 2-I and 2-S Groups

The average gain of the 2-8 group was larger each
generation than the average gain of the 2-I group. The
difference between the average gain in the two groups
tended to become larger as inbreeding progressed, just as

it did with the three-strain groups. Crossing of the

 

 

36

inbred lines apparently restored some of the productivity
lost during inbreeding, as the difference between the gain
averages of the 2-I and the 2-S groups in the fifth
generation was smaller than in the two previous genera-
tions.

The high.producing line of the 2-I group, Line
2-IB, exceeded the 2-S group average by 20 grams in the
first generation of 2-I inbreeding. In the third genera-
tion the average of line 2-IB was only two grams less than
the 2-S group. In the fourth generation Line B had an
average of 1A grams less than the 2-8 group.

None of the 2-I lines produced crosses which sur-
passed the average of the selected group. The average
gain of the highest producing line cross (Line C x Line B)
was 151 grams, which was only five grams less than the
average of the 2-S group. The average gain of all crosses
of Line 2-IC was only nine grams less than the average of
the selected group.

The average gain of the 2-1 three-line crosses in
the sixth generation was 2A grams less than the 2-S group,
and none of those crosses were superior to the average of
the selected group. The average of the best three-line
cross was only five grams less than the selected group,
and it was produced by mating a sire of Line A to cross-
line dome of Line B 1 Line C.

 

 

37

ZLitter Size at Birth

During the generations that inbreeding was prac-
ticed, average litter size at birth (Table 6) was larger
in the selected groups than in the inbred groups. The
difference in litter size averages between the selected
group and the inbred group ranged from a low of .3 rat
in the three-strain group to a high of .8 rat in the
two-strain groups. The differences which were non-signifi-
cant were apparently random, as no trend toward either
increased or decreased differences was observed in the

first three generations.

TABLE 6. .Average litter size at birth for each group by

 

 

 

generations
Generation 2-S 2-I 3-8 3-I
1 11.6 12.0 9.3
2 10.8 10.3 10.7 10.A
3 11.8* 9.0 9.6* 8.6
A 10.9** 8.2 10.3 8.7
5a 12.0 5.6 7.A 8.3
6 llt3 10.6 8.5 8.8

 

*Significant at P(.05
**Significant at r<.01
aData from.repeat mating only

The first significant difference (P(.05) came in
the fourth generation. It was in this generation that
inbred females were first used as dams; and, in both the
two-strain and three-strain groups, the selected group

average litter size was significantly higher than the

 

38

inbred group. Thus, inbreeding of the dam and litter
combined appeared to produce a significant decrease in
number of offspring born.

In.most of the generations only one litter per
line was represented, and for this reason no meaningful
comparison could be made of the separate lines with the
selected group.

The 2-I line crosses, as a group, had a signifi-
cantly (P .01) lower average litter size than did the 2-8
group. Data on the litter size at birth were incomplete
on the first line cross litters produced. Therefore, the
data used for litter size in the line cross litters were
only from.those litters produced from the repeat matings.
The difference of .9 rat larger litter size in the 3-I
than the 3-S group in this same generation was not sig-
nificant at Pb.05.

0n the basis of both the two-strain and the three-'
strain groups, it appeared that the single cross litters
out of inbred dams had a smaller litter size at birth
than the litters in the mass selected groups.

No significant differences were found in litter
size between either the 2-S and 2-I groups or the 3-S and
3-I groups in the sixth generation. This indicated that,
if the reductions in litter size were brought about
through inbreeding, these reductions were restored in the

three-line crosses in which outbred dams were used.

39

In both the 2-I and the 3-I groups, the only
generations in which litter size was significantly dif-
ferent from the selected group were those generations in
which inbred dams were used. The 3-I inbred dams in the
fifth generation did not produce cross-line litters sig-
nificantly different in size from the 3-S group that
generation; nevertheless, in three of the four possible
comparisons, significant differences did occur.

