• TABLE OF CONTENTS
HIDE
 Title Page
 Acknowledgement
 Table of Contents
 List of Tables
 Abstract
 Introduction
 Review of literature
 Materials and methods
 Results and discussion
 Summary and conclusions
 Bibliography
 Biographical sketch














Title: Selection for 42-day weight in mice on high and low fiber diets
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098178/00001
 Material Information
Title: Selection for 42-day weight in mice on high and low fiber diets
Physical Description: vii, 100 leaves : ; 28 cm.
Language: English
Creator: Rodriguez, Rafael Efren, 1939-
Copyright Date: 1974
 Subjects
Subject: Breeding   ( lcsh )
Animal Science thesis Ph. D
Dissertations, Academic -- Animal Science -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis (Ph. D.)--University of Florida, 1974.
Bibliography: Bibliography: leaves 93-98.
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Rafael Efren Rodriguez.
 Record Information
Bibliographic ID: UF00098178
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000401626
oclc - 24683425
notis - ACE7478

Downloads

This item has the following downloads:

PDF ( 2 MBs ) ( PDF )


Table of Contents
    Title Page
        Page i
        Page i-a
    Acknowledgement
        Page ii
    Table of Contents
        Page iii
        Page iv
    List of Tables
        Page v
        Page vi
        Page vii
    Abstract
        Page viii
        Page ix
        Page x
    Introduction
        Page 1
        Page 2
        Page 3
        Page 4
    Review of literature
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
    Materials and methods
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
    Results and discussion
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
    Summary and conclusions
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
    Bibliography
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
    Biographical sketch
        Page 99
        Page 100
        Page 101
        Page 102
Full Text

















SELECTION FOR 42-DAY WEIGHT IN MICE ON
HIGH AND LOW FIBER DIETS












By

RAFAEL EFREN RODRIGUEZ




















A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1974















ACKNOWLEDGMENTS


The author wishes to express his sincere apprecia-

tion to Dr. Marvin Koger, Animal Geneticist and Professor,

for his many hours spent in advising and supervising this

student. His personal friendship and support are most ap-

preciated.

To Dr. Walter Harvey, Biometrician with the Depart-

ment of Dairy Science, Ohio State University, the author

expresses his appreciation for his invaluable and most

generous advice. His patience and understanding of the

problems of this type of research made this endeavor

feasible.

To Dr. Robert Arrington, Professor, the author

extends his thanks for his unselfish help in procuring the

physical plant for this project.

The assistance of Dr. D.E. Franke, Associate

Professor, is acknowledged with sincere thanks. Also,

sincere thanks are extended to the other two members of

my supervisory committee, Dr. Alvin Warnick and C.J.

Wilcox for their help.

The author extends his deepest gratitude to his

wife and family who gave their support and understanding

during the course of this study.


-1















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS .................. ................. ... ii

LIST OF TABLES .......... .......................... v

ABSTRACT ........................................... iii

INTRODUCTION ......................................... 1

REVIEW OF LITERATURE ............................... 5

General Review ............................... 5
Results From Experiments ..................... 6
Weaning Weights ....................... ......... 8
42-Day Weights ......... ....................... 11
Fertility ...................................... 13
Litter Size .................................... 14
Literature Summary ............................ 16

MATERIALS AND METHODS .............................. 18

Collection of Data .................. ........... 18
Description of Data .......................... 21
Analyses of Data ......... ..................... 23

Pregnancy Rate .......... ................ 23
Litter Traits ......................... 23
Individual Weights ..................... 24
Random Variances ....................... 24

Genetic Parameters ........................... 27

Variance Components .................... 27
Estimates of h2 ...................... .. 29
Causal Components of Variance ......... 29
Realized Genetic Correlations ......... 31

RESULTS AND DISCUSSION ............................. 33

Pregnancy Rate ................................. 35
Litter Size at Birth ......................... 35
Litter Size at Weaning ....................... 41
Total Litter Weight at 42 Days of Age ........ 48
Weaning Weight (21-Day Weight) ............... 52



iii










Page

Individual 42-Day Weight ................... 57
60-Day Weight .............. ................. 63
Genetic Parameters ........................ 66

Realized Heritabilities for 42-Day Weight 67
Selection Intensities ................ 69
Relation to Natural Selection ......... 69
Observational and Causal Components of
Random Variances ................... 71
Heritability Estimates ................ 75
Realized Genetic Correlations ......... 77

SUMMARY AND CONCLUSIONS ........................... 87

Analyses of Data From Four Subgroups,
Generations 4 to 8 ....................... 87

Pregnancy Rate ........................ 87
Litter Size at Birth ................. 87
Survival and Litter Size at Weaning ... 88
Total Litter Weight at 42 Days of Age 88
Individual Weaning Weights ........... 89
Individual 42-Day Weight .............. 89
Individual 60-Day Weight .............. 90

Genetic Parameters ......... ................ 90

Growth and Maternal Components in Weight
Data ............................... .. 90
Realized Heritabilities ............... 91

LITERATURE CITED ............ .................. .... 93

BIOGRAPHICAL SKETCH .............................. 99


~ 1














LIST OF TABLES


Table Page

1 Mating design employed in developing the two
lines and in subsequent testing of the
two lines on the opposite diets ......... 19

2 Effects fitted, degrees of freedom and their
expected mean squares for model 1 employed
in analyzing pregnancy rate ............. 22

3 Effects fitted, degrees of freedom and their
expected mean squares for model 2 employed
in analyzing litter traits .............. 25

4 Effects fitted, degrees of freedom and their
expected mean squares for model 3 employed
in analyzing individual weight data ..... 26

5 Effects fitted, degrees of freedom and their
expected mean squares for model 4 employed
in analyzing individual weight data within
line and diet ........................... 28

6 Averages and standard deviations for
dependent variables ..................... 34

7 Least squares analysis of variance for
pregnancy rate, model 1 ................ 36

8 Least squares means and standard errors for
pregnancy rate (%) by diet, line and
generation, model 1 ...................... 37

9 Least squares analysis of variance for litter
size at birth, model 2 ................. 39

10 Least squares means and standard errors for
litter size at birth by diet, line and
generations, model 2 ................... 40

11 Least squares analysis of variance for
litter size at weaning, model 2 ......... 42

12 Least squares means and standard errors for
litter size at weaning by diet, line and
generation, model 2 ..................... 43











13 Number of observations and average length
in centimeters from anus to tip of snout
by line, sex and generation ............ 45

14 Least squares means for survival to weaning
(%) by diet, line and generation ....... 46

15 Least squares analysis of variance for total
litter weight at 42-days of age, model 2 49

16 Least squares means and standard errors in
grams for total litter weight at 42-days
of age by line, diet and generation,
model 2 ................................ 50

17 Least squares analysis of variance for
weaning weight, model 3 ............... 53

18 Least squares means and standard errors in
grams for weaning weight by line, diet
and generations, model 3 ............... 54

19 Least squares means and standard errors in
grams for individual weight parameters
by sex, line and diet, model 3 ......... 58

20 Least squares analysis of variance for
individual 42-day weight, model 3 ...... 59

21 Least squares means and standard errors in
grams for individual 42-day weight by
line, diet and generations, model 3 .... 60

22 Least squares analysis of variance for 60-
day weight, model 3 .................... 64

23 Least squares means and standard errors in
grams for 60-day weights by line, diet
and generation, model 3 ............... 65

24 Least squares means in grams of individual
weight parameters by line and generation,
model 4 ................................ 68

25 Selection differentials and intensities for
line 1 and 2 by sex .................. .. 70

26 Estimates of observational components of
variance for weaning weight, 42-day
weight and 60-day weight, model 4 ...... 72


Table


Page












27 Estimates of causal components of variance
for weaning weight, 42-day weight and
60-day weight ........................ 73

28 Within-line heritability estimates for
weaning weight, 42-day weight, and 60-
day weight, model 4 .................... 76

29 Realized genetic correlations (rA) between
weight parameters, standard errors and
the genetic estimates needed for their
computation ............................. 79

30 Within line correlations between weights
at different ages .................... 80

31 Least squares analysis of variance for
weaning weight in line 1, model 4 ..... 81

32 Least squares analysis of variance for 42-
day weight in line 1, model 4 ......... 82

33 Least squares analysis of variance for 60-
day weight in line 1, model 4 ......... 83

34 Least squares analysis of variance for
weaning weight in line 2, model 4 ..... 84

35 Least squares analysis of variance for 42-
day weight in line 2, model 4 ......... 85

36 Least squares analysis of variance for 60-
day weight in line 2, model 4 ......... 86


Table


Page















ABSTRACT OF DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY




SELECTION FOR 42-DAY WEIGHT IN MICE ON
HIGH AND LOW FIBER DIETS

By

Rafael Efren Rodriguez

December, 1974


Chairman: Marvin Koger
Major Department: Animal Science


Direct and correlated responses were determined in

two lines of mice (Ll and L2) selected for increased 42-

day weight on two diets; Ll on Dl, a commercial laborato-

ry ration (PLC) and L2 on D2, a ration consisting of 70%

PLC and 30% nonnutritive bulk. After 3 generations eight

females from each of the original lines were transferred

to the opposite diet forming four subgroups. For the next

5 generations sires were selected only from original

lines, each being bred to females from his line on both

diets. Traits measured were individual weights at weaning,

42 and 60 days; litter size at birth and at weaning and

total litter weight at 42 days.









L1 females were heavier on both diets and pupped

larger litters due to a positive effect of maternal weight

on litter size. L2 surpassed L1 for weaning weight, lit-

ter size at weaning, and litter weight at 42 days. L1 in-

dividuals were heavier at 42 and 60 days on both diets.

Diet x line interaction approached significance (P < .075)

for litter weight at 42 days as a result of a reversal in

line rank with L2 surpassing L1 in the later generations.

Diet effects were significant for all traits. Gener-

ation effects were significant for individual 42- and 60-

day weights with means for both weights increasing in re-

sponse to selection for 42-day weight. Diet x generation

effects were significant for litter size at birth, litter

weight at 42 days and individual weight at all ages. These

interactions were explained by generation means increasing

in D1 while D2 means decreased or remained unchanged.

There was a significant negative regression of individual

weight on number weaned for weaning weight and for 42-day

weight. Sex differences were significant for all weights,

with males being generally heavier than females. A signif-

icant sex x line effect for 60-day weights resulted from a

larger difference between sexes in LI. All weights showed

significant sex x diet interactions caused by smaller sex

differences in D2 than in Dl. Diet 2 did not allow males

to express their superior growth potential.

Genetic parameters for individual weights were esti-

mated from 8 generations of data from L1 Dl and L2 D2









Sire effects were significant only for 42- and 60-day

weights in LI. Dam effects were significant for all weights

in both lines. All weights had a larger additive genetic

variance for growth (VAo) in Ll and a larger within-litter

variance (VWL) in L2. Maternal effect (VMe) accounted for

most of the phenotypic variance (Vp) for all weights in L2.

Heritability estimates from L1 and L2 data were for weaning

weight .21 and .13; for 42-day weight .34 and -.17 and for

60-day weight .48 and .04. Cumulative selection differen-

tials and responses in grams per generation were 2.71 and

.87, respectively, for Ll versus 2.40 and .19 for L2.

Realized heritabilities and selection intensities were .33

and 1.35 for L1 compared with .08 and 1.20 for L2. In Ll,

42-day weight had a realized genetic correlation with wean-

ing weight 1.04 and with 60-day weight of .57.

Selection for 42-day weight in D1 yielded large

growth responses on both diets and a negative correlated

response for maternal ability in D2. Selection in D2.

resulted in a small direct response, a large correlated re-

sponse for 42-day weight in Dl and a positive correlated

response for maternal ability on Dl. Growth rate behaved

as a maternal trait in D2.


I I___















INTRODUCTION


One of the basic premises of biometrical genetics

concerns the nature of the relationship between genetic

and environmental variations which together give rise to

variation in phenotypes.

Since most breeds of livestock originated in tem-

perate zones, their performance in the tropics may be

profoundly influenced by genetic and environmental vari-

ations and their interrelationships in the expression of

phenotype.

Selective cattle breeding historically has been

practiced predominantly in temperate zones. The temperate

zones with their short growing seasons produce forages of

low fiber content and high nutritional value. As western

man began to colonize and use the tropical and subtropical

zones for agriculture and animal husbandry, he generally

used existing breeds of cattle from the temperate zones as

foundation stock.

