‘.::A|a‘(l q:-

 

 

   

 

 

 

 

 

    
  

va‘ns

 

 

 

 

.u-uuuu

 

 

 

 

I.

 

 

 

       

 

 

 

 

 

 

   

   

 

 

     

 

       

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ZHEB$
3
[£fo

llllllllllllllllMINI“Illlllllllllllllllllllllllllllllllll

31293 01835 3999

This is to certify that the

thesis entitled

Genetic and composition effects on
reproductive fitness traits in yearling

Limousin cattle.

presented by

‘Z\ DuoPc‘fNE "b. CHVbUc-Y

has been accepted towards fulfillment

of the requirements for
MS degree in MMC
,57/ ‘fl

Major professor

 

 

Dmé/g/W/

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

 

 

 

 

 

memes!
Michigan Siam
University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M wmm.“

 

 

GE

Subr

 

GENETIC AND COMPOSITION EFFECTS ON REPRODUCTIVE FITNESS

TRAITS IN YEARLING LIMOUSIN CATTLE

By

Dwayne D. Faidley

ATI-IESIS

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

MASTER OF SCIENCE
Department of Animal Science

1999

GE

improx
moderz
combir

enhanc

 

Indust ‘

measm
heifers
hOVk‘ex-
in thei:
methoi
SimulL

Geneti

ABSTRACT

GENETIC AND COMPOSITION EFFECTS ON REPRODUCTIVE FITNESS
TRAITS IN YEARLING LIMOUSIN CATTLE

By
Dwayne D. Faidley

Fertility, fleshing ability, and muscularity are three traits vital to breeders of Limousin
cattle. While muscularity is a recognizable breed strength, fertility traits and fleshing ability
are considered breed deficiencies in many production settings. Breeders are striving to
improve inherent fertility through genetic selection, realizing reproductive traits are low to
moderately heritable. Expressed reproductive efficiency is being stressed through a
combination of genetic selection and management using body condition scores. Although
enhanced reproductive fitness is a primary objective, breeders have determined beef
industry necessities dictate preservation of Limousin muscularity advantages.

Age, weight, and increased body condition favorably affected many phenotypic
measures of reproductive traits in both bulls and heifers. Moderate and heavy muscled
heifers were superior to light muscled heifers for yearling fertility traits. Muscular bulls,
however, had the smallest scrotal circumferences. Genetic effects of sire were important
in their male progeny and comparable to similar analyses. Bivariate analyses provided
methods to estimate sire effects on male and correctly account for female traits
simultaneously. Heritability estimates were low, but in the range found in the literature.

Genetic correlations between fertility indicators in yearling heifers and bulls were low.

 

apara.

credit

Kc

 

taught
and f0]

and {hi

0f the
Di

fifiVe

belief

COHVQ

me, E“
Temp

quant:

ACKNOWLEDGMENTS

The following people have made a significant contribution to my life and education and
share credit for all I have learned at Michigan State University.

Dr. B. Dennis Banks is a teacher, scientist and practical man, but most importantly the
type of advisor and friend I attempt to emulate, but have faint hope to equal. He has been
a paragon of character, loyalty, understanding, and patience. Nobody deserves more
credit for my completion.

Ken Geuns knows me as well as anyone east of the Mississippi, and he has probably
taught me as much about cattle as anyone, anywhere. Thanks, Ken, for your friendship
and for the always-welcome opportunity to flee the office and reconnect with livestock
and the land.

Thanks to all the women in these two great men’s lives, for making me feel like part
of the family.

Drs. Harlan Ritchie and John Edwards are beef industry giants who also happen to
serve on my committee. I greatly appreciate your endorsement as I pursued a dream,
belief in my abilities, and enthusiasm for the results. As stockmen, and gentlemen,
conversations with you are always too brief.

Dr. Rob Tempelman has been an invaluable committee member who has challenged
me, yet never withheld assistance regarding the statistical analyses. Thanks to Dr.
Tempelman, I have a deeper understanding and appreciation for all those who seek to

quantify and explain what others merely speculate about.

iii

SUP?“
sari m

student

Doroth;
farm fa,
Fin,

You ft“,

 

There are many who have made the graduate experience more enjoyable through their
support and positive attitude. Thanks to Drs. Ivan Mao, John Shelle and Russel Erickson,
staff members Kim Dobson, Mary Lehnert and Faye Watson, the barn crews and all the
students who brightened my day from time to time.

All the best parts of me spring from the love and direction of my parents, Don and
Dorothy Faidley. Words cannot express my gratitude for growing up part of a devoted
farm family that lived by the principles of faith, hope, and love.

Finally, I need to thank the independent farmers and ranchers out there in the country.

You folks are my inspiration.

 

 

LBT(

LBIt

CHAP

l lNl

2 RE
Re

TABLE OF CONTENTS

LIST OF TABLES -- viii

LIST OF FIGURES ...... - -- - .............
CHAPTER
1 INTRODUCTION and OBJECTIVES -

2 REVIEW OF LITERATURE

 

Reproductive Traits in Beef Females

 

Age at Puberty

 

Pregnancy and Reproductive Performance
Productivity and Longevity

 

Heritability

 

Relationships with Growth
Relationships with Body Condition
Relationships with Muscularity
Reproductive Tract Score
Evaluation System
Pregnancy and Reproductive Performance
Productivity
Genetic Aspects of RTS
Relationships with Growth
Other Factors which Influence RTS
Timing and Appropriate Use
Discussion
Summary

 

 

 

 

 

 

 

 

 

 

 

Threshold Trait Considerations

Logistic Regression
Obtaining odds and odds ratios from logistic regression
Interpretation of coefficients from the logit form of the logistic model .........

 

 

Reproductive Traits in Beef Males
Bull Fertility
Puberty -
Scrotal Circumference and Bull Fertility
Scrotal Circumference Relationships with Age at Puberty
Breed Effects
Heterosis Effects
Inbreeding Effects

 

 

 

 

 

 

 

Rt

.\1

Sir

BiiI

 

Surl

SPH

 

Relationships with Growth A 33

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Relationships with Composition A A AA - - 34
Heritability 35
Relationships of Scrotal Circumference to Female Reproductive Traits ............. 36
Mixed ModelsAAA A AA A A A A A A _- - _ 38
Sire Models A 40
Bivariate Models A AA 41
Summary of Literature ....... . ..... ........ A AA AA AA AA A AA AA 43
Implications A A A A - 44
3 PHENOTYPIC RELATIONSHIPS BETWEEN REPRODUCTIVE FITNESS
AND COMPOSITION TRAITS OF LIMOUSIN CATTLE 47
Abstract 47
Introduction and Objectives A A 48
Materials and Methods 50
Heifer Population 50

Data 50

Heifer Management 51
Reproductive Tract Scoring System 51

Muscle Score 53

Data Editing 53

Bull Population 56

Data 56

Bull Management 56

Data Editing A A 56
Reproductive Tract Score Analysis 58
Scrotal Circumference Analysis 59
Results A A -- 61
Probability of cycling 61
Effects of composition on probability of cycling 61

Scrotal Circumference 69
Composition effects on scrotal circumference 69
Discussion A 73
Implications 75

4 GENETIC AND COMPOSITION EFFECTS ON REPRODUCTIVE FITNESS
TRAITS IN YEARLING LIMOUSIN CATTLE 76
Abstract 76

 

Introduction and Objectives 77

vi

 

BIE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Materials and Methods A 79
Heifer Population 79
Data -- 79

Heifer Management 79
Reproductive Tract Scoring System 80
Muscle Score 82

Data Editing _ 82

Bull Population - 85
Data 85

Bull Management 85

Data Editing - 85
Univariate Analysis - 87
Reproductive Tract ScoreA A 87
Scrotal Circumference 89

Sires 90
Bivariate Analysis 93
BIVARB A 93
Bivariate Model Specification 94
Results 95
Univariate Analysis 95
Fixed Factor Effects on Reproductive Tract Score - 95
Effects of Composition on Reproductive Tract Score 95

Fixed Factor Effects on Scrotal Circumference 96
Composition Effects on Scrotal Circumference 97
Summary of Univariate Fixed Effect Results - 97
Genetic Effects on Reproductive Tract Score - - 100
Genetic Effects on Scrotal Circumference 100
Bivariate Analysis Fixed Effect Estimates 101
Age and Weight Adjusted - BIVARB 101

Age Adjusted - BIVARB - 101
Weight Adjusted - Bivarb - 102
Bivariate Analysis Random Effect Estimates 106
Genetic Effects 106
Environmental Effects 106
Heritability and Genetic Correlations Estimates 107
Discussion 108
Implications 113
5 SUMMARY and CONCLUSIONS AA ...... 114

 

BIBLIOGRAPHY

 

vii

- 116

Table 2.
Table 2.
Table 2.
ka2

Table 2

 

Table 2‘
Table 2.

ka2
Table 2
Table 2

Table 3
Table 3
Table a
Table 3
Table 3

Table 4

 

Table .1
Table 4

Table 4

 

Table 2.1

Table 2.2

Table 2.3

Table 2.4

Table 2.5

Table 2.6

Table 2.7

Table 2.8

Table 2.9

Table 2.10

Table 3.1

Table 3.2

Table 3.3

Table 3.4

Table 3.5

Table 4.1

Table 4.2

Table 4.3

Table 4.4

LIST OF TABLES

 

 

 

 

 

 

 

 

 

 

 

 

 

Correlation of heifer age at puberty with productivity traits ............... 7
Heritability estimates for age at puberty (AP) A A - 8
Genetic correlations of age at puberty with growth traits AA AAAAA 9
Description of reproductive tract scores 12
Effect of reproductive score on reproductive trait means ................... 15
Regression coefficients for heifer performance traits - 16
Heritability estimates for reproductive tract score (RTS) ................... 17
Heritabilities and genetic, phenotypic, and environmental

correlations for heifer growth and reproductive traits A 18
Genetic correlations between scrotal circumference and

measures of growth 34
Heritability estimates for scrotal circumference in yearling

beef bulls ' 35
Description of reproductive tract scores A AAAAAAA - 52
Data description for Running Creek Ranch females 55
Data description for Running Creek Ranch males AA 57
Parameter estimates for probability of cycling 63
Parameter estimates for yearling scrotal circumference 70
Description of reproductive tract scores 81
Data description for Running Creek Ranch females — 135 sires ........ 84
Data description for Running Creek Ranch males - 127 sires ........... 86
Data description for Running Creek Ranch females - 110

paired sires 91

 

viii

Table

Table -

Table .

Table -

Table -

Table -

Table l

 

Table
Table .

Table .

Table

Table

Table 4.5 Data description for Running Creek Ranch males - 110 paired
sires A -- - - - -- - 92

Table 4.6 Models and data sets used for bivariate analysis modeling both

 

 

 

 

 

 

SC and RTS as response variables AAAAA 94
Table 4.7 Univariate type 3 tests of fixed effects A -- AA 98
Table 4.8 Univariate analysis fixed effect parameter estimates - AA - -- -- 99
Table 4.9 Variance parameter estimates - RTS- - 100
Table 4.10 Variance parameter estimates - SCA A AA AA A A 100
Table 4.11 Age & weight adjusted bivariate analysis fixed effect estimates ...... 103
Table 4.12 Age adjusted bivariate analysis fixed effect estimates 104
Table 4.13 Weight adjusted bivariate analysis fixed effect estimates ................. 105

Table 4.14 Genetic (co)variances for scrotal circumference and cycling
status in yearling Limousin cattle 106

Table 4.15 Environmental variances for scrotal circumference and cycling
status in yearling Limousin cattle 107

Table 4.16 Heritabilities and genetic correlations for scrotal
circumference and cycling status in yearling Limousin cattle .......... 108

ix

 

figure

Figure

Figure

figure

Figure

Figure

Flgure .

 

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

LIST OF FIGURES

Probability of cycling for heifers with composition differences
as both age and weight are increased

Probability of cycling for body condition score six heifers with
muscle and age differences

 

Probability of cycling for average muscled heifers with body
condition and age differences

 

Probability of cycling for heifers with composition and age
differences

 

Probability of cycling for heifers with composition differences
as both age and weight are increased

 

Age adjusted composition effects on predicted scrotal
circumference

Weight adjusted composition effects on predicted scrotal
circumference - -

65

66

67

68

71

72

 

Ere
summc
Breed".
efficien
relative
trials an
accepta;

cr0p_ H

 

their ab.
R‘produ.
Bree

for fem]
allows f,
DOt Obse
record f;
of breed,
‘BOUI
envimnn
expregsic

fai'orabi t

CHAPTER 1

INTRODUCTION and OBJECTIVE

Breeders of Limousin cattle took a hard look at breed strengths and weaknesses in the
summer and fall of 1990 through their Directions Symposium. Known as the "Carcass
Breed", both domestically and in their native France, the relative leanncss, muscularity and
efficiency advantages of Limousin cattle are well documented. Likewise, among other
relative weaknesses of the breed, fertility was identified as a major concern through research
trials and production experience. The breed has experienced rapid growth and industry
acceptance due to the ability of Limousin bulls to add muscle and feed efficiency to a calf
crop. However, Limousin enthusiasts realized continued breed expansion depended on
their ability to overcome breed shortcomings. Therefore, Limousin breeders have targeted
reproductive fitness for improvement while maintaining breed muscle advantages.

Breeding for improved fertility is complicated by the fact that underlying genetic merit
for fertility is often not expressed on a continuous scale. Pregnancy status, for instance,
allows for only two possible outcomes: pregnant or nonpregnant; degrees of pregnancy are
not observable. The side of the pregnant/nonpregnant threshold on which an individual
record falls also depends on environmental influences, such as plane of nutrition and length
of breeding season.

'Bourdon and Brinks (1986) reported potentially more important factors are genotype x
environment and genotype x genotype x environment interactions. All else being equal, the
expression of inherent fertility differences is more pronounced in poor environments than in

favorable environments. Differing levels of genetic merit for other traits, such as milk

 

 

prairie:
instance
environ?
than inr.
Gen

vs. care.
cattle b.’
econom;
are affec
muscul»
a
selectior.

Titer

 

 

production and calf growth rate, often obscure inherent and expressed fertility. For
instance, heavy-milking cows with high inherent fertility will calve regularly in good
environments, but in stressful environments they will have poorer reproductive performance
than inherently less fertile, but lighter-milking cows.

Genetic antagonisms between production traits, such growth vs. calving ease, lean yield
vs. carcass quality, and maintenance cost vs. milk/growth pose major challenges to beef
cattle breeders. Fertility traits are likewise related, often adversely, with other traits of
economic importance. What sets fertility traits apart, however, is the degree to which they
are affected by environment. Inclusion of composition traits, such as body condition and
muscularity, among those receiving emphasis will undoubtedly complicate multiple-trait
selection efforts.

Therefore, the objectives of this project were to:

1) model the phenotypic effects of composition on indicators of fertility in yearling

Limousin cattle, .

2) determine if composition effects differ between males and females,

3) account for composition differences in a genetic analysis of reproductive trait
indicators in yearling cattle,

4) estimate a genetic correlation between scrotal circumference and cycling status
utilizing a bivariate threshold - continuous model in an attempt to correctly account
for the discrete nature of the female fertility trait,

5) provide practical information to the industry that merits consideration as selection

decisions are made.

 

 

 

Age 2

A;
abilit}
manag
yearlir
greatl)
pubert

A

f m

PTOdUC
level 0
her AP
be a Set
Potenti;
tTait to ,
fOHOWir
Breeder:
adt-amaé

to idem

CHAPTER 2
REVIEW OF LITERATURE

REPRODUCTIVE TRAITS IN BEEF FEMALES

Age at Puberty

Age at puberty (AP) of heifers is an important early indicator of potential reproductive
ability, casting significant implications on cow longevity. Provided adequate nutrition and
management, most heifers have the potential to reach puberty and breed satisfactorily at
yearling ages. However, management and nutritional levels required to support estrous vary
greatly among breeds and heifers within a breed. Heifers with the inherent ability to reach
puberty at early ages may breed at less cost than heifers with later inherent AP.

Age at puberty is the only known fertility trait expressed in a heifer before she is in
PTOduction, and therefore relatively immune from many interactions with other traits. Her
level of milk production in a given environment, for example, is obviously unlikely to affect
her AP because at the time of puberty she has yet to lactate. Consequently, AP appears to
be a sensible measure of inherent fertility in young cattle, providing the earliest indication of
POtcntial reproductive performance. However, AP in females is a time and labor intensive
trait to measure. Other indicators of reproductive fitness, such as pregnancy status
followmg a fixed breeding season or age at first calving, are more easily measured traits.
Bree'Clers wishing to place selection pressure on fertility may prefer the convenience
advantages of those other traits over AP, realizing the additional expense and labor required

t° identify heifers which may ultimately prove to have inferior and unacceptable fertility

 

 

levels.
loneew

the end

[0 CODCt

 

research
both a b
breed m
Percent;
and mor
Correlar
Bnnk5(
Concept
“Speed
each ye.
indiCate

Nun
Weight a

GrfiEOn

levels. The objective here is to summarize AP, its influences on fertility, productivity,
longevity, and relationships with growth and composition traits. For a complete review of

the endocrinological mechanisms of puberty, see Kinder et al. (1987).

Emmy an_d Reprodgtive Perfrmngngg

Age at puberty (age of first behavioral estrus) has been shown to be favorably and at
least moderately associated with pregnancy percentage (Laster et al., 1979), percent calving
during the first 25 days (Laster et al., 1979; Doombos et al., 1983), and estrous cycles prior
to conception for the first, second, and fourth lactation (W erre and Brinks, 1986). Several
researchers have reported a favorable correlation of AP with yearling pregnancy rates on
both a between and within-breed basis. Lester et al. (1979) reported correlations among AP
breed means with percentage calving the first 25 days of -.75 and for AP with pregnancy
percentage of -.42. These values indicate that earlier AP resulted in earlier conception dates
and more numerous pregnancies between breeds. Doombos et al. (1983) reported a residual
correlation of -.40 between AP and percentage pregnant for Hereford cattle. Werre and
Brinks (1986) reported correlations among line of sire AP means with heat cycle of
conception (l = early, 3 = late) of .54, .34, -.O6, and .47 for the first four lactations
respectively. Thus, heifers from lines with early puberty also tended to conceive earlier
each year, except for the third lactation (Table 2.1). Furthermore, these correlations may
indicate a genetic relationship.

Numerous beef cattle studies reported between and within-breed differences in age and
weight at puberty related to subsequent reproduction in beef cattle (Laster et al., 1976;

Gregory et al., 1979a; Dow et al., 1982; Cundiff et al., 1986; Gregory et al., 1991; Gregory

 

 

et al.. 1
Germ i
enter.
preys;
three C)
in yearl
breed g.
1979; i
were grr
(approx
Observer
on diets
lOWer pr
[Needing
MOFCOW
Significa
refieCIch

the lfittest

”“6 in all

of F" CTOS.

et al., 1992; Cundiff et al., 1993). Most of the work was conducted as a component of the
Germ Plasrn Evaluation project (GPE) at the Meat Animal Research Center (MARC), Clay
Center, Nebraska. Even so, contradictions exist over time regarding the influence of AP on
pregnancy percentage, percent calf crop born and weaned, etc. For instance, during the first
three cycles of the GPE project AP differed significantly among breeds, but pregnancy rate
in yearling heifers did not differ between breed groups that were oldest at puberty and those
breed groups that reached puberty at the youngest ages (Laster et al., 1976; Lester et al.,
1979; Young et al., 1978; Gregory et al., 1979a). Heifers in all breed groups at MARC
were grown and developed under drylot conditions on a moderately high-energy diet
(approximately 2.2 Mcal MFJkg) and pregnancy rates were not limited by variation
observed among breed groups in AP. It has been shown that heifers developed more slowly
on diets with lower energy density, reached puberty at significantly older ages, and had
lower pregnancy rates than did heifers developed more rapidly when both were exposed to
breeding as yearlings (W iltbank et al., 1966; Wiltbank et al., 1969; Ferrell, 1982).
Moreover, Gregory et al. (1991 , 1992) and Cundiff et al. (1993) have subsequently reported
significant differences in pregnancy percentage following a fixed breeding season that were
reflective of differences in AP. In all cases, Limousin cattle have been shown to be among
the latest breeds to reach sexual maturity.

Gregory et al. (1992) reported the correlation between AP in heifers and reproduction
rate in all ages was -.79 among pure breed means. Relationships among purebred means are
higher than those among F1 crosses in the GPE program. This is due to a higher percentage

of F1 crosses observed in estrus by 490 days of age in the GPE program (Laster et al., 1976;

Laster et al., 1979; Gregory et al., 1979) than were observed in estrus among pure breeds by

490 days (Gregory et al., 1991), possibly because of favorable effects of heterosis on AP.

Prodictivitv and Longevity

Cundiff et a1. (1986) also found AP to have favorable relationships with milk
production and progeny weaning weights. Lesmeister (1973) reported early calving heifers
had higher average annual lifetime calf production than late calving heifers, demonstrating
reproductive performance of primiparous cows is dependent on the time of first calving.
Age at puberty has a significant effect on beef production when heifers are bred to calve as
2-year-olds, particularly in systems that utilize restricted breeding seasons (Ferrell, 1982).
In addition, heifers that calved as 2-year-olds produced more calves during their lifetime
than did heifers that calving for the first time as 3-year-olds (Nfifiez-Dominquez et al.,
1991). Werre (1980) reported favorable correlations (by line of sire means in Hereford
cattle) between AP and adjusted weaning weight and most probable producing ability
(MPPA) through four lactations (Table 2.1). Thus, earlier age at puberty was associated
with heavier offspring weaning weights and higher accumulative MPPA values, presumably

through higher milk production.

 

 

 

been ob

correlat:

 

bem'een
for milk
mature s
knowled
importer:

Optimtm

Hen”-
bUIIS, bu
EstimateE

remitted

TABLE 2.1: Correlation of heifer age at puberty with productivity traits'

 

 

Age at Puberty Hcc" AWWR‘ MPPA“
lst Lactation .54 -.65 -.62
2nd Lactation .34 -.l 1 -.38
3rd Lactation -.06 -.24 -. 10
4th Lactation .47 -.11 .25
“Werre, 1980.

 

bHeat cycles of conception (1 = early, 3 = late)
CAdjusted 205 day weaning weight ratio
dMost probable producing ability

A favorable relationship of earlier age at puberty with higher milk production has also
been observed from correlations among breed means. Laster et al. (1979) reported a
correlation of -.88 between AP and milk production, indicating a favorable relationship
between breeds. Similarly, Gregory et al. (1991) reported that breeds historically selected
for milk production reached puberty at significantly younger ages than breeds of comparable
mature size and retail yield that had not been selected for milk production. Thus,
knowledge of the relationships between puberty traits and measures of productivity is

important for effective utilization of selection, heterosis, and complementarity to achieve an

optimum production level.