Although inbreeding of the litters reached its
.maximum in the fourth generation in each group, the change
from the previous generation in the inbreeding coefficient
of the litters was much less than the change in the
inbreeding coefficient of the dams. That larger change in
the inbreeding of the dams, in addition to the fact that
there were differences in litter size at birth in the only
generations in which inbred dams were used, lends support
to the conclusions that litter size is more affected by the
inbreeding of the dam than by the inbreeding of the litter
itself.

Litter Size at Weaning
It has already been indicated that litter size at
weaning was not an absolute measure of the performance of
a dam, There were two reasons for this. In the first
place, the attempt to equalize the numbers in each litter
by reducing the litters to eight eliminated some of the
variation among litters which would normally be expected.

Secondly, the dams used in developing the inbred lines

#0

were used as dams for several generations. Thus, the
variation of.maternal influence on litter size at weaning
was considerably reduced between generations. The exact
effect of maternal influence on litter size at weaning is
unknown, but it was assumed to be somewhat constant,
especially in the inbred groups, and at least to average
nearly the same each generation in each of the selected
groups.

Table 7 shows the average number of rats weaned
per litter in each generation. These averages were com-
puted from among those litters which actually had rats
born and no attempt was made to consider those females
which were exposed to the male but did not produce a
litter. The selected groups generally, although not
always significantly, had larger litters at weaning than
did the inbred group counterpart. The 3-I line cross
litters of the fifth generation were non-significantly
larger than the 3-S litters. The 2-S litters, in contrast,
were significantly (P(.05) larger than the 2-I line cross
litters.

In the line cross litters of the 3-I group a
recovery of the loss of litter size at weaning was made
when compared with the selected group. It was recognized
that litter size at weaning was not used as a criteria
for selection; and thus, unless there was genetic correla-
tion between litter size at weaning and post-weaning

gains, the primary cause of any difference between the

#1

groups would be produced by mothering ability of the dams.
It appeareimore likely that sampling errors rather than
poor mothering ability made the 3-S litter size small in
the fifth generation.

TABLE 7. Average litter size at weaning for each group
by generations

——__.

 

 

Generation 2-S 2-I 3-S 3-I
1 7.87 7.75 6.83
2 7.63 7.75 7.73 7.00
3 8.00 6.63 7.52 6.29
1. 6.91 6.11 6.90* 5.60
51) 6.1.7'“ A.66 4.09 5.77
a

8Data on generation 6 not used because of small
number of litters weaned

bAverage of first and second (repeat) matings
*Signigicant at P<.05

 

A.reduction in litter size at weaning in the
inbred groups could have occurred as a result of reduced
litter size at birth or from.reduced livability from birth
to weaning or a combination of those factors. Because
litter size was restricted to a maximum of eight rats, an
analysis was made of the differences between the number
allowed to live and the number actually weaned. That dif-
ference was taken as the best estimate of pro-weaning
death loss or, the reverse, livability to weaning.

Table 8 shows the average litter death loss.
The analysis of litter death loss showed no signi-

ficant difference between either the 2-S and 2-I groups

#2

or the 3-3 and 3-I groups. Thus the significant dif-
ferences found in litter size at weaning between the
inbred and selected groups were apparently the result

of the differences in litter size at birth.

TABLE 8. Average litter death loss of the groups by

 

 

 

generationsa
Generation 2-S 2-I 3-S 3-I
1 .50
2 .13 .25 .09 .70
3 .50 .18 .57
A 1.00 .83 1.00 1.30
5b 1.60 2.00 4.20 2.10

 

aCalculated as the difference between the number
of rats allowed to live and the number weaned

bBased on repeat mating only
Variance of Weight Gains _a__s_ a_ Measure
‘72; Per ormance

Variance Components of weight Gains
Between Litters

The variance components of weight gains between
litters are shown in Table 9. The components for the
inbred group in generations 1, 2, and 3 were actually
components between lines as there was only one litter per
line in those generations. The variance component
between lines was expected to be larger than that between
litters of the selected group the same generation. In
one generation of the 2-I group and in two of three
generations of the 3-I group, the between litter (line)
components were larger that the between litter components

of the respective selected groups.