Cattle breeding techniques in the temperate zone have

become increasingly more sophisticated with developments

such as performance testing, progeny evaluation, artifi-

cial insemination, heat synchronization and embryo trans-

plants. These techniques usually are practiced under in-

tensified management systems with optimum nutritional re-

gimes.

1









Progressive livestock breeders from the tropical and

subtropical zones have engaged in importation of temperate

zone cattle into their areas for improvement of foundation

stock. Controversy and doubts exist as to whether the geno-

types best suited for temperate zones will be most produc-

tive in the tropics. Tropical forages due to the interrela-

tionships between plants, soil nutrients, moisture and tem-

perature are high in lignin and fiber resulting in lower

quality than temperate-zone forages. One school of thought

is that animals selected under a low plane of nutrition for

increased growth rate will maintain or excel their level of

performance when transferred to a good plane of nutrition.

The study of interactions of environments and geno-

types in large animals is a project of long duration and

great expense. The small animal model provides a short

generation interval, requires only modest facilities, and

lends itself better to selection experiments due to the

large number of young.produced per mating allowing for a

higher selection intensity. Data from the small mono-

gastric animal model, however, may not be altogether char-

acteristic of responses from ruminant populations.

Wright (1939) showed that the relationship between

genotype and environment need not be additive. He sug-

gested that if nonadditive genotype-environment interac-

tions existed a breed would have to be developed for every

ecological niche capable of supporting one. In 1947

Hammond suggested that animals should be bred in a highly


1









favorable environment in order to maximize expression of the

trait.

Falconer and Latyszewski (1952) hypothesized that for

Hammond's thesis to be valid there should be no genotype-

environment interaction present. They also pointed out that

heritability in a good environment must be higher than in a

poor environment in order for selection in a good environ-

ment to be superior to selection under a poor environment

when both lines are measured under the good environment. It

was further stated that a good environment should reduce

the random environmental variation which normally lowers

the correlation between genotype and phenotype. Falconer

(1960a), expanding upon his earlier work, demonstrated that

selection for growth in mice under a poor environment was

superior to selection under a good environment when per-

formance was measured under both environments.

The interrelationship of diet and genotype in the

tropics is a condition that is present during the animals'

complete life cycle. Most experiments in the literature

have been concerned with temporal diet effects where

animals did not reproduce or lactate on the poor diet.

They were exposed to it for a short postweaning growth pe-

riod after which the selected individuals were returned to

the optimum diet to breed, gestate and raise their young.

Experiments in which diets were not temporal in action

were reported by Fowler and Ensminger (1960) in swine and

by Hardin and Bell (1967) working with the flour beatle






4


Tribolium castaneum.

This project was designed using diet as a continuous

environment with the mice spending their entire life in the

diet designated as environment. The objectives were to

study the response to selection under two nutritional en-

vironments, to assessthe performance of the lines both en-

vironments and to measure the correlated responses in

traits that would be of economic interest if measured in

livestock.


~














REVIEW OF LITERATURE


General Review

Genotype-environment interactions have been studied

from different approaches by many workers. McBride (1959)

presented an extensive review of opposing arguments. As

early as 1939 Wright recognized that the relationship be-

tween genotype and environment enhanced the effectiveness

of selection, while Lush (1945) recommended that the en-

vironmental conditions for a breeding herd be similar to

those prevailing for the general population or breed under

improvement. James (1961) postulated that an unfavorable

environment accompanied by increased physiological stress

would accelerate genetic improvement.

In order for selection to be more effective in a good

environment, the good environment should be capable of re-

ducing the random environmental variation which normally

lowers the correlation between genotype and phenotype; in

other words, the heritability of the character should be

higher in a good environment than in a poor one.

Haldane (1946) made the first attempt to classify

genotype-environment interactions. He postulated four

types of interactions between two genotypes and two environ-

ments. He gaveno attention to the type of differences

between environments and genotypes. Lerner (1950) intro-










duced the term "nonlinear genotype-environment interaction"

to define Haldane's 2, 3, 4a and 4b type interactions while

he termed the remainder as "linear."

McBride (1959) postulated a method for classifying

genotype-environment interactions. He partitioned environ-

mental differences into two types, micro-and macro-macroen-

vironmental, and treated intra and interpopulation differ-

ences separately. Different climates, diets or even dif-

ferent management practices are examples of macroenviron-

mental differences, whereas the microenvironmental differ-

ences are those fluctuations which occur even when all

animals apparently are treated alike. These differences

were referred to as "intangible" by Wright (1921). An

example would be peck order in domestic fowl.


Results From Experiments

Falconer and Latyzewski (1952) reported that the im-

provement made by selection under good conditions may not

be expressed when animals were transferred to poor condi-

tions but that the improvement made under less favorable

conditions was retained when animals were transferred to

favorable conditions.

Falconer (1960a) showed that mice selected for in-

creased growth on a low plane of nutrition, but later

reared on a high plane, were similar to those produced by

selection on a high plane. Their growth was the same but

they were heavier, had less fat and more protein and were

better mothers.










The interaction between temperature and lines was

studied by Baker and Cockrem (1970). Mice selected for high

42-day weight under cold, medium and high temperatures for

four generations were switched at 3 wees of age to the

other two temperatures. No significant interaction for

lines x temperature was found.

Dalton (1967) reported a selection experiment with

mice on two diets. His design was very similar to that of

Falconer (1960a), except that dietary differences were

greater. His measurement of growth was the weight gain

between 3 to 6 weeks of age. The low plane consisted of 70%

cellulose (crude fiber) and 30% regular lab ration. Selec-

tion was within litters standardized at birth to eight. He

found no adverse effects from natural selection by compar-

ing selection differentials. Estimates of realized herita-

bility were .23 and .03 in the upward direction for good

and poor diets, respectively. When both upward selected

lines were tested by means of second litters on the diet

opposite to that in which they were selected, he found that

on the good diet the good line was superior, while on the'

poor diet the poor line was better. Differences were not

significant, however. He concluded that there was no

genotype by environment interaction present.

Park et al. (1966), working with rats, measured post-

weaning growth (3 to 9 weeks) under full and restricted

feeding regimes. Response to selection was higher under

the full-feed regime with realized heritability estimates


___









of .127 .02 and .058 .02, respectively, for the two re-

gimes. When the performance of both lines was evaluated

under the two feeding regimes it was found that a signifi-

cant line x feeding regime interaction occurred in females,

but not in males. He concluded that selection was best

practiced under the regime where the animals were to per-

form. Correlated responses resulting from selection for

postweaning gain are reviewed later in this paper.


Weaning Weights

Falconer (1960a) found that female mice selected

under a poor nutritional plane produced heavier offspring

at 21 days of age when lactating on the good plane of

nutrition, both in the upward and downward selection groups.

Falconer (1955), selecting within litters, reported that

upward selection for increased 42-day weight resulted in a

slight increase in weaning (21-day) weight up to generation

12 (.75 grams), followed by a decline to the end of the ex-

periment (30 generations). At the end, the large line had

a 21-day weight 3 grams below the control line. He further

stated that the regression of 21-day weight of the off-

spring on 42-day weight of the mother was .09 .05, as

compared with -.09 .10 for the unselected group.

Carry-over effects of weaning weight on postweaning

growth are to be expected since this represents a part-

whole relationship. However, Falconer (1960a) introduced

controversy by reporting a higher regression of growth (3









to 6 week weight gains) on weight at 3 weeks for the poor

environment than for the good diet. This higher regression

was more evident in males than in females, due possibly to

a greater growth potential. Sutherland et al. (1970) reported

a decrease in weaning (28-day) weights for lines selected

for 4 to 11-week weight gains, feed intake and feed effi-

ciency. Decreases per generation were about -.30 grams per

generation. He postulated that the decrease might have been

due to an increase in inbreeding, as even the unselected

line accumulated 56.8% inbreeding.

Selecting for high 42-day weight under hot, medium

and cold temperatures, Baker and Cockrem (1970) reported a

positive correlated response for weaning weights under the

medium and cold environments and a negative one for the hot

environment.

Park et al. (1966), working with rats and using post-

weaning gains (3 to 6 weeks) as the trait under selection,

obtained a negative correlated response for weaning weights

under both full feed and restricted regimes.

Maternal environment acting on growth of the offspring

through the prenatal and postnatal influences is an import-

ant source of variation in body weight of mice. Maternal

effects are still evident at 60 days of age. Brumby (1960)

studied this area extensively, using techniques of ova

transplant and cross-fostering. His experimental material

consisted of three lines, one selected for high 42-day

weight, one for low 42-day weight and the third an unse-


~










elected control. He found a deterioration in lactational

performance of both selected lines as compared with controls.

His results agreed with Bateman (1954) who stated that 73%

of the variance associated with 21-day litter weights was

due to prenatal influences. The effect of maternal environ-

ment was evident at 60 days of age, even when adjustments

were made for litter size. These results emphasize the im-

portance of prenatal environment.

Cox, Legates and Cockerham (1959) looked at the rela-

tionship of prenatal versus postnatal environments on 12-

day, 21-day and 42-day weights of mice. Cross fostering

was used in a fashion similar to Bateman (1954). He found

no effect of sex on 21-day weights. Postnatal environment

accounted for 60% of the variance of 21-day weights, even

when litters were standardized to six young at birth. He

disagreed with Brumby (1960) by stating that postnatal en-

vironment was the main source of variation in weaning

weights.

Young and Legates (1965), using a cross-fostering

technique with mice, reported upon relationships between

postnatal maternal performance and subsequent weights. The

genetic correlation between postnatal maternal performance

and 21-day weight was .46 using standardized litters of

six young. However, they reported negative correlations

between preweaning and postweaning gains. Apparently dams

that milked well produced young which had lower gains fol-

lowing weaning.










MacDowell et al. (1930) stated that the increase or

differences in weights incurred during the nursing period

persisted at maturity in mice. Zucker et al. (1941a, b)

found the opposite with rats. Butler and Metrakos (1950)

working with mice reported data which agreed with MacDowell

et al. (1930).

Neville et al. (1962), working with Hereford cattle,

reported a lower partial regression between preweaning growth

and milk yield on high nutritional levels than for cattle on

lower levels of nutrition, indicating less dependence of

calf gains on milk supply as nutrition improves.


42-Day Weights

Sudden increases in the trait under selection as a

selection experiment progresses are not uncommon. Roberts

(1966) found a marked increase in 42-day weight between

generations 43 and 44, from 32 to 35 grams. He believed at

first that the change was environmental since the small

line (SC) also showed an increase at this point. In addi-

tion, the diet had been modified at the time. He concluded

that the changes probably were genetic in origin, basing his

hypothesis on the fact that no such increases were observed

in four other lines derived from the same line. Further-

more, the SC line was notorious for its fluctuations be-

tween generations, also the diet modification was of a minor

nature.

The number of young born or weaned in a litter has

been reported to have a marked effect on individual weight










of the mice at 42 days of age. Roberts (1966) reported a

regression of -.6 grams for each mouse born, whereas

Falconer (1960b) reported a lower value of -.34 for an un-

selected line. However, Roberts (1966) further stated that

correction of 42-day weights using within-generation regres-

sions resulted in such a small correction that no experimen-

tal conclusion could possibly be affected.

Baker and Cockrem (1970) reported results of a selec-

tion experiment with mice under three temperatures, high

(29C to 33C), medium (19"C to 23*C) and low (6C to 8C).

The trait under selection was 42-day weight and the corre-

lated response was tail length. After four generations,

mice from the second litters of all three lines of genera-

tion four were allotted at random to the other two tempera-

tures at weaning (3 weeks) and their growth to 6 weeks re-

corded. He obtained realized heritabilities of .64 .05,

.38 .09 and .52 .07 for hot, medium and cold tempera-

tures, respectively. Since litter size was not adjusted at

birth, the experiment afforded an opportunity to study ef-

fects of litter size on 42-day weight. All regressions were

negative. The maximum was -.5 grams per mouse weaned. How-

ever, since litter size at weaning differed by only .7 mice

between the three lines, there were no significant differ-

ences in the regressions between lines.

Cox et al. (1959), using a cross-fostering technique,

reported an estimate of the effects of postnatal environ-

ment on 42-day weight. Using standardized litters of six









mice, he reported that 26% of the variation in 42-day

weight could be attributed to postnatal environment.

A limit was reached by Roberts (1966) after 20 gener-

ations of selecting for increased 42-day weight in mice.

Neither reversed nor relaxed selection yielded change, in-

dicating an exhaustion of additive genetic variance.