Heritabilig

Heritability estimates for AP average about 40%, slightly lower than SC in yearling
bulls, but still relatively high compared with many other reproductive traits. Heritability
estimates for age at puberty range from .07 i .17 reported by McInemey (1977) to .67

reported by Werre and Brinks (1986). Heritability estimates for AP are listed in Table 2.2.

 

 

Table I

’-

Source

______
Arije at
Srr‘ith e
Mclner
Laster:
Lunsrra
lQng et
MacNe
Werre 2
Smith e
Gregor

MFR

 

Relate

Table 2.2: Heritability estimates for age at puberty (AP)

 

 

 

Source Heritability 1 SE
Arije and Wiltbank, 1971 .20 i .16
Smith et al., 1976 .64 i .31
McInemey, 1977 .07 i .10
Laster et al., 1979 .41 _-l_- .17
Lunstra, 1982 .41
King et al., 1983 .48 i .18
MacNeil et al., 1984 .61 i .18
Werre and Brinks, 1986 .67 i .68
Smith et al., 1989a .10 i .09
Gregory et al., 1995 .31 j; .04
AVERAGE .39
Relationships with Growth

 

The genetic correlations of AP with growth traits are similar to those between SC and
growth traits in yearling bulls. Estimates of genetic correlations of AP with various
weights are listed in Table 2.3. ' Negative correlations with AP (favorable) indicate that
heifers with more genetic strength for growth from birth to yearling reach puberty earlier.

Genetic relationships of AP with growth traits indicate a relatively low, but favorable,
correlation with birth weight, whereas the relationships with weaning and yearling weights
are slightly stronger and complementary. These genetic correlation estimates suggest an
advantageous relationship between measures of puberty and the growth curve (i.e.,
increasing early growth rate and reaching a higher percentage of mature weight at earlier

ages without increasing birth weight).

Table 2.3: Genetic correlations of age_at puberty with growth traits.
Birth Weaning Yearling Mature

 

Source weight weight weight weight
Smith et al., 1976 -.07 -.52 -.29 -.20
Werre & Brinks, 1986 -. 16 -.31 -.25

Smith et al., 1989a .58 -.04 -.14

Gregory et al., 1995 .03 -.14 -.05 -.05
Average .1 -.25 -.18 -.13

 

Other studies have reported a positive genetic correlation between age and weight at
puberty. Arije and Wiltbank (1971), Smith et al. (1976) and Laster et al. (1979) reported
values of .36, .67 and .52 indicating that, genetically, heifers reaching puberty at later
ages were also heavier, or vice versa‘ These studies seem to be in conflict with the
findings described above. However, the latter relationships were calculated at the point in
time when first behavioral estrus was observed. When one considers that the correlations
reported in Table 2.3 were reported for weights adjusted to an age constant, the favorable
relationship of AP with the growth curve remains plausible.

Similarly, Martin et a1. (1992), reported heifers of faster gaining breed groups with large
mature size (e.g., Charolais, Chianina) tended to be older and heavier at puberty than did
heifers sired by breeds with smaller mature size (e.g., Hereford, Angus). Correlations
between mature size and AP are .57 for Bos taurus breeds and .25 when 803 indicus breeds
are included. Both between and within-breed sources of genetic variation were large and
important for age or weight at puberty. Heavier milking breeds seem to reach puberty
earlier than breeds of similar mature size and retail product that do not have a history of

selection for milk production (e. g. Simmental and Gelbvieh vs. Charolais and Limousin).

Relationships with Body Condition

Body condition (degree of fatness) is reportedly favorably correlated with rebreeding
performance of beef cattle. Gregory et al. (1995) reported a favorable genetic correlation
for body condition score and age at puberty in yearling females of -.09. The phenotypic
relationship between body condition and age at puberty, although favorable, was small
and unimportant. Furthermore, Limousin heifers were shown to be the leanest of nine
breed groups in the GPE Project and among the latest for age at puberty.

Ritchie et al. (1992) has described the 9-point body condition scoring system that is
the most widely adopted in industry:

1. Severely emaciated. Rarely seen. All ribs and bone structures easily visible. Very
little visible muscle tissue. Physically weak.

2. Emaciated. Similar to Condition Score 1, but not weakened. Little visible muscle
tissue.

3. Very thin. No fat over ribs or in brisket. More apparent muscling than on Condition

Score 2. Backbone easily visible.

4. Thin. Ribs usually visible with shoulders and hindquarters showing modest
muscling. Backbone visible.

5. Moderate. Last two or three ribs can be seen. Little evidence of fat in brisket, over
ribs or around tailhead.

6. High moderate. Smooth appearance throughout. Slight fat deposition in brisket and
over tailhead. Ribs covered, and back appears slightly rounded.

7. Good. Brisket full. Tailhead shows pockets of fat. Back appears well rounded due

to fat. Ribs appear very smooth.

10

8. Obese. Back square due to fat. Brisket distended. Heavy fat pockets around
tailhead. Neck thick.
9. Very obese. Rarely seen. Similar to Condition Score 8, but more extreme. Heavy

disposition of udder fat.

Relationships with Muscularity

Relationships between muscularity and female age at puberty are not well defined.
Objective methods of measuring differences for either muscle or fertility traits in live
animals are inherently difficult, expensive and often imprecise. Coupling two such traits
in any analysis requires additional time, labor, and assumptions regarding the validity of
the raw data. While not a perfect solution, subjectively assigned muscle scores were used
prior to the onset of more advanced technology and continue to be in use. Gregory et al.
(1995) reported an intermediate genetic relationship of a subjective muscle score with
measures of carcass composition, suggesting visual evaluation of differences in muscle
thickness has value in contributing to making changes in carcass composition.
Characterization of biological types by researchers at the MARC indicates tendencies
among Bos taurus breeds with greater retail product to be later maturing cattle that are

significantly older at first detectable estrus (Gregory et al., 1992; Gregory et al., 1995).

ll

Reproductive Tract Score

Several researchers have studied AP in beef heifers when defined as the age at first
behavioral estrus. However, because AP is difficult and labor-intensive to measure, direct
selection for age at puberty in females is seldom practiced. A method for evaluating the
reproductive tract of yearling heifers has been developed at Colorado State University
(Andersen et al., 1991). The reproductive tract scoring (RTS) system was designed to
estimate pubertal status via rectal palpation of the uterine horns and ovaries. Each heifer is

assigned a score of l (immature) through 5 (cycling) as described in Table 2.4.

Table 2.4: Description of reproductive tract scores“l

 

 

 

Ovaries (approx. size)
Repro.
tract Uterine Length Height Width Ovarian
scores horns (mm) (mm) (mm) structures
1 Immature < 20 mm No palpable
diameter - no tone 15 10 8 follicles
2 20-25 mm diameter - no
tone 18 12 10 8 mm follicles
3 25-30 mm diameter - 8-10 mm follicles
slight tone 22 15 10
< 10 mm follicles
4 30 mm diameter - good corpus luteum
tone 30 16 12 possible
>10 mm follicles
5 > 30 mm diameter - corpus luteum
good tone, erect >32 20 15 present

 

aReproductive tract score was determined approximately one month pre-breeding by rectal
palpation. Andersen et al., 1987.

12

Evaluation System

The reproductive tract scoring system estimates pubertal status via rectal palpation of
the uterine horns and ovaries as described in Table 2.4 (LeFever and Odde, 1986). A RTS
of 1 is assigned to heifers with infantile tracts as indicated by small, toneless uterine horns
and small ovaries devoid of significant structures. Heifers scored as 1 are likely the furthest
from cycling at the time of examination. Heifers given a RTS of 2 are closer to cycling than
those scoring 1 due primarily to the presence of small follicles and slightly larger uterine
horns and ovaries. Those heifers assigned a RTS of 3 are on the verge of cycling based on
slight uterine tone in addition to the presence of follicles. Heifers assigned a score of 4 are
presumably cycling as indicated by good uterine tone, uterine size, and follicular growth.
However, heifers with tract scores of 4 lack an easily distinguished corpus luteum due to the
stage of the estrous cycle. Heifers with tract scores of 5 are similar to those scoring 4 except

for the presence of a palpable corpus luteum.

Preggancy and Reproductive Performing

Synchronization trials and other studies at Colorado State University indicate that
reproductive tract scoring one month prior to breeding can help identify heifers less likely to
become pregnant during a short breeding season (LeFever and Odde, 1986; Brown, 1986;
Andersen, 1987; Andersen et al., 1987; Odde et al., 1989). This prompted study of the
genetic aspects of RTS and other traits used in selecting replacement heifers. Genetic
analysis results prompted an investigation of the relative importance of RTS, condition
score, age, and weight to predict estrous response to synchronization (RS), pregnancy rate to

synchronized breeding (PS), pregnancy rate at the end of the breeding season (PBS), and

13

 

 

 

studies
iii PS ii

S}Tl'Ci'llr

end of
rate at
bar he
0i heil
31mg:-
The
espetia
rates. 5

6«tier.

conception date (Andersen, 1987). The reproductive tract scoring system was more useful
as an indicator of pregnancy rate (PS and PBS) than as an indicator of RS and conception
date as evidenced by significance levels (Andersen, 1987). Inclusion of condition score,
age, and weight along with RTS in models used for analysis provided little additional
accuracy over RTS alone for the reproductive traits analyzed.

The relationship of RTS to reproductive performance information from a series of
studies is presented in Table 2.5. Generally, means in Table 2.5 indicate a 5-22% advantage
in PS for tract scores of 4 and 5 versus tract scores of 3. Rates of pregnancy to
synchronized breeding for RTS 1 were 41-48% lower than for RTS of 4 and 5.

Differences in breeding season pregnancy rates for RTS 4 and 5 versus scores of 3
suggest anywhere from a 5-20% advantage. Heifers with RTS 3 prior to the breeding
season seem to have little apparent advantage over RTS 2 in terms of pregnancy rates at the
end of the breeding season. Heifers scoring 1 tended to have a 31-67% lower pregnancy
rate at the conclusion of the breeding season than heifers scoring 3. These results indicate
that heifers scoring 1 should be eliminated from the breeding group or managed differently.
Of heifers that conceived, conception dates for heifers with scores of 3, 4, and 5 were, on
average, about 10 days earlier than heifers with a RTS of 1 and 2.

The RTS values were predictive of reproductive performance of yearling heifers,
especially for pregnancy rates to synchronized breeding and to breeding season pregnancy
rates. Heifers with more mature reproductive tracts had higher pregnancy rates and calved

earlier.

14

Table 2.5: Effect of reproductive tract score on reproductive trait means'

 

 

Reproductive tract score
Reproductive measure 1 2 3 4 5
Response to synchronization, %b
LeFever and Odde, 1986 50.0 79.7 79.3 93.1 96.8
Brown, 1986 (MGA-PGFZ)c 3.4 63.6 82.2 95.3 92.4
Brown, 1986 (SMB)c 76.0 100.0 92.0 95.4 93.2
Andersen, 1987 56.0 63.0 69.0 79.0 76.0
Average 46.3 76.6 80.4 90.7 89.4
Pregnancy rate to synchronized breeding, %b
LeFever and Odde, 1986 0.0 17.0 37.0 62.1 54.0
Odde et al., 1989 0.0 23.0 32.0 55.0 54.0
Brown, 1986 4.5 23.4 35.8 52.3 50.0
Andersen et al., 1987 58.8 56.6 70.9
Andersen, 1987 6.0 27.0 34.0 48.0 46.0
Average 2.6 22.6 39.5 54.6 55.0
Pregnancy rate at end of breeding season, %b
Odde et al.. 1989 38.0 61.0 70.0 93.0 85.0
Brown, 1986 (MGA-PGFz)c 18.8 70.7 85.8 95.2 89.6
Brown, 1986 (SMB)° 16.2 100.0 80.2 90.9 82.4
Andersen et al., 1987 _ 74.1 99.4 86.1
Andersen, 1987 40.0 65.0 74.0 92.0 82.0
Average 28.2 74.2 76.8 94.1 85.0
Conception dated
Odde et al.. 1989 22.0 12.0 4.0 2.0 0
Andersen et al., 1987 1.0 9.0 0
Andersen, 1987 16.0 8.0 3.0 2.0 0
Average 19.0 10.0 2.0 4.3 0

 

“Means presented by LeFever and Odde (1986; 1989) are raw means. All other means are
least-square means.

bNumber responding or conceiving divided by the number treated at the start of the
breeding season.

cSynchronization method.

dAverage conception dates represent the average number of days into the breeding season
that conception occurred compared to the average number of heifers which had
reproductive tract scores of 5.

15

Productivity

Considerable apparent breed differences were reported between both purebreds and
composite crossbreds for RTS as well as for other traits (Andersen et al., 1988). Purebred
Herefords exhibited the lowest mean RTS (2.78), while Angus and Simmental had the
highest means (4.14 and 4.07, respectively). Generally, the purebreds known for higher
milk production also had higher RTS. These findings agree with Cundiff et al. (1986), and
others who have reported the favorable relationship between age at puberty and milk

production.

Genetic Asyts of RTS

Reproductive tract scores were included in a study that evaluated genetic aspects of
several performance traits in beef heifers (Andersen et al., 1988). For RTS, the effects of
line of sire, sire within line of sire, birth year, inbreeding, and age were all significant (P <
.05) sources of variation while age of dam was not significant. For all dependent variables
except birth weight, increased inbreeding had detrimental effects while increased age had

favorable effects on all traits analyzed as indicated by regression coefficients (Table 2.6).

Table 2.6: Regression coefficients for heifer performance traits'

nw‘wwvwnrscsrA PH‘PW‘
0b.) 0b.) 0b.) (unis) (units) (cm’) (cm) (cm)

 

Regression on
inbreeding -.16" A221“ -245“ A019“ -013“ A27“ A011“ -.016"
Regression on age .03 1.57“ 1.67“ .02“ .009“ .37“ .012“ .018“

 

‘Andersen et al., 1988
bBW = birth weight; W = weaning weight; YW = yearling weight; RTS = reproductive tract
score; CS = condition score; PA = pelvic area; PH = pelvic height; and PW = pelvic width.
"Pelvic height and width regression coefficients were calculated from a subset of the data used for
lvic areas because of missing data from 1987.
< .05.

16

Heritability estimates for RTS were lower than AP estimates, as might be expected
because this measure of puberty is less precise and the distribution of RTS depends on
when the heifers are examined. Although moderate, the average heritability estimate of
.28 (Table 2.7) for RTS seems feasible (Andersen et al., 1991; Andersen, 1991). The
heritability estimate for RTS is within the range of estimates for age at puberty and
suggests that RTS should respond favorably to selection. Heifers in the studies cited were
measured approximately one month prior to breeding.

Table 2.7: Heritability estimates for reproductive tract score (RTS).

 

 

Source Heritability i SE

Andersen et al., 1991 .32 i .17

Andersen, 1991 .24 -_+-_ .13

AVERAGE .28
Relationshfis with Growth

 

Positive correlations are favorable for RTS with measures of growth. Heifers in these
studies were raised under limited feed resources from weaning until breeding.

Genetic correlations (T able 2.8) between RTS and growth and pelvic traits were all
favorable (Andersen et al., 1988). These correlations suggest that favorable correlated
response would occur in all traits evaluated, given selection for improved RTS. However, a
follow-up study also conducted by Andersen (1991) reported genetic relationships between
RTS and birth weight, weaning weight, and yearling weight as .05 i .39, -.01 i .43, and .24
j; .35, respectively. Phenotypic correlations between RTS and other traits were all positive

and moderate in magnitude, except for birth weight, which was slightly negative.

17

Table 2.8: Heritabilities and genetic, phenotypic, and environmental correlations for
heifer ggowth and reproductive traits '

BW WW YW RTS CS PA PH PW

 

Birth Weight
0" .09 .98 .74 -.37 .14 -.06 .62
SE (G)c . 16 .95 .75 .95 .79 .61 .59
E“ .10 .36 .07 .18 .18 -07
P' .30 .33 -.02 .07 .15 .05 .23
Weaning Weight
G .58 .93 .20 .16 .64 .53 .75
SE(G) .20 .07 .39 .44 .23 .29 . 18
E .67 .60 .62 .27 -.83 .08
P .82 .41 .39 .47 .25 .58
Yearling Weight
G .87 .3 1 .40 .72 .47 .95
SE(G) .21 .32 .31 .17 .28 .13
E .94 1.60 .32 ~05 -.56
P .44 .45 .57 .35 .63
Reproductive Tract Score
G .32 -.06 .53 1.01 .30
SE(G) . 17 .49 .34 .77 .55
E .71 .31 -.35 1.08
P .37 .39 .28 .39
Condition Score
G .42 .61 .59 .77
SE(G) .22 .37 .47 .36
E - -. 10 -.64 .72
P .26 .18 .36
Pelvic Area
G .53 .78 .75
SE(G) . 19 .12 .13
E .81 1.08
P .78 .78
Pelvic Height
G .82 .16
SE(G) .26 .29
E .84
P .22
Pelvic Width
G .91
SE(G) .26

 

‘ Anderson et al., 1988.
l’G = genetic (correlation between breeding values); cSE(G) = standard error of genetic parameter; 61?. =

environmental (correlation between environmental effects on two traits); °P = phenotypic (correlation between
phenotypic values).

18

Other Fme which Influence RTS

Another study was undertaken to evaluate how contemporary group (location, year, and
synchronization treatment combinations), age, weight, and condition score influence RTS
(Andersen, 1988; unpublished data). Approximately 45% of the variation in RTS was
accounted for by the above four factors. Contemporary group, condition score, and weight
were highly significant (P < .005) sources of variation. In this study, age did not account for

a significant portion of the variation in RTS.

Timing and Appropriafilse

The distribution of RTS for a group of heifers depends upon when the heifers are
examined. Variation within a group of heifers is temporary. If taken before a year of age,
most heifers will not be cycling and receive tract scores of 1 and 2. Conversely, if scores
are taken too late, most heifers will be cycling and scored 4 or 5. The time or age at which
heifers are examined depends upon the desired use of the scoring system and the particular
group of heifers to be evaluated.

It is my belief that when tract scoring is to be used as an indicator of a heifer's ability to
conceive early during the first breeding season, scoring should be done about one month
before breeding or earlier. If RTS is used as a tool to place selection pressure on age at
puberty, the best time to evaluate is when 25-50% of the heifers are thought to have begun
cycling based on age, weight, and casual estrus observation. Depending upon management,
maturing rate, and age uniformity, heifers may only have to be scored once to take full
advantage of the system. If heifers are cycling more than one month before the start of the
breeding season, they should be scored earlier while greater variation exists. A RTS taken

19

30 to 60 days prior to the start of the breeding season can also serve as a check for the
heifers nutritional development program. Then, according to the resultant scores, the ration

or start of the breeding season may be adjusted.

Discussion

Widespread use of the reproductive tract scoring system depends upon overcoming
special problems associated with this trait. Training personnel that can palpate and assign
scores proficiently is perhaps the major obstacle. It seems likely that veterinarians should
be able to adapt to the scoring system and provide clients with a useful service.

The usefulness and resulting economic value of this scoring system in specific selection
and management systems is difficult to assess, at least partially because of the threshold
nature of most measures of reproductive performance. Breeds, lines, or crossbreds
characterized as later maturing may profit more from scoring than earlier maturing types.
Because of the favorable relationship between sire scrotal circumference and daughters age
at puberty (Brinks et al., 1978; King et al., 1983; Lunstra, 1982), operations that have
practiced sire selection for scrotal circumference may already have high levels of fertility
built into their herds. In such cases, RTS may provide little additional benefit. Regardless
of how inherent fertility is improved, through decreased age at puberty from the male and/or
female side, it can be thought of as an insurance policy for expressed fertility. This should
provide greater flexibility when matching mature size and milk production to management
and environmental constraints.

Breeding heifers several weeks before the cow herd permits concentration of time and

labor during breeding and calving and allows for a longer postpartum interval the following

20

year. Unfortunately, this practice may be futile if a large proportion of heifers are expected
to conceive at pubertal estrus, which is often less fertile than subsequent estrus (Byerley et
al., 1987). The utility of this management practice can be evaluated using RTS to predict
pubertal status before starting the breeding season. If non-cycling heifers can be identified
and eliminated from the breeding group, higher first service pregnancy rates and other
benefits of this management practice can be more fully realized.

There is justification for feeding heifers so that less than a maximum proportion are
allowed to express their genetic potential for cyclic activity one month before breeding. If
less total weight has to be maintained during winter feeding months, the associated feed
costs could be reduced and weight could be gained more cost effectively on grass after
breeding. A one-time tract score would then allow selection for both age at puberty and
reproductive performance during the first breeding season. Also, only heifers capable of
adequate consumption and (or) efficient conversion of feed will achieve weights that allow
them to cycle. Heifers that cycle at lighter weights or lower condition scores on less feed

relative to their age could be assumed the most inherently fertile.

Summg

In summary, RTS appears to be moderately heritable and favorably related to lower
birth weight as well as pelvic and higher growth traits. Tract scores have also been found to
be favorably associated with response to synchronization, pregnancy rate to synchronized
breeding, pregnancy rate at the end of the breeding season, and conception date. Usefulness
of the tract scoring system depends upon timing, previous selection practices, management,

and environmental factors. Tract scores can be used to evaluate the status of heifer

21

development, time synchronization programs and the start of the breeding season, and place
selection pressure on age at puberty and related traits. The scoring can be done in
conjunction with collection of yearling weights, condition scores, pelvic measures, and

general processing as part of a yearling heifer evaluation and health program.

Threshold Trait Considerations

Reproductive tract score, like most reproductive traits expressed in the female, is an
example of a threshold trait. Threshold traits are those which vary in a discontinuous
manner but aren’t inherited in simple Mendelian fashion. Because characteristics such as
reproductive tract score, pregnancy, etc. are typically measured in discrete terms; they
appear to be indescribable as quantitative traits. Genetic analysis reveals however, these
traits are indeed inherited in the same was as continuously varying characters.
Reproductive tract score has an underlying continuous distribution of both genetic and
environmental origin and a threshold that imposes discontinuity on the observable scale.
When the underlying variable effects on the animal are greater than the threshold for
puberty, the animal scores a 4 or 5 in the previously described RTS system. By contrast,
if the sum of effects is less than the threshold value, the heifer is scored 1, 2, or 3 as she
has not yet completed her first estrous cycle.

Incidences of the threshold for puberty being met and the trait being expressed are
coded as “1” and a failure to have cycled is coded as “0”. Thus, despite the fact a
continuous distribution is assumed for the factors that control age at puberty, observable

differences and corresponding codes allow for only binomial or multinomial

22

 

1151?."