1+3

TABLE 9. variance components of weight gains between
litters of the group by generations

 

 

 

Generation 2-3 2-1 3-S 3-I
‘1 95 80 1588
3 111. 115El 1,9 26a
1" 71+ 27b 93 59b
5“ A 2 9A 131
5B #72 121 1.1 110
6 “9 #7 O o

 

aBetween lines (one litter to each line)
bBetween litters within lines
9A negative component was calculated.

The between litter within line variance compo-
nents of the fourth generation were smaller in both the
2-I and the 3-I groups than the between litter compo-
nents in the selected groups, which was expected, As
inbreeding in a line increases, genetic differences
between animals in that line decrease. Thus, the variance
between individuals is reduced. The variance between
litters within that line is also reduced, as the differ-
ences between non-sibs tend to become no greater than
differences between sibs.

The variance components between litters in the
fifth generation of the two-strain group were not consis-

tent with those of the three-strain group. The variance

component of both matings of the 2-I group was smaller

than the 2-8 group; whereas, the component of both matings

in the 3-1 group was larger than that of the 3-8 group.

A4

The very small components in both the 2-S and 2-I groups
the first mating and the very large components of the
2-S group in the repeat mating indicate that environment
or sampling errors, or both, had a large influence.

The mean square between litters in both the 3-3
and 3-I groups the sixth generation was smaller than the
within litter mean square; therefore, the estimate of
the variance component would be negative. .A negative
variance component is impossible; thus, no valid estimate
was available.

Examples of analysis of variance tables used in
calculating the variance components are shown in Table 10.
Tables similar to those in Table 10 were made for each
group in each generation. For sake of brevity only three
examples are shown, but the results from all the 27 tables
used are shown in Table 9. The analysis of variance
tables shown are of the 2-I group in generations 3, A, and

5 and represent each of the types of component estimates.

Variance of Weight Gains Within Litters

.A trend was noted in the 2-1 group for the var-
iance of gain within litters to be reduced as inbreeding
increased, as shown in Table 11. In the 3-I group this
trend was not noted; instead, both increases and decreases
in the variance occurred without much change. The var-

iance in the fourth generation, in which inbreeding was a

maximum, was about the same in both the 2-I and the 3-I

TABLE 10.

of the 2-I group in generations 3, A, and 5
.making variance component estimates

1+5

Generation 3

Analysis of variance tables for weight gains
A used for

a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source Degrees Mean Expected
of Variation of Freedom. Square Mean Square
Between lines 7 91,1, 02 + 6,), (3’;
Within lines A3 213 (f 2
G‘ -213 VB-ll5

Generation A

Source Degrees Mean Expected
of Variation of Freedom Square Mean Square
Between lines A 1058 0’2 + 6 0'12; +110:
Between litters 2 2

within lines 5 318 0' + 6 UL
Within litters 50 157 62
*2. " 2. 2 2 68

(r 157 (IL 27 GB #
Generation 5A

Source Degrees Mean Expected

of Variation of Freedom Square Mean Square_
2 2

Between litters 13 320 U f 5’8 0’13,
2

Within litters 67 308 U

A A

02-308 egg-2 _

 

aEstimateg variance component for between litters

within lines -

L
Estimated variance component for between

A2
lines -- GB

Estimated variance component for within litters -

 

 

 

#6

group. The variance of the first inbred 2-I group was
twice that of the first 3-I group. '

TABLE 11. Variance of weight gains within litters of the
7 groups each generation

W

 

 

Generation 2-S 2-I 3-S 3-I

Base 367 156 H
1 379 193 150
2 18h 318 11.3 179 '
3 166 213 138 138 _
A 150 157 117 155 :
5A 352 308 950 197
5B 288 180 161 238 I
6 3A7 122 521 1A0 l

 

.A reduction in the within litter variance occurred
in both the 2-S and 3-8 groups in the first four genera-
-tions. The variance in the fourth generation.was similar
for all groups, and differences in reduction were from
differences in the early generations rather than the later
generations. Differences between the selected and the
inbred groups were the largest in the sixth generation in
which the three-line crosses were produced.