Falconer (1955) found a high correlation of .90 between

postweaning growth and 42-day weight.

Young and Legates (1965) reported a genetic correla-

tion between postnatal maternal performance and 42-day

weight of .37 using standardized litters of six young. The

phenotypic correlation between 21-day and 42-day weights

was .68.

Working with Hereford cattle, Neville et al. (1962)

reported a lower partial regression coefficient between

milk yield of dam and postweaning performance of the calf

under a high nutritional level than under a low nutritional

level, indicating a higher dependency on maternal perform-

ance under poor nutritional regimes.


Fertility

Roberts (1966) demonstrated that cessation of selec-

tion for large or small size at 42 days of age resulted in

increased fertility of exposed females.

The effect of temperature on fertility was briefly

discussed by Baker and Cockrem (1970). They reported a

longer interval between pairing and birth of first litter










in cold or hot environments than for normal temperatures.

The interval from first exposure to birth of first

litter was used by Park et al.(1966) as a fertility crite-

rion in rats. When the response of the correlated trait

was regressed on the cumulative selection differential of

the selection trait (3 to 6 weeks growth), estimates

obtained were .00.01 and -.005.01 for full-fed and re-

stricted-fed lines, respectively.


Litter Size

MacArthur (1949) with mice, and Hetzer and Birer

(1940) with swine, found a direct association between body

weight at maturity and litter size. Wright (1922) and

Eaton (1941), working with guinea pigs, observed that a

decline in growth from inbreeding was correlated with re-

duced litter size, while growth heterosis was associated

with large litters. Dickerson et al. (1954), working with

swine, found that selection for increased litter size and

growth was not effective in improving genetic worth of

these traits. They postulated a negative genetic correla-

tion between the traits concerned. Cockerham (1952) esti-

mated that the genetic correlation between growth rate and

litter size did not differ significantly from zero.

Rahnefled, Boylan and Comstock (1962) reported a

heritability based on paternal half-sib correlation of

.11.04 for litter size in mice. The genetic correlation

between litter size and growth rate was .15, but not










significantly different from zero. The genetic change in

average litter size for the line selected for growth rate

during 13 generations, expressed as a linear regression,

was .082.04 per generation.

Falconer (1955) reported a dam-daughter regression

for litter size of virtually zero. However, the standard-

ized regression coefficient, holding the daughter's 42-day

weight constant, gave a value of .10.05 for a heritability

of 20%.

It has been demonstrated in swine by Dickerson et al..

(1954) that inbreeding was detrimental to prolificacy, even

when upward selection for that trait was applied. Coef-

ficients of variation of 35% for number born per litter and

54% for number weaned were reported.

The effect of temperature on litter size in mice was

reported by Baker and Cockrem (1970). No significant dif-

ference between hot, cold and medium environments for litter

size was found.

The correlated response in litter size to selection

for postweaning growth (3 to 9 weeks) in rats was studied

by Park et al. (1966). It was expressed as the regression

of the correlated trait on the weighted selection differ-

ential for postweaning gains. The value was -.006.004

under the full-fed regime and .002.005 under the re-

stricted regime.


___









Literature Summary

Conflicting reports are found in the literature con-

cerning the results of selection in mice and rats for in-

creased body weight on two nutritional environments.

Falconer and Latyzewski (1952) and Falconer (1960a) re-

ported an advantage for selection under the poor environ-

ment. Dalton (1967), Korkman (1961), Park et al. (1966)

and Bateman (1971) recommended that selection should be

practiced on the nutritional environment in which the off-

spring are expected to perform.

Females from a line selected in a low nutritional en-

vironment were found to be better mothers (Falconer, 1960a)

than those selected on a high plane when both were reared

on the high plane.

Selection for growth rate was found to be negatively

associated (Roberts, 1966) with fertility in mice. Park

et al. (1966) in rats found no significant effects on fer-

tility due to selection for growth rate.

A positive relationship between increased body weight

and large litters in mice was reported by MacArthur (1949),

Rahnefeld et al. (1962), Falconer (1960b), Eisen (1970),

Falconer and King (1953) and Dinsley (1966). Cockerham

(1952) estimated a genetic correlation between growth rate

and litter size not significantly different from zero.

Korkman (1961) found a decrease in litter size in mice se-

lected for increased growth rate.

Kownacki (1971) reported higher survival to weaning









in mice selected in a high plane of nutrition as compared

to those selected on a low plane.

A negative effect on weaning weights as a result of

selection for increased 42-day weight was reported by

Falconer (1955) and Brumby (1960). No significant sex ef-

fects at weaning were reported by Cox et al.(1959). A

larger difference between the sexes in a good environment

was reported by Butler and Metrakos (1950) and Korkman

(1957).

Roberts (1966) and Falconer (1960b) reported a

marked negative effect of the number weaned per litter on

individual weight of mice at 42 days of age. Postnatal en-

vironment was found to be a highly significant source of

variance for preweaning and postweaning weight by Cox et

al. (1959), Young and Legates (1965), Jinks and Broadhurst

(1963), Rutledge et al. (1972) and Moore et al. (1970).

Small responses to selection for growth rate in the

low plane were reported in mice (Korkman, 1961- Dalton,

1967; Kownacki, 1971) and in rats (Park et al., 1966).

Phenotypic variances were found to be larger in the

low plane as compared to the high plane by McLaren and

Michie (1956) and Bateman (1971).

Non-Mendelian transmitted maternal effects were re-

ported by Morton (1970), Brumby (1960) and Reutzel (1970).


~














MATERIALS AND METHODS


Collection of Data

Two lines were formed in 1967 allocating eight males

and twenty-four females randomly from a population of

random bred ICR mice from Dublin Laboratories to each of

two diets. Diet one (Dl) consisted of Purina Laboratory

Chow. Diet two (D2) was 70% Purina Laboratory Chow and 30%

Alphacel, a nonnutritive bulk. Both diets were fed ad

libitum with feed and water present at all times.

Mass selection was practiced from first litters only

for increased individual weight at 42 days of age for eight

generations. Mice were weighed after feed had been with-

held for approximately 24 hours. Selections were made only

from litters in which at least six mice were weaned. Litters

were reduced to 10 mice at birth by sacrifice of all young

exceeding 10. Mice were bred at 65 2 days of age. No

full- or half-sib mating were permitted, nor was it allowed

for two full-sib females to be bred to the same male. The

mating design employed is presented in Table 1.

Eight third-generation females were selected from

each diet and placed on the opposite diet at 60 days of age,

thus forming four groups. Mice selected originally on diet

1 were referred to as line L (Ll), and those from diet 2 as

line 2 (L2). The four groups consisted of the two original


18














Z
H
H
en
E-A

H
mE-
z













On
H
U)

0











WH




AM


pm
ZP
H
















QH


HO

w Q

O
HO
0
E-i
w















H
Z
H














-l
H
03






H

Q E4


CN
-a
em


a) E
(N *O a)







H









--4
U)
N
tC











4-1 4
0 E
-rH )




a)
*-H



0)
'-4
*-




r-4

aW



C^ a
1-H C


en en n m m n m












I I I H H- r- r- rI




0 W (N CD 9 0 U) N
r-4 Nq -W Ln rN C0 (N


ID i ii i I I
-i M Ln -4 N o en 4D l
i-4


n en en n m mn n




Ln CD m C 0 ( c -
-I m en Ln 10 0 ( r-l
r-l

v-I o N e CD Ln N
H- (N v in r- m 0
r-







r-I (m m 'rn LA D rN C


__









groups, Ll/D1, L2/D2, and the two new groups, L1/D2 and

L2/D1.

Breeding males were selected only from the original

two groups in all succeeding generations, whereas breeding

females were selected from all four groups. Females were

bred only to males of their own line, i.e., Ll/Dl and L1/D2

females were bred to Ll/D1 males whereas L2/D2 and L2/D1

females were bred to L2/D2 males. This was accomplished by

changing males from diet 1 to diet 2, or from diet 2 to

diet 1, every 24 hours during the breeding season so that

the females underwent no change of diet. Offspring from

the four groups were grown, selected and bred with no change

of diet for the next five generations with the exception of

males during the breeding season. Selection criteria were

the same for females in all four groups.

Each breeding male was mated to three females from

his own line-diet group and to one female from his line on

the opposite diet. The mating period continued for 18 days

after which all females were placed on individual solid

bottom cages using treated corncob grounds as bedding.

All litters were weighed at birth and if larger than

ten in size they were reduced to 10 by sacrificing at

random except for weak or deformed pups.

The mice were weaned at 21 days of age. At this time

they were weighed, ear marked and separated by sex. They

were raised in 3 x 3 mesh wire cages in groups of approxi-

mately 10. The cross-over mice (LlD2, L2D1) were inter-









mingled with the selection group on the same diet. At 41

days of age feed was withdrawn from the mice for a period

of 24 hours. The mice then were weighed at day 42 on an

empty basis. The same procedure was followed in obtaining

60-day weights.


Description of Data

The data were partitioned into two distinct sets.

One set consisted of 8 generations from Ll/Dl and L2/D2

and is referred to as the selection experiment. The other

set involved the last 5 generations which compared contem-

poraneously Ll/Dl, L1/D2, L2/D1 and L2/D2. This set is

referred to as the genotype by environment interaction (GEI)

experiment.

A portion of the results from the selection experi-

ment have been reported by Rodriguez, Koger and Franke

(1970).

The data represented two types of traits, dam traits

and individual weight data. Dam traits included pregnancy

rate (a trait common to all selected females) and the lit-

ter traits, litter size at birth, litter size at weaning,

average within litter 42-day weight and total litter weight

at 42 days of age. The individual weight data included

weaning weights, 42-day weights and 60-day weights.

Pregnancy rate, litter traits and individual weights

were analysed by fitting different models to the GEI data

(Tables 2 to 4). Genetic parameters for individual














TABLE 2. EFFECTS FITTED, DEGREES OF FREEDOM AND THEIR
EXPECTED MEAN SQUARES FOR MODEL 1 EMPLOYED
IN ANALYZING PREGNANCY RATE



SOURCE df E (MS)


Generations (G) 4 02w + k15(2 + k1602

Lines (L) 1 02, + k1302 + k14o2

G x L 4 02w + k11ll2 + k1202gl

Sires/GL (S) 70 a2w + kl002s

Diet (D) 1 02, + k802ds + kg2d

D x G 4 02w + k602ds + k702dg

D x L 1 o2w + k402ds + k502dl

D x G x L 4 2w + k2 2ds + k302dgl

D x S 70 02 + klG2ds

Within 160 2w









weights were obtained from the selection experiment (Table

5). Complete confounding of sires and dams in L1/D2 and

L2/D1 precluded the unbiased estimation of random variances

in these two groups.


Analyses of Data

The data were analyzes employing least squares tech-

niques described by Harvey (1960), utilizing the Ohio State

University Computer Least Squares Program written by Dr.

Harvey. The data were analyzed according to procedures

described by Harvey (1964) using absorption techniques in

the analysis of split-split plot designs.


Pregnancy Rate

Pregnancy rate was expressed as the presence or

absence (recorded as 1 or 0, respectively) of a litter re-

sulting from the pairing of 240 females with 80 sires.

Model 1 was fitted and the effects, degrees of freedom and

expected mean squares appear in Table 2. Sires in the mod-

el represent males mated with females, not sires of females.


Litter Traits

Litter traits are those in which the litter measure-

ment is the observation. Litter size at birth and at

weaning, were unadjusted observations. Total litter weight

at 42 days of age was obtained by totaling individual

records of litter mates adjusted for sex. Least squares

constants obtained from fitting a model containing sex as a









single effect were employed to adjust data. Average within-

litter 42-day weight was obtained by dividing total litter

weight by the number of litter mates at 42-days of age.

The data included 175 litters by 54 sires. There

were 74 litters by 23 sires in line 1, and 101 litters by

31 sires in line 2. Sires were considered random in model

2, with all other effects (Table 3) assumed to be fixed.


Individual Weights

Weaning weight (21-day weight), 42-day weight and 60-

day weight were the individual weights analyzed. Unadjusted

observations were analyzed by fitting model 3 (Table 4).

The number of pups weaned per litter was used as a covari-

ate to remove environmental effects due to variations in

litter size.

The data consisted of 1312 records, representing 51

sires and 162 dams. Only sires with progeny on both diets

and litters of at least three young at weaning were

included. This restriction was imposed to prevent distor-

tion of sire x diet effects by insuring at least three

observations in sire x diet subcells.