 

$111

in;

line

met
lCSp
H1011:
Stand

56! of
ll811101
timbre
mill lDOr‘

Homing],

distributions. Therefore, linear model procedures are inappropriate for estimating the
variance components for threshold traits such as pubertal status because their use violates
the assumption of a continuous, normal distribution required by these methods. Gianola
and Foulley (1983) and Hoeschele et al. (1995) noted that linear model procedures
couldn’t account for the non—normal distribution of error commonly associated with
categorically observed traits. Threshold model procedures that assume the existence of
an underlying continuous variable (Gianola and Foulley, 1983) have overcome this
dilemma. Analytical methods that account for the binomial nature of female fertility traits

will be discussed further in the next section.

Logistic Regression

Most genetic analyses of reproductive traits with discrete categories have utilized
linear model procedures for variance component estimation. As previously stated, such
methods are inappropriate due to the violation of the required assumption that the
response variable be normally distributed. Logistic regression is a form of statistical
modeling that is often appropriate for categorical outcome variables (Selvin, 1996). Like
standard linear modeling, it describes the relationship between a response variable and a
set of explanatory variables. The response variable is usually dichotomous; in which case
it follows the binomial distribution. However, logistic regression also fits polytomous
variables, that is variables having more than two response levels, which then follows a
multinomial distribution. These multiple-level response variables can be scaled

nominally or ordinally. The explanatory variables may be categorical or continuous.

23

‘r—HV

this '

3101.

fire
u l\

Till

fro

f 1'

pm“

come
descn'
mcisur
the mic
l‘CSponSe

Li

 

On a phenotypic level, researchers may be interested in the likelihood that a fertility
trait of interest is expressed given an array of explanatory variables such as weight, age,
body condition, etc. Logistic regression is advantageous to standard linear modeling in
this instance as it guarantees the probability of success will range between 0 and 1.
Moreover, the logistic function has a sigmoid curve, which more accurately reflects the

1

"Z

 

threshold nature of fertility traits. The logistic function is described as: f (z) = l

The value of logistic function is dependent on the value of 2. To obtain the logistic model
from logistic function 2 is written as a function of independent variables (X’s).

z = or + B1X1+B2X2 + + [3ka

The logistic model for the conditional probability of cycling given characteristics
X1,...,Xk, i.e. P(1IX.,...,Xk) is written as follows:

1 e(a+fl,X‘+B2X2+...+fikxk) 1

f(z) = 1+e-(a+B,X|+32‘X2+...+BtXp) = 1+8

 

(“+fi‘X|+flzxz+...+flka) = 8 Where a? BI)
-(a+zpixr)
III

1 + e
B;,..., 13;. are parameters to be estimated (e.g. age, season). Parameter estimation is
performed through maximum likelihood techniques.

An important aspect of logistic regression to remember is that it was developed in the
context of follow-up studies. Therefore, in fields such as epidemiology it is used to
describe the probability of developing disease as a function of independent variables
measured ‘at the start of some follow-up period. There may be some concern regarding
the validity of parameter estimates calculated on traits measured at the same time as the
response variable. However, logistic regression appears to be the best analysis tool for

categorical outcome variables

24

The;

Iii-J
'11-.

“1113, 11

3” lndix;

Obtaining odds and odds ratios from logistic reggession

Another feature of the logistic regression function is the interpretation of the
coefficient [3, through estimation of the odds in favor of an event and the odds ratio. The
odds ratio is interpreted as the increase in odds for a unit increase in the independent
variable(s). To do this the logistic model must be rewritten in the logit form. By
definition, if p is the probability than an event will occur then:

Odds is defined as the probability an event will occur divided by the probability the

P

and,
1-p

event will not occur, i.e.

 

 

logit (p) = 1n[1 p ]=1n(odds). In logistic model

1

-(a+fi,x,+p,x, +...+fi, X‘) = 1)(Cl)()

 

PCIX ,...,X
( l k)l-i-e

e -(a+plx‘+p2XI +...+p. xi)

 

... 1" P(C|X' ""’Xk) = 1+ e-(a+p,x,+fizxz+-.+fi.xg) = 1_ P(C|X)

. Odds = P(C|X) = 1 = ea+Ble+p2X2+m+fl.X.
o e - P(Clx) 1 + e-(a+fl'x'+p2xZ+m+fl‘x‘)

 

logit(P) = ln(odd) = ln( p J = ln(e“*”'x‘*”2x2*'"*”*x* )
1- p
= a +fi,X, + p,x,+...+,6,x,

The most common correct way of writing the model in the logit form.

ln(TL] = a + 61X, +132X2+...+/3,X,
“P

Thus, the logit form of the logistic model offers an expression for log odds of an event for

an individual with a specific set of X’s.

25

law"
1,.»-

la)

i.e. ‘

in

Del

Interpretation of coefficients from the logit form of the logistic model
(a)

Ifall X’s are set to zero then logit(P) = a + [3,Xl + 62X2+...+,6,‘X,‘ = a

i.e. Otis the log odds for an individual with zero values for all X’s. This interpretation has
limitations for naturally occurring continuous variables such as age or weight.
A more appealing alternative is to interpret the intercept as the baseline odds i.e. odds
that would be estimated if no X’s were considered at all.
(31’s)
Interpretable in terms of odds and/or odds ratios. Consider the following example of
a model that fits both weight and muscle score. If weight remains fixed and muscle
scores vary from 1 to 2, the data would appear as follows:
y = yearling reproductive tract score (0 = non-cycling, l = cycling)
X1: muscle score 1 (1: MS =1, 0: MS $1)
X2 = muscle score 2 (1: MS = 2, 0: MS at 2)
X3 = muscle score 3 (1: MS = 3,0: MS ¢ 3)
X4 = weight
Denote by Al, the collection of X’s for a light muscled individual (MS 1), and let A2
specify the collection of X’s that characterize an average muscled individual (MS 2).
A1 =(X11=1.X12 = 0. X13 = 0. X14: 0)
A2 =(X21= 0. X22 =1. X23 = 0. X24= 0)

If it is assumed to be known that:

ln(i—Ef") = ln(OddSAr) = a + filxll + fizxiz + fi3X13 + 34X”
" Al

26

lien

01

then

“litre
533153
For a
beat-{er t!

Mace

 

ln(lf: ]= ln(oddsAz) = 03+ ,8an +BZX22 +£33X23 +E,X24
A2

then

eln(oddsA2) _ ea+fl1X21+fi2X22+fi3X23+I34X24

and

810(0dd341) _ ea+fi1X11+52X12+53X13+f54xi4

0ddS(X ) ea+fl1X21+52X22rfisxzpp‘xu
- 2 —
I

 

 

Rules of algebra indicate that :5- = e
e

a-b

Therefore, 0R( ,2 .An can be rewritten as

k k k
((61+):l fling )-(at+.}:l flixn )l '21 fii(x21“xli)
I. I8 I.

OR(A2,A‘) :6 :6

Thus, for the example
A1 =(X11=1.X12=0.X13 =0.X14=0)
A2 =(X21= 0. X22 =1. X23 = 0. X24= 0)
then

OR ___ [/31 (0.1)+52(1-0)+133(0-01+pa(0-0) _ -p,+pz
(A2 A1) 6 - e

Where BI and B; are coefficients in the model: logit P(C.x) = 0t + BIMSI + BZMS2 +
B3MS3 + B4WI.
For a continuous variable such as weight, the odds of cycling for a heifer 45 kg

heavier than the mean are four times the odds of cycling for a heifer 45 kg lighter than the

Bit (difference) 0.0156(90)
8

mean according to the following calculation e = = 4.071 .

27

 

 

 

 

 

 

 

 

Bulll

 

denier

bird's

stand;

males

Yearl:
Somfl

i‘earlii

bull fe

Mon
fertility

“Elude.

new}

 

REPRODUCTIVE TRAITS IN BEEF MALES
Bull Fertility

Brinks (1991) stated that nearly 90% of the genetic improvement within a cowherd are
derived from the sire side of the pedigree. Thus, the largest potential improvements in a
herd’s inherent fertility levels are made through sire selection. One method to determine a
bull’s reproductive capacity is the breeding soundness evaluation (BSE). The exam is a test
standardized by the Society of Theliogenology and is used on yearling bulls to identify
males unfit for breeding. It consists of a reproductive organ examination, semen evaluation,
and structural soundness test. A bull that fails to satisfy all exam requirements fails the test.
Yearling bulls are often re-evaluated as their additional nutrient requirements for growth
sometimes prevent reproductive maturation, and not all bulls reach reproductive ability as
yearlings. The BSE requires a minimum scrotal circumference of 30 cm.

Genetic factors, which will be discussed in subsequent sections that influence inherent
bull fertility, are additive effects, breed, heterosis, inbreeding and correlations with other
performance traits, such as growth, maternal, and composition. Indicator traits for bull
fertility, genetic merit and breeding capacity that will be described in forthcoming sections
include: age at puberty, semen quality, daughter age at puberty, and yearling scrotal

circumference.

28

 

sir.

381

Hall
rein
W61.
11165181
Blink
3113351

mm:

“We

Pubem
Male puberty is defined and evaluated by the following criteria: 50 x 106 spermatozoa

per ejaculate, a minimum of 10% motility, an average scrotal circumference of 27 to 28 cm,
and sheath detachment (Lunstra et al., 1978; Lunstra et al., 1982). Davis (1993) described
an alternative evaluation for puberty using increased serum testosterone with GnRH
stimulation. Either this method or the previously described criteria will describe a bull’s
sexual maturity status, although the hormonal array may change as sexual maturity
approaches. Luteinizing hormone and testosterone serum levels increased in bulls prior to
reaching puberty, and breeds known for higher testosterone levels reached age at puberty
earlier (Lunstra et al., 1978).

Obviously, it is necessary for a bull to reach puberty prior to breeding. Scrotal
circumference and semen evaluations are the standard indicators for bull fertility and age at
puberty status. Neely et al. (1982) reported that 365 d scrotal circumference measurements
were better indicators of yearling bull fertility than those taken at other ages were.
Heritabilities of 205 and 365 d scrotal circumference were .08 i .20 and .44 i .24,
respectively. The 205 d scrotal circumference was associated with changes in a bull’s live
weight rather than developed semen-producing tissue. The 365 d scrotal circumference
measurement was favorably correlated with semen quality and testes size. Smith and
Brinks (1989) showed that as age increased, seminal quality improved. A linear regression
analysis on age for all traits showed significant results for motility, percent normal, scrotal
circumference, breeding soundness exam, and primary abnormal percentage. Thus, aging
appears to increase semen quality, motility, concentration, and percent normal sperm and

decrease abnormalities (Abadia et al., 1976).

29

 

am

16811

1105

pro

out

Scrotal Circumference a_nd Bull Fertilig

Scrotal circumference is a standard indicator of fertility in yearling bulls. Palasz et al.
(1994) showed that bulls with scrotal circumference measurements greater than 31 cm had
greater sperm producing capability than those bulls with less than 30 cm. Smaller scrotal
circumference bulls (<30 cm) exhibited a higher incidence of germinal epithelial loss and
damaged testicular tissue. Madrid et al. (1988) reported that bulls with scrotal
circumferences less than 32 cm had a higher incidence of semen abnormalities and
testicular lesions. However, other research has shown that the relationship between scrotal
circumference and bull fertility has not been established. Thompson et al. (1992) reported
no significant associations between the loss of germinal epithelium and spermatozoa
production. Additional studies found scrotal circumference to be a poor predictor of semen
output (Carter et al., 1980; Knights et al., 1984; Makarechian et al., 1985; Thompson et al.,
1992). While these differing results may be disconcerting from a practical standpoint, the
relevant question involves the general relationship between yearling scrotal circumference
and inherent fertility in collateral relatives and offspring. These issues will be addressed in

a following section.

Scrotal Circumference Relationships with Age at Puberty
Results from the Roman L. Hruska Meat Animal Research Center, (Lunstra, 1982)

indicated that SC was a more accurate predictor of bull puberty than either age or weight
regardless of breed or breed cross. Martin et al. (1992) reported that among breed group

means, high absolute correlations were observed between AP and SC in males (-.92).

30

cslr

51110;

 

 

 

an
.L\
N

Spay
old:
Bro
uslr
clrc
Scro
1199

358‘

Smith et al. (1989b, c) reported that for each centimeter increase in a sire’s SC there was
a .31 cm increase in sons’ SC, which is more than expected based on average heritability
estimates. Hough (1991) reported preliminary results from a high versus low selection
study utilizing scrotal circumference EPD in Hereford cattle. High SC sires had 2.4 i .03
larger EPD values than low SC sires. Bull progeny of the high-EPD sires averaged 2.2 i .7,
3.5 i .9, and 2.6 :l: 1.0 cm'larger SC than their contemporaries by low-EPD sires at weaning,
yearling and 15 months, respectively. Consequently, it seems apparent that sire SC impacts

AP in his male progeny.

Breed Effects

There are differences between breeds in age at puberty and the ability to produce viable
spermatozoa at an early age. Lunstra et al. (1978) reported that Hereford bulls had the
oldest age at puberty and the lightest weight at puberty compared to Red Poll, Angus, and
Brown Swiss. Gregory et al. (1991) reported breed differences for scrotal circumference
using MARC data Breed group was a highly significant fixed effect for scrotal
circumference and paired testicular volume least square means. Least square means for
scrotal circumference ranged from 29 cm for Limousin to 34 cm in Gelbvieh. Gregory et al.
(1995) stated there was a reduction in variation for scrotal circumference with 368 d weight
as a covariate, but there were still differences between breeds.

Fields et al. (1979) reported the breed effect to be a significant source of variation for
testicular volume, semen volume, motility, sperm concentration, and seminal fructose. The
breeds evaluated were Hereford, Angus, Santa Gertrudis, and Brahman. Brahman bulls had

smaller scrotal circumferences and later ages at puberty than the other breeds. Sperm

31

concentration and motility were lower for the Brahman and Santa Gertrudis, and higher for
Angus and Hereford. Smith and Brinks (1989) reported that breed was a highly significant
factor for percent motility and primary abnormalities. Red Angus bulls had the highest

breeding soundness results while Brangus bulls charted the lowest BSE scores.

Heterosis Effects

Heterosis has been shown to have positive effects on bull fertility. Composite MARC
data showed that scrotal circumference was partially influenced by breed group and showed
heterosis as an additional source of variation. Heterosis differences were significant for
scrotal circumference and paired testicular volume in all composite-breeding generations,
and reflected heterosis values for weight (Gregory et al., 1991). However, Gregory et al.
(1995) followed up by stating that heterosis effects on scrotal circumference occur because
of additional factors in addition to weight. When the covariate 368 d weight was included,

a high percentage of heterosis effects were removed (Gregory et al., 1991).

Inbreeding Effects

In contrast to heterosis, inbreeding can be detrimental to bull fertility and has been
unfavorably associated with semen quality traits and scrotal circumference. Abadia et al.
(1976) reported that inbred bulls were at greater risk for testicular and seminal vesicular
inflammation, and specific inbred lines showed a higher incidence. Palasz et al. (1994)
showed results of increased testicular hypoplasia and damaged serniniferous tubules
because a high percentage of the testicle was composed of germinal epithelium. This

decreased the percentage of normal sperm because of the significant negative correlation

32

with degree of germinal epithelium loss. Smith et al. (1989b) reported that inbreeding was

significant and detrimental for semen quality traits, except for secondary abnormalities.

Relationships with Growth

Scrotal circumference can be favorably associated with most measures of growth
according to genetic and phenotypic correlations reported in the literature. Bourdon and
Brinks (1986) reported weight had a greater effect on scrotal circumference than the other
covariates, age and height. Furthermore, any factor that caused an increase in weight
tended to increase scrotal circumference. Knights et al. (1984) reported that bulls with
increased growth sired male progeny with larger scrotal circumferences. Smith et al.
(1989b) showed favorable genetic correlations for scrotal circumference with weaning
and yearling weights of .56 and .63, respectively. Knights et al. (1984) reported a .68
genetic correlation between SC and yearling weight, a 0 correlation with weaning weight,
and .15 for birth weight. A partial listing of genetic correlations found in the literature
are presented in Table 2.9.

The genetic correlation between SC and birth weight appears to be relatively low
whereas the correlation between SC and yearling weight is relatively high. This suggests
that larger SC (earlier puberty) and faster growth rate is compatible in young bulls. These
relationships suggest a favorable growth curve, i.e., reaching a higher percent of mature
weight at earlier ages while maintaining or increasing early growth rate and possibly

holding mature weight in check.

33

TABLE 2.9: Genetic correlations between scrotal circumference and measures of
_growth

 

 

Birth Weaning Post-Weaning Yearling
Source Weigfl Weigh: Gain Weigm
Bourdon & Brinks (1985) .18 .29 .35 .44
Bourdon & Brinks (1986) .22 .20 .39
Knights et al. (1984) .10 .00 .68
Neely et al. (1982) .86 .22 .52
Smith (1989a) . .08 .56 .59 .63
Lunstra, Gregory &
Cundiff (1988) -.02 .00 .00 .10
Keeton et al. (1996) .14
Kriese et al. (1991) .02 .08 .35
Kriese et al. (1991) -.04 .34 .12

 

Relationships with Commsition

Moser et al. (1996) studied correlated responses in progeny reproduction and growth to
selection for scrotal circumference in Limousin bulls. Bulls were grouped initially into
large (mean = 36.3 cm) and small (mean = 28.5 cm) lines, and subsequently regrouped
according to non-parent SC expected progeny differences into low, average and high EPD
categories. Weaning ribeye area was greater for progeny of sires with average scrotal
circumference EPD’s than those of low SC EPD’s. No other differences in composition
were reported. Gregory et al. (1995) reported genetic and phenotypic relationships
between 368 d SC, body condition score and muscle score. Scrotal circumference genetic
and phenotypic relationships with body condition were .1 and .08i.07 respectively.
Genetic and phenotypic relationships between SC and muscle score were .0 and -. 171- .07.

Keeton et al. (1996) and Neely et al. (1992) reported that ultrasonic measures of

backfat were not significant sources of variation in SC.

34

Heritability

There are several reports in the literature relating to the heritability of SC in yearling
beef bulls on both an age and weight adjusted basis (Table 2.10).

These studies indicate that SC in yearling bulls is a moderate to highly heritable trait

(approximately 50%) and that selection should be very effective in changing SC.

TABLE 2.10: Heritability estimates for scrotal circumference in yearling beef bulls

 

 

Source Heritability SE
Age Adjusted

Bourdon and Brinks (1986) .49 06
Coulter and Foote (1979) .78 .07
Keeton et al. (1996) .46 .04
Knights et al. (1982) .36 06
Latimer et al. (1982) .38 16
Lunstra (1982) .52

Lunstra, Gregory and Cundiff (1987) .41 .06
Neely et al. (1982) .44 .24
Smith et al. (1989a) .40 .09
Nelsen (1986) .55 .21
Average .48

Weight Adjusted

Bourdon and Brinks (1985) .46 .06
Lunstra (1982) .69

Lunstra, Gregory and Cundiff (1987) .50 .06
Neely et al. (1982) .44 .24
Kriese et al. (1991) .16

Kriese et al. (1991) .53

Avegge .46

 

35

 

 

S‘s‘Veral .

group me

Relationships of Scrotal Circumference to Female Reproductive Traits

A heifer’s underlying fertility is associated with male reproductive measures. Until
recently, heritability estimates for direct measures of fertility were very lowly heritable
and unworthy of selection pressure as a means of improving reproductive fitness.
Therefore, indicator traits such as scrotal circumference, which is more highly heritable
and correlated with female fertility measures, were used in an effort to make genetic
progress. The genetic correlation between heifer age at puberty and sire yearling scrotal
circumference and has been estimated at -.71, -1.07, -.39, and -.91 (Hoehenboken et al.,
1971; King et al., 1983; Gregory et al., 1991; Morris et al., 1992). Lunstra (1982)
reported a correlation of .98 among breed means (8 breeds) for SC of bulls with age at
puberty in heifers. Hough (1991) reported a 60 _-1_- 28 day advantage in AP for heifer
progeny of sires with SC EPD values 2.4 i .3 cm. larger than those of other sires. Genetic
correlation estimates between SC in yearling bulls and age at puberty in half-sib heifers of
-.71 and -1.07 (favorable) have been reported by Brinks et al. (1978) and King et al.
(1983). In all mentioned studies, heifers related to larger scrotal circumference bulls were
more likely to attain early puberty than heifers related to bulls with smaller scrotal
circumferences. These very strong genetic relationships, coupled with Lunstra’s (1982)
data has led others to speculate that heifer age at puberty and SC are essentially the same
trait.

Toelle and Robinson (1985) also reported that SC was favorably related genetically to
several measures of female reproductive traits. Martin et al, (1992) reported among breed

group means, high absolute correlations were observed between SC in males and pregnancy

36

rate in heifers (.97). Smith et al. (1989b, c) reported regressions of -.80 days in age at
puberty, -.67 days in day of first calving, and -.83 days in age of first calving of female
offspring per centimeter of sire SC. The change in age at puberty is less than that expected
based on genetic parameter estimates. Possibly, there is a greater change in inherent age at
puberty than in expressed age at puberty due to seasonal or daylight length effects of other
environmental limitations such as nutrition. King et al. (1983) reported that although
heifers born later in the calving season reached puberty at an earlier age, heifers born earlier

in the calving season reached puberty at an earlier date the following year.

37

 

 

illCC- .

are

lndltl

 

subsri

lest

{Pp};

MIXED MODELS

The general mixed model embraces the vast majority of estimation problems
encountered from highly unbalanced family sizes and fragmentary data from numerous
types of relationships common to many agricultural species. Data sets from which
individuals have been eliminated by natural and/or artificial selection may deviate
substantially from the base population about which one wishes to make inferences.
Further culling of the data to accommodate a conventional statistical technique, such as
analysis of variance, even if nonselective, still leads to an inefficient use of information.
General mixed models allow the efficient estimation of quantitative genetic parameters
under arbitrary settings, including those involving extended pedigrees, unequal family
sizes, assortative mating, and selection.

Mixed model theory has existed for a number of years (Eisenhart 1947), and generally
describes a model that jointly accounts for fixed and random effects. Fixed effects
typically include the population mean and other factors such as birth year, gender,
experimental treatment and so forth. Random effects are usually genetic effects such as
additive genetic values. The mixed linear model with one random factor is generally
describedas: y=Xb+Zu+e
where: y is an N x 1 vector of phenotypic observations,

b is a p x 1 vector of fixed effects associated with y,
u is a q x 1 vector of random effects associated with y,
X is a known incidence matrix of order N x p relating elements of b to

elements of y,

38

 

Eli‘i

 

The

j)»
1»

Safer

7131* -

lll

Z is a known incidence matrix of order N x q that relates elements of u to
elements of y, and
E is an N x 1 vector of residual effects.
Additional attributes of the general form of mixed linear models include the

expectations of the random variable, which include:

50’) = Xb.
E(u) = 0, and
E(e) = 0.