The variances within litters indicated that

breeders should not expect to get completely uniform

litters with regard to growth rate either by inbreeding

or by selection.

 

 

SUMMARY AND CONCLUSIONS

The heterozygous populations of the laboratory rat
were developed from.the crosses of two highly (F=.8)
inbred lines. The backcross to one of the lines produced
the population called the two-strain group. The other
population was developed by crossing the offspring of the
original line crosses with a third inbred line; thus, it
was called the three-strain group.

Both a selected and an inbred group were formed
from each of the base populations. Selection was made on
the basis of weight gains in a post-weaning feeding period
from the twenty-first to fiftieth day of age. Inbred
lines were formed by mating sons, double sons, and triple
sons, back to their outbred dams. Weight gain of inbred
.males were found to average A0 grams more than that of
inbred females. In the selected groups males were found
to average 5A grams more in weight than the females. The
differences were based on average differences between the
sexes within litters.

.A comparison of the weight gains in the selected
groups and the inbred groups indicated that as inbreeding
increased the difference in average weight gain between

the groups increased from A grams to BA grams in the

A7

48

three-strain group and from 12 grams to 21 grams in the
two-strain group. Differences in the high producing lines
of each group also increased but to a lesser degree.
Crosses between the inbred lines restored part of
the loss in weight gains which was incurred during
inbreeding; however, no hybrid vigor in excess of the
selected group was observed in either the 2-I or 3-I
line crosses, although there was less difference between

the average gain of the selected group and the line crosses

 

than between the selected groups and the inbred groups the
previous generations.

A regression of average weight gain on the inbreed-
ing coefficient indicated a reduction of A.A grams for
each 10 per cent of inbreeding in the 2-I group and a loss
of 2.5 grams for each 10 per cent of inbreeding in the
3-I group.

Average litter size at birth was significantly
higher in both the 2-8 and 3-8 groups in the fourth genera-
tion than in the corresponding inbred group and higher in
the 2-3 group the fifth generation also. As inbred dams
were used in both of these generations, it was concluded
that inbreeding of the dam.had more influence on the size
of the litter at birth than did the inbreeding of the
litter itself.

Litter size at weaning was significantly larger
in the 3-8 group than in the 3-I group in the fourth

generation” and it was also larger in the 2-8 group than

A9

in the 2-I group in the fifth generation. An analysis of
the average litter death loss showed no difference in
birth to weaning livability between the selected and
inbred groups. Thus, it was concluded that the differ-
ences in litter size at weaning were due primarily to
differences in litter size at birth.

.Estimates of the variance components for weight
gains between litters within inbred lines were smaller
than the between litter estimates for the selected groups.

Variance of weight gain within inbred litters in
the 2-I group tended to be reduced as inbreeding pro-
gressed. The reduction was from a variance which was
larger than the selected group in the early generations
to a variance which was apparently equal to the selected
group in the fourth generation.

The amount of heterozygosity in the base popula-
tion apparently had no effect on selection in either the
selected or the inbred groups. Average weight gains,
litter size at weaning and at birth, and livability from
birth to weaning were similar for both selected groups and
for both inbred groups including the line crosses.

One of the major drawbacks to the development of
superior inbred lines by the son-to-mether matings system
used in this experiment was the reduced prolificacy of
the inbred dams. Of course it was understood gwpgigg;
that numbers within the lines would be restricted by the

use of a single dam. Reduced prolificacy, in addition to

50

the already restricted size of the inbred lines, increased
the necessity of developing more lines if this system
were to be used to its best advantage.