Random Variances

Estimates of random variances were obtained for

weaning weight, 42-day weight and 60-day weight by analyz-

ing individual weight data from 8 generations. Line 1

data consisted of 1113 mice representing 50 sires and 125

dams, while line 2 had 878 records by 46 sires and 115
















TABLE 3. EFFECTS FITTED, DEGREES OF FREEDOM AND THEIR
EXPECTED MEAN SQUARES FOR MODEL 2 EMPLOYED
IN ANALYZING LITTER TRAITS



SOURCE df E (MS)


Generations (G) 4 o2 + k1502s + k1602g

Lines (L) 1 02 + k13o2s + k14a21

G x L 4 02w + kllos + k1202gl

Sires/GL (S) 44 0o2 + k10O2s

Diet (D) 1 02w + kgo2ds + k9g2d

D x G 4 OZw + k602ds + k7O2dg

D x L 1 o2w + k402ds + k502dl

D x G x L 4 o2w + k2O2ds + k302dgl

D x S 44 a02 + kl02ds

Within 67 0 w


- -













TABLE 4. EFFECTS FITTED, DEGREES OF FREEDOM AND THEIR EXPECTED
MEAN SQUARES FOR MODEL 3 EMPLOYED IN ANALYZING
INDIVIDUAL WEIGHT DATA



SOURCE df E (MS)


Generations (G) 4

Lines (L) 1

G xL 4

Sires/GL (S) 41

Diet (D) 1

DxG 4

D x L 1

Dx G xL 4

D xS 41

Dams/SDGL (I) 58

No. Weaned (Linear) 1

No. Weaned (Quadratic) 1

Sex (Z) 1

Z x G 4

Z xL 1

Z xD 1

Within 1141


k2802i + k29a2s

k25 G i+ k2602s

2 2
k222 i + k23 s

k202i + k210 s

k17a2i + kl8 2ds

k14 G2i + kl5y2ds
k 02 +k o2

kllo2 + kl2G2ds

k 2i + k902ds +

k62i + k7o2ds
k5a2
i5
K 02
31 bl

K3202bq

k42z
4 z

k3zg

2 zl
k202z
kl2zd


k30 2g

k272 1

k2402gl


+ k19 d

+ kl602dg

+ kl30 dl

kl0 2dgl









dams. Data from line 1 and line 2 were analyzed separately

by fitting model 4 (Table 5).


Genetic Parameters

Variance Components

Estimates of the components of random variances were

obtained using the method of King and Henderson (1954) for

unequal number of progeny per dam and unequal number of

dams per sire.

K2
MSS MSw K1 (MSD MSw)
S K3



MSD MSW

1



O2W = MSW


K1 = number of offspring per dam

K2 = number of offspring per dam/sire

K3 = number of offspring per sire

o2 = sire variance

2D = dam variance

02W = within-litter variance


The K values were computed as suggested by Becker

(1964).
Total No. E (number/dam)2
number/sire
K1 =d.f. dams







28












SH
0 n

H Z +


S+ +


t0 0 I
U H m H f N N N
H 0 b b 0 ID 0 0)

MO En ,








XE + I I+ + +









O *H H ) H H
cI

00
M ac







HH rM N
O0 0
E-i E E
0 V







<, H \ I 3 I
End
O w














ci a) U)
U $ n *

H -I o E En En
w P w 0 q Ul fl En En n
E 0 0 H En E
i- 0 En : w R









E (number/dam)2 (number/dam)2
number/sire total number
K2 = d.f. sires


Total No. E (number/sire)2
K total number
Sires 1



Estimates of h

The estimates of h2 and their standard errors were

computed as suggested by Dickerson (1960).

4o2s VA
h2 =
OW + 2D + 2S Vp



MS2 MS2
4 2/K23 S + 1 + D-S-2

s.e.h2 =
2 + D + 2S



s.e.h2 = standard error of the h2

s = number of sires

D = number of dams


Causal Components of Variance

The causal components were computed from.the observa-

tional components as described by Falconer (1964), with

modifications introduced by Willham (1963) and Miller,

Legates and Cockerham (1963).

The phenotypic variance (Vp) of mice weights can be

expressed in terms of its components as









V =02 +0 +0 +02 2 2 2 +2
p = Ao Do AoAm Am Dm + Ec +Ew + I


02Do = nonmaternal dominance variance


oAoAm = covariance of maternal and nonmaternal
components


o2Ao and G2Am = additive nonmaternal and maternal
variances


S2Dm = maternal dominance variance


o2Ec = random environmental variance common to litters


o2Ew = random environmental variance within-litter


o21 = epistasis variance


Miller et al. (1963) found no nonadditive genetic

variance due to dominance or epistasis in 21- and 42-day

weights in mice. Similar results were reported by Cox,

Legates and Cockerham (1959). On the basis of these findings

it was assumed that dominance and epistasis effects are not

different from zero. The expression then reduces to a

simpler form.

V = 02Ao + oAoAm + 02Am + o2Dm + o2Ec + o2Ew
P

This expression can then be defined in terms of the

observational components.

o2S = 1/2 0 Ao


o2 = 1/4 02Ao + GAo Am + 02Am + 02Dm + a2Ec
D


o2 = 1/2 2Ao + o2Ew










The causal components can then be computed from the

observational components.

Nonmaternal additive variance (VAo) = 4 2S

Maternal effect (VMe) =2D 2

Within-litter variance (VWL) = o2W 2 S

Phenotypic variance (V ) = o2 + 2D 2W

Therefore:

Ao = 2Ao

V = O2 + AmAo + 02 + 02
Me Am AmAo Dm Ec

VWL = Ew


p = VAo + VMe + VWL



Realized Genetic Correlations

Falconer (1952) developed the first genetic correla-

tion formula based on direct and correlated responses:

CR oax
rA RXay


Where:

rA = realized genetic correlation


CRy= correlated response in trait y to selection for
trait x

RX = direct response to selection for trait x


o = additive standard deviation of trait x
ax


cay= additive standard deviation of trait y






32


Standard errors of the realized genetic correlations

were calculated according to Robertson (1959).


-r2A (h2 ) a (h2y)
S(rA) h2h2y



Where:

o(rA) = standard error of the genetic correlation
estimate

r2A = square of the genetic correlation

o(h2x) = standard error of the heritability estimate
of trait x

o(h2 ) = standard error of the heritability estimate
of trait y


__















RESULTS AND DISCUSSION

Both the direct and correlated responses to selection

enter into total response which is generally the item of

greatest interest in selection experiments. In single

trait selection, the aggregate of correlated responses gen-

erally are of greater impact than direct response. This is

true especially of experiments involving genotype by environ-

ment interactions. The present trial fits into this catego-

ry.

To gain the best comprehension of results from the ex-

periment requires the presentation of correlated responses

both within and across diets. Consequently, the GEI data

set will be presented first. The data include those from

generations 4 through 8. Estimates of direct and important

correlated responses along with total productivity in the

two dietary environments will be presented first.

Data from the two selection groups, D1/L1 and D2/L2,

across all eight generations will then be presented along

with certain genetic parameters having a bearing on the

results from the GEI phase of the trait.

The unadjusted means and standard deviations for the

traits measured in the project are presented in Table 6.




















0-r
4-J O in i) Ln tn in (N 0(N c(
(a c0 r- r0 r rN rN- N-H 1-1
U) ( > (e i iH 4 r- ,- rH m m
k rl. H r-l
t1 mE m)

H O 0
wQ

"C 0









0 4-) M
w i in N rq rq m H o
Z (0 4( o\ D (N (N Hl o CO o










W o
~Q -N iD ( ( (N c 0 (N in
fa 0> m in

4W

U) 4 U U) U















-) 0' 0' 0'
U) Ul tol










o0 I>n(0 -
0 LO o a) I m -V
H On in in ui u- N,
H r



4 -3 3 4- ( R







*- N- N H O- 0 H i4
,H 0I - 0 (N (N




(0 H 0 '



N0 N H
(1) (N)




a4- 4 4 .4 ca Q n
41 -14 (N > a)

P E-1 -

n C >i 4-i 0



m >B -H *Hl -1 4 4- ) (0






pi 2 -l il -4 E-4 .- ri r io









Pregnancy Rate

Pregnancy rate was analyzed by fitting model 1. The

analysis of variance is shown in Table 7. Generation and

sires were the only significant (P < .05) sources of varia-

tion. Least squares means for pregnancy rate are presented

in Table 8..

While generation effects were evident, they did not

appear to follow a trend attributable to a response to

selection for body weight. The first four generations were

suggestive of a negative trend. During the last generation,

however, pregnancy rate returned approximately to the level

of generation 4. The regression of generation means on

generation number was not significant.

When generation and sires were considered together, it

appeared that random sire differences were the major cause

of fluctuations in generation means. Sire differences ap-

preared to be the result of an occasional sterile sire

within a particular generation. These results disagree with

those obtained by Roberts (1966) with mice and Park et al.

(1966) with rats who reported a decrease in pregnancy rate

as a result of selection for increased weight.


Litter Size at Birth

Physiological aspects of reproductive responses in

this project were reported previously by Dickens (1970).

Litter size at birth, weaning and 60 days of age were in-

cluded in the present study as dependent variables to help















TABLE 7. LEAST SQUARES ANALYSIS OF VARIANCE FOR PREGNANCY
RATE, MODEL 1



SOURCE df SS MS F


Generations (G) 4 1.86 .46 2.52*

Lines (L) 1 .15 .15 .83

G x L 4 .89 .22 1.21

Sires/GL (S) 70 12.76 .18 1.62*

Diet (D) 1 .04 .04 .33

D x G 4 .19 .05 .42

D x L 1 .07 .07 .58

D x G x L 4 .61 .15 1.25

D x S 70 9.20 .12 1.11

Within 160 18.00 .11


* P < .05





























o










H
010
cN0




H






WE-












Lo
at
E-4














0 Q
0

K 0
W


-H NN -1
1 -l 1-1 10 1


-I I H CN


h Q Q


__1~









characterize the litters used in these analyses. Litter size

was analyzed using Model 2.

The variance analysis for litter size at birth is shown

in Table 9. Two main effects (line and diet) and one inter-

action (generation x diet) were found to be significant

(P < .05). The subclass least squares means for litter size

at birth are shown in Table 10..

Larger litters were born on diet 1 than on diet 2. With

successive generations of selection, diet 1 litters became

larger, while diet 2 litters became smaller, resulting in a

significant (P < .05) diet x generation interaction. The

positive effect of maternal body weight on litter size at

birth has been reported for mice by Falconer and King (1953),

Dinsley (1966), Rahnefeld et al. (1962) and MacArthur (1949);

by Hetzer and Brier (1940) for swine, and by Robens (1968) in

golden hamsters. The 60-day weights of mice taken just prior

to pairing appear in Table 23. It is evident that an in-

crease both in 60-day weight and litter size at birth occurred

in diet 1. It also appears that L1/D2 mice were heavier and

had larger litters than L2/D2.

Significant (P < .05) line effects on litter size may

also be attributed to the relationship between body weight

and litter size at birth. Line 1 mice were bigger at 60 days

of age and had larger litters than line 2 mice in both diets.

The regression coefficient of litter size on dam's weight in

L1/D1 for all 8 generations of the selection experiment was

.37 (P < .001). This was similar to results with mice
















TABLE 9. LEAST SQUARES ANALYSIS OF VARIANCE FOR LITTER
SIZE AT BIRTH, MODEL 2



SOURCE df SS MS F


Generations (G) 4 33.54 8.38 1.23

Lines (L) 1 40.71 40.71 5.95*

G x L 4 8.43 2.11 .31

Sires/GL (S) 44 300.83 6.84 1.25

Diet (D) 1 34.37 34.37 5.89*

D x G 4 60.57 15.14 2.59*

D x L 1 12.44 12.44 2.13

D x G x L 4 48.48 12.12 2.08

D x S 44 256.77 5.84 1.07

Within 67 366.33 5.47


*P < .05

















H
H










E-I
0




H





N





E-i








oa




CO









H
0w







H


t-l




of

M


ull

HO
0 0
i-l cN l
Q Q rt; U


H (N H N



Hl HlN N rH (N
~ili-l i-ll il-l l










reported by Eisen (1970). He concluded that the sum of addi-

tive covariance between maternal effects for litter size and

maternal effects for body weight; and for additive covariance

between maternal effects for litter size and direct effects

for adult weight must be positive. He hypothesized that co-

variances must have been the result of pleiotropic effects

of genes influencing litter size and body weight. Korkman

(1961) reported that selection for body weight in crosses

between certain inbred mice lines resulted in a decrease in

litter size and an increase in weight.