The (co)variance structure is:

y ZGZ'+R ZG R
Var u = GZ' G 0
e R 0 R

The Best Linear Unbiased Prediction (BLUP) developed by Henderson (1953) utilizes
the mixed model to predict random effects, such as breeding values. A similar procedure,
adopted to estimate fixed effects, is referred to as the Best Linear Unbiased Estimator.
The procedures are similar in the sense that they share three important functions. They
are best by minimizing sampling variance, linear as they are linear functions of observed
phenotypes y, and unbiased in the sense that E[BLUE(b)] = b and E[BLUP(u)] = u.
BLUP is primarily used to identify individuals with high estimated genetic merit in
selection programs, monitor response to selection and has become the dominant
methodology for estimating breeding values. The literature collection describing BLUP
methodology is voluminous and extensive (Henderson, 1977a, 1984a, 1988a; Schaeffer,

1991; Kennedy, 1991; Searle et al., 1992; and Mrode, 1996). BLUP methodology is

39

useful to estimate breeding values for a single trait in a strictly additive model,
expandable to include estimation of dominance values and maternal effects, and can

accommodate repeated records and multiple traits (Lynch and Walsh, 1998).

SIRE MODELS

General linear mixed models form the fundamental framework for BLUP analysis, yet
there are numerous ways in which this model can be formulated and applied. Of note to
animal breeders is the animal model, which predicts breeding values for each measured
individual, the gametic model, which describes breeding values of measured in terms of
parental contributions. The reduced animal model developed by Quaas and Pollak (1980)
combines aspects of both the animal and gametic models in specific applications in which
parental breeding values are of primary interest. The objective here is to focus on the
gametic model, specifically as. it is used to evaluate sires for important traits.

The gametic model is often used when parental breeding values are of more concern
than offspring values, as when one is attempting to estimate the breeding value of bulls
. from large arrays of descendants (Lynch and Walsh, 1998). In this model, the additive
genetic value of each offspring is expressed in terms of its parents’ breeding values. The
sire model is a variation of the gametic model wherein the dam contribution is ignored
and incorporated into the error term.

Variations in the sire model primarily concern whether relationships among sires are
considered. Kemp et al. (1984) indicated additive genetic relationships between sires are

important as their inclusion reduces standard errors of prediction.

40

BIVARIATE MODELS

Several traits of interest in animal breeding appear as categorical variables. Analysis
of such variables by linear methodology violates several assumptions of the linear model
and is not optimal (Gianola 1982). A more satisfactory method is based on the threshold
model concept (Wright, 1934). Statistical treatment of the threshold model has been
developed for sire evaluation by Gianola and Foulley in a Bayesian setting, and
equivalently by Harville and Mee (1984) from a classical viewpoint.

J anss and Foulley (1993) attempted to perform a genetic evaluation of French beef
bulls for calving difficulty which includes a bivariate analysis of birth weight in an
attempt to increase accuracy of prediction. Birth weight was included due to its high
genetic and residual correlation with calving difficulty (Philipsson, 1976; Meijering,
1985). Simianer and Schaeffer (1989) have described bivariate analyses for evaluation of
a binary and continuous variable. However, their methods lack general applicability
when missing records exist and/or different fixed effects influence each trait. J arms and
Foulley (1993) described general bivariate analyses that successfully accommodated
unequal design matrices for a binary and continuous variable. The usefulness of the
information matrix to obtain standard errors, loss of information when records with
missing data are not included, and ability of this analysis to correct for selection on one of
the traits were investigated by simulation.

Observations for the continuous trait followed the linear model:

Model 11 Y1 = lel + Z1111 + 81

41

 

 

w?

1";

where X and Z were incidence matrices, relating records to fixed effects (b) and to sire-
effects (u) for birth weight. The error variance is given as 62:1.

Observations for the threshold trait calving difficulty was modeled as:

Model 2: y2 = x262.» Z2“; + E(ez I e.) + e‘2

where all terms are similar to those in Model 1 and E denotes expectation. Model 2
cohditions observations for calving difficulty on birth weight, so that the remaining error,
e}, is uncorrelated with the error for the continuous trait. This enables writing the joint
likelihood of the data. The variance on the underlying scale is arbitrary (=1).

The variance structure for sire effects is given as:

V ‘11]: G = [G11 G12] = [102111 10;"2]...and... “ = [G: G:]
[“2 G11 G22 10' 021 10' 022 G G G

Sires in this analysis were assumed unrelated, but an extension to include relationships
between sires is straightforward by defining G differently.

Accuracy of evaluation improved for both traits in bivariate analysis. The increase in
accuracy is particularly substantial for the discrete trait when all data are included. The
relative increase in accuracy ranged between 15 and 30%. The most significant aspect of
this methodology is the ability to account correctly for the information content of a

discrete observation.

42

SUMMARY OF LITERATURE

Significant genetic and phenotypic variation exists within and between breeds of beef
cattle for age at puberty (AP). In general, faster-gaining breed groups of larger mature size
reach puberty at a later age than do slower—gaining breed groups of smaller mature size.
Breeds selected for milking ability reach puberty at younger ages than do those breeds not
selected for milk production. Heterosis, independent of heterosis effects on weight,
influences most measures of puberty in females and scrotal circumference (SC) in males.
Crossbred heifers reach puberty at younger ages and heavier weights than their straightbred
counterparts. Age at puberty in both bulls and females appears to be a good indicator of
inherent fertility and expressed reproductive efficiency, especially in the early productive
years. Scrotal circumference is an excellent measure of age at puberty in yearling bulls.
Furthermore, a favorable genetic relationship exists between SC in bulls and AP of female
offspring. Scrotal circumference is relatively easy to obtain, highly heritable and is
favorably related to seminal characteristics and measures of early growth. Reproductive
tract score, although not as specific a measure as age at puberty, possesses many of the
same attributes regarding measurability and repeatability as scrotal circumference in
yearling bulls. Thus, AP appears to be a feasible and good selection criterion in yearling
heifers. Limousin cattle are consistently older at puberty and less reproductively efficient
throughout their lifetime than many other breeds.

Selection to improve threshold traits, such as fertility, becomes particularly challenging
because of the strong influence of environment and importance of genotype x environment

interactions. Furthermore, genetic antagonisms that exist between expressed fertility and

43

milk/growth in certain environments create still greater challenges. Age at puberty,
however, is expressed in a female before potential interactions with traits like milking
ability have the opportunity to be expressed. Therefore, age at puberty may be the most
reliable indicator of inherent fertility. Consequently, selection for age at puberty or more
easily measured traits such as age of first calving or scrotal circumference may provide the

most feasible strategy for making genetic improvement at this time.
IMPLICATIONS

Limousin breeders can choose from multiple strategies to attempt selection for
improved fertility. The first involves direct selection for fertility traits. Scrotal
circumference, because it is relatively highly heritable and easy to measure, is a likely
candidate for selection. Other possibilities include traits mentioned earlier, as well as age at
first calving, calving interval and longevity. The utility of any prospective fertility trait
depends on its ease of measure and relationship with inherent fertility. Furthermore, breed
associations that gather complete and accurate reproductive data from breeders are able to
provide substantial information and service to aid producers in their genetic improvement
efforts. Obviously, the value of the information generated is highly dependent on the extent
of breeder participation.

The indirect approach to breeding for improved fertility involves selection for an array
of traits that have favorable correlated effects on fertility. These include milk production,
calving ease, growth rate, and body condition. By selecting for optimum combinations of

these traits, breeders create a favorable " genetic environment" for fertility. The objective

44

would be simply to set the levels of other traits in such a way that expressed fertility is
optimized. This is the point at which genotype x environment interactions are of major
concern, and producers must be aware of management level required to achieve desired
fertility levels.

The success of either approach to seedstock breeding will depend on the extent to which
traits of interest are incorporated into national genetic evaluations. With potentially vast
amounts of family and progeny data available, it is possible to accurately identify sires with
exceptional breeding values for even lowly heritable traits. Many breed associations
currently report expected progeny differences (EPD) for scrotal circumference and
additional indicators of fertility traits, such as longevity (Limousin, Gelbvieh, Red Angus).
Obviously, whole herd reporting and comparable efforts by breeders regarding data
collection and submission to the breed association are critical if reliable accuracies are to be
generated. Used correctly, this information accommodates a rate of progress inconceivable
through traditional phenotypic selection.

As with other traits, genetic improvement in fertility traits will come largely through sire
selection. Strict culling regimens based on expectations of heifer and cow reproductive
performance will satisfactorily eliminate the problem females from the herd. However, darn
selection, as a method for improving fertility, suffers from poor accuracy of breeding value
prediction and low selection intensity. The key to genetic improvement of fertility is
identifying and using superior sires.

The direct and indirect approaches to breeding for improved fertility are appropriate for
comrrlercial herds as well. Commercial breeders generally have more flexibility, however,

of being able to apply each approach on a between and within-breed basis. Using the direct

45

approach, they can choose outstanding sires for SC and AP within breeds, utilize them
according to knowledge of breed differences for fertility and other economic traits, and reap
the additional advantages of heterosis on reproductive fitness. Commercial breeders can
also use the indirect approach, once the proper economic balance of milk, growth, mature
size and fertility has been determined. More importantly, although reliance on heterotic
effects in the crossbred cow to provide ample expressed fertility is justified, sire selection
remains critical.

Limousin breeders have the opportunity to improve the reproductive ability of their
cattle and maintain breed strengths of muscling and efficient growth. The extent to which
they prioritize this task will largely determine the role Limousin cattle play in the beef

industry of the future.

46

Chapter 3

Phenotypic relationships between reproductive fitness and composition traits of

Limousin cattle.

D. D. Faidley‘, B. D. Banks“, R. J. Tempelman“, H. D. Ritchie‘, K. J. Andersenz, and
D. G. LeFever3
1Michigan State University, E. Lansing, 2North American Limousin Foundation,

Englewood, CO, 3Colorado State University, Fort Collins.
Abstract

Phenotypic relationships between composition traits and indicators of fertility were
evaluated from Limousin field data collected at Running Creek Ranch, Elizabeth, CO.
Yearling reproductive tract score (RTS) of 1,575 heifers and scrotal circumference (SC,)
of 1,247 bulls were analyzed separately. Traits measured included yearling weight (WT,
kg), RTS of heifers, SC (cm) of bulls, body condition score (BCS), and muscle score
(MS). For RTS analyses, heifers were either grouped as cycling (n=1134, RTSZ4) or
non-cycling (n=442, RTSS3). Logistic regression analyses were performed on heifers
using SAS" to model the dichotomous response variable (cycling vs. non-cycling) as a
function of the main effects of heifer age (AGE, days), age of dam (AOD), WT, percent
Limousin (PL), year (YR), MS, BCS. The model for SC was similar, with the addition of

the WT by MS interaction. Percent Limousin, additional interactions, and quadratic

47

 

mUSt

WI 1
Thcs

imp:

main:

Care a1

 

 

terms for WT and AGE were not important in either model (P>.1). In general, increases
in WT, AGE and BCS favorably affected both the probability of cycling and SC. Odds
ratio estimates for cycling of BCS 6 and BCS 7 heifers relative to BCS 5 were 8.272
(p<.001) and 3.333 (p<.001), respectively. Regression coefficients for SC on WT and
AGE were .021, cm/kg and .015, cm/day, respectively. Solutions for SC on BCS
increased .6 cmfor a BCS increase from 5 to 6 and 1.2 cm for an increase from 5 to 7
(p<.05). There were sex effects of muscularity on reproductive fitness traits. Favorable
odds ratio estimates for MS 2 and MS 3 heifers on RTS were 1.905 and 1.896 relative to
MS 1 (p<.05). Average and heavy muscled heifers did not differ in probability of cycling
(p>. 5). However, heavier muscled bulls had significantly smaller scrotal measurements
than average and light muscled bulls (p<.05). Solutions of SC for light and average
muscled bulls were .783 :1: .335 and .485 :l: .136, cm. Regression coefficients for SC on
WT by M8 were .018 :l: .005 and .007 :t .003, cm/kg for MS 1 and MS 2, respectively.
These results pose selection and management concerns for Limousin breeders wishing to

improve fertility and maintain muscularity of their cattle.

Introduction and Objectives

Breeders of Limousin cattle have targeted reproductive fitness for improvement while
maintaining inherent breed muscle advantages. Relationships between fertility and
carcass composition traits are not well defined, but represent important considerations in

breeding stock selection.

48

The dichotomous nature of fertility traits (e.g. cycling vs. non-cycling, pregnant vs.
non-pregnant) does not facilitate distinction of the subtle differences in inherent fertility
that exist between females. Furthermore, poor accuracy of breeding values for fertility
traits coupled with lower selection intensity in female populations restrict progress
through dam selection. Consequently, sire selection for fertility traits, such as scrotal
circumference, is considered the most rapid method for improving inherent fertility in
beef cattle.

Some have speculated that age at puberty in heifers is influenced by the same genes as
scrotal circumference in yearling bulls due to strong, favorable genetic correlations
(Brinks et al., 1978; King et al., 1983). Correlations among breed means (Lunstra, 1982;
Gregory et al, 1991; Gregory et al, 1992; Cundiff, et al., 1993) have supported this
conclusion, and indicated phenotypic expression of scrotal circumference, age at puberty
and other fertility traits were lower in Limousin cattle than other breeds. Moreover, both
female age at puberty and scrotal circumference have been shown to be favorably related
to growth, reproduction and maternal performance. (Laster et al., 1979; Werre, 1980;
Bourdon and Brinks, 1986; Werre and Brinks, 1986; Smith et al., 1989)

Gregory et al. (1995) reported small favorable genetic and phenotypic correlations for
body condition score with scrotal circumference in yearling bulls and age at puberty in
yearling females. Limousin heifers were shown to be the leanest of nine breed groups in
the Germ Plasm Evaluation Project (GPE) at the Meat Animal Research Center, Clay
Center, NE, and among the latest for age at puberty (Gregory, et al., 1991).

Subjective muscle scores (1 = extremely light muscled to 9 = extremely heavy

muscled) were assigned to yearling bulls in the GPE project (Gregory et al., 1995) and

49

shown to have a negative genetic relationship (-. 17307) with scrotal circumference, but
no phenotypic relationship. Moser et al. (1996), reported no difference in composition
among yearling crossbred progeny of Limousin sires with high and low adjusted yearling
scrotal circumferences (difference=8 cm). However, sires with average SC expected
progeny differences (EPD) had progeny with significantly larger weaning ultrasonic
ribeye area measurements than those of low SC EPD sires.

Therefore, the objectives of this study were to:

1) model the effects of composition on indicators of fertility in yearling cattle and,

2) determine if those relationships (if any) were different between males and

females.

Materials and Methods

Heifer Population.

Data. Field data collection was initiated in 1990 on 2179 yearling Limousin heifers
as part of ongoing research at Running Creek Ranch, Elizabeth, CO. Traits measured
included yearling weight (WT, kg), body condition score (BCS) as described by Ritchie et
al. (1992), reproductive tract score (RTS as described by Andersen et al. (1991), and
muscle score (MS) which will be described later. All data collection was performed by
the same, trained technician. Herdbook information for Running Creek Ranch was
provided by the North American Limousin Foundation (NALF), Englewood, CO. Heifer
age (AGE, days), age of dam (AOD, years), and percent Limousin (PL) information was

obtained from the NALF herd file.

50

Heifer Management. The heifers included in this analysis were recorded in mid-May
when the majority were approximately 13 months of age. Calves were not creep fed.
Following weaning, heifers were wintered in a single contemporary group under dry-lot
conditions. No healthy heifers were culled or removed from the replacement heifer group
until after they had been measured as yearlings. Heifers were fed a silage-based ration
adequate for maintenance and approximately .8 kg/day gain. The herd objective is to
retain or sell as many replacement females as possible. Consequently, the feeding
objective was to maximize the number of heifers cycling at the time of measurement.
Therefore, heifers were grown at a rate sufficient to ensure that optimum weight and body
condition levels for cycling were reached. Cycling heifers were implanted with Syncro-
Mate B at the time of recording to synchronize estrus at the start of the breeding season.

Reproductive Tract Scoring System. Several researchers have studied age at puberty
(AP) in beef heifers when defined as the age at first behavioral estrus. However, because
AP is difficult and labor-intensive to measure, direct selection for age at puberty in females
is seldom practiced. Andersen et al. (1991) described the reproductive tract scoring (RTS)
system developed at Colorado State University which estimates pubertal status via rectal
palpation of the uterine horns and ovaries. Each heifer is assigned a score of l (infantile)
through 5 (palpable corpus luteum) as described in Table 3.1.

A RTS of 1 is assigned to heifers with infantile tracts as indicated by small, toneless
uterine horns and small ovaries devoid of significant structures. Heifers scored as l are
likely the furthest from cycling at the time of examination. Heifers given a RTS of 2 are
thought to be closer to cycling than those scoring 1 due primarily to the presence of small

follicles and slightly larger uterine horns and ovaries. Those heifers assigned a RTS of 3 are

51

thought to be on the verge of cycling based on slight uterine tone in addition to the presence

of follicles. Heifers scoring 4 are presumably cycling as indicated by good uterine tone,

uterine size, and follicular growth. However, heifers with tract scores of 4 lack an easily

distinguished corpus luteum due to the stage of the estrous cycle. Heifers with tract scores

of 5 are similar to those scoring 4 except for the presence of a palpable corpus luteum.

Table 3.1. Description of reproductive tract scoresa

 

 

 

Ovaries (approx. size)
Repro.
tract Uterine Length Height Width Ovarian
scores horns (mm) (mm) (mm) structures
1 Immature < 20 mm No palpable
diameter - no tone 15 10 8 follicles
2 20-25 mm diameter - no
tone 18 12 10 8 mm follicles
3 25-30 mm diameter - 8-10 mm follicles
slight tone ' 22 15 10
< 10 mm follicles
4 30 mm diameter - good corpus luteum
tone 30 16 12 possible
>10 mm follicles
5 > 30 mm diameter - corpus luteum
good tone, erect >32 20 15 present

 

3 Reproductive tract score was determined less than 1 mo. prior to start of breeding season.

52

Muscle Score. Numerous studies have reported use of a scoring system to describe
muscularity differences between cattle (Gregory et al., 1995; Koch et al., 1995; Wolfe et
al., 1990). Subjective scores were assigned utilizing visual appraisal of an animal’s
topline, hip, rear quarter, and base width. For these analyses ordinal scores ranged
between 1 and 3.5 with higher levels indicating greater muscularity as follows:

1.0) Light muscled. Narrow loin, hip and stance and tapered through the quarter. Little
visible muscle tissue.

1.5) More apparent muscling than in muscle score 1, yet flat quartered and narrow
throughout.

2.0) Average muscling. Moderate muscle shape in topline and rear.

2.5) Above average muscularity.

3.0) Heavily muscled. Wide stance, thick and expressive quarter and hip, broad loin
with visible muscle shape throughout topline.

3.5) Very heavily muscled; Abundant visible muscular shape and dimension

throughout.

Data Editing. For these analyses heifers were either grouped as cycling (72%,
RTSZ4) or non-cycling (28%, RTSS3). Thus, RTS was used to simply distinguish
between pre-pubertal heifers and heifers who had reached puberty. Yearling MS ranged
from 1 to 3.5 and were grouped as light (MS<2), average (MS 2) or heavy muscled
(MS>2). All heifers had BCS between 5 and 8. Percent Limousin (PL) was included to
account for potential heterosis effects. Initially, PL was grouped according to NALF

specifications as: 37-50, 51-75, 76-87, 88-93, 94-100. However, due to the existence of

53

predominantly high percentage cattle and resulting small class sizes in the lower
percentage categories, PL was regrouped: < 75, 75-87, > 87. Year and PL effects were
reparameterized to sum to zero, so reported solutions are for the mean YR and mean PL.
Age of dam (AOD) was stratified for 2, 3, 4-10, and greater than 10 year old dams. The
mean age and actual weight at time of recording were subtracted from each record on age
and weight respectively, to minimize computer round off error. Heifers with missing
responses or explanatory covariates (n=603) were removed from the analysis. Only those

females with a complete array of response and explanatory variables remained.

54

Table 3.2. Data description for Running Creek Ranch females

 

 

Trait Mean Percent Range
Age (days) 398 308-473
Weight (WT, kg) 380 210-525

Percent Limousin (PL)

< 75% 15%
76-87% 28%
> 87% 57%
Age of dam (AOD)
2 18%
3 16%
4-10 58%
>10 8%
Reproductive Tract Score (RTS) 4.6 1-5
Cycling (2 4) 72%
Non-cycling (< 4) 28%
Muscle Score (MS) 2.2 1-3.5
Light (5 1.5) 15%
Average (= 2) 60%
Heavy Q 2.5) 25%
Body Condition Score (BCS) 6.2 5-8
5 2.5%
6 57%
7 39%
8 1.5%

 

55

Bull Population.

Data. Data collection on 1,472 yearling Limousin bulls at Running Creek Ranch,
Elizabeth, CO was also initiated in 1990. Traits measured were the same as those for the
heifer analysis with the substitution of SC for RTS.

Bull Management. Bulls were reared under similar conditions as their heifer mates,
though an unknown number were culled at weaning, presumably due to lack of
performance. The important distinction is that after weaning, bulls were grown at
proportionately slower rates relative to their gain potential than were females. Running
Creek Ranch has a substantial number of repeat and potential customers who prefer 18-
month old bulls for use in fall calving programs, and two-year old bulls that can be
exposed to more cows. Cost of gain is a more important economic factor than maximum
early growth to their bull production efforts. Consequently, there were no apparent rate
of gain targets for the bulls. Therefore, bull development strategies utilized lower quality,
more cost effective feeds. Bulls that met NALF recommendations for adjusted yearling
scrotal circumferences (232 cm.) could be sold in the spring as yearlings or later. Bulls
that failed to meet Society of Theriogenology criteria for breeding soundness
examinations at sale time (SC<30 cm.) were culled.

Data Editing. Procedures followed when grouping, AOD, PL, and MS for RTS
analysis were replicated prior to modeling SC. Ill this case, however, final PL classes
were 51 to 75, 76 to 87, 88 to 93, 94 to 100. All bulls had BCS between 4 and 7. Bulls
with missing data (n=225) were removed from the analysis. Only those males with a

complete array of response and explanatory variables remained.