Another drawback to the development of inbred
lines by a strict method of son-to-dam.matings, equally
as important as reduced prolificacy, was the small number
of animals from.which to select for maintenance of the

individual lines and for making line crosses. Expansion :

 

‘I‘.in

of the lines by brother-sister matings earlier in the

‘.F..*b

development of the lines, perhaps even as early as the
first generation of inbreeding, along with continued son-
dam.matings would have probably allowed sufficient num-
bers to maintain all of the proposed eight lines in each
group as well as supply sufficient numbers of females for
more complete testing of the possible line crosses. Early
expansion of the lines would slow the rate of the increase
in inbreeding, but selection cannot be effective if there
are not sufficient numbers from.which to select.
Traditionally the cost of producing inbred lines
has been high, and specific or general combining ability
had to be found in the inbred lines if inbreeding were to
be practical. Thus, it was postulated that this system
could be used in swine breeding with the recommendations
that many lines he formed and these lines be expanded
early to assure a large number of lines from which to
select the superior lines and to assure a large number of

animals within the lines for testing the line crosses.

 

APPENDICES

52

APPENDIX.A. Formula for rat feed prepared at Michigan

State University

 

Ingredient

 

Ground shelled corn . . . . .

Sucrose . . . . . . . . . . .

Soybean oil meal (AA% protein).

Fish meal (58% protein) . . . . .

Alfalfa meal (17% protein, dehydrated).

Dried skim.milk . . . . . . .
Corn oil . . . . . . . . . .
Super trace mineral salt a .
B vitamin supplement b. . . .

Vitamin.A and D concentrate .

Vitamin B12 supplement (Pfizer's 9+) d

Pounds
e e e e 960
. . . . 100 1

e e e e 400
e e e e 200
e e e e 100

 

O O O O 200
O O O O 60
O O O 0 lo

 

aTrace mineral content (per cent):

0 O C O 2

e e e e l

e e e e 05
N801. 97;

I, .007; Fe, .33; Cu, .0A8; Mhm .AO; Co, .022; Zn, .500.

bRiboflavin, 2000 mg. per pound; pantothenic acid,
#000 mg. per pound; niacin, 9000 mg. per pound; choline
(choline chloride equivalent), 10,000 mg. per pound.

cContains A,A50,A80 I.U. vitamin.A; 1,261,074 I.U.

vitamin D per pound.

dPfizer supplement No. 9, containing 9 mg. B12

per pound.

53

 

 

 

 

 

on o 8 Iain .841 m 2 Tom

sen w on a-ma «on w an m-~H

sea 8 Ha msma «an m as puma

one m an «-ma and m on sue

one m as «non sea m ea m-m men w ma mums
owd o o NIHN 05H m CH Nlm m4H m 4H Nlmd
men m m «.8 own m as puss “on s u «use
mmd w NH HIQH 40H m 4H :35 mmH m Ha Alum
man m we Hum can m m sued mad 8 ma m-s~
and m an m-sn use m on m-oa men m an ~-o~
one m m mus can m on «use men m an snnm
one o s nuns sen m m H-H~ 45H m on mwmm
.made common. anon eon anew douse; duom son snow senses. duom son

noassz noaasz cease: nonssz noaasz noaasz
Ham mum mum

 

 

.a donuonoaow on mason» on» no snow noppna owonopo one cane noppnq

.mHHHQZEBmd

 

 

 

 

 

 

 

 

mod m Ha Has: A
son 4 an mama sea m 4H H-4m
mma m ma mnma Aha o o «no:
can u 5 «son «on w an mnmm and a ma samm son m as sum“
and m m muam use m w a-mn «mg m an «-58 can A 8 much
mod o o «10 00H m Ha ~1Am mod 5 as «non ooa w an Alum
.» one m on a-ma «on m on «-44 and m on same Hen m on a-mm
a, and m an n-m men m ma sum: «as m on n-8m men m an a-mh
men s ma n-4n and m an a-mm are m a m-~o man 8 ma a-mh
one m h mus can w o Huam “an m on mums and m ma sums
“NH m «a n-aa men s ma 4-4m «an m m a-mm ass m on 4.58
INHN IN N MN IN N
mew: mew: 2mm? :2“;
pI I pI I “HI I I I
1111. Hum mum Hum mum