Litter size at birth for L1/D1, L2/D1, L1/D2 and L2/D2

were consistent with ovulation rates reported by Dickens

(1970) for these same groups.


Litter Size at Weaning

According to the analysis shown in Table 11, the only

significant effects influencing litter size at weaning were

generation (P < .05) and diet (P < .01). Least squares

generation means were variable and seemingly without pattern

(Table 12). Generation 7 means were low for all groups due

to an unspecific pathogen causing death in preweaning pups.

Diet had a marked effect on number weaned, with means

of 8.42 for diet 1 and 7.04 for diet 2. At least two cum-

mulative effects were involved. Diet 1 females had larger

litters than diet 2, as discussed in the previous section.

The survival rates of pups to weaning, Table 14, in diets 1

and 2 were 88.6 and 74.3%, respectively. The combination
















TABLE 11. LEAST SQUARES ANALYSIS OF VARIANCE
SIZE AT WEANING, MODEL 2


FOR LITTER


SOURCE df SS MS F


Generations (G) 4 67.99 17.00 2.93*

Lines (L) 1 .56 .56 .10

G x L 4 9.86 2.47 .42

Sires/GL (S) 44 255.66 5.81 1.60

Diet (D) 1 66.32 66.32 15.72**

D x G 4 22.36 5.59 1.32

D x L 1 8.56 8.56 2.03

D x G x L 4 16.02 4.01 .95

D x S 44 185.65 4.22 1.16

Within 67 243.50 3.63


* P < .05

** P < .01


- ------------








43








N a. 00 o e n in i m
fl m N N N N N H
H 1C
(a +1 +1 + + +1 +1 +1 +1 +I +


co H "- 0 o r




H oi N i a 0

CO +1 +1 +1 +1 +1 +1 + ++1 +1
- 0 O 0 H N r-Hl iD N
N (n 0co io I' N N 0 i

E CO ,D G i- [' 00 : "
HQ


0 x i ln No 'T Cl "n 00


-) N +1 +1 +1 +1 +1 +1 +1 +1 +1
O v H w CN m -1 in H o
OH m kH H m i N o N co





N 1 n a) in in N- ') in in

H
oo 0 o r N H m in










Q *H N) -i N in "D 4 in in m
S n +1 +1 +1 +1 +1 +- +1 +1 +1










no c co m C co oo o
- i in H in N im



0 a 0 m C h I .


Q 1 m I i0 00 0) r- (- CO C1
w m 3 H 00












m 00 01
J 4
E i m N in N N N N
S 0 in in i n N- Q
Q+ +i +i +i +i +i +i +I +1 +i
Cl in i n N in ri di in C \o






EH Hn No Hm N (


IMci) ^D Oi oo









< *HI H l I H N iI NM NI N HIN -
f- ilo~~ i~l i-) i- - s










of larger litters and better survival in diet 1 resulted in

a highly significant difference between diet means. Kownacki

(1971) reported larger litters at birth and higher survival

for mice selected on a high protein diet than for a line

selected on a low protein level.

The data showed some interesting nonsignificant trends

as the project progressed. Some of these trends, however,

were evident enough to merit discussion.

As selection progressed from generation 4 to genera-

tion 8, line 1 increased in litter size at birth. A nega-

tive trend, however, was evident in L1/D2. In generation 8,

litter size at weaning in line 2 (8.21) was superior to that

of line 1 (7.61). Litter size at weaning in L1/D2 responded

negatively to selection for increased body weight in Ll/Dl,

while litter size at weaning in L2/DI responded positively

to selection for body weight in L2/D2. Oddly enough, litter

size at weaning in the parental lines L/D1 and L2/D2 showed

little change.

Table 14 indicates superior survival for pups of line

2 in both diets. If survival to weaning is a function of

maternal ability, then line 2 females were superior to

mothers from line 1. As females of line 1 became heavier,

their maternal performance suffered, especially in diet 2.

Visual observation of carcasses of these mice indicated an

increasing predisposition to fatness in Ll/Dl animals as

body weight increased in response to selection. Anus to

tip of snout measurements, Table 13, revealed no increased











TABLE 13. NUMBER OF OBSERVATIONS AND AVERAGE LENGTH IN
CENTIMETERS FROM ANUS TO TIP OF SNOUT BY LINE,
SEX AND GENERATION


Males
L1/D1 L2/D2

Number of Obsel

19 20

20 20

14 13

17 15

18 16

14 14

17 17

119 115


Length

9.88

9.89

9.80

9.66

10.37

9.97

9.84

9.89


Females
L1/D1 L2/D2


ovations


20

20

15

11

18

14

19

117


13

19

115


in Centimeters

9.27 9.54

9.18 9.29

9.57 9.41

9.55 9.20

9.59 9.71

9.65 9.24

9.30 9.20

9.44 9.37


8.76

8.85

8.60

8.45

8.90

8.24

8.75

8.65


Generation


2

3

4

5

6

7

8

Total


2

3

4

5

6

7

8

Aver.



















Z M
C o N N o r %D (n o

W m 0 r N Ca r H
M () M 0 W N N C m N in
Z w oo o o) r- r o0 0o r oo
H


1--1
Ec


S r o in C CN O Co 0 0








H 0 i C < C-
in O r (N H a i (N O a
n N m o) tN rN i co ID. 0n






HCo o 0 o ti N co -H in r-

N CIN WD in o o N 4 O4 N
HD n om D I. r- t- i) to

0N N
o
HZ C
0 0





ai)





S N ( N "N m ( N 0m

C) ID m 0a0m C
> o ii i o i i o o n















U )
S(N m iN H (N i N I







4 0 C ( 40 0 H 0 0 0
0)






0

H L H ( >N H (N


0 Hq' H (N l r( l H N H (N HD
Icr o o^ ii Oi OI o o










skeletal length in L1/D1. Korkman (1961) has reported an

increase in length of mice selected for increased 40-day

weight on a high plane of nutrition. It is a prima facie

assumption that the increase in body weight which occurred

in L1/D1 was the result of a genetic ability to accumulate

fat. L2/D2 mice visually were leaner and more active than

L1/D1 individuals. It appears reasonable, then, to

theorize that selected breeders from LI/D1 were in effect

selected on the basis of ability to accumulate fat under

an optimum nutritional regime, while in the L2/D2 group

selected individuals were heavier because they or their

dams were better adapted to the adverse nutritional regime

under which they were selected.

When the line 1 females were forced to grow and

reproduce under the adverse nutritional regime of diet 2,

there was a tendency to utilize available nutrients for

fat deposition rather than for milk for the young, result-

ing in pup survival of 18.5% less than that for line 2.

Inability of a line selected under a good plane of nutri-

tion to adapt to a low plane was encountered also by

Falconer (1960a). He referred to it as the "susceptibili-

ty" of the good line to the low plane.

On diet 1, differences between line 1 and line 2

were mainly a function of initial litter size and sur-

vival. Both lines showed increased litter size at birth,

Table 10. For example, in generation 8, litter size for

line 1 was 12.01 and for line 2, 10.59; however, after










adjustment for litters to a maximum of 10 pups the two lines

had nearly an equal starting point. Survival to weaning in

line 1 was 89.3% and in line 2, 91.5% resulting in line 2

having an advantage over line 1 (8.81 vs. 8.70) in litter

size at weaning.


Total Litter Weight at 42 Days of Age

Litter weight is a good indication of total perform-

ance since it combines litter size, survival and growth

rate. The least squares analysis for this trait is

presented in Table 15. Factors affecting this trait sig-

nificantly were generation (P < .05), diet (P < .01) and

their interaction G x D (P < .01).

Diet effects again were highly significant with least

squares means of 208.96 grams for diet 1 and 119.10 grams

for diet 2. Generation effects also were significant as

were G x D interaction effects. Means for generations

(Table 16) in diet 1 became progressively larger while

means for generations in diet 2 declined, particularly those

for L1/D2. The L x D interaction component approached sig-

nificance (P < .10).

As discussed in the section on litter size at wean-

ing, line 2 females appeared to be superior in maternal

performance to line 1 females. Even though line 2 females

gave birth to smaller litters adjustment to 10 pups per

litter at birth and superior survival to weaning resulted

in larger litters at 42 days of age for line 2 than for

line 1.


___














TABLE 15. LEAST SQUARES ANALYSIS OF VARIANCE FOR TOTAL
LITTER WEIGHT AT 42-DAYS OF AGE, MODEL 2



SOURCE df SS MS F


Generations (G) 4 14,878.29 3,719.57 2.69*

Lines (L) 1 9.33 9.33 .01

G x L 4 5,529.31 1,382.33 .76

Sires/GL (S) 44 80,141.40 1,821.40 1.37

Diet (D) 1 281,456.04 281,456.04 173.66**

D x G 4 33,993.65 8,498.41 5.24**

D x L 1 6,046.94 6,046.94 3.73***

D x G x L 4 9.033.30 2,258.33 1.39

D x S 44 71,311.20 1,620.71 1.22

Within 67 89,154.54 1,330.66


* P < .05

** P < .01

***P < .10








50


I in 0 --I N H 0 H
N n o N Cm m r- -o -i

Ic.0 o 0 r- Ln in m m
S+1 +1 +1 +1 +1 +1 +1 +1 +1
w cO (N CO o Wo o 0o m
a r- o m r- 4l o
E-, 4 . . ..
In in N oin m co n
H N H N H -rl rH -l

Nm o m ON 0m CO in i

H 0 -i 3 r-I r-H r-i CO
E-H r r- r-l H- rH l
H +i' +1 +1 +1 +1 +1 +1 +1 +1
N mco a I' in in o N

-l 04 -- . *
0 N
O- m "a cr m m -,

OW N N H H H N
0
O
. H- m H CO o O( O in
O o m i CO CO H-

'o co H o0 C H (N t
c/) 0 r r-- i-l O- < l rOl
I r +1 +1 +1 +1 +1 +1 +1 +1 +1
Ni IHN H(N Nin H

(N N Nr Ho m O (N rH o C0

H N N H H H N r r

0 0
mO Q ui m r- IO N N
O1 c! CD m 0 r N m m r-l
*H O in o m co rl o o
EH -4 l r Al rH r-
w m o +1 +1 +1 +1 +1 +1 +1 +1 +1
OH p 0 0 (N (N O (N o 0 o


S a) o o N aco in o Hl m
Q I U (N OH 0 H CO C4 rH H C"
r H N C) H HIl
ElH
S (N i i H in 0
V cin m -i imn CD m
I'D 00 m o r co in Co co

, *' I r- H
c o i +1 +1 +I +1 +1 +1 +1 +1 +1
Z 0 N C. CN in in NO N o
< W CN cO N CO in H N (N
O N o ( 0 N 0 a r
cn cN N- ( N o M in







EH v9 +1 + 1 +1 +1 +1 +1 +1 +1 +1
c0 co T 0 .0 Lo Ci m i co
< n in i in -l r 0 m
in in N L in H H in Co
S in c in co r- o r-
(N lrl l H Hl HM H ( Hl

rCl
ia rH N Hr. N U)
H cia a a)
Hm 0 0 0 o
O "* iIQ rl ^I M Mr- i 11 ( -










Selection for 42-day weight in L1/DI resulted also in

an increase in total litter weight in diet 1. The

correlated response in diet 2, however, was negative and

large. The responses were linear in both cases with a

gain of 15.74 grams in diet 1 and a loss of 68.06 grams in

diet 2.

As line 1 mice became heavier, litter size increased

in both diets. Survival of the pups, however, decreased

(Table 14). The resulting decreased litter size at wean-

ing explains why at generation 8, line 1 was below line 2

in total litter weight in both diets. This occurred even

though individual 42-day weights were heavier in line 1

than in line 2.

L2/D2 mice maintained a near constant level of per-

formance during the five generations. The lack of direct

response in this line apparently resulted from the

nutritional level being below the threshold required for

phenotypic expression of genotype. The effectiveness of

selection in line 2 in changing genotype, however, was

evident from the correlated response in L2/D1. The

superior survival to weaning for L2/D1 over Ll/D1 and

their similarity in average 42-day weights enabled L2/D1

to surpass the performance of L1/D1 in generation 8.

Superior mothering ability of low plane females was

encountered by Falconer and Latyszewski (1952) who

described them as heavier with less fat and more protein

when reared on the high plane.