56

Table 3.3. Data description for Running Creek Ranch males

 

Trait Mean Percent Range
Age (days) 383 308-473
Weight (WT, kg) 429 260-732
Percent Limousin (PL)
< 75% 15%
76-87% 21%
> 87% 64%
Age of dam (AOD)
2 13%
3 16%
4-10 63%
>10 8%
Scrotal Circumference (SC, cm) 32.1 22-42.3
2 34 25%
5 30 25%
Muscle Score (MS) 2.3 1-3.5
Light (5 1.5) ’ 3%
Average (= 2) 56%
Heavy (2 2.5) 36%
Body Condition Score (BCS) 5.8 4-7
4 .3%
5 27%
6 70%
7 2.4%
n = 1247

 

57

Reproductive Tract Score Analysis.

Logistic regression analyses were conducted using SAS® (Proc Logistic, 1995) to
model the dichotomous response variable (RTS) grouped as cycling (1) and non-cycling
(0) as a function of continuous and fixed categorical explanatory variables. Cycling
status was assumed to follow a binomial distribution with the probability of the i’h heifer
cycling equaling P,~. Logistic modeling is advantageous over standard linear modeling in
this instance as it guarantees the estimates obtained for probability of cycling will range

between 0 and l and allows a more tenable distributional assumption. The logistic

function is described as follows: f (z) = 1—1_-?- The value of logistic function is
+ e

dependent on the value of 2. To obtain the logistic model from the logistic function, 2 is
written as a linear function of independent variables (X’s).

z = 01 + [31X] + Bzxz + +13ka
The logistic model for the conditional probability of cycling given characteristics

X1,...,Xk, i.e. P(CY|X1,...,Xk) is written as follows:

1 e(a+p,x,+p,x,+...+p,x,) 1

= = = w ,
f(Z) 1 + e—(a+l31Xt+fl2X2+...+Ban) 1+ e(a+p‘x|+p2x2+m+ptxk) -(a+:p‘x‘) here (1, Bl
i-l

 

 

l+e

[32,..., 8., are parameters to be estimated. Parameter estimation is performed through

maximum likelihood techniques.

58

The final model included the main effects of AGE, WT, PL, year (YR), MS, and BCS.
P(Cycling) = natural log probability of cycling

0 a intercept

- YR year (1990, 1991, 1992, 1993, 1994, 1995, 1996)

0 PL percent Limousin (< 75, 75-87, > 87)

o AOD age of dam (2, 3, 4-10, >10)

0 MS muscle score (1, 2, 3)

o BCS body condition score (5, 6, 7, 8)

o AGE age at recording deviation from the mean age

0 WT actual weight at recording deviation from the mean weight

Note the absence of a normally distributed error term. Variability is defined by the

binomial distribution as:
Vaf(Pi) 5 P1 (1- Pi)
Scrotal Circumference Analysis.

Regression analyses were conducted using PROC MDfED of SASG. No random
effect was fitted, but PROC MIXED was selected for two reasons. First, general linear
models produced the same results as mixed model methodology in the absence of random
effects. Second, mixed models account for fixed and random effects, such as genetic
effects, and thus this phenotypic analysis serves as a precursor to a genetic analysis. The
fixed effects model used in this analysis is given by the following expression: y = XB + e

where y denotes the vector of observed scrotal circumferences, X is the known matrix of

59

explanatory variables, B is the unknown fixed effects parameter, and e is the unknown

vector of independent and identically distributed Gaussian random errors.

The final model included main effects: year (YR), PL, AOD, AGE, WT, MS, BCS

and YR by MS interaction. Additional interactions and PL were not important (P>.1).

Keeton et al. (1996) suggested either bull age or bull weight should be fit to a scrotal

circumference model, as the traits were highly and favorably correlated (.91). In these

data however, bull AGE and WT were related only a modest r=.37 (p<.0001).

Linear Model - Bulls

SC:

PL

AOD

MS

BCS

AGE

WT*MS

intercept

year(1990, 1991, 1992, 1993, 1994, 1995, 1996)

percent Limousin (< 76, 76-87, 88-93, >93)

age ofdarn (2, 3, 4-10, >10)

muscle score (1, 2, 3)

body condition score (4, 5, 6, 7)

age at recording deviation from the mean age

actual weight at recording deviation from the mean weight
weight by muscle score interaction

random error N ~ (0, 162.)

60

Rams

Probability of cycling.

Parameter estimates for probability of cycling are given in Table 3.4. The
baseline predicted probability of cycling for heifers was .27. This value corresponds to a
heifer of mean percent Limousin, born in the mean year to a >10 year old dam, of average
age and weight, with body condition score 5 and muscle score 1. As expected, heifer age
was the most significant source of variation (p<.000), and had the largest chi-square
value. Although WT was not as important (p = .067), the authors felt it was questionable
enough to remain in the model. Furthermore, each kg increase in weight increases
probability of cycling by one-tenth of a percent. The effects of age and weight are
depicted graphically in figure 3.1.

Efiects of composition on probability of cycling. In general, heavier muscled and
fatter heifers were more likely'to be cycling than their lighter muscled and/or leaner
counterparts. Predicted cycling status was not different among average and heavy
muscled heifers according to Wald tests (p>.1), but both groups were advantageous to
light muscled heifers regarding probability of cycling (p<.05). These muscle score effects
are depicted in figure 3.2.

Although muscle score effects were significant, body condition score was a more
important factor affecting predicted probability of cycling. Increases in BCS had favorable
affects on reproductive tract maturity. Obese heifers (BCS 8) were the exception, as their
odds for cycling were lower numerically than those heifers in good condition (BCS 7).

However, Wald tests for significant differences between body condition scores revealed

61

BCS 8 heifers were not different than BCS 6 or BCS 7 (p>.05). Heifers recording body
condition scores of 8 were a relatively small percentage of the population and associated
parameter estimates had a larger standard error than other reported solutions. All other
body condition score classes were significantly different from each other (p<.05). Body
condition score effects are plotted in figure 3.3.

The combined effects of muscularity and body condition indicate the relative
importance of each trait on predicted probability of cycling. Fatter and heavier muscled
heifers were the most likely to be cycling. The leanest and lightest muscled heifers were the
least likely to be cycling. Figure 3.4 indicates that the combination of fatter and lighter
muscled heifers is preferable to lean, heavy muscled heifers regarding reproductive tract
maturity.

Some may be concerned about possible confounding of muscle score and body
condition scores. Logic dictates muscle and fat accretion accompany increases in age and
weight under conventional management practices. Subjectively assigned scores such as
these are certainly more prone to bias than are objectively measured traits. Step-wise
selection procedures were utilized to test interactions between composition and continuous
traits. As previously mentioned, no interactions explained significant variation above and
beyond that described by the main effects model. Likewise, model fit statistics indicated the
main effects model sufficiently described the relationships of age, weight, body condition
and muscularity with reproductive fitness. The combined effects of age, weight, body
condition and muscularity are described in figure 3.5.

These data suggest that selection of heavier muscled replacement heifers, given

adequate weight and body condition score, is not detrimental to reproductive fitness.

62

Table 3.4. Parameter estimates for probability of cycling.a

 

 

Variable Parameter SE Chi- Pr < Chi- Odds
Estimate Square Square Ratio
Intercept -1012 .378 7.177 .007 -
Weightb (kg) .004 .002 3.348 .067 1.004
Ageb (days) .018 .003 29.634 .000 1.018

Muscle Scorec

3 .640 .262 5.947 .015 1.896
2 .645 .231 7.822 .005 1.905
1 - - - - -
Body Condition Scored
8 1.862 1.1 1 1 2.827 .093 6.476
7 , 2.1 13 .395 28.549 .000 8.272
6 1.204 .355 12.912 .000 3.333
5 - - - - -
Age of Dam
2 -.085 .135 .391 .532 .919
3 .408 . 146 7.797 .005 1 .504
4-10 .176 .100 3.087 .079 1.193
>10 - - - - -

 

“ Year and percent Limousin were reparameterized to sum to zero; thus estimates are
provided for the mean year calved and percent Limousin.
Expressed as deviations from the mean.
c MS; 3 = heavy muscled, 2 = average muscled, 1 = light muscled.
“ BCS; 8 = Obese, 7 = Good, 6 = High Moderate, 5 = Moderate

63

me

 

Booze ca<+£9o>>ll mfioto om<1|r 9023 2903131 J

 

:38 05 52— 26.6.30 3.: 2203 new Agony en<

mm mm m_. m m. m —. mm. mm- m?

 

 

 

 

 

 

 

 

 

 

..d e5 E, :85 .$.uoo< ..nwz .mumom $5.26 .o no.8 2:83 a 23.2 .o. 859:8 so 85332..

 

 

6338:. 2a 2925 new one
52. we 30:23.5 ...oEuonEoo 5.3 9.3.2. ..2 05.23 .0 3:330... .pd 0.52".

§. s
O O
5011vo )9 Arlllqoqmd

 

_ 8.22221? 5.21.1 q

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:33: 5:53.. eu<
9. mm mm 9 m m- m..- 3. mm- m?
.E eco ¢> .2; coo... .o.-vuoo< .enmom ”22.2. .2 8:558 2o 8.53326.

 

 

03:28:“. one new 0.02::
5.3 9.8.2.. 0 Soon :oEucoo >02. ..o. 9523 .0 32330.... .u.n 0.50.“.

cod

o...o

0N6

and

.. ovd

omd

and

056

cod

cad

ooé

6111195910 6:111:10de porcupine

65

 

omomll Bowl-1 88' mmomllm

 

Eon. Season o3
mv mm mm mp m m. 9. mm- mm-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.E can m> .>>> cmoE .oTvnDO< .Nqu ”20:0: 3. 859:8 2a 82.5395

 

 

moo—3.6:... one new
coineoo >uon 5.3 9.8.2. 00.33:. ounce; .6. 05.2.0 .0 3.3395 .m.» 0.59...

cud

ONO

61111vo 10 Mllqaqmd

66

 

Tmombwzlel 38.8.2111 58.5211 38.5211;

mv mm mm mp

333 eozegeu ou<
m m.

p 1P -

m?

 

 

 

 

 

 

 

 

 

 

'
|
| 1
|
' ‘
| ‘
' ‘
II 41‘
I'I ‘I‘
I'l ‘al‘
"|' 41“
"" “‘ I
I ‘ ‘ I
4‘ .1
1

 

 

 

 

..o use 5 .2; =85 .8433 ”use; 3. 3.358 2o 82.532“.

 

 

one new :oEeanoo 5.3 9.032.. ..o. 9526 .0 3:32.95 6.» 2:2“.

30:23::

cod

opd

cad

o
".
o

o
‘9
o

O
In
0
5911959 to MINIMUM

owd

cad

co.—

67

mv

 

fl 38.8.2101 88.85.11 58.5.2l 88.55.1314

:35 05 E9: 30.3.30 3.: 2.0.03 new .33. ou<
mm mm m. m m- 3. mm- 8. me.

#1 P h 3 b p h P F b n h h

 

 

 

 

 

 

 

 

 

 

 

 

\

 

 

 

\
\\

op-vuoo< new m> ..5 :35 8 92.3 3. 8.388 93 8.53305

 

.3398... 2a 30.0.... new one 52.
3 30:20:... 20:32.53 5.! 28.3 3. 05.03. .0 3.3395 .md 9505

000.0

00F.0

00N.0

000.0

0010

000.0

000.0

Buuafio #0 Alumnae“

005.0

000.0

000.0

000. p

68

Scrotal Circumference.

Parameter estimates for fixed factor effects on yearling scrotal circumference are
given in table 3.5. The mean scrotal circumference was 31.9 cm., which is close to
Limousin breed recommendations for yearling bulls (Directions, Limousin Breeders
Symposium, 1990). Both weight and age were significant sources of variation in scrotal
circumference (p<.05). As expected, scrotal circumference increased with increasing age
and weight. Regression coefficients for WT and AGE were .0224, cm/kg and .0075 ,
cm/day respectively. AOD effects were important for sons of first calf heifers and
consistent with estimates and adjustment factors reported in literature. Sons of 3-year old
and mature dams did not have significantly different scrotal circumferences, but
parameter estimates tended to follow literature estimates.

Composition eflects on scrotal circumference. Somewhat contrary to estimates of
heifer fertility, fatter and lighter muscled bulls had larger scrotal circumferences (Figure
3.6). Excluding BCS 4 (n=4),bulls with higher BCS values had larger yearling SC.
Solutions for SC on BCS increased .6 cm for a BCS increase from 5 to 6 and 1.2 cm for
an increase from 5 to 7 (p<.05). Heavier muscled bulls had significantly smaller scrotal
measurements than average and light muscled bulls (p<.05). Solutions of SC for light
and average muscled bulls were .808 i .338 and .479 i .136, cm. Additionally, there was
a weight by muscle score interaction which contributed to the negative effect of increased
muscularity on scrotal circumference (figure 3.7). Coefficients for SC on WT by MS

were .018 :l: .005 and .007 :t .003, chkg for MS 1 and 2, respectively.

69

Table 3.5. Parameter estimates for yearling scrotal circumference.

 

 

Variable Parameter SE T Pr < ITI
Estimate
Intercept 31.928 .485 65.84 .000
Weighta (kg) .021 .002 8.75 .000
Agezal (days) .015 .004 3.71 .000
Muscle Scoreb
1 .783 .335 2.33 .019
2 .485 .135 3.58 .000
3 - - - -
Wt*MSc
1 .018 .005 3.27 .001
2 .007 .003 2.33 .020
3 _ - - -
Body Condition Scored
4 -.816 1.155 -.71 .480
5 -l.252 .430 -2.91 .004
6 -.621 .394 -1.58 .116
7 - - - -
Age of Dam
2 -.569 .275 -2.07 .039
3 -.161 .256 -.63 .529
4-10 .225 .223 1.01 .314
>10 - - - -

 

3‘ Expressed as deviations from the mean.

b MS; 1 = light muscled, 2 = average muscled, 3 = heavy muscled
° Interaction between weight and muscle score class.

d BCS; 4 = Thin, 5 = Moderate, 6 = High Moderate, 7 = Good.

70

 

_ enmom a $9.17 53¢ a «62'. muwom a .umzlol mung a mums. 1.14

 

Ago... 5.533 ou<

 

 

8
fl

 

8
8

 

 

I
1

8

8

 

1
\
. 8. 8. '
a 5
(we) aouemwmauo mores pampord

 

 

8.
3

07vuoO< 0:0 m> ...n. .>>> ...on “a 3.32 3. 02530.8 20 950302.”.

 

 

3:238:26
.220» 3.33.... :0 33:0 eoEmonEoo 03mins 00¢ d.» 950.“.

71

 

nnmom .u pumilll numom _u mum—2+ manm a pumzl muwom a mums—Isl.
a... 5.3.25 2.22...

 

 

 

 

 

 

 

 

 

 

2.39 o5 2; .... .93 =85 a 2.3 .o. 8.28.8 so 3828:.

 

 

3:235:26 .223 :0 23:0 eoEmonEoo 02253 20.25 K.» 2:2“.

00.50

(ma) aouerqwneua moms pezogperd

72

Discussion

The relationships between composition and indicators of reproductive fitness in yearling
cattle presented here challenge the perspective of many involved with beef cattle breeding.

It is commonly acknowledged that reproductive fitness traits of Limousin cattle generally
require improvement, yet the amount of progress required to meet industry demands is
debatable. Heifers cycling at a year of age will undergo at least three heat cycles before
being bred to calve at 24 months. Certainly, favorable relationships have been well
established between reduced age at puberty and lifetime fertility and productivity of a cow.
However, the role of the Limousin breed in the beef industry mandates genetic advancement
of fertility traits be conducted in the context of maintaining breed strengths. The data
indicate Limousin breeders appear to have that opportunity.

Commonly accepted philosophy dictates maternal ability, including fertility and milk
genetics, is of primary importance in replacement heifer selection and that muscularity can
be ignored in the female and addressed through sire selection. However, if greater muscling
is not detrimental to reproductive fitness in yearling Limousin heifers, given adequate
weight, body condition, and a comparable environment, then breeders should retain heavier
muscled females. Doing so allows Limousin breeders to accomplish two important goals.
First, breed muscle advantages will be maintained without reliance upon utilization of the
heaviest muscled bulls in the breed. At the same time, intense sire selection pressure can be
directed towards fertility trait improvement and other breed weaknesses. Therefore,
Corrective mating schemes can be implemented in the manner that will produce the most

rapid genetic improvement, through sire selection. The implications of this are substantial

73

for Limousin seedstock producers wishing to supply the beef industry with muscular
terminal sires and retain or merchandise female siblings as replacements.

The cause of the sex effects of muscularity on reproductive fitness traits is disconcerting
and unknown at this time. Conceivably, composition traits that appear positively correlated
with phenotypic expression of fertility may in fact have negative, underlying genetic
relationships expressed differently in males and females. Perhaps the discrepancy is a
function of the differing maturing rates for bulls and heifers. One might theorize that
composition differences have varying effects depending upon the point on the growth curve
at which they are measured. Perhaps herd management is the essential element
distinguishing between bulls and heifers. Heifers at Running Creek Ranch must breed to
calve as two-year-olds to maximize production efficiency. Therefore, heifers are fed so that
as many as possible can achieve cycling status prior to the breeding season. Two-year-old
virgin bulls, however, are in sufficient demand that there is not as much pressure placed on
reproductive maturation rate in the male population.

Many will contend the scrotal circumference results justify a belief that heavy muscled
bulls, specifically muscular Limousin bulls, are useful solely as terminal sires. These
convictions appear well founded as documentation states that daughters of sires with
smaller yearling scrotal circumferences are older at puberty and less reproductively efficient

throughout their lifetimes. My belief is the results suggest against attempting to utilize
extremes, i.e. compensating for light muscled heifers by utilizing the heaviest muscled

bulls.

74

Implications

Limousin breeders have the opportunity to retain heavy muscled replacement females
with confidence in their reproductive ability, provided heifers possess the genetic
potential required and receive management levels sufficient to achieve adequate weight
and body condition prior to breeding. Moreover, selection efforts directed towards

fertility trait improvement through sire selection, in which maximum genetic response is .

 

expected, without sacrificing breed advantages in muscularity is possible. A paradigm
shift may be required as Limousin seedstock producers look to females in their attempt to
maintain muscularity and rebalance fertility trait and composition trait emphasis among

potential sires.

75

Chapter 4

Genetic and composition effects on reproductive fitness traits in yearling Limousin

cattle.

D. D. Faidleyl, B. D. Banks', R. J. Tempelmanl, H. D. Ritchie‘, K. J. Andersenz,

and D. G. LeFever3

1Michigan State University, E. Lansing, 2North American Limousin Foundation,

Englewood, co, 3Colorado State University, Fort Collins.

Abstract

Genetic and composition effects on indicators of fertility were evaluated from Limousin
field data as part of ongoing research at Running Creek Ranch, Elizabeth, CO. Yearling
reproductive tract score (RTS) of 1,578 heifers and scrotal circumference (SC) of 1,248
bulls were used in the analysis. Traits measured included yearling weight (WT, kg), RTS of
heifers, SC of bulls, body condition score (BCS), and muscle score (MS). Dam age (AOD)
was categorized for dams 2, 3, 4 to 9, and 210 yr. Percent Limousin (PL) was classified
S75, 7687, 88-93 and >93%. Yearling MS were grouped as light (MSd), average (MS 2)
or heavy muscled (MS>2). Heifer BCS ranged between 5 and 8, while all bulls had BCS
between 4 and 7. Reproductive tract ordinal scores of heifers ranged from 1 to 5. Higher

RTS scores indicate greater reproductive tract maturity. For RTS analyses, heifers were

76

either grouped as cycling (n=l 134, RTSZ4) or non-cycling (n=444, RTSS3). Univariate
and bivariate analyses were conducted to model RTS and SC as functions of main effects:
animal age (AGE, days), WT, AOD, PL, year (YR), MS, BCS and random sire effect.
Reduced models in which the covariates AGE or WT were removed were also tested. All
fixed effects except PL were important (p<.05) in at least one model. In general, increases
in WT, AGE and BCS favorably affected both the probability of cycling and SC. Average
and heavy muscled heifers were favorable to light muscled heifers for RTS, while their
heavy muscled bull mates had the smallest SC estimates. Heritability estimates ranged from
.04 to .13 for RTS and .23 to .37 for SC. Surprisingly, genetic correlation coefficients

between the response variables ranged between —.2 and .02.
Introduction and Objectives

Breeders of Limousin cattle have targeted reproductive fitness traits for improvement
without sacrificing inherent breed muscle advantages. Relationships between fertility and
composition traits are not well defined, but constitute important breeding stock selection
considerations.

The binary nature of many fertility traits (e. g. cycling vs. non-cycling, pregnant vs.
non-pregnant) does not facilitate distinction of the subtle differences in inherent fertility
that exist between females. Furthermore, poor accuracy of breeding values for fertility
traits coupled with lower selection intensity in female populations restrict progress

through darn selection. Consequently, sire selection for fertility traits, such as scrotal

77

circumference, is considered the most rapid method for improving inherent fertility in
beef cattle.

Others have speculated age at puberty in heifers is essentially the same trait as scrotal
circumference in yearling bulls due to strong, favorable genetic correlations (Brinks et al.,
1978; King et al., 1983). Correlations among breed means (Lunstra, 1982; Gregory et al,
1991; Gregory et al, 1992; Cundiff, et al., 1993) have supported this conclusion. These
studies also suggest phenotypic levels of performance of Limousin cattle for scrotal
circumference, age at puberty and other fertility traits are lower than those of other
breeds. Moreover, both female age at puberty and scrotal circumference are favorably
related to growth, reproduction and maternal performance. (Laster et al., 1979; Werre,
1980; Bourdon and Brinks, 1985; Werre and Brinks, 1986; Smith et al., 1989)

Gregory et al. (1995) reported small, but favorable genetic and phenotypic
correlations for body condition score with scrotal circumference in yearling bulls and age
at puberty in yearling females. Limousin heifers were shown to be the leanest of nine
breed groups in the Germ Plasm Evaluation Project at the Meat Animal Research Center,
Clay Center, NE, and among the latest for age at puberty (Gregory, et al. 1991).
Subjective muscle scores (1 = extremely light muscled to 9 = extremely heavy muscled)
were assigned to yearling bulls in the GPE project (Gregory et al., 1995) and shown to
have a negative genetic relationship (-. 17:.07) with scrotal circumference, but no
phenotypic relationship. Moser et al. (1996), reported no difference in composition
among yearling crossbred progeny of Limousin sires with high and low adjusted yearling

scrotal circumferences (difference=8 cm). However, sires with average SC expected

78

progeny differences (EPD) had progeny with significantly larger weaning ultrasonic
ribeye area measurements than those of low SC EPD sires.
Therefore, the objectives of this study were to:
1) Account for composition differences in a genetic analysis of reproductive trait
indicators in yearling cattle,
2) Estimate a genetic correlation between scrotal circumference and cycling status in

half-sibs.