 

 

m nonponooow on unsoaw on» no snow ascend omsnohs was owns Havana .o HHQZHth

55

 

 

 

 

mud w (ea canon

and o w «145 Ned w NH anmoa

45H m NH Alum mad w a mlhoa

and m m much mud w ma muaoa

and m 0H mlmo and 0 ma Slam bod w ma mlmoa

00H m m Nudw mod w NH mums bmd w ad «the 00H w «A mined

and m o «no bud w a dumb mad m w «lam Aha w ma Humoa

m4H m ma anon and m on duos mad 4 m nine and w NH auaoa

bad m 0 Hum med m HH naom odd m «a Hahn Hod m ma m140d

and 4 s mind mod m 0H Alec mad 5 AH mime mod m m Alsoa

men n 4 man own w m mush mod n m mnmm mod m ma mumoa

med 5 0H HIHH «on m AH Nuow 44a 5 b dawn and w Ha Humoa
MN N KN N IN N IN N

momma :mmmm :2: :2:
pt. I PI I D...» I D... I
Ham mum Ham mum

 

 

 

 

 

m nonponoqom on meson» one no snow second ewonopo one owns noppnn .Q Hanmmm4

56

 

 

 

 

 

 

 

 

 

and m m Humma
end d Ha simmd
cod m o Hawma sad m m sumo
HNN N m Him 00H m ma HImNH find 0 m HIOm
NmH m HA MINNH Nod m ma NIQNH and b b HIH4H ova w Ha m:#4d
HOH m m HINNH mud w ma MIMNH and m 5 mlNo 04a b 4H Hanna
and m 0H m14H sea m m mIJNH find 0 Ha .AIOJH mud b m Janna
and w m HIHNH 05H m NH Juana mmH m ma Nlosa Fwd m o mléna
4nd w a nun Nod m 0H mnomd omH m Ha Minn 00H m a 4|N4H
mNH h o MIONH Hod o b NINNH méd o h Janna Hod m 0H le4H
OSH m o NIONH 4nd 0 m mINmH find 5 b mined mod N Ha Juana
MMH m m HIHH mud m 0H mIoNH mNH m 4 Huwm and m ma Jlozd
IN N MN N IN N MN N
:mmm m Wmmmmm :mmm m :mmmm
P... I P... 1 PI 1 PI 1
Him 11! min HIN MIN

 

 

4 nonponcaow an mason» one no snow Rowena owonohs one owns Havana .HUHanmmmd

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A _ and a sumo: _ A
bNH o Nnfloa 04d m ouNos NJH 4 mafia: m

00H m NIHbN add u Nnnmm
and 4 alwa mNH m HIHSN and b NINmm o

oNH w nunms

3..” m Team

on." m suwmm

7 «.3 m Tame 3a a Team
5 one N «#3 «on m Team and a. nun? m
04H m minnm and o Nana4 .4
undo dense: awn anew condo: and 38 doses: and 5am dense: Imam} sea
1 nope—dz nee—:2 Shana noes—oz no
a o m 4 can.”

sham no scan _

 

 

on? no on: on m consensus»

on someone on: H1N on» no flow omens: one wands: as on: nepanq .m NEE

58

 

 

 

 

 

 

 

 

 

 

 

 

44H m N-Nmm ; ; n
and m munw
Ana 5 mlflba
Hon m «-una Hon m «tum “on p m-n~
man m ~-uua «ma m a-npa a4a m «nupa o
nan m «-maa
can 4 m-~m m
and a Human
4nd 0 Huxma find a H.Nw .4
mwoa oodomz__adn anon voqook__a¢n once dodoos. adn gnaw condos. _adn adn
nonaoz nonadz nonaoz nooaoz no
o o m .4 oqnq
onnm no coda

 