Productivity as measured by total litter weight at 42

days of age responded well to selection for 42-day weight

in the optimum environment, diet 1. There was a negative

correlated response in productivity when the line selected

under optimum environment was reared under the unfavorable

environment of diet 2.

In the unfavorable environment, productivity was not

measurably responsive to selection for increased 42-day

weight on diet 2. There was a highly positive correlated

response in productivity when the line selected for

increased 42-day weight on diet 2 was reared in diet 1, the

optimum environment. The correlated response in diet 1 to

selection for increased 42-day weight in diet 2 was as ef-

fective as the response obtained from selection for increased

42-day weight in diet 1.

Among the most interesting results from.this trial

was that selection for increased body weight under an un-

favorable nutritional regime resulted in better total pro-

ductivity in a favorable regime than selection for the same

trait under an optimal nutritional plane.


Weaning Weight (21-Day Weight)

Individual weaning weights were analyzed using model

3, Table 17. No significant differences were found for

lines, generations or their interaction (L x G). Sire ef-

fects within line and generation approached significance

(P < .10).

Table 18 contains least squares subclass means and














TABLE 17. LEAST SQUARES ANALYSIS OF VARIANCE FOR WEANING
WEIGHT, MODEL 3



SOURCE df SS MS F


Generations (G) 4 250.60 62.65 2.19

Lines (L) 1 15.91 15.91 .55

G x L 4. 58.03 14.51 .51

Sires/GL (S) 41 1174.52 28.65 1.52***

Diet (D) 1 5703.98 5703.98 302.40**

D x G 4 309.38 77.34 4.10**

D x L 1 22.88 22.88 1.21

D x G x L 4 38.33 9.58 .60

D x S 41 773.30 18.86 1.18

Dams/SDGL (I) 58 927.16 15.98 24.97**

No. Weaned (Linear) 1 149.62 149.62 233.80**

No. Weaned (Quadratic) 1 51.07 51.07 79.80**

Sex (Z) 1 9.07 9.07 14.20**

Z x G 4 2.76 .69 1.07

Z x L 1 .49 .49 .76

Z x D 1 12.17 12.17 19.01**

Within 1141 734.48 .64


* P < .05

** P < .01

***P < .10








54







SC H C.. N
S0 1-1 0 o o
+1 +1 +1 +1 +1 +1 +1 +1 +1
3- H ( m I n co




H r m H (N H H N Hn




co +1 +1 +1 + +1 +1 +1 +1 +1

N nu 0 0 (m m o H




O
1 -




co H ( N H (N H H (N H

N +1 +1 +1 +1 +1 +1 +1 +1 +1
O (N (N N H N C) H C)
0 co . 0 0
>-)- rAl l rl -l








0
ScoN I m r m ( )n



H M H N r l
H Z +1 +1 +1 +1 +1 +1 +1 +1 +1
0 M N C- C- r- C) CO (- o
E -i I n (N Co N t 'r C4 O C D oo



W6

0 r8 en rl ] t o 4 rI ^>








(nN C m o O I C H
Hi 04 lo U) m r o L a
H (N (U H H H H H








n +1 ;4 +' +* +1 +1 +1 +1 +1
C)00 C C) H m C 0 m n
H H H N Hr N
( C m en Cn Cn a)
E-4 d N Ln o i m m m rl






W +i1 +1 +1 +i +1 +i +i +1 +1
o c ( No ow N N


id 0 rl r l)
01











* -i H H ( N H (N -I (N H
H l



< -d Q l- r- M M l- M ~) TI
E-i i-) ia i- a ? il Q Q a










standard errors for weaning weights by line, diet and gener-

ation. Diet effects were highly significant (P < .01) as

would be expected because of the superior nutritional plane

provided by diet 1.

A significant (P < .01) D x G interaction resulted

from weaning weights in diet 1 increasing from generation 4

through generation 8, while in diet 2 weights were decreas-

ing. The increase in weaning weights in diet 1 was larger

than the one obtained by Falconer (1955), where in 12 gener-

ations of selection for increased 42-day weight the total

increase in weaning weight was .75 grams. L1/D1 increased

2.83 grams from generation 1 to generation 8. The corre-

lated response in L2/DI was positive also and slightly

larger for generations 5' to 8 than in L1/D1. Falconer

(1960a) obtained similar results when he reared a low-plane

selection group in a good environment. Weaning weight in

L1/D2 showed a negative response to selection for 42-day

weight in diet 1. These results are in agreement with those

reported by Sutherland et al. (1970) in rats, Falconer (1955)

in mice and by Park et al. (1966) in rats. The only nega-

tive correlated response in this study occurred for weaning

weight in L1/D2. The L2/D2 mice exhibited no apparent re-

sponse in weaning weight to selection for increased 42-day

weight.

Superior weaning weights for line 2 in both diets at

generation 8 most likely are explained in part by a progres-

sive loss of milking ability in line 1. This loss probably


__










was the result of increased fatness in line 1. M.G. Godwin

(unpublished data) compared milk yield of females from gen-

eration 9 of this project with that of females from the

random-bred base population. She found that L1/D1 females

produced 25% less milk per day than base population females.

L2/D2 females produced more gross milk than L1/D1 even

though they were on a lower plane of nutrition. This loss

in milk yield in line 1 apparently was more pronounced in

diet 2 where both survival and the weaning weight of the

surviving pups were reduced drastically during the latter

generations. Brumby (1960) reported a loss of lactational

performance in lines of mice selected for increased or

decreased 42-day weights. Cox et al. (1959) reported that

postnatal environment accounted for 60% of the variance of

21-day weights, even when litters were standardized to six

young at birth.

Regression of weaning weight on number weaned, both

linear and quadratic, was highly significant (P < .01).

These significant regressions.(-1.93.27 and .07.02) indi-

cate a dependency of weaning weight on the amount of milk

available to the pup. Significant dam effects which re-

mained with number weaned in the model suggest significant

individual differences between dams for milk production.

Effects for sex and sex x diet interaction were sig-

nificant (P < .01). Male pups were heavier than female

pups across diets; but within diets, male pups were heavier

in diet 1 while females were slightly heavier in diet 2,









resulting in the significant interaction, Table 19. This

interaction apparently was due to the inability of the sexes

to express their differences in growth potential on the low

plane of nutrition provided by diet 2. In diet 1, energy

intake was higher and differences in growth potential were

expressed. This diet induces sexual dimorphism is common

(Korkman, 1957; Butler and Metrakos, 1950) in mice selection

experiments under two planes of nutrition.

These results disagree with Cox et al. (1959) who

reported no sex effects on 21-day weight of mice on a regu-

lar diet.


Individual 42-Day Weight

The trait under direct selection in this project was

42-day weight, equivalent approximately to that of puberty

in farm animal species. This trait is a composite of pre-

weaning and postweaning growth, being influenced by both

the maternal performance of the dam and the genetic growth

potential of the individual. It corresponds roughly to

yearling weight in cattle. This trait was analyzed by fit-

ting model 3. Results are presented in Table 20. Signifi-

cant (P < .05) effects were found for line, generation and

the linear regression on number weaned. Significant (P < .01)

effects included diet, diet x generation, dams, sex and sex

x diet.

Least squares means for 42-day weight are presented

in Table 21. Line 1 mice were heavier than line 2 mice in

both diets by an average of .75 grams. Generation means
















m n in in 0 CN

0 0 .
H a +1 +1 +1 +1 +1
S H n0 rH 0
ids o mo C H



( m
IM 44 CM CN (N C1 (4



OH > a~ in 0 (N



S H
H I +1 +1 +1 +1 +1
0 en in T m














P: H H
O n m r, (N in





0 H
I 0 0), in r,










00 H
H +1 +1 +1 + +1







0 H



4 30) N C
SrN n en N H
W 0 *
F<;






ca 9
H I +1 +1 +1 +I +I

N m a m N
W gi < N H M
Ua H 7


a (N O (N H CO


Ho +1 +1 +I +i +1
H 4J ) en 0 ip co ko

S* i in r- i 0 m









p N en o i
C ch o 0N iN cn





Cd Id

Hi i-l i-l 0 0












TABLE 20. LEAST SQUARES ANALYSIS OF VARIANCE FOR
42-DAY WEIGHT, MODEL 3


INDIVIDUAL


SOURCE df SS MS F


Generations (G)

Lines (L)

G x L

Sires/GL (S)

Diet (D)

Dx G

D x L

Dx GxL


Dx S

Dams/SDGL

No. Weaned

No. Weaned

Sex (Z)

Z x G

Z x L

Z x D

Within


(I)

(Linear)

(Quadratic


4

1

4

41

1

4

1

4

41

58

1

) 1

1

4

1

1

1141


402.57

117.77

18.00

1204.39

12818.19

1104.61

42.99

108.45

716.22

1393.10

16.75

.31

3497.79

21.59

11.13

283.07

3466.17


100.64

117.77

4.50

29.38

12818.19

276.15

42.99

27.11

17.47

24.02

16.75

.31

3497.79

5.40

11.13

283.07

3.04


3.42*

4.01*

.15

1.22

533.65**

15.81**

2.46

1.55

.73

7.90**

5.41**

.10

1150.60**

1.78

3.66

93.12**


* P < .05

**P < .01








60







> m oN n U i (Y) m Co 01 co
S< ,-4 cN -l -l o -l H- o

S +1 +1 +I +i +i +1 +1 +1 +1


0 rU ) n r-N C D1 r-i U N o


H 0) H 0 0H m H m 0 )


H m m N m N N (N

H 0c +1 +1 +1 +1 +1 +1 +1 +1 +1
H H N H 0 f H m o
a
2 N W H N 0(
)D r -I ko ko o r

0 C m






(N m N (N N m (N
N r-- C N r-- H

S r- +1 +i +i +i +i +i +i +1 +1
m V m en oV m V4 u
0) C I rHl Hl N o Io rH ao

;O N -H CN iH N (N N Hl 0
Z H
W E-4

d C H o N O C N C9 D
C rHI O m N m N 0N N
0
Z LD +1 +1 + + 1 +1 +1 +1 +1 +1
4J 0 a) H m H re) (N
V a ) m (r- N N O r-H

d r( N r(Hl c in N H
Q N H CN r (N N (N -1 N



Z C ( H N E)
Z I Nf e m ( N 1-1 04 ( H
L i +1 +1 +1 +i +i +1 +1 +i +I
H C N (N C ( ko Cm m H-l
ul k_ co m O 4 m M 00

K N C N H C) M C)
co rM n rc rml o M o o





(N iH N -H N N N H N
r: +1 +00 + + 0 1 + 4 0+ +i -M



4 (q N H N H H N (N (N H +N




(I) + l 0 H









increased steadily from generation 4 to generation 8. Diet

1 mice averaged 7.72 grams heavier than diet 2 mice.

A significant diet by generation interaction (P < .01)

was caused by a linear increase for generation means in diet

1 while a slight decrease occurred in diet 2.

Dams had a persistent significant (P <..01) effect on

42-day weight of their progeny. Since regressions for

number weaned (linear and quadratic) were in the model, this

effect for dams was independent of litter size. Rutledge

et al. (1972) concluded that growth in mice until about 7

weeks of age is influenced more by postnatal maternal dif-

ferences than by direct genetic differences. Jinks and

Broadhurst (1963) stated that maternal effects in rats are

more important at 50 days of age than at weaning (21 days).

The significant (P < .01) linear regression of 42-

day weight on number weaned was -.65 .37 grams for each

mouse weaned in the presence of a nonsignificant quadratic

of -.005 .03. This agrees with Roberts (1966) and

Falconer (1960b) who reported regressions of -.6 and .34

grams per mouse born. Roberts (1966), however, reported

that corrections for this regression were so small that no

conclusions were affected. At 42 days of age, males were

3.48 grams heavier than females (P < .01).

Significant (P < .01) sex x diet interaction was the

result of sex difference being larger in diet 1 than in

diet 2. In diet 1, males and females weighted an average









of 27.49 and 22.98 grams, respectively, for a difference of

4.51 grams. In diet 2, the respective means for the two

sexes were 18.74 and 16.29 for a difference of only 2.45

grams.

Direct response to selection for 42-day weight in

line 1 was positive. Correlated response in L1/D2, however,

was slightly negative. In L2/D2, the direct phenotypic

response to selection was slightly negative. Correlated

response in L2/D1, however, was positive and almost as ef-

fective as direct selection in diet 1. L1/D1 had a direct

response of 4.48 grams while L2/D1 had a correlated

response of 4.26 grams. As was the case for 21-day weights

the nutritional regime in line 2 apparently prevented the

expression of any increase in genotype for growth rate. The

positive correlated response in L2/DI, however, indicates

that animals with superior growth rates were selected as

breeders.