Materials and Methods

Heifer Population.

Data. Field data collection was initiated in 1990 on 2179 yearling Limousin heifers
as part of ongoing research at Running Creek Ranch, Elizabeth, CO. Traits measured
included yearling weight (WT, kg), body condition score (BCS) as described by Ritchie et
al. (1992), reproductive tract score (RTS as described by Andersen et al. (1991), and
muscle score (MS) which will be described later. All data collection was performed by
one, trained technician. Herdbook information for Running Creek Ranch was provided
by the North American Limousin Foundation (NALF), Englewood, CO. Heifer age
(AGE, days), age of dam (AOD, years), percent Limousin (PL) and sire information was
obtained from the N ALF herd file.

Heifer Management. The heifers included in this analysis were recorded in mid-May
at approximately 13 months of age. Calves were not creep fed. Following weaning,

heifers were wintered in a single contemporary group under dry-lot conditions. No

79

healthy heifers were culled or removed from the replacement heifer group until after they
had been measured as yearlings. Heifers were fed a silage-based ration adequate for
maintenance and approximately .8 kg/day gain. The herd objective is to retain or sell as
many replacement females as possible. Subsequently, the feeding objective was to
maximize the number of heifers cycling at the time of measurement. Therefore, heifers
were grown at a rate sufficient to ensure that adequate weight and body condition levels
for cycling were reached. Cycling heifers were implanted with Syncro-Mate B® at the
time of recording to synchronize estrus at the start of the breeding season. Non-cycling
heifers were culled.

Reproductive Tract Scoring System. Several researchers have studied age of puberty
(AP) in beef heifers when defined as the age at first behavioral estrus. However, because
AP is difficult and labor-intensive to measure, direct selection for age at puberty in females
is seldom practiced. Andersen et al. (1991) described the reproductive tract scoring (RTS)
system developed at Colorado'State University which estimates pubertal status via rectal
palpation of the uterine horns and ovaries. Each heifer is assigned a score of 1 (infantile)
through 5 (palpable corpus luteum) as described in Table 4.1.

A RTS of l is assigned to heifers with infantile tracts as indicated by small, toneless
uterine horns and small ovaries devoid of significant structures. Heifers scored as 1 are
likely the furthest from cycling at the time of examination. Heifers given a RTS of 2 are
thought to be closer to cycling than those scoring 1 due primarily to the presence of small
follicles and slightly larger uterine horns and ovaries. Those heifers assigned a RTS of 3 are
thought to be on the verge of cycling based on slight uterine tone in addition to the presence

of follicles. Heifers assigned a score of 4 are presumably cycling as indicated by good

80

uterine tone, uterine size, and follicular growth. However, heifers with tract scores of 4 lack
an easily distinguished corpus luteum due to the stage of the estrous cycle. Heifers with

tract scores of 5 are similar to those scoring 4 except for the presence of a palpable corpus

luteum.

Table 4.1. Description of reproductive tract scores'

 

Ovaries (approx. size)

 

 

 

Repro.
tract Uterine Length Height Width Ovarian
scores horns (mm) (mm) (mm) structures
1 Immature < 20 mm No palpable
diameter - no tone 15 10 8 follicles
2 20-25 mm diameter - no
tone 18 12 10 8 mm follicles
3 25-30 mm diameter - 8-10 mm follicles
slight tone 22 15 10
< 10 mm follicles
4 30 mm diameter . good corpus luteum
tone 3O 16 12 possible
>10 mm follicles
5 > 30 mm diameter - corpus luteum
good tone, erect >32 20 15 present

 

a Reproductive tract score was determined approximately 1 mo prior to breeding.

81

Muscle Score. Numerous studies have reported use of a scoring system to describe
muscularity differences between cattle (Gregory et al., 1995; Koch et al., 1995; Wolfe et
al., 1990). Subjective scores were assigned utilizing visual appraisal of an animal’s
topline, hip, quarter, and base width. For these analyses ordinal scores ranged between 1
and 3.5 with higher levels indicating greater muscularity as follows:

1.0) Extremely lacking in muscularity. Narrow loin, hip and stance and tapered
through the rear quarter. Little visible muscle tissue.

1.5) More apparent muscling than in muscle score 1, but still fiat quartered and narrow
throughout.

2.0) Average muscling. Moderate visible muscle shape in topline and rear quarter.

2.5) Above average muscularity.

3.0) Heavily muscled. Wide stance, thick and expressive quarter and hip, broad loin
with visible muscle shape throughout topline.

3.5) Very heavily muscled. Abundant visible muscular shape.

Data Editing. For these analyses heifers were either grouped as cycling (72%,
RTSZ4) and coded as 1, or non-cycling (28%, RTSS3) coded as 0. Thus, RTS was used
simply to distinguish between pre-pubertal heifers and those who had achieved pubertal
status. Yearling MS ranged from 1 to 3.5 and were grouped as light (MS<2), average
(MS 2) or heavy muscled (MS>2). All heifers had BCS between 5 and 8. Percentage
Limousin (PL) was included to account for potential heterosis effects. Initially, PL was
grouped according to NALF specifications as: 37-50, 51-75, 76-87, 88-93, 94-100.

However, due to the existence of predominantly high percentage cattle and resulting

82

 

small class sizes in the lower percentage categories, PL was regrouped: 375, 76—87, 88-92
and >92%. Age of dam (AOD) was stratified for 2, 3, 4-9, and dams 10 years old and
older. The mean age and actual weight at time of recording were subtracted from each
age and weight record to minimize computer round off error. Heifers with missing data
(n=601) were removed from the analysis. Only those females with a complete array of

response and explanatory variables remained (n=1578, Table 4.2).

83

Table 4.2. Data description for Running Creek Ranch females-135 sires

 

 

Trait Mean N Percent S. D. Range
Age (days) 398 22.81 317-455
Weight (WT, kg) 380 4.91 216-525

Percent Limousin (PL)

< 75% 115 7.3%
76-87% 59 3.7%
88-92% 113 7.2%
> 92% 1291 81.8%
Age of dam (AOD)
2 283 17.9%
3 246 15.6%
4-10 911 57.7%
>10 138 8.8%

Reproductive Tract Score (RTS)

Cycling (1) ' 1134 71.9%
Non-cycling (0) 444 28.1%

Muscle Score (MS) 2.2 1-3.5
Light (5 1.5) 127 8.1%
Average (= 2) 985 62.4%
Heavy (2 2.5) 466 29.5%

Body Condition Score (BCS) 6.2 5-8

5 85 5.4%
6 1098 69.6%
7 381 24.1%
8 14 .09%

 

84

Bull Population.

Data. Data collection on 1,474 yearling Limousin bulls at Running Creek Ranch,
Elizabeth, CO was also initiated in 1990. Traits measured were the same as those for the
heifer analysis with the substitution of SC for RTS.

Bull Management. Bulls were reared under similar conditions as their heifer mates,
though an unknown number were culled at weaning, presumably drie to lack of
performance. The important distinction is that after weaning, bulls were grown at
proportionately slower rates relative to their gain potential than were females. Running
Creek Ranch has a substantial number of repeat and potential customers who prefer 18-
month old bulls for use in fall calving programs, and two-year old bulls they can expose
to more cows. Cost of gain is a more important economic factor than maximum early
growth to their bull production efforts. Consequently, there were no apparent average
daily gain targets for the bulls. Therefore, bull development strategies utilized lower
quality, more cost effective feeds. Bulls that met NALF recommendations for adjusted
yearling scrotal circumferences (232 cm.) could be sold in the spring as yearlings or later.
Bulls were culled that failed to meet Society of Theriogenology criteria for breeding
soundness examinations at sale time (SC<30 cm.).

Data Editing. Procedures followed when grouping, AOD, PL, and MS for RTS
analysis were replicated prior to modeling SC. All bulls had BCS between 4 and 7. Bulls
with missing data (n=224) were removed from the analysis. Only those males with a

complete array of response and explanatory variables remained (n=1248, Table 4.3).

85

Table 4.3. Data description for Running Creek Ranch males-127 sires

 

 

Trait Mean N Percent Range
Age (days) 383 298-471
Weight (WT, kg) 429 368-567
Percent Limousin (PL)
5 75% 86 6.9%
76-87% 46 3.7%
87-92% 77 6.2%
_>_ 93% 1039 83.2%
Age of dam (AOD)
2 164 13.1%
3 203 16.3%
4- 10 785 62.9%
>10 96 7.7%
Scrotal Circumference (SC, cm) 32.07 24-42.25
z 34 ' 27.6%
_<__ 30 25%
Muscle Score (MS) 2.3 1-3.5
Light (5 1.5) 102 8.2%
Average (= 2) 701 56.2%
Heavy (2 2.5) 445 36.6%
Body Condition Score (BCS) 5.8 4-7
4 9 .7%
5 339 27.2%
6 870 69.7%
7 30 2.4%

 

86

Univariate Analysis
Reproductive Tract Score. A sire model was used to examine random genetic effects
on cycling status. 135 sires were considered unrelated and fitted to the main effects
model that included AGE, WT, PL, AOD, year (YR), MS and BCS. The resulting linear
mixed model is generally described as:
y = Xb + Zu + e
where: y is a N x 1 vector of phenotypic observations,
b is a p x 1 vector of fixed effects associated with y,
u is a q x 1 vector of random effects associated with y,
X is a known incidence matrix of order N x p relating elements of b to elements
of y,
Z is a known incidence matrix of order N x q relating elements of u to elements
of y, and
e is a N x 1 vector of residual effects.
Additional attributes of the general form of mixed linear models include the expectations

of the random variable which include:

E(y) = Xb
E(u) = 0, and
E(e) = 0

Henderson (1953) developed the Best Linear Unbiased Prediction (BLUP) which is
used to predict random effects, such as breeding values, utilizing the mixed model. BLUP
is primarily used to identify individuals with maximum genetic merit in selection programs,

monitor response to selection and has become the dominant methodology for estimating

87

breeding values. The literature collection describing BLUP methodology is voluminous and
extensive (Henderson, 1977a, 1984a, 1988a; Schaeffer, 1991; Kennedy, 1991, Searle et al.,
1992; and Mrode, 1996)

Sire models are used to estimate breeding value of bulls from large arrays of
descendants (Lynch and Walsh, 1998). In this model the additive genetic value of each
offspring is expressed in terms of its sire’s breeding values and the dam contribution is
incorporated into the error term.

The final model for analysis of cycling status is as follows:

RTS = Cycling vs. non-cycling (1 or 0)

o a intercept

- YR year(1990, 1991, 1992, 1993, 1994, 1995, 1996)

0 PL percent Limousin (< 75, 75-87, > 87)

o AOD age ofdam (2, 3, 4-10, >10)

0 MS muscle score (1, 2, 3)

- BCS body condition score (5, 6, 7, 8)

0 AGE age at recording deviation from the mean

0 WT actual weight at recording deviation from the mean

0 SIRE random genetic effect of sire N ~(0, 102.)

0 e random error N~(0, I02.)

The author realizes the binary nature of RTS violates assumptions of a normally
distributed error term. Methodology to correctly account for this limitation of mixed

linear models will be discussed in a later section.

88

Scrotal Circumference. Mixed model analysis on the bull data was performed
utilizing the same methodology as the RTS analysis described above, with the
substitution of scrotal circumference as the response variable. The final model included
main effects: year (YR), PL, AOD, AGE, WT, MS, BCS and the random effect of sire.
Keeton et al. (1996), suggested either bull age or bull weight should be fit to a scrotal
circumference model, as the traits were highly and favorably correlated (.91). In these
data however, both bull AGE and WT proved to be significant sources of variation.

The final model for analysis of scrotal circumference is as follows:

SC = Scrotal Circumference

o a intercept

0 YR year(1990, 1991, 1992, 1993, 1994, 1995, 1996)

0 PL percent Limousin (< 76, 76-87, 88-93, >93)

0 AOD age of dam (2, 3, 4-10, >10)

0 MS muscle score (1, 2, 3)

o BCS body condition score (4, 5, 6, 7)

o AGE age at recording deviation from the mean

0 WT actual weight at recording deviation from the mean

0 SIRE random genetic effect of sire N~(0, 162,)

o e random error N~(O. 162.)

89

Sires. Data collection was performed on progeny of 152 sires. Sire identification was
conducted in such a manner that a single number denoted each sire regardless of analysis
method or model. However, not all sires had both male and female progeny records. For
univariate analyses, 135 sires were used to model RTS, while 127 sires were represented
in the bull population. As a matter of definition, “paired” sires are those 110 sires with
both male and female progeny records. The bivariate analyses were performed separately
on 110 paired sires, those sires with both male and female progeny records, as well as the
full data set including all 152 sires. Paired data sets included 1491 females and 1225

males as described in Table 4.4 and 4.5.

90

Table 4.4. Paired data description for Running Creek Ranch females-110 paired sires

 

 

Trait Mean N Percent S. D. Range
Age (days) 398 22.39 318455
Weight (WT, kg) 380 4.19 216-525
Percent Limousin (PL)
< 75% 108 7.2%
76-87% 49 3.3%
88-92% 101 6.8%
> 92% 1233 82.7%
Age of dam (AOD)
2 247 16.6%
3 235 15.8%
4-10 876 58.8%
>10 133 8.9%
Reproductive Tract Score (RTS)
Cycling (1) ‘ 1087 72.9%
Non-cycling (0) 404 27.1%
Muscle Score (MS) 2.2 1-3.5
Light (5 1.5) 117 7.9%
Average (= 2) 929 62.3%
Heavy (2 2.5) 445 29.8%
Body Condition Score (BCS) 6.2 5-8
5 69 4.6%
6 1040 69.8%
7 369 24.8%
8 13 .09%

 

91

Table 4.5. Paired data description for Running Creek Ranch males-110 paired sires

 

 

Trait Mean N Percent S.D. Range
Age (days) 383 298453
Weight (WT, kg) 429 368-513
Percent Limousin (PL)
5 75% 84 7.9%
76-87% 46 4.7%
87-92% 77 6.3%
2 93% 1018 83.1%
Age of dam (AOD)
2 163 13.3%
3 198 16.2%
4-10 770 63.8%
>10 94 8.7%
Scrotal Circumference (SC, cm) 32.06 2.64 24-42.25
2 34 ' 525%
5 30 525%
Muscle Score (M8) 2.3 1-3.5
Light (5 1.5) 101 8.2%
Average (= 2) 687 56.1%
Heavy (2 2.5) 437 36.7%
Body Condition Score (BCS) 5.8 4-7
4 9 .7%
5 333 27.2%
6 858 70%
7 25 2%

 

92

Bivariate Analyses.

Janss and Foulley (1993) described a general bivariate analyses (BIVARB) which
successfully accommodated unequal design matrices for a binary threshold and
continuous variable. In their evaluation, birth weight was included in an analysis of
calving ease scores due to its high genetic and residual correlation with calving difficulty
(Philipsson, 1976; Meijering, 1985). Many have reported scrotal circumference in males
and age at puberty in females are associated (Brinks et al., 1978; King et al., 1983). Thus,
it seems appropriate that to predict of sire breeding values for a threshold trait such as
RTS, scrotal circumference of male progeny should be included in a multiple trait
analysis. Observations for the continuous trait, SC, follow the linear model:

SC = X1131 + Zrlh + 91
where X and B are incidence matrices, relating scrotal circumference records to fixed
effects (B) and to sire effects (11). The error variance is given as 0'2“. Observations for
the threshold trait, RTS, are modeled as:

RTS = X23; + Z2112 + e;
where all terms are similar to those in the continuous trait model. The variance on the

underlying scale is arbitrary (=1). The variance structure for the sire effects is given as:

2 2 ll 12
v.11 _G_ G11 G12 _ 10011 10012 d '1: G G

- — - z 2 ...an ...G 21 22
“2 G21 G22 10' 021 10' 022 G G

As in the univariate analyses, sires were assumed unrelated in the bivariate model.
Several attempts were made to fit the relationship matrix to the BIVARB model, but

resulting (co)variance estimates were incalculable and convergence criterion was never

met.

93

Bivariate Model Specification. Three different models and six separate data sets were
used for these analyses. Differences between data sets are found in the number of sires
analyzed and/or the number of covariates included. The full data set includes progeny
records of 152 sires. The corresponding model fits the same seven fixed effects as the
single trait model, including both AGE and WT as covariates. The reduced models
included either AGE or WT in an attempt to remove collinearity from the model. As
previously mentioned, the data were also paired by sire to include only those 110 sires
with both male and female progeny records. A full description of each bivariate model
used is given in table 4.6.

Age and Weight adjusted bivariate analyses were also conducted on unpaired and
paired data sets using multiple-trait-restricted-maximum-likelihood (MTDFREML) for

estimation of genetic variances and covariances (Boldman, et al., 1993).

Table 4.6. Models and data sets used for bivariate analyses modeling both SC and RTS
as response variables.

 

 

 

152 Sires - Unpaired 110 Sires - Paired
Fixed Effect Full Age Weight Full Age Weight
YR X X X X X X
PL X X X X X X
AOD X X X X X X
MS X X X X X X
BCS X X X X X X
AGE X X X X
WT X X X X

 

94

Results
Univariate Analysis

Fixed factor effects on reproductive tract score. Type 3 (SAS®) F-tests of fixed
effects are given in Table 4.7. Parameter estimates for fixed effects are given in Table
4.8. The intercept cycling status value for heifers was .827. This value corresponds to a
heifer of >92% Limousin, born 1996 to a 210 year old dam, of average age and weight,
with body condition score 8 and muscle score 3. Heifer age was the most significant
source of variation (p<.001), and had the largest F value. WT was also important (p =
.012). Neither AOD (p = .425) or PL (p = .707) were significant explanatory variables as
fixed effects.

Efi’ects of composition on reproductive tract score. In general, heavier muscled and
fatter heifers were more likely to be cycling than their lighter muscled and/or leaner
counterparts. Cycling status was not different among average and heavy muscled heifers
according to t-tests (p = .95 1),. but both groups were advantageous to light muscled
heifers regarding cycling status (p < .05).

Although muscle score effects were significant, body condition score was a more
important factor affecting predicted probability of cycling. Increases in BCS had favorable
affects on reproductive tract maturity. Obese heifers (BCS 8) were the exception, as their
estimates for cycling were lower numerically than those heifers in good condition (BCS 7).
However, t-tests for significant differences between body condition scores revealed BCS 8
heifers were not different than BCS 6 or BCS 7 (p>.05). All other body condition score

classes were significantly different from each other (p<.05).

95

The combined effects of muscularity and body condition indicate the relative
importance of each trait on cycling status. The combination of fatter and lighter muscled
heifers is preferable to lean, heavy muscled heifers regarding reproductive tract maturity.

Some may be concerned about possible confounding of muscle score and body
condition scores. Logic dictates muscle and fat accretion accompany increases in age and
weight under conventional management practices. Subjectively assigned scores such as
these are certainly more prone to bias than are objectively measured traits. Step-wise
selection procedures were utilized in a prior analysis to test interactions between
composition and continuous traits. As previously mentioned, no interactions explained
significant variation above and beyond that described by the main effects model. Likewise,
model statistics indicated the main effects model sufficiently described the relationships of
age, weight, body condition and muscularity with reproductive fitness.

Fixed factor efiects on scrotal circumference. Type 3 (SAS®) F-tests of fixed effects
are given in Table 4.7. Parameter estimates for fixed effects are given in Table 4.8. The
adjusted mean scrotal circumference was 32.6 cm., which is slightly higher than
Limousin breed recommendations for yearling bulls (Directions, Limousin Breeder
Symposium, 1990). This intercept value corresponds to bulls of >92% Limousin,
measured in 1996 to a 210 year old dam, of average age and weight, with body condition
score 7 and muscle score 3. Both weight and age were significant sources of variation in
scrotal circumference (p<.05). As expected, scrotal circumference increased with
increasing age and weight. Regression coefficients for SC on WT and AGE were .057,
cm/kg and .016, cm/day respectively. Age of dam effects were important for sons of first

calf heifers and mature dams and consistent with estimates and adjustment factors

96

reported in literature. Sons of 3-year old, and cows 210 did not have significantly
different scrotal circumferences, but parameter estimates tended to follow literature
estimates.

Composition effects on scrotal circumference. As with estimates of heifer fertility,
fatter bulls had larger scrotal circumferences, excluding BCS 4 (n=9), which was not
significantly different than other BCS classes (p>.05). Solutions for SC on BCS increased
.65 cm for a BCS increase from 5 to 6 and 1.23 cm for an increase from 5 to 7 (p<.05).
Heavier muscled bulls had significantly smaller scrotal measurements than average muscled
bulls (p<.05). Light muscled bulls were not different than average or heavy muscled bulls
(p>.05). Contrary to the heifer analysis, however, SC solutions for light muscled bulls were
numerically higher than those of heavy muscled bulls. Solutions of SC for light and
average muscled bulls were .144 i .251 and .480 :1: .131, cm.

Summary of univariatefixed efi'ect results. These data suggest that selection of heavier
muscled replacement heifers, given adequate weight and body condition score, is not
detrimental to reproductive fitness. Moreover, yearling scrotal circumference can be
improved while muscling is maintained at average levels. Minimum body condition levels
are required to achieve desirable phenotypic expression of fertility traits. Of note is the
comparison between F-values for BCS and WT, which indicated a greater importance of

BCS than WT for heifer cycling status.

97

Table 4.7. Type 3 (SAS®) Tests of Fixed Effects

 

 

 

Cycling Status Scrotal Circumference
Effect Num Den Num Den

DF“ DF" P Value Pr > F Dr“ DP" F Value Pr > F

BCS 3 1424 1 1.69 <.001 3 1102 6.44 <.001
MS 2 1424 4.53 .011 2 1102 7.32 <.001

AOD 3 1424 .93 .425 3 1102 I 5.4 .001
PL 3 1424 .47 .707 3 1 102 1.03 .380
Year 6 1424 3.46 .002 6 1 102 23.66 <.001
Age 1 1424 26.52 <.001 1 1 102 13.55 <.001
Weight 1 1424 6.26 .012 1 1 102 24.99 <.001

 

a F-test numerator degrees of freedom.
b F-test denominator degrees of freedom.