 

an mommono odna Hum

onnm no odna kn m donnonodow
on» no gnaw owonoba can wandooz no ounm Havana

.¢_Hanmmm4

59

APPENDIX E. Litter size at. weaning and average litter

gain or the selected grou

pa in generation 5

 

 

 

 

 

2-S 3-8
Number Number
Dam weaned Gain Dam 'Weaned Gain
311-3 8 168 l8x-h 8 lhl
321-3 8 161 161-h 7 lb?
281-2 A 161 hx-l 5 162
251-1 8 170 61-1 6 135
29x-l 3 159 7192 8 157
361-2 8 1&2 12x-2 3 189
301-3 8 157 61-1 5 110
2Lx-l 8 161 11-2 1 155
251-h 7 159 91-h h 161
le-h 8 1&5 161-h 2 156
281—3 6 162
311-3 6 177
25x-l 8 175
361-2 h 13h
2hx-1 7 137
28x-2 7 17h
301-3 6 13h
321-3 7 158
291-1 h 113 ##_

 

 

 

APPENDIX 1.

6O

Litter size at weaning and average litter

gain of the three-line crosses in generation 6 by line

of sire

2-I

M

Line of Sire

 

 

 

 

 

 

 

 

 

 

 

Line A D
Cross Number Number
or Dam. Dam weaned Gain Dam weaned Gain
B x 0 951-1 8 150

921-3 7 153
C x D 911-2 8 135

661-2 8 111
‘A x B 94x-h 8 1A1
‘A x C 931-1 7 133

3-I
Line of Sire
C D

Line
Cr Number Number
ofogzm Dam Weaned Gain Dam ‘Weaned Gain
‘A x C 6Ax-2 6 123
C 1 D 681-1 8 126
B x 0 581-3 3 127

 

 

 

NJ I! - '«

APPENDIX J .

61

Litter size at weaning and average litter

gain or the selected groups in generation 6

 

 

 

 

 

2-S 3-S
Dam figgfizg Gain Dam gggggg Gain
671-1 8 155 571-3 6 118
781-4 h 156 631-1 3 113
711-6 6 181 591-2 3 166

 

 

LICIERATURE CITED

 

LITERATURE CITED

Chai, C. K. 1956. Comparison of two inbred strains of
.mice and their F hybrid in response to androgen.
.Anat. Rec. 126:269.

Chambers, D. and J. A. Whatley, Jr. 1951. Heterosis in ;?§
crosses of inbred lines of Duroc swine. J. Animal '
Sci. 10:505.

Comstock, R. E. and L. M. Winters. 1966. A comparison .
of the effects of inbreeding and selection on 4
performance in swine. J. Animal Sci. 3:380.

 

 

 

Craft, W. A. 1963. Swine breeding research at the
regional swine breeding laboratory. U.S.D.A.
Misc. Pub. No. 523.

Craft, W. A. 1953. Results of swine breeding research.
U.S.D.A. Circular No. 916.

Craft, w. A., A. B. Chapman, 0. F. Sierk, L. M. Winters,
G. E. Dickerson, C. E. Terrill and J. L. Lush.
1951. Effectiveness of selection. J. Animal Sci.
10:3.

Craig, J. V. and A. B. Chapman. 1953. Experimental test
of predictions of inbred line performances in
crosses. J. Animal Sci. 12:12h.

Dickerson, G. E., J. L. Lush and C. C. Culbertson. 1916.
Hybrid vigor in single crosses between inbred
lings of Poland China swine. J. Animal Sci.

5:1 e .

Eaton, 0. N. 1932. (A quarter-century of inbreeding in
guinea pigs. J. Expt. Zool. 63:261.

Garwood, V. A. 1956. .A comparison of inbred lines, two-
line crosses, three-line crosses, rotational
line crosses, and outbreds within the Duroc breed
of swine. Ph. D. thesis. Univ. of Nebr.

63

6h

Godbey, E. G. and L. V. Starkey. 1932. A genetic study
of the effects of intensively inbreeding Berkshire
swine. So. Car. Agr. Exp. Sta. Ann. Rept. (forty-
fifth).

Greenwood, A. W. and J. S. S. Blyth. 1951. A repeated
cross between inbred lines of poultry. J. Agr.
Sci. 41:367.

Hetzer,-H. D., W. V. Lambert and J. H. Zeller. l9h0.
Influence of inbreeding and other factors on
litter size in Chester White swine. U.S.D.A.
Circular No. 570.

Jones, D. F. 1918. The effects of inbreeding and cross-
breeding upon development. Conn. Agr. Exp. Sta.
Bul. No. 207.

.King, H. D. 1918. Studies on inbreeding. I. The effects
in inbreeding on the growth and variability in the
bgdy weight of the albino rat. J. Expt. Zool.

2 :1. ‘

King, H. D. 1919. Studies on inbreeding. IV. A fur-
ther study of the effects of inbreeding on the
growth and variability in the body weight of the
albino rat. J. Expt. 2001. 29:71.

Lush, J. L. 1948. The genetics of populations. Mimeo-
graph. Ames, Iowa.

Lush, J. L. 1958. Animal Breeding Plans. The Iowa State
College Press, Ames,Iowa.

McPhee, H. 0., E. Z. Russel and J. Zeller. 1931. An
inbreeding experiment with Poland China swine.
J. Heredity 22:393.

Nelson, R. H. and J. Lh.Lush. 1950. The effects of
mild inbreeding on a herd of Holstein-Friesian
cattle. J. Dairy Sci. 33:186.

Regan, W. M., S. W. Mead and P. W. Gregory. l9h7. The
relation of inbreeding to calf mortality. Growth
11:101e

Shultz, F. T. 1953. Concurrent inbreeding and selection
in the domestic fowl. Heredity 7:1.

Sierk, C. F. 1948. A.study of heterosis in swine.
J. Animal Sci. 7:515 (Abstract).

65

Sierk, C. F. and L. M. Winters. 1951. A study of
heterosis in swine. J. Animal Sci. 10:10h.

Snedecor, G. W. 1956. Statistical Methods. Iowa State
College Press, Ames, Iowa.

Stewart, H. A. 19A5. An appraisal of factors affecting
prolificacy in swine. J. Animal Sci. #:250.

Tantawy,.A. 0. 1956. Selection for long and short wing
length in Drosophila.me1anqgaster with different
systems of mating. Genetica 28:231. 5%

Tantawy, A. 0. 1957. Genetic variance of random inbred
lines of Drosophila melanogaster in relation to
coefficients of inbreeding. Genetics A2:12l.

— ~34..- .

Tyler, W. J., A. B. Chapman and G. E. Dickerson. l9h7.
Sources of variation in the birth weight of l
Holstein-Friesian calves. J. Dairy Sci. 30:h83. =

 

Warwick, E. J. and J. R. Wiley. 1950. Progress in inbred
lines of swine and their uses in crosses. Purdue
Agr. Exp. Sta. Bul. No. 552.

Willham, 0. S. and W. A. Craft. 1939. An experimental
study of inbreeding and outbreeding swine. Okla.
Agr. Exp. Sta. Tech. Bul. No. 7.

Winters, L. M., P. S. Jordan, R. E. Hodgson, 0. M. Kiser
and W. W. Green. 19AA. Preliminary report on
crossing of inbred lines of swine. J. Animal Sci.
3:371.

Wright, S. 1922a. The effects of inbreeding and cross-
breeding on guinea pigs. I. Decline in vigor.
II. Differentiation among inbred families.
U.S.D.A. Bul. No. 1090.

Wright, S. 1922b. The effects of inbreeding and cross-
breeding on guinea pigs. III. Crosses between
highly inbred families. U.S.D.A. Bul. No. 1121.

Wright, S. 1922c» Coefficients of inbreeding and
relationship. Am. Nat. 56:330.

an“:
A? USE as: Y