Diet 2 apparently provided a more severe nutritional

stress than Falconer and Latyszewski (1952) provided in

their poor environment. They were able to obtain a response

in the poor environment; however, the mice were on the poor

environment for only 3 weeks (3 to 6 weeks of age). Failure

to show response to selection under a low plane has been

demonstrated in mice by Korkman (1961), Bateman (1971) and

Dalton (1967), and in rats by Park et al. (1966). The

high correlated response in L2/DI to selection for increased

42-day weight in diet 2 was very similar to the one reported

in mice by Falconer (1960a).









The lack of significant line x diet and sire x diet

interaction effects together with heavier weights for line

1 in both diets indicates an absence of genotype x environ-

ment interaction for this trait.


60-Day Weight

This is the weight at which the mice in this project

were bred. It was assumed to be sufficiently removed from

maternal environmental effects to yield a good estimate of

additive nonmaternal effects for weight.

Table 22 shows the least squares analysis for this

trait. Line, generation, diet, diet x generation, dam,

sex, sex x line and sex x diet all had highly significant

(P < .01) effects. Sire effects were significant (P < .05).

Generation means increased steadily for the duration

of the project. Line 1 mice were significantly heavier than

line 2 mice in both diets. Diet 1 mice were 7.56 grams

(Table 23) heavier than diet 2 mice. The large increase in

generation means in diet 1, while generation means in diet

2 exhibited a small increment,resulted in a highly signifi-

cant diet x generation interaction.

The effect of dams persisted at 60 days of age, even

though the effects due to regressions of 60-day weight on

number weaned (linear and quadratic) were small.

Sex effects were highly significant (P < .01) with

males being heavier than females. Sex effect differed with

line resulting in a sex x line interaction (Tables 19 and














TABLE 22. LEAST SQUARES ANALYSIS OF VARIANCE
WEIGHT, MODEL 3


SOURCE df SS MS F


Generations (G) 4 1384.60

Lines (L) 1 261.93

G xL 4 9.94

Sires/GL (S) 41 1363.18

Diet (D) 1 12308.52

D x G 4 1113.25

D x L 1 30.57

D x G x L 4 55.41

D x S 41 835.33

Dams/SDGL (I) 58 1608.58

No. Weaned (Linear) 1 6.56

No. Weaned (Quadratic) 1 1.02

Sex (Z) 1 8691.02

Z x G 4 17.72

Z x L 1 32.83

Z x D 1 352.46

Within 1141 4605.77


346.15

261.93

2.48

33.25

12308.52

278.31

30.57

13.85

20.37

27.73

6.56

1.02

8691.02

4.43

32.83

352.46

4.04


* P < .05
**P < .01


FOR 60-DAY


10.41**

7.88**

.07

1.63*

604.20**

13.66**

1.50

.50

.73

6.86**

1.62

.25

2151.20**

1.10

8.13**

87.20**








65








CQ N C "I 0 -H 1
U0 (N' H H H H H r- 0

S+1 +1 +1 +1 +1 +1 + 1 + +1 +1

0 Gm m a m

SnN N N N N n CM N


0 H i I m y Ha%
I ( m c m N m rN H

S+1 +1 +I +I +1 +1 +1 +1 +1

Kl) in N H if W H

mm a ) m a NM
SN m IN N N IN N






Pi .1o m r4 C co r m o
C0 m W m IN N N N N









r 0 r- r m i
0 OD m m N N
4. 0- +1 +i +I +1 +1 +I +1 +1 +-
S i n O in Cl c N -I Hl



-W Cl I I co NO C ) co






00 N In C N N N m IN (N
ID D N H

S 0 N IV l IN l iN N IN H






i 1l +1 + +1 +1 +1 +1 +1 +1
0N Iz I' cO 0 o n NO Co Co No co










Sn, ~ l N 0 10
a; Q a) . .
HZ W 0 IN 0) IN in in 0 IN in











Cl N N4N N N IN N IN CN


C a O m am C 0
0 r







5H
Lain in H IN -4 C

KI
El i N in CO CN il
H H H in N- in n tr' Oi i


N co m Cl N C3 i in i N C Ln
1N IN IN IN IN r' IN IN







0 iI Hr HIN IN Hl N H IN rHl-
EH









22). The difference between male and female weights in line

1 was larger than in line 2. Moore, Eisen and Ulberg (1970)

encountered a similar sex x line interaction involving lines

of large mice. The response may be the result of scale ef-

fects increasing the magnitude of the differences between

the sexes in the larger lines. A significant (P < .01) sex

x diet interaction likewise resulted from a larger sex dif-

ference in diet 1 than in diet 2.

Sire effects for 60-day weight were significant

(P < .05). Absence of sire effects on growth traits taken

at younger ages (weaning and 42-day weight) points to the

higher maternal components present in those traits. Variance

due to dams is the within variance of sire groups. The

magnitude of this variance in the younger weights in combi-

nation with small number of dams per sire'was responsible

for nonsignificant sire effects. The presence of signifi-

cant sire effect on 60-day weights is the result of a

reduction of within sire variance due to a smaller dam ef-

fect at this age. This point will be discussed further in

the section on genetic parameters.

As was the case for 42-day weight, there was no indi-

cation of diet x genotype interaction present for this

trait.


Genetic Parameters

Response to selection and genetic parameters can be

ascertained by various techniques. The three approaches

used in this study were, (1) to compute selection differ-









entials, selection intensities and realized heritabilities

for the traits under selection, (2) to obtain estimates of

observational and causal components of random variances and

(3) to obtain estimates for heritabilities and realized

genetic correlations for individual weights at weaning, 42

days and 60 days of age.

Since sires and dams were completely confounded in

L1/D2 and L2/D1, the estimates for random variances were

obtained only from the eight generations of L1/D1 and L2/D2.

Data from each line was analyzed by fitting model 4.


Realized Heritabilities for 42-Day Weight

This parameter is the ratio of the response to selec-

tion divided by the cumulative selection differential. Re-

gression of least squares generation means on generation

number (Table 24) was considered the response for each

trait. Regression of generation means of cumulative selec-

tion differential on generation number was the measure of

selection differential. The responses per generation for

the two lines were .87 and .19 grams, respectively. The

corresponding selection differentials per generation (Table

25) were 2.61 and 2.39 grams. Realized heritabilities were

.33 and .08, respectively for the two lines. These esti-

mates are similar to those found in the literature. Korkman

(1961) obtained estimates of .22 and .06 on optimum and

restricted regime, respectively. Park et al. (1966) in rats

obtained a h2 estimate of .105.and .058 on a full fed and



















z








H










H






OO

H
co

-14




















H


H





0
H














El)
i-






Ha



2


(Q.
I d)
o H d
w hi hi









restricted diets. Kownacki (1971) obtained estimates of

.251 and .135, respectively,for high and low protein diets.

A limited response to selection under poor nutritional re-

gimes appears to be a general response.


Selection Intensities

Intensity of selection depends only on the proportion

of the population included in the selected group, provided

the distribution of phenotypic values is normal. Intensities

for this project appear in Table 25, and were obtained by

calculating the proportion of males and females selected.

These proportions yielded the values of i from the tables

of Fisher and Yates (1943). Sufficient offspring were pro-

duced in each generation by both lines to insure similar

selection intensities.


Relation to Natural Selection

Relation of natural selection to the trait under se-

lection is an important factor influencing response. Ratio

of effective over expected selection differentials, if dif-

ferent than 1, indicates a conflict between the direction

of natural and artificial selection (Falconer, 1964). The

effective selection differential was calculated by weight-

ing the selected parents by the number of offspring pro-

duced. If the more extreme selected individuals have a

tendency to be less viable or fertile, they will produce

fewer offspring than the remainder of the selected breeders.

This may result in a difference between expected and effec-

tive selection differentials.

















re m LfN 0 n m















mo HH H HHH Hm
. 0 N m re m ri










H_ H- H- Hd (N 1(1 H


x



>r


N





H


z
H






W,





0




H
W





H





(12
H





H





C-)


rei r4 r4 C


H (N m .T in rfl


CN Li dm

m <^ (r


~_










The ratios for this project were, 1.0266 and .9962 for

lines 1 and 2, respectively, suggesting no conflict between

natural and artificial selection in this trial.


Observational and Causal Components of Random Variances

Analyses of variance for individual weights appear on

Tables 31 to 36. In line 1, sire effects were positive for

all ages and significant for 42- and 60-day weights, where-

as in line 2 sire effects were not significant for any

weight with a negative sire component for 42-day weight.

Dan effects were significant (P < .01) for all weights in

both lines.

Nonsignificant and negative sire variances in line 2

indicated that sire components were not significantly dif-

ferent from zero. Observational components are listed in

Table 26.

Partitioned dam components were larger for line 2

for all traits, indicating a large common maternal effect

in line 2. Dam variance accounted for most of total random

variance in line 2.

Moore et al. (1970) indicated that maternal effects

in mice were significant at all ages reaching a maximum at

21 days and decreasing as the mice became older. Monteiro

and Falconer (1965) believed maternal effects in mice to

be the most important source of variance up to 4 weeks of

age. Jinks and Broadhurst (1963) reported that maternal

effects in that data were higher for 50-day weight than for













TABLE 26. ESTIMATES OF OBSERVATIONAL COMPONENTS OF VARIANCE
FOR WEANING WEIGHT, 42-DAY WEIGHT AND 60-DAY
WEIGHT, MODEL 4


Weaning Weight 42-Day Weight 60-Day Weight
Source Line 1 Line 2 Line 1 Line 2 Line 1 Line 2

Components

Sires .15 .07 .40* -.20 .73** .06

Dams/Sire 1.90**a 1.47** 1.56** 2.61** 1.71** 2.66**

Within .62 .50 2.70 2.38 3.67 3.06

Total 2.66 2.04 4.65 4.99 6.11 5.77



As % of Total

Sires 5 3 8 0 12 1

Dams/Sire 71 72 34 52 28 46

Within 23 24 58 48 60 53

Total 100 100 100 100 100 100


aSignificance of

*P < .05

**P < .01


tests of ratios of mean squares















TABLE 27. ESTIMATES OF CAUSAL COMPONENTS OF VARIANCE FOR
WEANING WEIGHT, 42-DAY WEIGHT AND 60-DAY
WEIGHT


Weaning Weight 42-Day Weight 60-Day Weight
Line 1 Line 2 Line 1 Line 2 Line 1 Line 2

Components

Phenotypic
(Vp) 2.66 2.04 4.65 4.99 6.11 5.77

Additive
(VAo) .58 .27 1.59 0 2.93 .24

Maternal Effect
(V ) 1.75 1.40 1.16 2.61 .97 2.60

Within-Litter
(VWL) .33 .37 1.90 2.38 2.20 2.94


As % of V

Vp 100 100 100 100 100 100

VAo 22 13 34 0 48 4

VME 66 69 25 52 16 45

VWL 12 18 41 48 36 51










21-day weight. This study agrees with these authors in

regard to line 1; however, line 2 presents a case not cov-

ered in the literature. Maternal effects persisted up to

60-day weight as 45% of the total phenotypic variance. Line

1 behaved similarly to the mice of Rutledge et al. (1972),

who concluded that up to 7 weeks of age growth was influ-

enced more by postnatal maternal differences than by genetic

differences. VAo for line 1 was 34% and VME 25% at 42 days

of age or of about equal magnitude.

Within-litter variances were smaller in line 2 as a

result of scale effects since all traits in line 2 were

smaller in magnitude. Within-litter variance as a percent

of total random variance increased as mice became older.

Causal components were estimated from observational

components, Table 27, using the method of Falconer (1963).

Additive variance (VAo) components in line 2 were

small. In line 1, VAo components were larger and accounted

for a greater percentage of total phenotypic variance (Vp)

as mice became older. Schmitz (1970) working with the

Goodale large strain of mice estimated values of VAo/VP at

28, 42 and 60 days of age to be 17.6%, 21.6%, and 33.9%,

respectively. Line 1 in this project had values of 22%,

34% and 48%, respectively.

Maternal effect (VME) accounted for most of the total

phenotypic variance in line 2. Maternal effects were large

in line 1 but decreased markedly with age. Within-litter

environmental variance (VWL) increased with age in both

lines.









Total phenotypic variance was smaller for line 2 than

for line 1 in all traits with the exception of 42-day weight.