98

Table 4.8. Univariate analysis parameter estimates

 

 

 

Cycling Status Scrotal Circumference
Effect Level Estimate Pr > ltl Estimate Pr > ltl
Intercept .827 4: .120 <.001 32.638 3 .493 <.001
BCS'“
4 -.034‘;t 1.015 .973
5 4301* i .126 .017 -1233! i .419 .003
6 -.052"3;.112 .641 -.58lfi.384 .131
7 .055c :1: .112 .623 0f
8 0'”
MS“
1 -425" :1: .046 .007 .144“i :1: .251 .567
2 -002" i .025 .951 .480i i .131 <.001
3 0° 0h
AOD
2 -.O38 :1: .046 .412 -.651j i .321 .043
3 -.046 :1: .044 .303 -.21 11* :1: .260 .418
4-9 -.061 :1: .038 .108 .2431 :1: .217 .263
10+ 0 0“
PL
<75 -.042 i .040 .291 -. 147 :1: .228 .519
75-86 .001 :1: .056 .990 .424 :1: .314 .177
87-92 -024 i .040 .551 -. 188 :1: .238 .429
>92 0 0
Year°
1990 -.062:1: .041 .133 -1.066:1: .312 <.001
1991 -.007 :1: .038 .864 -.709 i .302 .019
1992 -.1403: .048 .004 -1752: 1.177 .137
1993 .015 3; .040 .712 -.670 :1: .269 .013
1994 .074 :t .041 .070 1.315 t .248 <.001
1995 .040 i .034 .242 -.095 i .263 .719
1996 0 0
Age” .003 :1: .001 <.001 .016 :1: .004 <.001
Weight” .009 :1: .000 .012 .057 i- .004 <.001

 

a”Variables with different superscripts are different (P<-05)

m BCS; 8 = Obese, 7 = Good, 6 = High Moderate, 5 = Moderate, 4 = Thin
“ MS; 3 = heavy muscled, 2 = average muscled, 1 = light muscled.

° Year effects were not tested for differences
P Expressed as deviations from the mean.

99

Genetic ejfects on reproductive tract score. Reproductive tract score analysis
variance parameter estimates are given in Table 4.9. Sire effects were not a significant
source of variation in the univariate analysis. The heritability estimate of .04 is lower
than those of RTS (Andersen, etal., 1991, Andersen, 1991) and many heritability
estimates for heifer age at puberty (Smith et al., 1976, Werre and Brinks, 1986; MacNeil
et al., 1984). McInemey (1977) and Smith et al. (1989a) reported heritability estimates
for age at puberty of .07 i .1 and .1 :1: .09, respectively, which encompass the estimates
calculated here.

Table 4.9. Variance parameter estimates — reproductive tract score

 

 

 

Cyclinggtatus
Covariance Parameter Estimate Z Value Pr 2
Sire .002 i .002 .85 .198
Residual .165 :1: .006 27.17 <.001

h2 .039

Genetic effects on scrotal circumference. Scrotal circumference variance parameter
estimates are given in Table 4.1. Sire effects were a significant source of variation in the
univariate analysis. The heritability estimate of .37 is lower than the average of scrotal
circumference estimates found in the literature (5.5) Literature heritability estimates
range from .16 (Kriese et al., 1991) to .78 (Coulter and Foote, 1976).

Table 4.10. Variance parameter estimates — scrotal circumference
Scrotal Circumference

 

 

Covariance Parameter Estimate 2 Value Pr Z

Sire .386 :t .128 3.02 .001

Residual 3.840 3: .162 23.76 <.001
h2 - .366

 

100

Bivariate analysis fixed efi‘ect estimates.

Age and weight adjusted model. Age and weight adjusted fixed effect estimates and
associated standard errors for unpaired and paired data are given in Table 4.11.
Regression coefficients calculated in the bivariate analysis were in general agreement in
both order and scale to estimates derived from univariate analyses. Regression
coefficients for RTS or SC on either AGE or WT were positive. Increases in BCS had
favorable effects on SC and RTS, except for the leanest bulls (n=9) and the fattest heifers
(unpaired n=14, paired n=l3). MS effects were consistent with single trait analysis
estimates. AOD, PL, and YR effects mirrored results from univariate analysis.
Differences between analyses of paired and unpaired data, occurred on some fixed effect
factor levels that had relatively high standard errors (i.e. bull BCS 4).

Age adjusted model. Age adjusted fixed effect estimates and associated standard
errors for unpaired and paired data are given in Table 4.12. As expected, regression
coefficients for SC and RTS on age increased with WT removed from the model,
particularly for SC. YR effects appear considerably different from the full model.
Solutions of SC and RTS on BCS increased in absolute value, acknowledging a strong
correlation between WT and BCS. Of note is the difference in MS effects on SC when
WT is removed from the model. In the AGE and WT adjusted model MS 1 had a less
detrimental effect on SC than MS 3. In this model, however, the effect of MS 1 on SC is
negative, and the effect of MS 3 on SC is positive and comparable to regression
coefficients for average muscled bulls. These results indicate MS is strongly correlated
with WT, agreeing with earlier findings of a purely phenotypic analysis on the same data

that required fitting a WT x MS interaction.

101

Weight adjusted model. Weight adjusted fixed effect estimates and associated
standard errors for unpaired and paired data are given in Table 4.13. Regression
coefficients for SC and RTS on weight increased with age removed from the model.
Composition effects resembled those from the full model more closely than results from
the model in which WT was removed. AOD, PL, and YR effects were similar to the full

model as well.

102

Table 4.11. Age & weight adjusted bivariate analysis fixed effect estimates

 

 

 

 

Cycling Status Scrotal Circumference
Estimates Estimates
Effect Level Unpaireda Paired” Unpaireda Pairedb
Mean .499 :1: .343 .485 j: .344 31.825 i .380 31.945 5 .380
BCS
4 .408 i .621 -.238 :L- .709
5 -.757 :1: .101 -.756 :1: .106 -.760 i .144 -.652 i .151
6 -.006 :1: .082 .022 :1: .083 -.108 i .143 .003 :1: .150
7 .471 :1: .086 .507 i .087 .460 i .215 .887 i .243
8 .292 :1: .209 .227 i .214
MS
1 -.267i.120 -.311:1:.121 -.O62:1:.137 -.027:1:.137
2 .120i.114 .14Zi.115 .274:t.120 .248:1:.12O
3 .147i.116 .168:t.117 -.212:1:.124 -.221;1:.124
AOD
2 -.002 :1: .071 .018 :1: .072 -.487 :1: .102 -.476 :1: .101
3 -.044 :1: .070 -.048 :1: .070 -.052 :1: .084 -.040 :1: .084
4-9 -.O99 :1: .066 -. 100 i .067 .391 :1: .075 .380 i .074
10+ .1443: .074 .13021: .074 .148 :1: .095 .1365 .094
PL
<75 -.088 :1: .076 -.120 :1: .077 -.171 :1: .102 -.204 i .101
75-86 .074 :1:- .084 .094 :1: .088 .409 i .127 .424 i .124
87-92 -.032 i .076 -.005 :1: .078 -.217 :1: .105 -.223 i .103
>92 .046 :1: .068 .031 i .069 -.021 :1: .079 .002 i .078
YR
1990 -.160 i .030 -.176 :t .031 -.639 :1: .088 -.863 :1: .096
1991 -.016 i .029 -.O36 :1: .029 -.302 :1: .086 -.509 :1: .095
1992 -.361 :1: .035 -.229 i .040 -l.275 :1: 1.121 -.O66 :1: 1.507
1993 .043 i .031 .027 :1: .032 -.263 i .075 -.417 :1: .086
1994 .309 :1: .034 .276 :1: .032 1.736 :1: .069 1.526 i .078
1995 .166 :1: .029 .156 i .024 .326 :1: .074 .117 :1: .086
1996 .018 i .028 -.018 :1: .022 .418 :1: .088 .212 :1: .098
Age6 .0103: .000 .012i .000 .0163: .000 .01721: .000
WTc .003 i .000 .003 i .000 .056 :t .000 .058 :1: .000
a 152 sires
b 110 sires

° Expressed as deviations from the mean.

103

Table 4.12. Age adjusted bivariate analysis fixed effect estimates

 

 

 

 

 

Cycling Status Scrotal Circumference
Estimates Estimates
Effect Level Unpaireda Pairedb Unpaired“ Pairedb
Mean .478 :1: .343 .464 i .345 31.743 i .391 31.805 i .391
BCS
4 -l.278 i .719 -1.225 i .839
5 -.831:1:.100 -.835:l:.104 -.938i.161 -1.027:l:.168
6 -.025 :1: .082 .005 :1: .084 .395 i .158 .305 :1: .168
7 .501 :1: .086 .538 i .087 1.821 i .236" 1.947 t .275
8 .355 i .211 .292 :1: .216
MS
1 -.323:l:.119 —.371:1:.120 -.514i.141 -.521i.l4l
2 .123:t.114 .145i.115 .2692t.122 .260:t.122
3 200212.116 .226i.1l6 .245:l:.126 .261 :1:.126
AOD .
2 -.021 :1: .071 -.005 :1: .072 -.775 i .106 -.770 :1: .106
3 -.041 :1: .069 -.045 i .070 .052 :1: .088 .065 :1: .088
4-9 -.077 :1: .066 -.078 :1: .067 .571 i .077 .559 :1: .076
10+ .139 :1: .074 .127 i .074 .153 :1: .101 .145 :1: .100
PL
<75 -.074 i .076 -.106 :1: .077 -.249 i .110 -.291 :1: .109
75-86 .076 j: .084 .098 i .089 .519 :1: .140 .542 i .136
87-92 -.050 :1: .076 -.024 :1: .078 -.174 i .114 -.171 :t .111
>92 .048 i' .068 .032 :t .069 -.096 :t .082 -.080 :1: .081
YR
1990 -.137 j: .030 -.153 :1: .030 -1.674 :1: .094 -l.673 i .107
1991 -.036 :1: .029 -.055 :1: .029 -.933 :1: .096 -.901 :1: .108
1992 -.385 :1: .034 -.251 :1: .040 1.288 i 1.303 1.131 :1: 1.801
1993 -.010 :1: .031 -.030 :1: .031 -1.013 :1: .082 -.926 :1: .097
1994 .333 i .034 .300 :1: .034 1.060 :1: .075 1.071 :1: .088
1995 .196 :l: .029 .187 :t .029 .623 :1: .083 .628 :1: .098
1996 .038 i .028 .001 i .028 .650 :1: .099 .670 :1: .111
Age6 .012 :l: .000 .013 :1: .000 .039 :1: .000 .042 :1: .000
a 152 sires
" 110 sires

° Expressed as deviations from the mean

104

Table 4.13. Weight adjusted bivariate analysis fixed effect estimates

 

 

 

 

Cycling Status Scrotal Circumference Estimate
Estimate
Effect Level Unpaired‘I Pairedb Unpaired“ Pairedb
Mean .502 i .343 .498 i .344 31.815 i .381 31.932 1.381
BCS 4 .474 i .626 -.268 i .715
5 -.851 i .101 -.848 j: .105 -.782 i .145 -.632 i .151
6 —.020 :1: .082 -.003 :1: .083 -.115 i .143 .039 :1: .151
7 .488 :1: .085 .514 :1: .087 .423 i .216 .862 i .245
8 .383 i .208 .337 :1: .212
MS 1 -.227 :1: .120 -.256 :1: .120 -.078 i .137 -.039 1.137
2 .112i.ll4 .128i.115 .278i.120 2511.120
3 .115i.116 .127i.116 -.200i.124 -.212i.124
AOD 2 .087 :t .071 .125 :t .072 -.361 i .102 -.326 :1: .102
3 -.039 i .070 -.044 :1: .070 -.049 :t .085 -.044 :t .085
4-9 -.149 :1: .066 -.159 :1: .067 .325 i .075 .307 :t .074
10+ .101 i .074 .078 :1: .074 .084 i .095 .063 :1: .094
PL <75 -.062 i .076 -.093 i .077 -.158 :l: .102 -.187 :t .101
75-86 .026 :1:- .084 .029 i .088 .368 :1: .127 .380 :l: .125
87-92 -.022 :1: .076 .015 :1: .078 -. 199 :1: .105 -.205 :t .103
>92 .058 i .068 .049 i .069 -.010 :1: .079 .011 :1: .078
YR 1990 -.215 :1: .031 -.235 i .031 -.485 :1: .088 -.730 :1: .097
1991 .009 :1: .029 -.012 i .029 -.198 i .087 -.438 :1: .097
1992 -.487 :1: .034 -.370 i .039 -1.391 :I: 1.133 .045 i: 1.524
1993 .143 :1: .031 .141 :1: .031 -.237 i .076 -.434 :1: .087
1994 .357 :1: .034 .327 i .034 1.739 i .070 1.492 i .079
1995 .150 :1: .029 .136 :1: .029 .327 :1: .075 .079 :t .087
1996 .042 i .028 .012 :1: .028 .245 i .087 -.015 :1: .096
W'I‘c .005 :1: .000 .005 :1: .000 .061 i .000 .063 i .000
a 152 sires
b 110 sires

° Expressed as deviations from the mean.

105

 

Bivariate Analysis Random Effect Estimates

Genetic efifects. Sire (co)variance parameter estimates for all bivariate models are
given in Table 4.14. Sire variance estimates for SC and RTS were greatest in the WT
adjusted models analyzed using BIVARB. With AGE effects added, sire variance
decreased, and decreased further when WT was removed from the model. MTDFREML
estimates of sire variance were the same as previously calculated in the single trait
analyses. Genetic covariances calculated from all models including AGE as a covariate
were negative. When AGE is removed from the model, the genetic covariances

calculated from WT adjusted models are essentially zero.

Table 4.14. Sire (co)variances for scrotal circumference and cycling status in yearling
Limousin cattle

 

 

Method Model sc‘ RTS“ scars"
MTDFREML Age & Weight Adjusted .386 .002 -.005
MTDFREML Paired Age & Weight Adjusted , 364 .002 -.004
BIVARB Age & Weight Adjusted .340 .025 -.019
BIVARB Paired Age & Weight Adjusted .341 .023 -.017
BIVARB Age Adjusted .309 .023 -.016
BIVARB Paired Age Adjusted .313 .023 -.012
BIVARB Weight Adjusted .407 .035 .003
BIVARB Paired Weight Adjusted .407 .032 -.002

 

aGenetic variance
bGenetic covariance

Environmental eflects. Environmental variance estimates are given in Table 4.15.
Bivariate MTDFREML analyses yielded the same genetic and environmental variance
estimates as univariate analyses. Environmental variances for the continuous trait were

smaller in models including the weight covariate than models in which weight was

106

 

removed. Environmental covariance is set to zero as traits were not measured on the same

animal.

Table 4.15. Environmental variances for scrotal circumference and cycling status in
yearling Limousin cattle

 

 

Method Model SC“ RTSa
MTDFREML Age & Weight Adjusted 3,837 .165
MTDFREML Paired Age & Weight Adjusted 3,803 .163
BIVARB Age & Weight Adjusted 4,243 1
BIVARB Paired Age & Weight Adjusted 4,047 l
BIVARB Age Adjusted 5.148 1
BIVARB Paired Age Adjusted 4.919 1
BIVARB Weight Adjusted 4.254 1
BIVARB Paired Weight Adjusted 4,067 1

 

“Environmental variance

Heritability and genetic correlation estimates. Heritability and genetic correlation
estimates are given in 4.16. Heritability for the discrete trait (RTS) in the full models
increased from single trait, and MTDFREML analysis (.04) to the analysis performed using
BIVARB (unpaired = .10, paired = .09). Thus, BIVARB analyses produced higher genetic
variances and corresponding herititability estimates for the discrete trait than those
calculated in MTDFREML. This is due to the fact that BIVARB models the discrete trait
on the underlying scale, and thus accounts for liability for RTS more correctly. Heritability
decreased for the continuous trait (SC) from the AGE and WT adjusted results (.30) to the
AGE only model (unpaired = .23, paired = .25). Heritability increased for both traits
from the AGE and WT adjusted results to the model in which AGE was removed.
Heritability estimates for the continuous trait ranged from .23 in the unpaired AGE adjusted

model to .36 (unpaired and paired MTDFREML; paired, WT adjusted BIVARB). Of note

107

 

is the range of negative genetic correlations of —. 16 to -.20 for models including AGE vs.

the genetic correlations calculated when AGE is removed (unpaired = .02, paired = -.01).

Table 4.16. Heritabilities and genetic correlations for scrotal circumference and cycling
status in yearling Limousin cattle

 

 

 

Method Model sc‘ RTS“ scarsb
Univariate SC .37 -'- --
Univariate RTS -- .04 --
MTDFREML Age & Weight Adjusted .36 .04 -.19
MTDFREML Paired Age & Weight Adjusted .36 .04 -. l6
BIVARB Age & Weight Adjusted .30 .10 -.20
BIVARB Paired Age & Weight Adjusted .31 .09 -.19
BIVARB A86 Adjusmd .23 .09 -. 19
BIVARB Paired Age Adjusted .25 .09 -. 15
BIVARB Weight Adjuswd .35 . l 3 .02
BIVARB Paired Weight Adjusted .36 , 12 -.01
a Heritability

b Genetic correlation

Discussion

It is widely accepted among beef cattle breeders that age at puberty and yearling
scrotal circumference are influenced by a number of the same genes in the different sexes.
The genetic correlation estimates reported here between scrotal circumference and
reproductive tract score seem to contradict such assessments. However, although
threshold-continuous bivariate analyses produced larger heritability estimates of RTS the
environmental variance component remained large. Thus, it seems likely that herd
management is an essential element creating expressed fertility trait differences between

sons and daughters of a sire. Heifers at Running Creek Ranch must breed to calve as

108

two-year-olds to maximize production efficiency. Therefore, heifers are fed so that as
many as possible can achieve cycling status prior to the breeding season. Two-year-old
virgin bulls, however, are in sufficient demand that there is less management emphasis on
early growth and reproductive maturation rate in the male population. Additionally,
Bourdon and Brinks (1986) reported genotype x environment, and genotype x genotype x
environment interactions affected fertility traits, perhaps to a largerextent than
environmental effects alone; findings which may be particularly applicable to these data.
Limousin breed recommendations for age adjusting scrotal circumference to 365 days
require bulls be measured between 10 and 14 months, as scrotal circumferences increase
at linear rates during that time (Directions, Limousin Breeders Symposium, 1991).
Conversely, reproductive maturation rates in females of similar ages are most definitely
not linear, but sigmoidal, or logistic in nature. The bivariate estimation method used in
these analyses is designed to accommodate this discrepancy, but perhaps the time of
reproductive tract score measurement was inappropriate. Heifers in these analyses were
measured just prior to being mated, with the primary aim being to cull those females
unlikely to respond to estrus synchronization or become pregnant within a fixed breeding
season. The objective was not to determine when a heifer first cycled (age at puberty),
but instead identify those heifers cycling at the time of measurement. Although,
reproductive tract scores can serve as estimators of age at puberty, their usefulness in that
regard is time dependent. Andersen (1991) theorized that, if the objective is to reduce age
at puberty and select for superior fertility trait genetics, reproductive tract scores should
be taken when no more than half the population is thought to be cycling. As previously

mentioned, genetic improvement of fertility traits was not the objective at Running Creek

109

 

 

Ranch beyond the inherent economic benefits of removing from the breeding herd those
females unlikely to conceive within a fixed mating season. Were reproductive tract
scores taken earlier or possibly even grouped differently, the genetic relationship with
scrotal circumference may have been more similar to literature estimates of comparable
analyses. The reproductive tract scoring system may require modification for use in
Limousin cattle. All literature heritability estimates for RT S were calculated from
analysis on cattle of British origin rather than Continental breeds.

Summers et al. (1999) modeled Limousin heifer pregnancy data with the intent of
developing EPD for pregnancy. A corresponding study indicated relationships between
scrotal circumference and heifer pregnancy are similar to those reported here (Edwards,
1999, personal communication). Pregnancy status, like cycling status, is a binary trait,
and was estimated using a maximum a posteriori probit (MAP) threshold model (Gianola
and Foulley, 1983; Harville and Mee, 1984). The bivariate analysis described by J anss
and Foulley (1993), used in this analysis is an extension of the MAP threshold model as it
includes analysis of continuous traits. Therefore, the bivariate analysis method described
utilized here, BIVARB, should be examined further. Heritability estimates for the binary
trait were improved in these analyses utilizing BIVARB over univariate and
MTDFREML estimation methods. More importantly, fertility traits are considerably
more economically important (Melton, 1995) than other beef production measures and
producers need information regarding genetic and management opportunities. The
threshold nature of female age at puberty does create estimation difficulties, which is
among the reasons literature heritability estimates for most fertility traits measured in beef

females are typically low. Researchers need methods to analyze threshold traits in

110

conjunction with other, related traits, such as birthweight and calving ease scores, if
appropriate, and accurate information is to be supplied to producers.

Additionally, it is noteworthy that the bivariate programs utilized do not report
standard errors for genetic covariance parameters. Consequently, the authors are unsure
of the true strength of the genetic covariance/correlation estimate. Thus, although the true
correlation could be moderately negative for these data, it could be essentially zero as
well.

Theoretically, beef cattle producers can breed for improved expressed fertility by
selecting for traits that are positively correlated with fertility traits, such as growth and
body condition. Although genetic effects were not estimated for the composition traits
described in this study, it seems reasonable that joint selection for growth and fleshing
ability are likely to create favorable corresponding responses in yearling fertility trait
indicators.

While it is obvious that age is the primary covariate that should be included in any
such analysis, weight is also an important source of variation. Although age and weight
are highly related in most feasible and practical production schemes, the critical point is
that management prevails as perhaps the most essential component affecting expressed
fertility. While age allows for a more direct comparison across herds, breeds and regions,
weight is a trait which producers can manage for, and thus influence directly. Another
such trait is body condition score, which along with muscle score, appears to be, as
expected, highly correlated with weight. The importance of management is fortified
when the effects of weight and body condition are jointly considered. While breeders are

not advised by this author to select against growth, common sense indicates that retaining

111

 

higher body condition score heifers with lower weights will reduce mature weights and
maintenance costs.

Certain relationships between composition and indicators of reproductive fitness in
yearling cattle presented here challenge the perspective of many involved with beef cattle
breeding. It is commonly acknowledged reproductive fitness traits of Limousin cattle
generally require improvement, yet the amount of progress required to meet industry
demands is debatable. Heifers cycling at a year of age will undergo at least three heat
cycles before being bred to calve at 24 months. Certainly, favorable relationships have
been well established between reduced age at puberty and lifetime fertility and
productivity of a cow. However, the role of the Limousin breed in the beef industry
mandates genetic advancement of fertility traits be conducted in the context of
maintaining breed strengths. The data indicate Limousin breeders appear to have that

opportunity.

112

IMPLICATIONS

Commonly accepted philosophy dictates maternal ability, including fertility and milk
genetics, is of primary importance in replacement heifer selection and that muscularity
can be ignored in the female and addressed through sire selection. However, if greater
muscling is not detrimental to reproductive fitness in yearling Limousin heifers, given
adequate weight, body condition, and a comparable environment, then breeders should
retain heavier muscled females. Doing so allows Limousin breeders to accomplish two
important goals. First, breed muscle advantages will be maintained without reliance upon
utilization of the heaviest muscled bulls in the breed. At the same time, intense sire
selection pressure can be directed towards fertility trait improvement and other breed
weaknesses. Therefore, corrective mating schemes can be implemented in the manner
which will produce the most rapid genetic improvement, through sire selection. Breeders
are cautioned, however, as a result of these analyses that improvement in yearling female
fertility may not be maximized through use of sire scrotal circumference EPD in
environments comparable to the one maintained at this ranch. Additional research is
needed to delineate genotypic expression of reproductive fitness traits in a given
environment. Perhaps different sire lines or mating systems may be identified for specific

management practices.

113

SUMMARY and CONCLUSIONS

Composition effects are important sources of variation in measures of fertility
indicators in yearling Limousin cattle. Except for obese heifers, heavier conditioned
cattle appear to be more fertile. Muscularity effects cannot be explained as simply, as
they appear to be affected by sex and correlations with weight. Lighter muscled heifers
were not as fertile as heifers of average and even greater muscularity. Heavier muscled
bulls, on the other hand, had the smallest scrotal circumferences. In early phenotypic
analyses, phenotypic selection for larger scrotal circumference was equivalent to selection
against muscularity and vice-versa Fortunately, when genetic effects were considered,
average muscled bulls had the largest adjusted scrotal measurements, indicating that
optimums for scrotal circumference and muscularity exist.

Genetic analyses yielded unexpected results, as correlations between male and female
fertility traits were either not important or negative. This phenomenon contradicts most
literature estimates of genetic relationships between male scrotal circumference and
fertility traits in either daughters or half-sib sisters. However, recent reports of
relationships between pregnancy status and sire scrotal circumference EPD in Limousin
cattle support the estimates calculated in this study (Edwards, 1999).

The bivariate program described by Janss and Foulley (1993) accomplished an
important objective by producing larger heritability estimates for the discrete trait than
those obtained through single trait analyses or MTDFREML. Heritability estimates for

scrotal circumference and cycling status were low, but in the range of literature estimates.

114

A paradigm shift may be required as Limousin seedstock producers look to females in
their attempt to maintain muscularity, balance fertility trait and composition trait
emphasis among potential sires. Limousin breeders who adopt this philosophy have the
opportunity to retain heavy muscled replacement females with confidence in their
reproductive ability, provided heifers possess the genetic potential required and receive
management levels sufficient to achieve adequate weight and body condition prior to
breeding. The implications of this are substantial for Limousin seedstock producers
wishing to supply the beef industry with heavy muscled terminal sires and retain or

merchandise female siblings as replacements.

115

 

LITERATURE CITED

Abadia, D., J .S. Brinks, and EJ. Carroll. 1979. Genetics of seminal traits in young beef
bulls. Proc. West. Sect. An. Sci., 27:30

Andersen, KJ. 1987. Reproductive tract scores, condition scores and performance traits in
beef heifers. M.S. Thesis. Colorado State Univ., Fort Collins, CO.

Andersen, K.J., J.S. Brinks, D.G. LeFever, W.C. Asbury, D.W. Schafer and IL. Moon..
1987. Reproductive performance of Angus heifers. CSU Beef Program Report, p. 71.
Colorado State Univ., Fort Collins, CO.

Andersen, K.J., J.S. Brinks, D.G. LeFever and K.G. Odde. 1988. Genetic aspects of
reproductive tract scores, condition scores and performance traits in beef heifers. Proc.
West. Sect. Amer. Soc. Anim. Sci., 39:265.

Andersen, K.J., 1988. Unpublished data

Andersen, K.J., 1991. Reproductive, calving ease, and growth traits in beef heifers. Ph.D.
Dissertation. Colorado State Univ., Fort Collins, CO.

Andersen, K.J., D.G. LeFever, J.S. Brinks, and K.G. Odde. 1991. The use of reproductive
tract scoring in beef heifers. Agri-Pract., 12:4.

Arije, GE and IN. Wiltbank; 1971. Age and weight at puberty in Hereford heifers. J.
Anim. Sci., 33:401.

Bourdon, R.M. and J .S. Brinks. 1986. Scrotal circumference in yearling Hereford bulls:
Adjustment factors, heritabilities, and genetic environmental and phenotypic
relationships with growth traits. J. Anim. Sci., 62:958.

Brinks, J .S. 1991. Genetic aspects of male and female reproductive traits. Directions,
Limousin Breeders Symposium. p. 54. North American Limousin Foundation.
Englewood, CO.

Brinks, J.S., MJ. McInemey, and R]. Chenoweth, 1978. Relationship of age at puberty in
heifers to reproductive traits in young bulls. Proc. West. Sect. Amer. Soc. Anim. Sci.,
29:28.

Brown, L.N. 1986. Progestogen-prostaglandin combinations for estrus synchronization in
beef cattle. M.S. Thesis. Colorado State Univ., Fort Collins, CO.

116

Burfening, P.J., D.D. Kress, D.C. Anderson, and R.L. Blackwell. 1979. Heterosis among
closed lines of Hereford cattle. II. Postweaning growth and puberty in heifers. J.
Anim. Sci., 49:598.

Byerley, DJ ., R.B. Staigmiller, J.G. Berardinelli and RE. Short. 1987. Pregnancy rates of
beef heifers bred either on puberal or third estrus. J. Anim. Sci., 65:645.

Carter, A. P., P.D. P. Woodward, and RA. Wright. 1980. Association between
reproduction, live weight and sperm output in cattle. J. Repr. Fert., 59:447.

Coulter, OH. and RH. Foote. 1979. Bovine testicular measurements as indicators of
reproductive performance and their relationship to productive traits in cattle: A review.
Theriogenology, 11:297.

Coulter, G.H., T.R. Rousaville, and RH. Foots. 1976. Heritability of testicular size and
consistency in Holstein bulls. J. Anim. Sci., 43:9.

Cundiff, L.V., K.E. Gregory, and R.M. Koch. 1974. Effects of heterosis on reproduction in
Hereford, Angus, and Shorthom cattle. J. Anim. Sci., 38:711.

Cundiff, L.V., K.E. Gregory, R.M. Koch, and GB. Dickerson. 1986. Genetic diversity
among cattle breeds and use to increase beef production efficiency in a temperate
environment. Proc. 3rd World Congr. on Genet. Appl. to Livest. Prod. IX, Lincoln, NE.
p. 282-282.

Cundiff, L.V., F. Szabo, K.E. Gregory, R.M. Koch, M.E. Dikeman, and JD. Crouse.
1993. Breed comparisons in the gerrnplasm evaluation program at MARC. Proc.
Beef Improvement Fed. Ann. Mtg., Asheville, NC.

Breed Recommendations. 1991. Directions; Limousin Breeders Symposium. North
American Limousin Foundation. Englewood, CO.

Doombos, D.E., C.A. Steffan, DD. Kress, and DC. Anderson. 1983. Beef heifers
differing in milk production. 111. Relationship of age at puberty, percent pregnant, and
day pregnant. Proc. West. Sect. Amer. Soc. Anim., Sci., 34:8.

Dow, J .8. Jr., J .D. Moore, C.M. Bailey, and W.D. Foote. 1982. Onset of puberty in heifers
of diverse beef breeds and crosses. J. Anim. Sci., 55:1041.

Edwards, J .E. 1999. Personal communication.
Eisenhart, C. 1947. The assumptions underlying the analysis of variance. Biometrics, 3:1.

Ferrell, CL. 1982. Effects of postweaning rate of gain on onset of puberty and productive
performance of heifers of different breeds. J. Anim. Sci., 55: 1272.

117

 

Fields, M.J., W.C. Burns, and AC. Warnick. 1979. Age, season and breed effects on
testicular volume and semen traits in young beef bulls. J. Anim. Sci., 48: 1299.

Gianola, D. 1982. Theory and analysis of threshold characters. J. Anim. Sci., 54: 1079.

Gianola, D., and IL. Foulley. 1983. Sire evaluation for ordered categorical data with a
threshold model. Annales de genetique et de selection evolution, 15:201.

Gregory, K.E., L.V. Cundiff, and R.M. Koch. 1992. Breed effects and heterosis in
advanced generations of composite populations for reproduction and maternal traits of
beef cattle. J. Anim. Sci., 70:342.

Gregory, K.E., D.B. Laster, L.V. Cundiff, R.M. Koch, and GM. Smith. 1978. Heterosis
and breed maternal and transmitted effects in beef cattle. IL Growth rate and puberty in
females. J. Anim. Sci., 47: 1042.

Gregory, K.E., D.B. Laster, L.V. Cundiff, G.M. Smith, and R.M. Koch. 1979a.
Characterization of biological types of cattle-Cycle HI: IL Growth rate and puberty in
females. J. Anim. Sci., 49:461.

Gregory, K.E., G.M. Smith, L.V. Cundiff, R.M. Koch, and DB. Laster. 1979b.
Characterization of biological types of cattle-Cycle III: 1. Birth and weaning traits. J.
Anim. Sci., 48:271.

Gregory, K.E., D.D. Lunstra, . L.V. Cundiff, and R.M. Koch. 1991. Breed effects and
heterosis in advanced generations of composite populations for puberty and scrotal traits
of beef cattle. J. Anim. Sci., 69:2795.

Gregory, K.E., L.V. Cundiff, and R.M. Koch. 1992. Breed effects and heterosis in
advanced generations of composite populations for reproduction and maternal traits of
beef cattle. J. Anim. Sci., 70:656.

Gregory, K.E., L.V. Cundiff, and R.M. Koch. 1995. Genetic and phenotypic
(co)variances for production traits of intact male populations of purebred and
composite beef cattle. J. Anim. Sci., 73:2227.

Gregory, K.E., L.V. Cundiff, and R.M. Koch. 1995. Genetic and phenotypic
(co)variances for production traits of female populations of purebred and composite
beef cattle. J. Anim. Sci., 73:2235.

Harville, D.A., and R.W. Mee. 1984. A mixed-model procedure for analyzing ordered
categorical data. Biometrics, 40:393.

118

Hohenboken, W.D. and J .S. Brinks. 1971. Relationships between direct and maternal
effects on growth in Herefords. II. Partitioning of covariance between relatives. J.
Anim. Sci., 32:26.

Henderson, C.R. 1977b. Best linear unbiased prediction of breeding values not in the
model for records. J. Dairy Sci., 60:783

Henderson, C.R. 1984a. Estimation of variances and covariances under multiple trait
models. J. Dairy Sci., 67: 1581

Henderson, C.R. 1988a. Progress in statistical methods applied to quantitative genetics
since 1976. In 3.8. Weir, E. J. Eisen, M. M. Goodman, and G. Namkoong (eds.).
Proceedings of the second international conference on quantitativee genetics. p. 85.
Sinauer Assoc. Sunderland, MA.

Hoeschele, I. 1995. Multiple-trait genetic evaluation for one polychotomous trait and
several continuous traits with missing data and unequal models. J. Anim. Sci.,
63: 1609.

Hough, J ., 1991. Selection for scrotal circumference in Hereford cattle. Presented at 1991
Beef Improve. Fed. Annul. Mtg, Reprod. & Growth Comm, May 17, San Antonio, TX.

Janss, L.L.G., and J .L.Foulley. 1993. Bivariate analysis for one continuous and one
threshold dichotomous trait with unequal design matrices and an application to birth
weight and calving difficulty. Livestock Production Science, 33: 183.

Keeton, L.L., RD. Green, S.P. Doyle, and K.J. Andersen. 1996. Estimation of growth
curves and age-of-dam adjustments for scrotal circumference in Limousin bulls. Prof.
Anim. Sci., 12:44

Keeton, L.L., R.D. Green, B.L. Golden, and K]. Andersen. 1996. Estimation of variance
components and prediction of breeding values for scrotal circumference and weaning

weight in Limousin cattle. J. Anim. Sci., 74:31.

Kemp, R.A., L.R. Schaeffer, and J .W. Wilton. 1984. Comparison of beef sire evaluation
models for an organized progeny test. J. Anim. Sci., 58: 1313.

Kennedy, B.W. 1991. C. R. Henderson: the unfinished legacy. J. Dairy Sci., 74:4067.

Kinder, J .E., M.L. Day and R.J. Kittock. 1987. Endocrine regulation of puberty in cows
and ewes. J. Repro. Fert., Suppl. 34:167.

King, R.G., D.D. Kress, D.C. Anderson, D.E. Doombos, and P.J. Burfening. 1983.

Genetic parameters in Herefords for puberty in heifers and scrotal circumference in
bulls. Proc. West. Sect. Anim. Sci., 34:11.

119

Knights, S.A., R.L. Baker, D. Gianola, and J .B. Gibb. 1984. Estimates of heritabilities and
of genetic and phenotypic correlations among growth and reproductive traits in yearling
Angus bulls. J Anim. Sci., 58:887.

Koch, R.M., L.V. Cundiff, and K.E. Gregory. 1982. Heritabilities and genetic,
environmental, and phenotypic correlations of carcass traits in a population of diverse
biological types and their implications in selection programs. J. Anim. Sci., 55: 1319.

Kriese, L.A., J .K. Bertrand and LL. Benyshek. 1991. Age adjustment factors, heritabilities
and genetic correlations for scrotal circumference and related growth traits in Hereford
and Brangus bulls. J. Anim. Sci., 69-478.

Laster, D.B., GM. Smith, L.V. Cundiff, and K.E. Gregory. 1979. Characterization of

biological types of cattle (Cycle II). Postweaning growth and puberty of heifers. J.
Anim. Sci., 48:500.

Laster, D.B., G.M. Smith, and K.E. Gregory, 1976. Characterization of biological types of
cattle. IV. Postweaning growth and puberty of heifers. J. Anim. Sci., 43:63.

Latimer, F.G., L.L. Wilson, M.F. Cain, and W.R. Stricklin. 1982. Scrotal measurements in
beef bulls: Heritability estimates, breed, and test station effects. J. Anim. Sci., 54:473.

Lefever, D.G. and K.G. Odde. 1986. Predicting reproductive performance in beef heifers
by reproductive tract evaluation before breeding. CSU Beef Program Report. Colorado
State Univ., Fort Collins, CO.

Lesmeister, J.L., P.J. Burfening, and R.L. Blackwell. 1973. Date of first calving in beef
cows and subsequent calf production. J. Anim. Sci., 36:1.

Littell, R.C., G.A. Milliken, w.w. Stroup, and RD. Wolfmger. 1996. SAs" System for
Mixed Models. SAS Institute Inc., Cary, NC.

Lunstra, D.D., J .J . Ford, and SE. Echtemkarnp. 1978. Puberty in beef bulls: hormone
concentrations, growth, testicular development, sperm production and sexual
aggressiveness in bulls of different breeds. J. Anim. Sci., 46: 1054.

Lunstra, DD. 1982. Testicular development and onset of puberty in beef bulls. Beef Res.
Prog. Prog. Rept., No. 1, US. Meat Anim. Res. Ctr., ARM-NC21, Clay Center, NE. p.
26.

Lunstra, D.D., K.E. Gregory, and L.V. Cundiff. 1988. Heritability estimates and

adjustment factors for the effects of bull age and age of dam on yearling testicular size
in breeds of bulls. Theriogenology 30:127.

120

Lynch, M., and B. Walsh. 1998. Genetics and analysis of quantitative traits. Sinauer
Associates, Inc. Sunderland, MA.

MacNeil, M.D., L.V. Cundiff, C.A. Dinkel, and R.M. Koch. 1984. Genetic correlations
among sex-limited traits in beef cattle. J. Anim. Sci., 58:1171.

Madrid, N., R.S. Ott, D.N.R. Veeramachaneni, D.F. Parrett, W. Vanderwert, and CL.
Willms. 1988. Scrotal circumference, seminal characteristics, and testicular lesions of
yearling Angus bulls. J. Vet. Res., 49:579.

Martin, L.C., J .S. Brinks, R.M., Bourdon, and L.V. Cundiff. 1992. Genetic effects on beef
heifer puberty and subsequent reproduction. J. Anim. Sci., 70:4006.

Makarechian, M.,A. Farid, and R.T. Berg. 1985. Scrotal circumference, semen
characteristics, growth parameters and their relationships in young beef bulls. Can. J.
Anim. Sci., 65:789.

McInemey, M.J.. 1977. Genetics of puberty in beef heifers. M.S. Thesis. Colorado State
Univ., Fort Collins.

Meijering, A. 1985. Sire evaluation for calving traits by best linear unbiased prediction and
nonlinear methodology. J. Anim. Breeding and Genetics, 102:95.

Melton, BE. 1995. Conception to consumption: The economics of genetic improvement.
In: Proc. Beef Improvement Federation 27th Research Symposium and Annual Meeting.
Sheridan, WY. p. 40-71.

Moser, D.W., J .K. Bertrand, L.L. Benyshek, M.A. McCann and TE. Kiser. 1996. Effects
of selection for scrotal circumference in Limousin bulls on reproductive and growth
traits of progeny. J. Anim. Sci., 74:2052.

Mrode, R. A. 1996. Linear models for the Motion of animal breeding values.
International, Wallingford, UK.

Neely, J .D., B.H. Johnson, E.U. Dillard, and O.W. Robinson. 1982. Genetic parameters for
testes size and sperm number in Hereford Bulls. J. Anim. Sci., 55: 1033.

Nelsen, T.C., C.R. Long, and TC. Cartwright. 1982. Post inflection growth in straightbred
and crossbred cattle 11. Relationships among weight, height, and pubertal characters. J.
Anim. Sci., 55:293.

Nelsen, T.C., R.E. Short, J .J . Urick, and W.L. Reynolds. 1986. Heritabilities and genetic

correlations of growth and reproductive measurements in Hereford bulls. J. Anim. Sci.,
63:409.

121

Nfifiez-Dominquez, R., L.V. Cundiff, G.E. Dickerson, K.E. Gregory, and R.M. Koch.
1991. Lifetime production of beef heifers calving first at two vs three years age. J.
Anim. Sci., 6922367.

Odde, K.G., D.G. LeFever and G.W. Boyd. 1989. Heifer and bull management for
optimum production. Proceedings from bull and replacement heifer selection
workshops. Colorado State Univ., Fort Collins, CO.

Palasz, A.T., W.F. Cates, A.D. Barth, R.J. Mapletoft. 1994. The relationships between
scrotal circumference and quantitative testicular traits in. yearling beef bulls.
Theriogenology, 42:715.

Philipsson, J. 1976b. Studies on calving difficulty stillbirths and associated factors in
Swedish cattle breeds. III. Genetic parameters. Acta Agric. Scand., 26:212.

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

Ritchie, H.D., D.A. Hawkins, B.D. Banks, and M. Kaercher. 1992. Evaluating body
condition and using it to improve beef cow reproductive performance. Michigan Beef
Production Extension Bulletin. Michigan State Univ., E. Lansing, MI

SAS Institute Inc., SAS® Technical Report P-229, SAS/STAT"lo Software: Changes and
Enhancements, Release 6.07, Cary, NC: SAS Institute Inc.. 1992. 620 pp.

Schaeffer, LR. 1991. C. R. Henderson: contribution to predicting genetic merit. J. Dairy
Sci., 74:4052

Searle, S.R., G. Casella, and CE. McCulloch. 1992. Variance components. John Wiley
and Sons, NY.

Selvin, S. 1996. Statistical Analysis of Epidemiolog'c Data. Oxford University Press Inc.,
New York, NY.

Simianer, H., and LR. Schaeffer. 1989. Estimation of covariance components between
one continuous and one binary trait. Genet. Sel. Evol., 21 :303

Smith, B.A., J .S. Brinks, and G.V. Richardson. 1989a. Estimation of genetic parameters
among reproductive and growth traits in yearling heifers. J. Anim. Sci., 67:2886.

Smith, B.A., J .S. Brinks, and G.V. Richardson. 1989b. Estimation of genetic parameters

among breeding soundness examination components and growth traits in yearling bulls.
J. Anim. Sci., 67:2892.

122

Smith, B.A., J.S. Brinks, and G.V. Richardson. 1989c. Relationships of sire scrotal
circumference to offspring reproduction and growth. J. Anim. Sci., 67 :2881.

Smith, B.A., and J.S. Brinks. 1989. Estimation of genetic parameters among breeding
soundness examination components and growth traits in yearling bulls. J. Anim. Sci.,
67:2892.

Smith, G.M., H.A. Fitzhugh, Jr., L.V. Cundiff, T.C. Cartwright, and K.E. Gregory. 1976.
A genetic analysis of maturing patterns in straightbred and crossbred Hereford, Angus,
and Shorthom cattle. J. Anim. Sci., 43:389. .

Summers, M.D., J .L. Evans, L.R. Hyde, C.R. Comstock, K.J. Andersen, and BL. Golden.
1999. Estimation of heritability for heifer pregnancy in Limousin cattle. Proc. West.
Sect. Am. Soc. Anim. Sci., In press.

Thompson, J .A., M.M. Buhr, and W.H. Johnson. 1992. Scrotal circumference does not
accurately predict degree of germinal epithelial loss or semen quality in yearling
Hereford and Simmental bulls. Theriogenology, 38: 1023.

Toelle, V.D. and O.W. Robison. 1985. Estimation of genetic correlations between
testicular measurements and female reproductive traits in beef cattle. J. Anim. Sci.,
60:89.

Werre, JP. 1980. Relationship of age at puberty in heifers to subsequent fertility and
productivity. M.S. Thesis. Colorado State Univ., Fort Collins, CO

Werre, IF. and J.S. Brinks. 1986. Relationships of age at puberty with growth and
subsequent productivity in beef heifers. Proc. West. Sect. Amer. Soc. Anim. Sci.,
37:300.

Wiltbank, J .N., K.E. Gregory, L.A. Swiger, J .E. Ingalls, J .A. Rothlisberger, and R.M. Koch.
1966. Effects of heterosis on age and weight at puberty in beef heifers. J. Anim. Sci.,
25:744.

Wiltbank, J.N., C.W. Kasson, and J .E. Ingalls. 1969. Puberty in crossbred and straightbred
beef heifers on two levels of feed. J. Anim. Sci., 29:602.

Wright, S. 1934. An analysis of variability in number of digits in an inbred strain of guinea
pigs. Genetics, 19:506.

Young, L.D., D.B. Laster, L.V. Cundiff, G.M. Smith, and K.E. Gregory. 1978.

Characterization of biological types of cattle. IX. Postweaning growth and puberty of
three-breed cross heifers. J. Anim. Sci., 47:843.

123

"‘lllllllllllllllll