This is in disagreement with results from other selection

experiments on different nutritional regimes. McLaren and

Michie (1956) in mice, Becker and Berg (1959) in poultry,

and Bateman (1971) in mice all observed higher phenotypic

variances on the low plane of nutrition.


Heritability Estimates

When sire variances are not significant, heritability

estimates are not reliable. This is evident from the

standard errors shown in Table 28. Only two estimates were

significant, those for 42- and 60-day weights in line 1.

Lack of significant sire mean squares in most traits

made it difficult to estimate additive genetic variances and

heritabilities. Large maternal effects between litters ac-

counted for most of total phenotypic variance in all traits

measured. Line 2 mice were particularly affected by

maternal environment.

Under a poor environment VME account for most of the

Vp. This indicates a major role played by maternal effects

in selection under a poor environment. The high correlated

response in diet 1 for selection in diet 2 demonstrates an

additive maternal component. The correlated response was

cumulative and nearly as effective as direct response in

diet 1. This mechanism is in general agreement with the

views of M. Koger (personal communication) who proposed
















TABLE 28. WITHIN-LINE HERITABILITY ESTIMATES FOR WEANING
WEIGHT, 42-DAY WEIGHT, AND 60-DAY WEIGHT, MODEL 4


Line 1
h2 s.e.


Line 2
h2 s.e.


Weaning Weight
21-Day Weight .2145 .3378 .1309 .3449

42-Day Weight .3422* .2210 -.1692 .2351

60-Day Weight .4798** .2240 .0407 .2334

K 3 8.8500 7.5382

K 4 8.9880 7.7749

K 5 22.2225 19.0556


* P < .05

**P < .01










that adaptability to unfavorable environments generally

behaves essentially as a maternal trait.

It has been postulated (Eisen, 1973) that rapid growth

enhanced by the mother's high lactational output is one way

in which the dam could increase the offspring's fitness.

Natural selection among dams for the ability to provide a

superior maternal environment (Naylor, 1964) may have

resulted in a reduction of additive nonmaternal genetic var-

iance. Other works with diallel crosses designed to parti-

tion maternal effects into its components and their inter-

actions have been published. Cock and Morton (1963)

demonstrated line maternal effects in poultry up to 20 weeks

posthatching. Morton (1970) in mice postulated an inter-

action between genotype of a sire strain and cytoplasm of

the female strain. This has been demonstrated with ova

transplants (Brumby, 1960) in mice where cytoplasmic factors

were apparent. Non-Mendelian transmitted maternal effects

were found in the Falconer and Goodale mice lines (Reutzel,

1970). Incorporation of genetic material from the cytoplasm

into the genome with a resulting additive genetic action can

not be discarded as a possibility for acquired fitness.


Realized Genetic Correlations

Estimates of realized genetic correlations (rA) from

the data are presented in Table 29. Only in line 1 could

the estimates be computed. The trait under selection, 42-

day weight, in line 2 had an estimated additive genetic

variance of zero.


__ __










The rA between 42-day weight and 60-day weight was

1.04.03, indicating that genes influencing 42-day weight

were also influencing 60-day weight. This was to be expected

since a part-whole relationship existed.

Genetic correlation between 42-day weight and weaning

weight was .57.48. Larger standard error and smaller mag-

nitude of the estimate as compared with that of 60-day

weight was the result of environmental variations in weaning

weight.

Phenotypic (.73 and .50) and genetic (1.04 and .57)

correlation coefficients were similar in magnitude when 42-

day weight was correlated with 60-day weight and weaning

weight in line 1. Phenotypic correlations appear in Table

30.

Positive phenotypic and genetic correlations in mice

between early postweaning growth (21 42 days of age) and

late post-weaning growth (42- 60 days of age) were reported

by Wilson (1973). He found a genetic correlation of .9

between early and late growth in mice selected for early

growth.

Data from GEI were not suited for genetic correlation

estimates. Additive variance was 0 for 42-day weight in

line 2, and sires and dams were completely confounded in

L1/D2 and L2/D1.






































m r-i -
m rN CN







r- CT




L -IA 1 1
H-- (N


tC











r-
(cl











m

0t

1-





Ui


N4 0





HO 0
oo
r-l d


.-







--4
-ri
.rl
*H

Q)
4)


H,
C




-s-I
Q)






















U4
Irl












.H


0



















0
U)
cU
a


Ul










4J
r-H











0)


4-)
iU



-)
.

(Ni







t)










*,
r a
A r
3l


NC






1-4








0 4
\ O0


0 4


am
S0












-,--

EH
'NO


















TABLE 30. WITHIN


LINE CORRELATIONS BETWEEN WEIGHTS AT
DIFFERENT AGESa


Weaning 42-Day 60-Day
Weight Weight Weight



Weaning
Weight -- .64 .54


42-Day
Weight .50 -- .82


60-Day
Weight .38 .73





Correlations within line 1 are to the left of the
diagonal, those for line 2 to the right.


----















TABLE 31. LEAST SQUARES ANALYSIS OF VARIANCE FOR WEANING
WEIGHT IN LINE 1, MODEL 4



SOURCE df SS MS F


Generations (G) 7 1376.53 196.65 9.40**

Sires/G (S) 42 878.23 20.91 1.20

Dams/S (I) 73 1270.58 17.40 28.06**

No.Weaned (Linear) 1 3.07 3.07 4.95*

No.Weaned (Quadratic) 1 1.10 1.10 1.77

Sex (Z) 1 34.36 34.36 55.42**

Z x G 7 4.93 .70 1.13

Within 980 610.96 .62


* P < .05

**P < .01














TABLE 32. LEAST SQUARES ANALYSIS OF VARIANCE
WEIGHT IN LINE 1, MODEL 4


FOR 42-DAY


SOURCE df SS MS F


Generations (G) 7 5342.99 763.28 29.88**

Sires/G (S) 42 1072.55 25.54 1.55*

Dams/S (I) 73 1202.95 16.48 6.10**

No. Weaned (Linear) 1 12.20 12.20 4.52*

No. Weaned (Quadratic) 1 .38 .38 .14

Sex (Z) 1 4687.10 4687.10 1735.96**

Z x G 7 140.48 20.07 7.43**

Within 980 2643.44 2.70


* P < .05

**P < .01
















TABLE 33. LEAST SQUARES ANALYSIS OF VARIANCE
WEIGHT IN LINE 1, MODEL 4


FOR 60-DAY


SOURCE df SS MS F


Generations (G) 7 8468.80 1209.83 34.27**

Sires/G (S) 42 1482.67 35.30 1.88**

Dams/S (I) 73 1370.98 18.78 5.12**

No. Weaned (Linear) 1 16.39 16.39 4.46*

No. Weaned (Quadratic) 1 5.00 5.00 1.36

Sex (Z) 1 10189.80 10189.80 2776.50**

Z x G 7 208.2 29.73 8.10**

Within 980 3598.76 3.67


* P < .05

**P < .01















TABLE 34. LEAST SQUARES ANALYSIS OF VARIANCE FOR WEANING
WEIGHT IN LINE 2, MODEL 4



SOURCE df SS MS F


Generations (G) 7 257.86 36.84 2.79*

Sires/G (S) 38 501.88 13.21 1.14

Dams/S (I) 67 776.74 11.59 23.18**

No. Weaned (Linear) 1 64.48 64.48 128.96**

No. Weaned (Quadratic) 1 25.85 25.85 51.70**

Sex (Z) 1 10.12 10.12 20.24**

Z x G 7 6.62 .94 1.88

Within 755 377.52 .50


* P < .05

**P < .01

















TABLE 35. LEAST SQUARES ANALYSIS OF VARIANCE FOR 42-DAY
WEIGHT IN LINE 2, MODEL 4



SOURCE df SS MS F


Generations (G) 7 649.20 92.74 4.93**

Sires/G (S) 38 714.53 18.80 .86

Dams/S (I) 67 1476.37 22.04 9.26**

No. Weaned (Linear) 1 31.64 31.64 13.29**

No. Weaned (Quadratic) 1 5.25 5.25 2.20

Sex (Z) 1 914.13 914.13 384.09**

Z x G 7 32.84 4.69 1.97

Within 755 1798.45 2.38


* P < .05

**P < .01


--
















TABLE 36. LEAST SQUARES ANALYSIS OF VARIANCE FOR 60-DAY
WEIGHT IN LINE 2, MODEL 4



SOURCE df SS MS F


Generations (G) 7 867.85 123.98 4.99**

Sires/G (S) 38 943.42 24.83 1.08

Dams/S (I) 67 1546.12 23.08 7.54**

No. Weaned (Linear) 1 66.36 66.36 21.69**

No. Weaned (Quadratic) 1 28.33 28.33 9.26**

Sex (Z) 1 2618.73 2618.73 855.79**

Z x G 7 54.77 7.82 2.56*

Within 755 2314.55 3.06


* P < .05

**P < .01


--- --















SUMMARY AND CONCLUSIONS

An investigation was performed to evaluate response

to mass selection in two lines of mice (L1, L2) for in-

creased 42-day weight on two diets (Dl, D2). Diet 1 was

Purina Lab Chow and D2 consisted of 30% nonnutritive bulk

and 70% Purina Lab Chow. After 3 generations of selection,

mice from the original two lines were transferred to the

opposite diet forming Ll/Dl, L1/D2, L2/D1 and L2/D2. Line

1 was originally selected in Dl and line 2 in D2.


Analyses of Data From Four.Subgroups, Generations 4 to 8


Pregnancy Rate

The overall least squares mean for pregnancy rate in

240 females mated to 80 males was 84.58%. The only vari-

ables having significant effects (P < .05) were generation

and sires. Sire effects appeared to be due to random

sterility while generation means were affected both by this

factor and by a nonspecific pathogen encountered in genera-

tion 7.


Litter Size at Birth

There were 175 litters by 54 sires involved in the

analyses of the three litter traits. The overall least

squares mean was 9.62. Significant effects (P < .05) were


__~ __










found for line (10.18 vs. 9.06), diet (10.12 vs. 9.12) and

D x G. Line differences probably were attributable to

larger weights in line 1 as suggested by a significant posi-

tive within-line regression of litter size on weight of dam.

Litter size in L1/D1 and L2/D1 increased over time while

decreasing slightly in L2/D2 and L1/D2, resulting in a sig-

nificant D x G interaction.


Survival and Litter Size at Weaning

Survival to weaning was higher in line 2 than in line

1 (82.7% vs. 79.6%). Thus, at weaning time the two lines

had changed rank in litter size with line 2 having a non-

significant advantage over line 1 (7.80 vs. 7.67). As ex-

pected, there was a significant difference (P < .01) in

litter size due to diet (8.42 vs. 7.04). Diet x line inter-

action was nonsignificant.


Total Litter Weight at 42 Days of Age

Lines averaged across diets did not differ signifi-

cantly for total litter weight (164.30 vs. 163.76 grams).

There was an interaction of line with diet (P < .075), how-

ever, with the lines changing rank on the two diets (L2/D1,

L1/D1, L2/D2, L1/D2 = 235.86, 231.32, 118.62, 87.48 grams).

During generations 7 and 8 line 2 females produced heavier

litters than line 1 in both diets for a complete reversal

from generation 4.










Individual Weaning Weights

Records from 1312 mice, 51 sires and 162 dams were

utilized in the analyses of the three individual weights.

Sub-line group means varied from a low of 6.71 grams for

L1/D2 to 12.17 grams for L1/D1, with average values of 9.44

for line 1 and 9.71 for line 2. Diet means were 12.15 and

7.00 grams (P < .01), respectively. Sire effects were sig-

nificant (P < .10) as were dam effects (P < .01). Sex x

diet interaction means were significant (P < .01) with males

being heavier on diet 1 (12.34 vs. 11.95 grams) and females

slightly heavier on diet 2 (7.02 vs. 6.98 grams). Diet x

generation was significant with generation means becoming

progressively heavier for diet 1 while for diet 2 generation

means declined.


Individual 42-Day Weight

Weight at 42 days showed trends similar to those at

weaning except that the effects of superior preweaning

maternal performance for line 2 had decreased and evidences

of genotype for slightly heavier weights for line 1 had

emerged. Line effects were significant (P < .05) with line

1 being heavier than line 2 on both diets and for all gener-

ations. Diet x generation interaction was significant

(P < .01) with mice on diet 1 becoming progressively

heavier over generations while diet 2 mice declined in

weight. There was a significant sex x diet interaction due

largely to scale effects in the two diets.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs