* CURRENT STATUS OF MANAGEMENT FACTORS WHICH
INFLUENCE GENETIC PROGRESS IN THE MICHIGAN
HOLSTEIN POPULATION

ThesIs for the Degree of M. 3‘
MICHIGAN STATE UNIVERSITY
G. LEHMAII METZUER
19??

 

 

 

 

 

 

 

 

LIBRARY

Michigan State
University

 

 

3.35.1,

..

33a. #4.. ":1 54.1.43.
,. 3 .. A I 1., .u~

a . LEV... v

ABSTRACT

CURRENT STATUS OF MANAGEMENT FACTORS WHICH
INFLUENCE GENETIC PROGRESS IN THE
MICHIGAN HOLSTEIN POPULATION

BY
C. Lehman Metzler

Progress by selection for a trait is determined by
the genetic variation in the specie under study, the
accuracy and intensity of selection with which the parents
of offspring are chosen. The rate of progress is determined
by generation interval.

A number of nongenetic factors tend to have some
influence on the accuracy and intensity of selection and
generation interval. The purpose of this study is to
examine the following identifiable factors: population
size, superiority of the bulls in use, calf mortality,
cow removals, calving interval, generation interval, sire
usage levels, useful life span of tested bulls in AI, and
progeny testing levels.

Data were obtained by a questionnaire survey of four
bull studs. Also, data were obtained from ineprogress DHIA

records, pedigrees of young sires, USDA sire summaries,

C. Lehman Metzler

Holstein herd books and current research by Spike (1972),
Erickson, g£_al. (1972), Erickson (1972), Speicher (1972),
and Oxender (1972).

It was found that the Michigan Holstein cow popula-
tion was 425,700 with 18 percent on official test. Cows
identified by sire represented 56.8 percent of those on
test. Sixty percent of Michigan Holsteins are bred arti-
ficially, 62 percent of which were sired by bulls proven
to be above breed average. Based on USDA summaries of
1158 bulls, AI sires of cows in the current herd had an
average predicted difference (PD) superiority of 144 and
seven pounds for milk and fat, respectively. Sires of
calves in high producing herds (17,694 pound average of
milk) had an average PD for milk of plus 808 pounds. In
herds with near breed average levels (13,397) of milk
production, sires of calves had an average PD for milk
of plus 694 pounds.

Thirty percent of all dairy cows left the herd
annually and were replaced by 30 heifers per 100 cows
successfully bred.

The average calving interval for Michigan Holsteins
over a 13 year period was 395 days.

Generation intervals for sires of sires, dams of sires,
sires of herd replacements, dams of herd replacements were

eleven, 7.5, nine, and 4.3, respectively.

C. Lehman Metzler

Two percent of the artificially inseminated Holstein
cows were used in progeny test programs. Proven bulls in
the artificial breeding program bred 7,614 cows each and

remained in the stud four years during their useful life.

CURRENT STATUS OF MANAGEMENT FACTORS WHICH
INFLUENCE GENETIC PROGRESS IN THE
MICHIGAN HOLSTEIN POPULATION

BY

C. Lehman Metzler

A THESIS

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

MASTER OF SCIENCE

Department of Dairy

1972

ACKNOWLEDGMENTS

I am especially indebted to Dr. C. E. Meadows for
his good natured guidance and encouragement through all
stages of the preparation of this thesis.

To Dr. Roger Neitzel, for his patient assistance
with the various aspects of the thesis which required com“
puter programming, to Mr. Alvin J. Thelen and his staff of
Michigan DHIA, Inc., in providing the desired records, and
to Lon D. McGilliard, for his suggestions and comments, I
am also grateful.

To my wife Alta, for her unwaning support during

this period of study, my most loving recognition and thanks.

ii

TABLE OF CONTENTS

INTRODUCTION .
REVIEW OF LITERATURE

Theoretical Basis for Progeny Testing .
Effective Population Defined .

Genetic Trends . . . . . . .
Fractions of AI Sires which are

Plus and Minus . . . . . . . . .
Available Herd Replacements . .
Cow Removals . . . . . . . . . . . . .
Calving Interval . . . . . .

Generation Intervals . .
Useful Life Span of Tested Bulls in AI

SOURCES OF DATA
METHODS AND RESULTS

Available Herd Replacements

Percentage of Cow Removals

Calving Interval .

Generation Intervals . . . .
Current Michigan Holstein Population . .

Effective Population for Progeny Testing .

Stratification of Michigan Holstein
Population by AI and Nat . . .

Fractions of AI and Natural which are
Plus and Minus . . . . . . . . . . .

Superiority of Bulls in Use . .

Useful Life Span of Tested Bulls in AI

Sire Utilization . . .

Minimum Number of Tested Bulls Needed

Present Progeny Testing Situation

in Michigan . . . . .
Progeny Testing Possible . . . . . . .
Suggested Improvement . . . . . . . . .
SUMMARY . . . . . . . . . . . . . . . . .
LITERATURE CITED . . . . . . . . . . . . . .
.APPENDIX . . . . . . .

iii

Table

10.
11.
12.

13.
14.
15.
16.

17.

18.

LIST OF TABLES

Annual genetic gains possible through AI

Observed annual genetic gains in populations
and herds of dairy cattle . . . . . . . .

Proven sires available through AI January 1,
1968 O O O O O I O O O O Q 0 O O O O Q

Percent cows culled for low production . . .
Cow removals during the first four lactations

Percent of heat periods recorded by days
after calving . . . . . . . . . . . . . . .

Generation intervals used by Thompson

and Freeman . . . . . . . . . . . . . . . . .
Tenure of Holstein dairy bulls in AI . . . . .
Percent of sire removals by reason . . . . .

Dairy cow removals reported by Michigan DHIA .
Average ages of sires of herd replacements . .
Average ages of dams and sires of young

bulls entering three studs between 1969

and 1971 . . . . . . . . . . . . . .

Michigan milk cows by year . . . . . . . . .
Percent DHI cows identified by sire . . . . .

Number of bulls by 200 pound PD intervals . .

Percent plus proven natural and AI sires and
daughters of the same . . . . . . . .

Average PD's of milk and fat for available
sires . . . . . . . . . . . . . . . .

Removal rates of bulls in Al . . . . . . . .

iv

Page

10

11
13
16

17

20
22
23
31
35

36
37
38
41

42

45
46

Table Page

19. Holstein bulls and inseminations by type of
service . . . . . . . . . . . . . . . . . . . . 47

20. Progeny testing parameters for Michigan . . . . 51

21. Young sires to be tested varying several
parameters . . . . . . . . . . . . . . . . . . 54

LIST OF FIGURES
Figure Page

1. Yearly trends in the percentages of U.S.
dairy cows bred by AI since 1945 . . . . . . . . 40

vi

INTRODUCTION

One goal of the dairyman is to improve the genetic
level of his herd. Several important management decisions
influence how rapidly and by how much the genetic level of
a dairy herd can be raised. Increasing the average produc-
tion depends on replacing the present herd with animals
that are genetically superior to their dams.

How much improvement can be made depends on three
things:

1. genetic variation

2. accuracy of the criterion for selection

3. intensity of selection.

Generation intervals determine how rapidly the
improvement is made. Researchers have calculated the
maximum progress possible within the biological limita-
tions of the animal to be approximately three percent of
the average yield [Robertson and Rendel (1950)]. It is
not likely that maximum progress will ever be achieved
because of the diffuse nature of the decision-making
group and the economic objectives held at certain levels
of the decision process.

The importance of certain factors which influence
progress is not always clear to management. Additive

genetic variation is not directly influenced by day to

day manageme
of the crite
are reflecti
The
the current
in Michigan
1.
2.

10.
11.

Have these nongenetic parameters influencing genetic progress

changed for
thesis work?
What are the
the changes?
or incorrect
dairy farms.

be discussed

nt practices. Selection intensity, accuracy
rion for selection and generation intervals
ve of several routine management decisions.
purpose of the present study is to evaluate
status of the following nongenetic parameters
which influence the rate of genetic progress;
Population size

Cows on official test

Cows on official test identified by sire
Fractions of bulls available which are pre-
dicted plus and minus

Superiority of bulls in use

Available herd replacements

Percentage of cow removals

Calving interval

Generation interval

Sire utilization

Progeny testing

the Michigan population since Specht‘s 1957
If there have been changes, what has changed?
management factors which have brought about
What needs to be done to correct improper
management practices of the AI units and the
To the extent possible, these questions will

in the body of the thesis.

w

 

Findings should be useful to university extension
personnel in recognizing particular areas in need of
emphasis, and for assisting dairymen and AI units in

obtaining maximum progress by selection.

REVIEW OF LITERATURE

Theoretical Basis for Progeny Testing

 

It has already been mentioned that there are four
basic factors determining the rate at which genetic pro-
gress can be made. Prior to entering into a discussion of
the management factors related to them, it is appropriate
to analyze briefly the relationship between these four
factors in the formula for the calculation of rate of
genetic improvement as outlined by Dickerson and Hazel
(1944). The general formula is AG = é§- where: AG =
average annual genetic gain

AP = average genetic superiority of parents of

offspring

T = time in years (average age of parents).

Average genetic superiority of those animals
selected to be parents has four sources in a large, artis
ficially inseminated population. Robertson and Rendel
(1950) referred to the four sources as:

l. Dams of replacements

2. Dams of bulls

3. Sires of bulls

4. Sires of replacements

Genetic superiority must be calculated for each of
these sources and is arrived at in the following manner:

AS or AD = ("i")rGI CG where:

(I) = selection differential in standard deviation
units
(r61) = accuracy (correlation between transmitting

ability and apparent merit)
0G = variation (the standard deviation of trans-
mitting ability).
Skjervold and Langholz (1963) suggest one additional I
factor for the formulav-a coefficient to correct for the
rate of inbreeding.

Robertson and Rendel (1950), in their discussion of i

 

the use of progeny testing with artificial insemination,
assumed the standard deviation of genotypic value for 305—
day yield to be ten percent of the average yield (Y).

Specht (1957) reported that the genetic standard deviation
(0.10Y) was almost exactly the same in estimates from Michi-
gan data as that assumed by Robertson and Rendel. Genetic
improvement is dependent on genetic variation.

Additive genetic variation is that portion of the
phenotypic variation which can be transmitted from parents
to offspring. The correlation between a sire‘s transmitting
ability and his apparent merit (rGI) is the square root of
the index used to measure heritability. Once the selection
differential has been found from tables [Pearson (1931) or
Nanson (1967)], it is reduced to reflect the inaccuracy by
multiplying by rGI’ the coefficient of accuracy.

Heritability of milk production has been estimated
to be 0.25 (Robertson and Rendel, 1950; Specht, 1957;
Deaton, 1964).

Because of culling which takes place in second and
later lactations, Rendel and Robertson (1950) demonstrate
that heritability becomes dependent on the number of
records; and based on that, use the following formula to

determine genetic superiority of cows:

ICC = Znhi (Y1 - T)
where: hi = heritability based on 1 lactations
Y = mean of the first lactations of the whole group
Y1 = mean of the lactations

number of daughters of each cow.

n

Since one cannot observe the important economic
traits in the dairy sire, information from his progeny is
relied upon as an indicator of his genetic worth.

Dickerson and Hazel (1944) reported that for a popu-
lation of 120 cows, progeny testing yielded slightly less
progress than with mass selection. Robertson and Rendel
(1950), with some variation of treatment, found "a very
slight advantage with progeny testing in the case of 305-
day yield in a herd of 120 cows." Specht (1957) found
progeny testing equal to mass selection in a fifty cow
herd sampling two, three, and four young sires on fifty,
forty, and sixty percent of the cows, respectively.

As pointed out by Specht (1957) and others, the two
basic limitations to progeny testing in a small herd are,
first, the number of bulls that can be tested is small,

which limits selection, and second, only a small number

1‘.- .. O
.9
g r

\ “flu;- '
t;- h" -

of cows is left to breed to tested bulls when so many are

bred to young bulls. It is also obvious that the high environ-
mental correlation between records of daughters and herd-

mates as well as the limited number of herdmates and the

small number of herdmates calving in the same season limit

the accuracy of the proof of the young sire. According to
Specht (1957) 6.66 percent of the cows in a Michigan study
were on official test. Thompson and Freeman (1971) used

0.15 as the faction of the population tested.

 

Effective Population Defined

 

Two major factors influence the population which is
"effective” for progeny testing. As previously mentioned,
the sire is evaluated on the basis of the performance of
his female offspring. In order to do that his offspring
must be correctly evaluated. Production testing represented
by DHI programs in the United States is the method used in
the evaluation of the production of the offspring. For
this reason cows not on test are not useful for sire
evaluation. The other limiting factor is records of cows
on test which lack sire identification.

A related but somewhat different problem is that of
misidentification. Van Vleck (1970), studied the effects
of misidentification in estimating the paternal sib correlas
tion. He concluded that misidentification of the sire can
lead to substantial underestimation of heritability from

the intransire correlation.

Further research reported by Van Vleck (1970) led
to conclusions that biases in evaluation of sires will
result from misidentified records in estimating the sire
and environmental variances, and that genetic progress

will be underestimated from misidentified records.

Genetic Trends

 

During the past decade several researchers have
studied annual genetic progress for various herds and
populations. Palmer, et a1. (1972), reported mean annual
phenotypic, genetic and environmental trends in milk pro—
duction of 23.4 i 4.6 kg, 37.5 i 6.4 kg, and ~14.0 kg,
respectively, in a Florida experiment station Jersey herd.
Dillon, §t_gl, (1955), reported a mean change in average
real producing ability between years of 706 pounds. The
latter study was based on records from the University of
Illinois dairy herd from 1901-1954. They concluded that
the average real producing ability in this herd had changed
very little since it was founded. Arave and Laben (1963)
studied twelve privately owned Jersey herds with environ«
mental conditions representative of central California
conditions. They reported variation between herds of from
16.4 i 54.8 to 186.2 1 30.1 lb. when FCM yield, corrected
for year effects was regressed on year. In only one herd
was this regression reported to be nonsignificant, indicat—

ing significant positive trends in eleven herds.

 

EV.

Burnside and Legates (1967), investigated genetic
trends using first lactation Holstein Herd Improvement
Registry records reported in 335 herds from 1953 through
1961. Few of the records studied were those of artificially
sired individuals. The estimated annual genetic improve-
ment was reported as being from 0.75 to 0.92 percent of the
mean milk yield. This approaches the expected rate of
improvement (0.01 7 for closed herds) reported by Rendel
and Robertson (l950)and Specht (1957).

Table 1 and Table 2, adapted from McDaniel (1968),

illustrate gains possible and gains achieved.

Table 1. Annual genetic gains possible through AI.

 

 

 

Source Percenta Actual (lb)
Hickman .86 to 2.36 120 to 330
Johannson 1.00 to 2.00 140 to 280
Skjervold 1.50b 210

Specht 1.70 to 2.30 238 to 322
Thompson 2.11 to 2.21 -- «-

 

aPercent of mean annual production.

bAssumes considerable selection for other traits.

Fractions of AI Sires which are
Plus and Minus

 

 

McDaniel (1968) noted that one of the reasons why

the superiority of AI has not been as great as it promised

I!
:‘.~'m
I

 

LI

10

Table 2. Observed annual genetic gains in populations and

herds of dairy cattle.

 

 

 

 

Source Milk Fat
%/yr %/Yr
Populations
Burnside (1964)* .83 --
Van Vleck (1961, Non AI)* .37 .44
Van Vleck (1961, AI)* .48 .65
Systrad (1966) .67 --
Single Herds
Branton (1967) 1.09 .74
Burnside (1967) (Ayr) 1.30 1.70
Burnside (1967) (Hol) 0.00 .32
Gaalaas (1967) «.18 .32
Gacula (1965) 1.09 .21
Legates (1966) 1.25 1.21

. 11 1.1.15!
I

 

 

*Data developed from those presented in source.

to be is because a substantial portion of the proven bulls

in A1 service are known to transmit below breed average
milk production. McDaniel used the following table to

summarize the percent of proven AI sires available which

are plus and minus (Table 3).

Available Herd Replacements

 

Johnson, et a1. (1970), in a study of calf losses

in Illinois Dairy Herd Improvement herds reported the

average loss of female Holstein calves at 13.1 percent

‘25,»; ‘

11

Table 3. Proven sires available through AI January 1, 1968.

 

 

Below Breed Average Above Breed Average

 

 

 

 

 

PD Milk EEETEH PD Milk 53${2n
(1b) (%) (1b) (95)
-l to -99 10.77 1 to 99 8.32
-100 to -199 6.24 100 to 199 9.91
-200 to -299 4.90 200 to 299 12.21
-300 to -399 4.65 300 to 399 6.85
-400 to -499 3.92 400 to 499 6.24
-500 to -599 2.08 500 to 599 6.61
-600 to -699 2.20 600 to 699 3.55
«700 to -799 .86 700 to 799 3.43
«800 or below 1.10 800 to 899 2.08
900 to 999 1.47
1000 or more 2.69
Total 36.72 65.27

 

*Based on studs reporting number of services to
individual bulls.

annually. There was a decline in the percent loss as the
average milk fat production increased.

The New Zealand Dairy Board (1970-1971) summarized
calf losses as a distribution of herds by main cause of
loss. Thirty-six percent of the herds reported scours as
the main cause. Forty-four percent (nearly one half)
reported that the main cause was either unknown (23 percent)

or "other" (21 percent) than those mentioned specifically.

Eelven, six, and three percent reported worms, salmonella

12

and redwater as their respective main causes. A void
exists in the scientific literature concerning calf losses
and their causative factors.

In artificially inseminated (AI) populations, selec-
tion on both the male and female sides are dependent on the
number of heifers which survive and calve. As a comparison
to the extreme case in which all heifers born (50%) survive
until they calve are the 0.35 to 0.40 mature heifers pro—
duced per cow in the herd per year as reported by Rendel
and Robertson (1950) (implies that 70-80% of all heifer

calves born calve in the same herd).

Cow Removals

 

Andrus, g£_al. (1970), in a study of Iowa dairy
herds, reported that the average herd life expectation was
3.12 years for all cows from herds actively enrolled in
the production recording program. Cows from inactive herds
averaged 3.09 years of herd life expectancy. It was
reported that the differences between herds actively test!
ing and those no longer active were significant.

White and Nichols (1965) studied the relationship
between first lactation, later performance, and length of
herd life in HolsteinsFriesian cattle. Average ages at
first and last calving were reported as 27.03 i 3.28 and
58.84 1 17.77, respectively. They reported a correlation
of 0.216 between first lactation M.E. milk and number of

lactations completed. This was significant at the one

3;.

 

IL-..

13

percent level. Also significantly correlated (P < 0.01)

were first lactation M.E. milk and age at last calving

(r 0.254).

Another study by White and Nichols (1965) revealed

36.9 percent of cow removals were for low production. A

brief table compares these results with those of other

researchers (Table 4).

Table 4. Percent cows culled for low production.

 

 

Source

Percent cows culled
for low production

 

White and Nichols
Asdell

O'Bleness, Van Vleck
Baltzer

Seath

36.9
33.5
27.1
41.1
30.5

 

Burnside, et a1.

(1971), reported that reporduction

(13‘24 percent) and low milk production (16—28 percent)

were major causes of voluntary herd removals in Canadian

herds. It was also reported that young cows were culled

more heavily for low production than their older herdmates.

This was offset, however, by increases in the percentages

of older cows removed for reporduction, diseases, and weak-

nesses in udders. Burnside, et a1. (1971), considered any

given reason associated with one percent or more of total

fem!

 

14

annual cow removals to be economically important. Percent
first lactation cullings were lower for the Holstein breed
(22 percent) than for the Ayrshire breed (37.2 percent).
Batra, g£_al. (1971), reported that in Canadian
herds of constant size, production level increases were
associated with increases in percent sold for breeding
purposes and increases in percent sold for beef becuase of
poor type (P < 0.01). Herd size had no effect on cow dis-
posals. They concluded that increases in the level of

milk production lead to increases in the percentage of the

 

herd inventory that changes each year in pedigree herds of
constant size.

According to Asdell (1951), the net loss of DHIA
cows to the industry is 16.8 percent annually. Fifty per-
cent of the cows in DHIA herds were reported to have been
removed for reasons other than for dairy sales in a little
over three years. It was concluded that 3.5 years of pro-
ductive life, as an average is substantially correct,
taking into consideration that in each year some of the
culling is among the new cows introduced as replacements.

Parker, gt_al. (1960), reported that differences in
longevity between individual cows are determined largely
by nongenetic influences. It was pointed out that cows
which die young leave less offspring than those which
remain in the herd longer. This they referred to as auto-

matic selection for longevity.

15

Parker's study also indicated that no deleterious
effects on longevity need be expected by concentrating
selection of progeny tested sires on the basis of first
lactation production records of their daughters. Non-
breeders were reported as accounted for 33.4 percent of
all Holstein cow disposals from 1918 to 1958. In addition
to T.B. and bangs reactors; udder trouble, infections, and
abortions were reported to be important reasons for dis-
posal.

Fosgate (1965) reported that for a herd of registered
Jerseys, the average age in years and the number of lacta-
tions completed by all cows at the time of disposal were
6.79 and 3.43, respectively.

Allaire and Henderson (1966) studied 34,000 lacta-
tions made by cows in 464 New York herds on continuous test
from 1957«1962. Using 195791960 averages they reported 0.788,
0.755, 0.717, and 0.672 as the fractions remaining in the
herd after the first, second, third and fourth lactations,
respectively. These represent animals selected during each
of the first four lactations. When these figures are con«
verted to represent removals during each lactation, they
create a useful comparison with those of various workers

in a table adapted from Specht (1967) (Table 5).

Calving Interval

 

Pelissier (1970) analyzed breeding records of 5,000

cows in ten herds to determine factors contributing to low

16

“I'D-lull
.. . n .. .1! .13.... .I...

OED”.
.. \ I». .
. r

.mcwaaso

msofi>onm mew>fi>hsm omocu mo coauwpowfi some co>oEon mzoo mo pqoopomm

 

 

 

 

m.Nm o.mm N.mm m.o~ o.m~ N.~m m.w~ a
m.w~ 4.0N H.mm N.AN N.mN N.mm N.SN m
m.4~ N.AH m.mH e.om m.Hm «.mm N.AN N

«N.HN mo.o mo.NH mo.Hm mfi.wN an.AN AN.mN H
moomflv .>.z mamcax mammax aon flHmmHM fiummfiv

comhovcom III .Hm pm unoomm :ofiumpomq

ogHaHH< flammav Hamem< floemfiv npamm .Hmuaom

 

 

. wCOMHmHUNH HDOW Hmhflmw

may mcfihsu mHm>oEoH Sou .m oHan

17

breeding efficiency in dairy herds. An average of 2.67
services was required per conception with 43.3 percent
representing conception on first service. Four or more
services were required by 18.3 percent of the cows. Those
cows sold open or conceiving too late to remain in the
herd represented 12.7 percent.

One of every six heat periods was reportedly either
missed or not recorded subsequent to the first recorded
heat period. Recorded heat periods are summarized by days

after calving in Table 6.

 

Table 6. Percent of heat periods recorded of those not
previously recorded by days after calving.

 

 

Days after

calving <60 60 90 120
Percent with

heat period 65.5 34.5 11.9 3.9
recorded

 

It was concluded that the large number of cows open for an
excessive period was primarily the cumulative result of low
conception rates and a high incidence of missed heat periods.

Sweden (1968) reported 62.9 percent 56 day nonreturn
rate to last insemination.

In its annual report for 1970«l971, the New Zealand
Dairy Board reported an average of 27 calves per 100 cows
were reared as dairy replacements. This corresponds exactly

with the number reared in 1969s1970. The average number of

18

calves per 100 cows in milk dropped from 30 calves per 100
cows in 1968-1970 In) 25 calves in 1970-1971. The average
number of dairy heifers born in 1969 was two lower than in
1969-1970.

Specht (1957), on the basis of information from 269
Holstein DHIA herds in Michigan, estimated that 70 percent
of all heifers ultimately enter the milking herd. The
number of heifers useful for progeny testing is further
restricted by the portion not having positive sire identi-
fication. Specht (1957) stated that his figure (47 percent
of all identified heifer calves enter the same herd as
first calvers) agreed closely with the 1956 findings of
the New Zealand Dairy Board (23 heifers will freshen in
the same herd for each 100 cows in calf). To account for

the normal sex ratio, the latter figure must be doubled.

Generation Intervals

 

Generally rGP (correlation between genotype and
phenotype) increases with age as more information is avail«
able, but progress is also a function of generation interval
and the ratio of :%E determines progress if age increases

both réP and the length of generation intervals T'. In-

r!
creased genetic progress can be expected only if —%$ exceeds
rGP
T 0
Of special interest to the dairyman is his own pro—

gress with reference to the progress which he could be mak—

ing. In a 200scow herd, Specht (1957) showed graphically

19

that genetic improvement with progeny testing was maximized
by saving one of four young sires sampled when 25 percent

or more of the cows are bred to young sires. Annual genetic
improvement reached 1.4 percent of the average yield when
approximately 80 percent of the cows were bred to young
sires. An additional 0.45 percent in improvement was realized
annually with progeny testing in an AI population of 150,000
cows (10,000 tested) saving five to fifty sires sampled on
fifty percent of the cows (tested). Thompson, gt_al. (1971),
using four methods of sire selection, evaluated the genetic
gain expected from four methods of sire selection:

1. The use of young, untested bulls, all of which
are replaced annually.

2. The use of young, untested bulls replaced
annually, all sired by progeny tested sires.

3. The use of naturally proven bulls.

4. The use of pedigree selection combined with
progeny testing.

Differences in the amounts of genetic progress pos-
sible among the various methods of sire selection in the
model used by Thompson and Freeman are largely reflections
of the generation intervals of the sires. The generation
intervals used by Thompson and Freeman (1971) are presented
in Table 7.

From this work they concluded that the use of young,
untested bulls, sired by progeny tested bulls resulted in
more rapid genetic gain than progeny testing in small popu«

lations (between 5,000 and 20,000 first services per year).

20

Table 7. Generation intervals used by Thompson and Freeman
in evaluation of four plans of sire selection.

 

 

 

 

Plan
Source
A B C D
LBB 2.5 6.0 8.4 8.0
LBC 2.5 2.5 8.4 8.0
LCB 4.75 4.75 4.75 4.75
LCC 4.75 4.75 4.75 4.75

 

Progeny testing resulted in more rapid gain in large popu-
lations, and in all population sizes when usage rates were
low. For a population of 100,000, annual genetic progress
was maximized using selection method D above with 15,000
first services per bull, selection at the five percent
level, and sixty daughters per young bull.

The importance of the differences in progress
achieved under the various plans should not be minimized.
Another important question is, what is the most efficient
method of identifying young sires which have the greatest
genetic potential. In 1969 Van Vleck reported three pedi~
gree sources as having primary importance. Van Vleck and
Carter (1972), by weighting these three pedigree sources
of inheritance, computed estimated daughter superiority
(EDS). Daughter superiority estimated as predicted dif-

ference (PD) or sire comparison (SC) was then regressed

21

on the pedigree estimate of daughter superiority. Accuracy
of the method used, as indicated by simple correlations
between EDS and PD or SC, was at or near the thirty percent
level for both bull studs included in the survey. It was
found that on the average, about 0.61 units of PD or SC
result from the increase of one unit of EDS. They con-
cluded that selection of high EDS young bulls is a very
effective method of gathering a superior group of young

bulls for further sampling in A1.

Useful Life Span of Tested Bulls in AI

 

According to Specht and McGilliard (1960) the
generation interval decreases 0.5 year for each ten percent
increase in the number of cows bred to young sires.

Roman, et_al. (1969), studied the tenure of bulls
from 1940 to 1964. Bulls were reported by ages at entry
and disposal, and unadjusted and adjusted tenure. Findings
for the Holstein breed are reported in Table 8. These ages
agree with Becker and Arnold (1953), especially those ages
and the length of tenure for corresponding years. While
the age at entry remained nearly constant over time, age
at disposal as well as tenure for all categories increased
over time.

Roman, §£_al. (1969), reported that data from
Germany showed 40.6 percent of 1,013 bulls discarded for

low fertility. Over 38 percent of a different sample of

22

Table 8. Tenure of Holstein dairy bulls in AI.

 

 

Age at
Tenure (yrs)

 

Period No. Sires
Entry Disposal

 

 

1940-49 703 5.31 7.26 1.953 2.05b 2.05C
1950—54 946 4.78 7.90 3.12 3.13 3.12
1955-59 1067 4.40 8.36 3.96 3.90 3.92
1960-64 733 4.47 8.80 4.33 4.29 4.30

a .

Unadjusted.

Adjusted for age at entry on an individual breed
basis.

CAdjusted for age at entry, removing year effects.
sires were reportedly culled for the same reason. In their
1969 study of 8,887 North American sires of six dairy breeds
leaving artificial service during 1939 through 1964, Roman,
g£_al., reported the following reasons for disposal and
their respective intensities in Table 9. Among the Holstein
bulls, 7.9 percent were culled for old age and 2.8 percent
for red factor.

Freeman (1970), reported that the average selection
intensity from 1951 through 1954 was 35 percent in a bull
stud which operates in the North Eastern United States. Of
the 196 young sires brought to the AI center, 23.5 percent
were not alive at the time of selection. Six percent of
these were discarded during the course of the progeny test.

Becker and Arnold (1953) analyzed tenure and turn-

over for 189 desirable dairy bulls in artificial studs,

23

Table 9. Percent of sire removals by reason.

 

 

 

Reason Intensity
Repro. inefficienty 36.3
Char. or performance of Daus. 15.8

low milk prod. 9.7

poor type 2.5

inbreeding 1.6

other 2.0
Diseases and infection 14.3
Problem-semem collection 11.8
Accidents or injury 4.2
Consolidation of studs 6.8
Poor type (bull's own) 1.2
Other reasons 9.6

 

born before January 1, 1937. Twenty-six started artificial
service between two and 5.9 years of age, 121 others at

six to 9.9 years and 42 at ten to fifteen years. For all
189 bulls, the average tenure was 2.68 years. Later work
by Becker and Arnold (1957) showed 56, 31, and 13 percent
of the culling of desirable bulls was for reasons connected
with reproduction, physical defects, and diseases, respec'
tively. Of the 691 bulls included in the 1957 study, all
of which were born before 1940, 25 percent were used less
than one year, 21 percent from one to two years, 17 percent

from two to three years, 11 percent from three to four years.

24

It was stated in the 1970-1971 annual report of the
New Zealand Dairy Board that 80 percent of the bulls in an
intake can be expected to be alive when selection is made.
In New Zealand bull centers, intensity of selection

increased during the period 1953 to 1970.

SOURCES OF DATA

Inaprogress DHIA records for February 15, 1972 were
used to develop a profile of the average Holstein cow milked
on that day. Her age was computed on the basis of 124,329
cows. Sires of 33,419 cows having official records (those
identified by sire and in milk at least 45 days) were iden-
tified from the records mentioned above. Of these, 20,678
were sired by a total of 1158 sires which were summarized
after April, 1967. The USDA sire summaries were used to
obtain the birth dates and the latest summary for fat and
milk of all 1158 sires. All bulls having a stud code
with their last summary were classified Al. Al and natural
bulls were then stratified into plus and minus.

Ages of sires and dams of young sires entering three
artificial insemination units from 1969 to 1971 were arrived
at in the following manner. Pedigrees for all young bulls
entering the three studs in the stated time period were
solicited and received. The registration numbers for the
sire and dam were obtained from the pedigree of each bull
and used to locate their birth dates in their respective
Holstein herd books. Identification of each young sire
along with his birth date and the birth dates of his
parents were punched on computer cards and the ages of

the parents at the time the young bull was born were computed.

25

26

A questionnaire survey (see Appendix) was conducted
of four bull studs which breed 90 percent of the artifi-
cially bred population in Michigan to determine the current
status of progeny testing in Michigan. Progeny testing
estimates are based on the number of first services per
stud in Michigan. Responses of varying degrees of clarity
were received from three of the organizations which bred an
estimated 83 percent of the Michigan Holstein population in
1971.

Recent research of Spike (1972), Erickson (1972),
and Erickson, gt_al. (1972) were sources of information
about calving interval, sire selection decisions, and
superiority of bulls currently in use. Michigan calf
mortality data is based on research findings reported by

Speicher (1972), and Oxender (1971).

RESULTS AND DISCUSSION

Available Herd Replacements

 

Specht (1957) used 70 percent as the fraction of
all heifer calves born that are available as herd replace-
ments. For purposes of the present study, an estimate of
the number of available herd replacements was desired.

The following formula was developed.

 

 

 

[(NCIL)] (l30x.85x.8fl 90

c 2 x °

2 2 or = 39

I 1.08

where: N = cows bred per 100 cows in the milking herd
C1 = percent bred which calve
L = percent of calves which live to breeding age
2 = to account for normal sex ratio

C2 = percent of heifers reaching breeding age which
also reach the milking herd

I = calving interval.
All heifers available as herd replacements (39)

expressed as a percentage of heifer calves born ((130§’85)-55)

 

results in 70.9 percent, or essentially the 70 percent
Specht (1957) used in his thesis.
Another basis which might be used to estimate the

number of available herd replacements is cow removal levels.

27

28

One limitation is that cows sold for dairy purposes are
either replacing cull cows in some other herd or enlarging
the size of some other herd.

Meadows (1968) refers to heifers not reaching the
milking string as "cow wastage." He cited McGilliard's
data as indicating that for each 100 cows successfully
bred, 30 heifers reach the milking string. Meadows also
cited Pelissier's reported 32 daughters reaching the milk-
ing string from every 100 cows based on California DHIA
data.

Reducing calculations made earlier in this section
to represent heifers reaching the milking string per 100
cows successfully bred [(%%%) = .77] an estimate which
agrees exactly with McGilliard's data is found
(.77 x 39 = 30).

Still another approach may be used to estimate
heifers born which reach the milking string. If an accurate
count were made of all first lactation females, the number
of available herd replacements should be the same except
for any culling which might have taken place. This was
done as part of the present study by examination of lacta«
tion codes of cows on the IPR file. Slightly more than 27
percent of the cows were found to be in their first lacta-
tion. Pennsylvania, in the 1968 Dairy Reference Manual,
reported that in a herd of 100 cows there are 23 first

calf heifers.

29

Calf mortality is the one most important factor
determining the number of heifers available as herd replace-
ments. According to Speicher (1972) several factors influence
calf mortality. Calf mortality is higher in Southern Michi-
gan (15 percent) than in Northern Michigan (12 percent).
During the summer months calf losses are lower (10.3 percent)
than for winter months (17.1 percent). Annual calf losses
increase as herd size increases. The greatest period of
hazard to the calf that was alive at birth came during the

first week. No significant differences were noted between

 

the breeds, but Jersey and Guernsey herds had slightly
higher calf losses than did Holstein herds. Stanchion,
switch, loose, and free stall housing systems were compared.
Losses were higher in loose (15.1 percent) and free stall
housing (14.2 percent) than in stanchion (12.0 percent) and
switch (11.3 percent).

Several management practices were also discussed.
Higher calf losses were reported by dairymen who turned
calf care over to hired labor. Feeding of colostrum within
the first five hours after birth helped reduce calf losses.
Survival rates were higher for calves kept in individual
pens. Oxender, gt_§1. (1972), reported an average calf
loss for the state of Michigan of 17.6 percent. Herds used
in the survey were reportedly larger than the average for

Michigan at the time of the study.

30

Percentage of Cow Removals

 

Cow turnover influences genetic progress in several
ways. The female generation interval is greatly dependent
on how many cows are culled and at what stages. Selection
achieved by the dairyman during two phases of the cow
removal process are important in determining genetic improve- I
ment which can be made. Phase I might be considered selec-
tion of replacement animals which, as evidenced in the

immediately preceding section, is reduced by sterility,

r “CELL.

abortions, losses of females which have not freshened, and
calf death losses. The second phase, then, is the selec-
tion achieved by voluntarily culling as large a portion of
the cows as possible. Discussion in this section will deal
mainly with phase 2. In Table 12 is a comparison of recent
(June 9, 1972 IPR files) cow removals with 1968 Michigan
DHIA removals and 13 year Michigan DHIA removal averages.
Some of the questions pertinent to the present study
are: changes which have taken place over time, levels of
involuntary culling, and factors which contribute most to
involuntary culling levels. Percent removals for the
various categories are very similar for those from the
current IPR file and those from the 1968 DHIA records.
Important differences exist between current removals and
those in the thirteen year average. The decrease in involun-
tary culling levels is an outstanding characteristic of the

differences between these two sets of data. Since death

31

.7 a
. .
.
PI .I I III

I... jflr

.mepouop <Hza camA;8az wsmHImmmHa

 

 

 

-- mm.~s Nm.mm «H.sm mucoupoao Haze spapcsflo>aH
o¢.H mm.m om.m mm.m wean ucoonom
me. Nw.H BH.H mm.H moa mansoeu mcfl>mao
mo. HH. II No. H meoxsa
co. em. II ma. 5 wflcoescom
so. ma. II No. H wcflcomwom ommhom
so. 0H. mm. ma. oH 0mm wHo
mm. oo.H wu.H Hm.H mu ucopfioow
mo. mm. mm. mo. m pecan
NH. Hm. up. mo. mm chested:
mo. ma. om. om. ON «MaocOpoom
om. mm. he. No.H um Ho>om xHHE wean

.. -- Ha.m .. I- Congo
ea. um. I. om. mm opmzopw;
om. Hm.H .. ao.H so was use
No. ov.oa om.OH ow.qH omm xpflafipoum
ma. om. mv.H mm. wH noxafie the;
ow. mm. II mm. om ucoEmhomEop
No. me. am. HA. s amass
OH.H No.e «H.o om.o cum mfiuflumme
Hm.H as.m sa.w ms.s omN sashes Hmuamsas
mm.mH no.0m Hm.we mm.nm nmam Zoo Haso
Hm.H Ho.w ss.a Hs.w awe shame - efiom

.mom mo w mHm>oEou Hmuou mo “coupon .02
<Hzm <Hmm . whoa Numfi
mama mesa a o>< Ham» mH «use mass

 

 

.<Hmm cmmfinowz xo poueomoa

mHm>oth 30o Spawn

.CH oHan

32

losses are similar, the difference must be in the sold
category. Physical injury was responsible for more than
three percent of the difference between the thirteen year
average and the current status as indicated by the June,
1972 study. In the present study, sterility claimed
nearly two percent less cows than reported in earlier
DHIA data. An additional 2.14 percent were culled for
"other" reasons in the thirteen year average as compared
to the current level (old age and "hardware" disease were
combined for purposes of this comparison).

According to Burnside, §£_al., any reason which
accounts for greater than a one percent loss is an econo«
mically important one. Physical injury, mastitis, steri-
lity, and death losses each represent more than one percent
of the cows in the 1968 DHIA cow population. Sterility,
physical injury, and mastitis are the three causes over
which dairymen would be expected to have greater amounts
of control. Based on the 1972 IPR data, losses due to
sterility appear to have declined since 1968. Mastitis
losses were lower in 1968 than they are currently. Based
on the thirteen year DHIA averages, total removals in
Michigan represent thirty percent of the cows in the
herd annually.

It is unlikely that too much emphasis will ever
be placed on reducing involuntary losses to effect an

increase in voluntary culling levels. The same argument

33

applies to both replacement heifers and cows in the milking

herd.

Calving Interval

 

Calving interval reflects the management level of
the dairyman, as much as, or possibly more than, any other
factor discussed herein. Not only is it a reflection of
heat detection, but also of general herd health, reproduc-
tive tract health, and the nutritional program. From a
management point of view, it is as much a question of
having the cow bred at the right time as it is of detect-
ing heat accurately. Spike (1972) found that the average
calving interval for Michigan Holsteins over a thirteen
year period (1953«1966) was 395 days.

Length of calving interval influences genetic
progress in at least three ways:

1. Long calving intervals lengthen the generation

interval.

2. The rate at which herd replacements come into

the herd is decreased by long calving intervals.

3. Voluntary culling levels are reduced as the

rate at which herd replacements coming into the

herd decreases.

Generation Intervals

 

As implied by the heading to this section, dairy

breeders must be concerned about the ages of more than one

34

group of animals. Robertson and Rendel (1950) outlined
the four sources of parents for the next generation as
sires of sires, sires of herd replacements, dams of sires
and dams of herd replacements.

Vinson and Freeman (1972) reported average ages
for sires and dams of young bulls of 136 months and 87.9
months,respectively. No ages were reported for parents
from the other two sources.

Ages of dams of herd replacements were taken from
IPR files on two different days. Average ages were found
to be 51.96 and 52.15 months for groups of 125,706 and
124,329 Holstein cows, respectively. Fiftystwo months is
an appropriate average age to use for Holstein dams of
herd replacements.

Sires listed by four AI organizations, in 1972
catalogues, were used to compute the age of sires of herd
replacements. It is clear that differences do exist
between the breeding organizations included in this study.
An average age of over nine years was calculated for
available sires in four bull studs.

Pedigrees of young sires entering three AI organi-
zations between 1969 and 1971 were used to compute average
ages of sires and dams of young sires. Results are listed
by stud in Table 11. Ages of sires and dams agree well
with ages reported by Vinson and Freeman (1972). Ages
point up a dramatic change which has occurred since

Specht's 1957 work. The nature of the formula for computation

4"-

 

35

Table 11. Average ages of sires of herd replacements.

 

 

 

 

No. of Ave. Age
StUd Bulls in Months Range
A 40 108.4 64—185
B 27 116.1 75-195
C 26 117.5 78-181
D 45 99.8 67-138
Overall 138 108.8

 

of genetic progress is such that doubling the generation
interval, for instance, requires that accuracy in the
computation of breeding values be increased two fold in
order to continue to make the same amount of progress as
before the increase occurred in the generation intervals.
A greater increase in generation interval than in accuracy
creates negative progress. Almost too obvious to require
mention, is the fact that this is a serious problem and
will receive much attention from AI organizations in the

coming decade.

Current Michigan Holstein Population

 

A knowledge of the cow population is necessary,
not only from the standpoint of the number of insemina-
tions required, but also from the standpoint of the number

of young bulls which must be tested annually.

36

 

 

 

 

 

 

 

 

5H.om no-“ Nam HHDmIEmo
nw.emfi moIHH mam HfismIouwm
mmnucoev AoEInxv mommpo>m
Hamuo>ow Hamuo>ok Hmu0pz was mHmuou HHmHo>o
“Hasn-eae
NoImH oaImH oquH +HH5nIoHHmV HmuOH
No-5 Ho oHIu «ea moIn Boa HHSmIEmo
ooIHH Ho ooINH qua moIOH AOH HfismIogflm
ooIm Ho HOIOH sea HHIm noa Ema
HHINH Ho moIea sea OHINH Boa oufim
OHIH Ho NoIN «ea uoIN noH HHsm
moEIhwi mz moEIthM Nz moEIhng Hz
0 m <
woum

 

 

.Humfl was mama coozuon
mpzpm oopnu wcflhopco maasn mcsox mo mopfim was wasp mo moms omwho>< .NH oflome

37

According to the July, 1972 Michigan Agricultural
Statistics, the dairy cow population (milk cows and heifers
that have calved) was 473,000 at the beginning of the year.
Changes in the dairy cow population which have occurred
during the past three years are reflected in the following
figures taken from the 1972 version of Michigan Agricul-

tural Statistics.

Table 13. Michigan milk cows* by year.

 

 

Year

 

1969 1970 1971 1972

 

Milk cows 469,000 462,000 463,000 473,000

 

*Milk cows and hiefers that calved.

According to Speicher (personal communication)
Holsteins constitute 86.9 percent of the Michigan dairy
cow population. Ninety«two and 93 percent of the cows
on the in«progress record (IPR) file were Holstein on two
different dates. Speicher's figure would indicate that a
smaller proportion of the cows of the minor breeds are on
test than are Holsteins. For the present study 90 percent
is used as the fraction of the dairy cow population repre-

sented by Holsteins or 425,700.

Effective Population for Progeny Testing

The 1971 DHIA annual summary reported 85,931 cows

on official test. Holsteins on official test represent

38

18.1 percent of all Holsteins.

A surprising proportion of grade Holstein cows are
not correctly identified by sire. It was reported by USDA
(1971) that of 1,907,042 lactation records submitted,
849,506 were used for genetic evaluations of bulls and
cows. Ninety-two percent of the 44.55 percent rejected
records were records of grade cows lacking sire identifi—
cation. The May, 1971 run by USDA found ninety percent

of 45.76 percent rejected for the same reason.

Table 14. Percent DHI cows identified by sire.*

 

 

 

 

 

Cows with sire States
1dent1f1cat1on 1968 1969 1970
% No No. No.
90 or more 0 0
80 - 89 2 3
70 - 79 8 6
60 « 69 10 12 15
SO - 59 17 10 11
40 - 49 8 15 12
30 - 39 4 2 2
20 « 29 0 0 0
10 - 19 0 0 0
l « 9 1 0 0
Less than 1 0 1 0
U.S. average % 59.0 57.1 59.1

 

*Table adapted from USDA DHI letter.

fl'n_._‘u-——-_ _- _ __.. ._ IA-

39

According to USDA (1971) only 56.8 percent of the
official Michigan DHI records reported in 1970 had valid
sire identification. The above table from USDA shows how
Michigan ranks in relation to other states and the national
trend over three years is also shown.

The current estimate of the Michigan population
effective for progeny testing is 43,762 Holstein cows.

Stratification of Michigan Holstein
Population by AI andfiNat.

 

 

In Figure 1, adapted from the June-July, 1971 issue
of the USDA DHI newsletter, the trend in the percent of
U.S. dairy cows bred by AI since 1945 is shown. In the
same newsletter it was reported that 46.7 percent of all
Michigan dairy cows and bred heifers were bred artificially
in 1970. In this study 46,376 Michigan in progress records
having method of breeding descriptions other than unknown,
showed 65.1 percent bred artificially. Based on 1972 USDA
estimates of firstsservices in Michigan, 59.1 percent of
the Holstein population is bred artificially. Sixty per-
cent is assumed in the present study to represent the
fraction of the population bred artificially.

'Fractions of AI and Natural Which
are Plus andFMinus"7T

 

 

Sires (summarized since April, 1967) of cows on the

IPR file which were in milk at least 45 days, were analyzed

 

40

Cows and heifers of breeding age

50‘

401

30.

20+

10 -

Percent of dairy cows bred artificially

 

 

V I V 1 U

45 50 55 60 65 70

Figure l. Yearly trends in the percentages of U.S. dairy
cows bred by AI since 1945.

 

41

to determine the fractions of plus and minus sires in each
of the strata discussed in the immediately preceding sec-
tion. Forty-six percent of the bulls analyzed were not
identified by stud code in the annual USDAvDHIA Sire Sum«
mary List containing their most recent summary. These bulls
were assumed to be natural while the remaining 54 percent
were known to be in bull studs at the time of their latest
summarization.

Five hundred ninetyxnine active AI Holstein bulls
were listed by USDA based on PD rankings with 200 pound
intervals. Percentages were calculated and are presented

in Table 15 along with the USDA data.

Table 15. Number of bulls by 200 pound PD intervals.

 

 

 

 

 

 

 

USDA AIa NATa

No. % No. % No. %

1000 and up 36 6 27 4 _- -—
800 to 999 42 7 38 6 2 1
600 to 799 83 14 54 10 7 1
400 to 599 86 14 66 ll 29 5
200 to 399 101 17 87 14 64 12

0 to 199 95 16 84 13 151 28

-l to -199 69 11 90 14 138 26
-200 to -399 56 9 80 13 90 17
-400 to —599 18 3 46 7 37 7
~600 to -799 5 1 27 4 ll 2
~800 to «999 5 l 8 1 5 l
-1000 and less 3 l 17 3 -- -—
Totals 599 100 624 100 534 100

 

aBased on current Michigan study.

42

The following general interpretation of the data was made

by USDA.

A zero predicted difference in the table indicates that

a sire with zero PD, used in breed average herds, has
about half of his daughters above and about half below
breed average. He can be expected to transmit somewhat
less than breed’average production to his future progeny
because of genetic improvement across years. In general,
the bulls with minus PD's sire daughters more than 50
percent of which are below breed average in production.
The bulls with higher PD's are expected to sire daughters
with higher production than bulls with lower PD's regard-
less of herd production level.

The current study included 1158 bulls. There were slightly

 

fewer natural sires with positive proofs than with negative ,
proofs; however, those bulls having positive proofs sired a
greater number of the naturally sired cows than did those
sires having negative proofs. Among the AI sires, a greater
percentage of the bulls were plus (57 percent) than minus.
A greater percentage of the AI daughters were also from
positive bulls (62 percent) than from negative bulls.
A summary of plus and minus sires and of daughters
of the same for AI and natural is presented in Table 16.

Table 16. Percent plus proven AI and natural sires and
daughters of the same.

 

 

Daughters of plus

Plus roven sires .
p proven Sires

 

 

 

No. % No. %
Natural 253 47 1785 57
A1 356 57* 10864 62*

 

*% + AI sires and % daughters of + AI sires were
significantly different from natural at the .001 level.

i

43

In Table 13 it is interesting to note that while a differ-
ence of nearly ten percentage points existed between plus
proven AI sires and natural sires in the same category,
only a five percentage point difference existed when dif«
ferences between the fractions of AI and natural daughters
from plus proven bulls were measured. It is also note—

worthy that daughters of minus proven AI sires represent

._..._. "I“?!

the smallest portion of either of the two strata.

Superiority of Bulls in Use I

 

 

Genetic improvement in the population depends on
the use of selection procedures which identify superior
sires and the differential use of sires.

Data presented in the immediately preceding section
represent sire selection decisions made by dairymen nearly
one generation of cows earlier or approximately four years
ago. The average superiority for milk and fat of natural
sires included in this study were +24 lb. and 91 1b.,
respectively (averages weighted by number of daughters for
each sire). For AI sires the average superiority for milk
and fat were +168 1b. and +6 1b., respectively.

According to data presented in the March, 1972 USDA
DHI newsletter, the weighted, average predicted difference
for milk and fat of active AI sires were +318 and +8,
respectively. Only in the Holstein breed were the weighted

predicted differences greater than the unweighted ones.

44

Researchers at USDA concluded that for those breeds other
than Holstein "this indicates selection for traits other
than yield and income." Other considerations related to
selection intensity for sires will be discussed in a
later section.

A 1971 survey of Michigan dairy farmers by Erickson,
g£_a1. (1972), revealed that farmers were using sires with .
predicted differences for milk of +826.3 and +822.3 for
grade and registered herds, respectively. Based on a more

recent survey, Erickson (1972) reported for sires of calves

.n‘n-cm‘J—go mn- c. ‘9

born in the last year, predicted differences of +808 in
herds with high levels (17,694) of milk production and
+694 in herds with near breed average levels (13,397) of
milk production. Of particular interest at this point is
the agreement between responses to Erickson‘s earlier
questionnaire and the response of owners of herds with high
levels in the more recent survey. There is an apparent,
however small, upward bias in the average response to the
earlier survey. A strong trend toward the selection of
higher predicted difference bulls is indicated.

Average predicted differences for milk and fat
were calculated for sires available to Michigan dairymen
through four studs. Active bulls listed in the 1972 bull
books were used for the calculations. Results are summarized

in Table 17.

45

Table 17. Average PD's for milk and fat for available sires.

 

 

No. Ave. PD* Ave. PD*

 

 

 

Stud Bulls milk Range m11k fat Range fat
A 40 414.0 9345 to 1892 12.3 -13 to 79
B 27 643.2 —82 to 1355 20.8 -6 to 55
C 26 543.1 -l49 to 1268 17.7 -2 to 52 P
D 45 753.8 197 to 1906 24.7 3 to 62 I
I
*Based on 1971 Sire Summary List.
Differences of 339.8 pounds of milk and 12.4 pounds L

of fat existed between the average predicted differences of
the high and low organizations. The overall average predicted
differences for milk and fat were 593.97 and 19.02, respec-

tively.

Useful Life Span of Tested Bulls in A1

 

A survey was conducted of four bull studs which
provided semen for 86.3 percent of the AI services to
Michigan dairy cows in 1971. There are three general stages
of culling in the life of an AI bull after he enters a stud:

1. Progeny test

2. Waiting for results from the progeny test

3. Active service.

The stages listed above are used for convenience sake,

recognizing that many differing programs exist.

46

Removal rates for each of the stages were requested
on a questionnaire. Table 18 summarizes the findings from

the four studs referred to for removals at two of the stages.

Table 18. Removal rates of bulls in Al.

 

 

 

. . Waiting % of permanent _
Stud Tfiit:2§e§§r1°d Period stud replaced *
° (% saved) annually
A 99.5 703 10-15
B 95 b 15-30
C 100. 80 15-20 a

 

 

aBased on paper presented to ADSA, 1970, by Walton.
bStud has no waiting period.

Sires in active service require a different approach.
Ten to 33 percent must be replaced each year with responses
for the individual studs (Table 18) differing, especially in
their upper limits. For purposes of this study, based on
New Zealand data (1970), current generation intervals,
Becker and Arnold (1953), and responses to the current sur-
vey, four years will be assumed to be the average useful

life span of tested bulls in AI.

Sire Utilization

 

Norman, et a1. (1970) reported on "utilization of
AI bulls in the United States in 1968." The report was

based on data supplied by 26 organizations on the source

47

of semen for 5,994,583 inseminations or 93 percent of all
dairy to dairy services to AI bulls in the United States
in 1968.

Number of bulls and the total of inseminations are
given for the Holstein breed by type of service in Table
19, adapted from Norman, §£_a1. The further breakdown to

inseminations by type of service and usage levels helps

clarify the intensity with which bulls are used. It is
noteworthy that a mere four percent of the inseminations

were in the progeny test category.

 

Table 19. Holstein bulls and inseminations by type of service.

 

 

Bulls Services*

 

Special Progeny All Regular Special Progeny All

 

Regular Mating Test Mating Test
Number
628 72 354 1,054 4,781,400 53,352 181,493 5,016,245
Percent
60 7 34 100 95 1 4 100

 

*Based on the average number of ampules required per
breeding.

USDA also reported usage levels within the various
categories. Only three percent of the regularvservice sires
are being used at nearnoptimum levels. THE USDA reported that
one Holstein bull had 61,835 services while the average for
Holsteins in regular service averaged 7,614 services per

bull.

48

Norman, et_al. (1970), also reported that 1967 New
Zealand data showed ten young Jersey bulls which were used
as extensively as possible averaging 43,188 inseminations,
the maximum for any one bull being 59,451 inseminations.

The USDA reported that elimination of the lowest

43 percent of the bulls active in AI (based on production)

 

would have been possible at 10,000 inseminations per regular :?
service bull. Since this figure is for all breeds, it is I
biased upward (Holstein bulls had higher usage levels than

bulls of the minor breeds) but is, non«the«less indicative ii I

of the genetic importance of usage levels. Mention should
be made at this point of the possible implications of
inbreeding when fewer sires are used. It is conceivable
that sampling more young bulls, using more pedigree selected
young bulls to sire herd replacements and limited (e.g.,
50,000 inseminations per bull) use of bulls in the latter
group as well as importation of semen will eliminate any

serious threat of extensive inbreeding.

’Minimum NUmber of Tested Bulls Needed

 

Realization of the largest possible selection dif-
ferential is contingent on the use of a minimum number of
tested bulls. This assumes that the bulls are accurately
ranked and that the best bulls are selected. Given that
the Michigan Holstein population consists of 425,700 cows,
that 60 percent of the cows are bred artificially, that

1.7 services are required per conception, and that 20 percent

49

of the cows are bred to young, untested bulls (Legates,
1971) the minimum number of services needed is 374,371.

If tested bulls can be expected to provide semen for

60,000 inseminations (Kucker, 1972, personal communication),
then 6.24 or seven bulls are needed.

According to USDA estimates of AI first services to
dairy animals (279,766 dairy x .90 Holstein = 251,789) the
above calculation overestimates the number of services
needed (251,789 x .80 x 1.7 = 342,433) but seven bulls
more than satisfy the semen requirements at this level.
Utilization of AI bulls in the United States will be dis-
cussed in more detail in a subsequent section.

Based on a Holstein population of 425,700, 60 per-
cent AI, and 7,614 services per Holstein bull in active
service the number of Holstein bulls currently needed in
Michigan is estimated at 34. Given that 25 percent of the
bulls in the active stud are replaced annually, nine new
bulls must be brought into the stud each year. To save
50, 20, and ten percent of those tested, progeny test
programs should include 23, 57, and 113 bulls, respectively.

Present Progeny Testing Situation
in Michigan

 

 

Van Vleck (1963) in a paper presented at the ADSA
annual meeting, referred to "fixed parameters" and "other
elements of the stud operation which can be adjusted."

Fixed parameters include population size, testing situation,

 

50

usage rate, and mortality rate. Three other elements which
he referred to as adjustable are:
1. The fraction of first services to young, untested
sires.
2. The number of first services to each young sire
before he is removed from service to undergo a
waiting period. ,

3. The number of young sires to sample each year. I
In this section and the following emphasis will be placed
on the three elements listed above.

Included in the survey of four bull studs operating
in Michigan were questions related to their progeny testing
activities in Michigan. Because of testing program differv
ences and differences in the responses to the survey ques-
tionnaire, estimates of the level of progeny testing activity
for two of the bull studs included in the survey are based
on first services to dairy in Michigan and progeny testing
levels of the other studs in the survey.

The desired parameter is the number of bulls cure
rently being progeny tested in Michigan annually. The
fraction of first services to young, untested sires and the
number of first services to each young sire before he is
removed from service to undergo a waiting period must be
determined. The number of cows and the number of herds
involved in the progeny testing program are also of interest.

ReSults of the survey are summarized in Table 20.

51

Table 20. Progeny testing parameters for Michigan.

 

 

r

 

 

f

Young Sires

 

Stud No. Cows No. Herds

 

 

% First Services ggéegf
Serv1ces Per Slre Sampled
A 748a 25 2.63C 500 1.5
B 3364a 172 2.47 500 6.73
c 704b -- 2.00 500 1.4
D 284b ~- 2.00 500 .56
Other 536 -- 2.00 500d 1.07
Total 11.26

 

aFirst services based on proportion of cows bred to
young sires.

Based on two percent of first services.

CAssumes 90 percent of first services are to Holsteins.

dAssumes same model as used for studs A and B.

Thompson and Freeman (1971) used three percent to
represent the fraction of first services resulting in produc-
tion tested daughters (fraction of population tested = 0.15,
times fraction of calves which are heifers = 0.50, times
fraction of heifer calves that make at least one record =
0.60). According to stud A, 40 daughters (eight percent)
result from distribution of 500 ampules of semen. Stud
B reported one two—year daughter record per eight
ampules of semen used which is twelve percent as

compared to the eight percent reported by stud A.

yo-
5'

52

Progeny Testing Possible

 

In a preceding section ”effective" population was
defined as cows on test identified by sire. Based on a
Holstein population of 425,700, eighteen percent on official
test, 56.8 percent identified by sire, 60 percent AI, and
500 first services per young bull; 52 young sires could be
tested annually if all cows were bred to young bulls. This, I
of course, leaves a large portion of the AI population to be 7

bred to progeny tested bulls. One point of practicality is I

 

that at least some progeny tested bulls should ultimately

be used in the "effective" population. The minimum reason—
able use of tested bulls in the effective population would
be for special mating purposes. Currently eleven bulls are
being tested which represents 20 percent of the "effective"

population.

Suggested Improvement

 

From the preceding section it appears that Legates
expectations are being met in Michigan; however, this 20
percent represents slightly more than two percent of the
artificially inseminated Holstein population.

It is not inappropriate to estimate the number of
young bulls which could be sampled if: (a) testing included
100 percent of Michigan Holsteins, and (b) all cows were
correctly identified by sire. Based on the present level

of identification by sire, if all cows were on official

53

test, 58 bulls could be tested. Using the current testing
situation, if all cows were correctly identified by sire,
approximately 18, 36, 55, and 73 young sires could be
sampled breeding 20, 40, 60, and 80 percent, respectively,
of the tested population to young sires. Still a third

situation to consider is that of all cows on test and

I."

identified by sire. This would permit sampling of 122, 204,

..h ‘0...- vm H

and 306 young bulls breeding 20, 40, and 60 percent of the
population to young bulls. A brief computer program was

developed. It computed the number of tested bulls needed :.

 

varying; (1) percent AI, (2) percent bred to proven sires,
(3) services per sire. Based on a constant annual removal
rate of 25 percent, it calculated the number of replacements
needed. The total number of bulls to be tested was then
computed based on a constant 80 percent remaining after the
waiting period and varied intensities of selection.

In Table 21 the number of young bulls to be tested
is computed for two levels of percent services to tested
bulls (60, 80) for varying intensities of selection. It
should be noted that for low intensities of selection too
few young bulls are being tested to breed the portion of
the population not bred to tested bulls (40 percent and 20
percent) at current usage levels for young sires. At
least two approaches or a combination of them could be
used to overcome this deficit. Increasing the number of

services per young, untested bull would insure greater

. II.I I E I-
'm}. J
1 o . I. .. . I"
. L I I) n

.ooHHom mcwpfimz may o>H>H5m wow moecmm<n

.oump m>oEoH macaw . a co omwm
a m

 

54

 

ONH Ho. N o ooo.om ow ow
NH H. N s coo.om ow cm
H m. N s ooo.cm ON cm
H N. N o ooo.om ow ow
N o.H N s ooo.om ow ow
Nmo Ho. s HN coo.OH cm as
so H. s HN coo.OH cm as
NH m. s HN coo.OH cm as
OH H. s HN coo.OH om so
A o.H s HN coo.OH cm as
mas Ho. H OH coo.OH so so
we H. H OH coo.OH so as
OH H. H SH coo.OH as as
N N. N SH coo.OH so co
m o.H H OH coo.OH so so
phon—ESZ GOHHUNHP whmInEBZ HOQEDZ Hmész uCQUHmm HQOUHOQ
wwwmww HUOHHDUOH wmwmmfi Ohwm mHHSD H< kn fimhfi
moHHm Hufimcoucfi mHH5o mHHsn Hon powwow coaumasmom
mcnm> cowuumHmm mc30> @mpmmh mGUH>H®m >0. Umhm mo nowuhom

.mHouoEmem Hmao>om mcHAHm> powwow on ow moHHm Mano» .HN oHomH

55

reliability of early USDA estimates of his genetic ability.
By increasing the selection intensity more young bulls
would be evaluated insuring selection of the very best
replacements available for the permanent stud. If one
combines these two factors with findings reported by Thomp-
son and Freeman (1971) and Hunt (1972), that a greater por~
tion of the AI population should be bred to young, untested
bulls (partly because of reduced generation intervals), then
the following conclusions seem appropriate:
1. More young bulls should be tested
a. to increase selection intensity
b. to reduce generation intervals
2. The. number of services per young bull should be
increased
a. to increase reliability of proofs
b. to reduce generation intervals
3. More cows should be on official test
a. to permit testing of more bulls
4. More of the cows on test should be identified
by sire
a. to be useful in sire evaluation
5. Increase fertility
a. less services to high PD bulls are wasted
repeats
6. Increase the use of AI
a. more bulls can be tested

b. to achieve more uniform herd production levels

 

 

7.

56

Select sires more intensely

a.

to limit the use of low PD bulls

 

SUMMARY .

A study was conducted of the management factors which
influence genetic progress and their current status in the
Michigan Holstein population. The sources of primary data
were pedigree records of 312 young bulls entering three
studs between 1969 and 1971 and a survey questionnaire of
four bull studs (75 percent response).

Secondary data was obtained from the following
sources:

1. In progress DHIA records.

2. USDA sire summaries (April 19675May 1971).

3. Holstein bull books.

4. Herdsbooks of the Holstein breed.

Results from research by Spike (1972), Erickson,
‘§£_al. (1972), Erickson (1972), Speicher (1972), and Oxender
(1972) represent still a third source of data.

On February 15, 1972 Michigan had 425,700 Holstein
cows, averaging 13,250 pounds of milk per year. A profile
of the average Holstein cow was developed.

1. She was 52 months of age at last calving.

2. It had been 395 days since she calved the last
time if natural service was used or 391 days if artificially

inseminated.

57

.7227

 

10.

58

Fourteen percent of her heifer calves will die

before six months of age.

Seventy percent of her heifer calves born are

available as herd replacements.

There is a thirty percent chance that she will

not be around to have another calf.

She was sired by one of 6000 possible sires.

Of these sires 1158 have been summarized by

USDA with a weighted (by No. of daughters) PD

for milk of +145 pounds.

a. Six hundred and twenty four of these bulls
are in A1 with an average PD of +168 pounds.

b. Five hundred and thirty four are natural
service sires with an average PD of +24
pounds.

If the cow was sired naturally, she was among

1785 cows whose sires were plus proven; 10,864

cows were sired by plus proven AI bulls.

She will be bred to AI bulls by:

 

 

Stud ‘Ave. PD No. Bulls
A 609 49
B 725 51
C 348 43
D 311 .32
Overall Ave. 556 Total 169

The average PD milk for sires of cows and sires

of calves in herds with high levels (17,694) of

milk production were +374 and +808, respectively.

mi- _-r
. _. raurfi’

 

11.

12.

13.

14.

15.

16.

17.

59

The average age of the bulls that she will be
bred by is nine years.

AI units serving Michigan will sample approxi»
mately 175 Holstein bulls this year.

Enough Michigan cows will be bred to sample
eleven bulls.

These young bulls will be from special mating
with sires that are eleven years of age and
dams that are eight years old.

The expected daughter superiority from pedigree
evaluation will be approximately +1000 pounds
of milk.

Eighteen percent of Michigan Holstens are on
official test (DHIAsDHIR).

56.8 percent of the cows on test are identified

by sire.

Improvement of these management factors would increase the

effectiveness of selection:

1.
2.

Reduce calving interval.
Increase fertility.

Reduce mastitis.

Reduce calf losses.

Increase DHIA.

Increase cow identification.
Increase AI usage.

Increase use of high PD bulls.

Test more bulls per year.

6O

10. Save a smaller percent of tested bulls.
11. Increase breeding efficiency.

12. Reduce generation interval.

H '41!

 

 

Literature Cited

LITERATURE CITED

Allaire, F. R., and C. R. Henderson. 1966. Selection
practiced among dairy cows: I. Single lactation
traits. Jour. Dairy Sci. 49:1426.

1966. Selection practiced among dairy cows:
II. Total production over a sequency of lactations.
Jour. Dairy Sci. 49:1435.

1967. Selection practiced among dairy cows.
III. Type appraisal and lactation traits. Jour.
Dairy Sci. 50:194.

Andrus, D. F., A. E. Freeman, and B. R. Eastwood. 1970.
Age distribution and herd life expectancy in Iowa
dairy herds. Jour. Dairy Sci. 53:764.

Arave, C. W., and R. C. Laben. 1963. Study of genetic
progress in California dairy herds. Jour. Dairy
Sci. 46:629.

Asdell, S. A. 1951. Variations in amount of culling in
DHIA herds. Jour. Dairy Sci. 34:529.

Batra, T. R., E. B. Burnside, and M. G. Freeman. 1971.
Canadian dairy cow disposals: II. Effective herd
size and production level of dairy cow disposal
patterns. Canadian Journal of Animal Science.
51:85.

Becker, R. B., and P. T. D. Arnold. 1953. Tenure and
turnover of desirable dairy bulls in artificial
studs. Jour. Dairy Sci. 36:575.

1957. Tenure of dairy bulls in artificial
service. Jour. Dairy Sci. 40:622.

Burnside, E. B., and J. E. Legates. 1967. Estimation of
genetic trends in dairy cattle populations. Journ.
Dairy Sci. 50:1448.

Burnside, E. B., S. B. Kowalchuk, D. B. Lambroughton, and
N. M. McLeod. 1971. Canadian dairy cow disposals:
I. Differences between breeds, lactation numbers,
and seasons. Canadian Journal of Animal Science.
51:75.

61

 

62

Deaton, O. W. 1964. Weighting information from relatives
to select for milk in Holsteins. Unpublished Ph.D.
dissertation. Michigan State University Library,
East Lansing, Michigan.

Dickerson, G. E. and L. N. Hazel. 1944. Effectiveness
of selection on progeny performance as a supplement
to earlier culling in livestock. J. Agr. Research.
69:459.

Dillon, W. M., Jr., W. W. Yapp, and R. W. Touchberry. 1955.
Estimating changes in the environment and real pro-
ducing ability in a Holstein herd from 1901 through
1954. Jour. Dairy Sci. 38:616.

Erickson, RH,E. Wright, C. Meadows, and P. Spike. 1972.
Sire selection by Michigan Dairymen. (Abstract)
Jour. Dairy Sci. 55:681.

Erickson, R. 1972. Analysis of high and average milk
production dairy farms. Unpublished Ph.D. disser+
tation. Michigan State University Library. East
Lansing, Michigan.

Fosgate, O. T. 1965. Rate, age, and criteria for culling
in a herd of registered dairy cattle over a 16
year period. Jour. Dairy Sci. 48:795.

Freeman, M. G. 1970. What has been realized from pedigree
selection of dairy bulls? Invitational paper,
ADSA.

Gaunt, Neimann, and Sorense. 1970. Relationship between
dairy sire proofs made at Danish progeny testing
stations and in field tests. Jour. Dairy Sci.
53:656.

Hunt, M. S., E. B. Burnside, M. G. Freeman, and J. W. Wilton.
1972. Comparison of artificial insemination pro«
grams using young bulls extensively with programs
using progeny tested bulls extensively. Paper pre-
sented at ADSA.

Johnson, R. V., and G. W. Harpstead. 1970. Calf losses
in Illinois dairy herd improvement herds. Jour.
Dairy Sci. 53:684.

King, G. J., B. T. McDaniel, and F. N. Dickinson. 1971.
Dairy herd improvement letter. Sept.-Oct., 1971.
ARS, USDA. Vol. 47, No. 8.

63

King, G. J., F. N. Dickinson, B. T. McDaniel, and H. D.
Norman. 1972. Distribution of active artificial
insemination (AI) sires summarized in the USDA+DHIA
sire summary by level of predicted difference for
milk and fat production of daughters. Dairy Herd
Improvement Letter. March, 1972. ARS, USDA. Vol.
48, No. 2.

King, G. J., B. T. McDaniel, F. N. Dickinson, C. A.
Rampendahl, J. J. Corbin, and A. H. Kienast. 1971.
Artificial breeding report. Dairy herd improvement
letter. June«July, 1971. ARS, USDA. Vol. 47, a
No. 6. R

King, G. J., B. T. McDaniel, F. N. Dickinson, C. A.
Rampendahl, V. H. Lytton, L. G. Waite, and J. J.
Corbin. 1970. Editing of lactation records used
in USDAaDHIA genetic evaluations of sires and cows. 3
Dairy herd improvement letter. December, 1970. Q
ARS, USDA. Vol. 46, No. 6. .

 

McDaniel, B. 1968. Production increases possible through
use of USDA—DHIA sire summaries. Dairy herd improve-
ment. AHRD, ARS, USDA. Beltsville, Maryland.

Meadows, C. E. 1968. Importance of traits other than
milk production in a breeding program. Jour. Dairy
Sci. 51:314.

Michigan Department of Agriculture. 1972. "Michigan
Agricultural Statistics," p. 24.

Nanson, A. 1967. Tables de la différentiele de sélection
dans la distribution normale. (0, 1).

New Zealand Dairy Board. 1970-'71. 47th farm production
report.

Norman, H. D., B. T. McDaniel, and F. N. Dickinson. 1970.
Utilization of artificial insemination bulls in the
United States in 1968. Dairy herd improvement
letter. April, 1970. ARS, USDA. Vol. 46, No. l.

Oxender, W. D. 1972. Personal communication.

Oxender, W.D., L. E. Newman, D. A. Morrow. 1972. Dairy
calf mortality on Michigan farms. Department of
large animal surgery and medicine. Michigan State
University, East Lansing, Michigan.

Palmer, J. B., C. J. Wilcox, F. G. Martin, O. G. Verde,
R. E. Barrantes. 1972. Genetic trends in milk
production in an experiment station Jersey herd.
Jour. Dairy Sci. 55:631.

64

Parker, J. B., N. D. Bayley, M. H. Forman, and R. D.
Plowman. 1960. Factors influencing dairy cattle
longevity. Jour. Dairy Sci. 43:401.

Pearson, K. 1931. Tables for statisticians and biometv
ricians: Part II. Cambridge University Press.

Pelissier, C. L. 1970. Factors contributing to low
breeding efficiency in dairy herds. Jour. Dairy
Sci. 53:684.

Penn State University. 1968. Dairy Reference Manual.
College of Agriculture, The Pennsylvania State
University. University Park, Pennsylvania.

Plowman, R. D. and B. T. McDaniel. 1968. Changes in
USDA sire summary procedures. Jour. Dairy Sci.
51:306.

Rendel, J. M., and A. Robertson. 1950. Estimation of
genetic gain in milk yield by selection in a
closed herd of dairy cattle. Jour. Genetics.
50:1.

Robertson, A., and Rendel, J. M. 1950. The use of progeny
testing with artificial insemination in dairy
cattle. Jour. Genetics. 50:21.

Roman, J., C. J. Wilcox, R. B. Becker, and M. Kroger.
1969. Tenure and reasons for disposal of artificial
insemination dairy sires. Jour. Dairy Sci. 52:1063.

Seath, D. M. 1940. The intensity and kind of selection
actually practiced in dairy herds. Jour. Dairy
Sci. 23:931.

Skjervold, H., and Langholz, H. J. 1963. Factors affecting
the optimum structure of AI breeding in dairy cattle.
Z. Tierz. Zuchtungsbiologie. 80:310.

Specht, L. W. 1957. Rates of improvement by progeny
testing in dairy herds of various sizes. Unpublished
Ph.D. dissertation, Michigan State University. East
Lansing, Michigan.

Specht, L. W., and L. D. McGilliard. 1960. Rates of im-

provement by progeny testing in dairy herds of
various sizes. Jour. Dairy Sci. 43:63.

Speicher, J. A. 1972. Personal communication.

v--

 

65

Spike, P. L. 1972.
dairy herds.
State University.

Thompson, G. M., and A. B. Freeman. 1971.

Calving interval trends in Michigan
Unpublished M.S. thesis.
East Lansing, Michigan.

Michigan

Expected

genetic progress in artificial breeding (draft

copy). Dept. Animal Science.

Sweden. 1968 annual report.

Animal Breeding Abstracts, 1969.

Misidentification
53:1697.

Van Vleck, L. D. 1970.
tion. Jour. Dairy Sci.

Van Vleck, L. D., and H. W. Carter. 1971.

Ames,

Iowa.

Svensk Husdjursskdtsel, Inc.
3288.

and sire evalua-

Comparison of

estimated daughter superiority from pedigree

records with daughter evaluation.

Jour. Dairy

Sci.

Van Vleck, L.

retrospect of selected young sires.

Sci.

55:214.

 

Relative selection efficiency in 5
Jour. Dairy

D. 1969.

52:768.

1970. Misidentification is estimating the
paternal sib correlations. Jour. Dairy Sci.
53:1469.

W. B., and A. B. Freeman. 1972. Rank correla-
tions between pedigree estimates of breeding
value and subsequent progeny test results of AI
Holstein bulls. Paper presented at ADSA.

Vinson,

R. E. 1970.
in symposium:
genetic progress.

Walton, A U.S. industry viewpoint. Presented
Designing AI programs for optimum

Invitational paper ADSA.

White, J. M., and J. R. Nichols. 1965. Relationships
between first lactation, later performance, and
length of herd life in Holstein Friesian cattle.
Jour. Dairy Sci. 48:468.

Reasons for dis-
Jour.

and J. R. Nichols. 1965.
of Pennsylvania Holstein cattle.
Sci. 48:512.

White, J. M.,
posal
Dairy

APPENDIX

 

 

QUESTIONNAIRE

Proportion of Holstein population in Michigan which
your organization breeds?

How many young sires do you sample in Michigan?
a. No. Holstein cows in young sire testing program
in Michigan.

b. Fraction of first services to young untested sires
in Michigan. ‘

c. Number of first services to each young sire before
he is removed from service to go into waiting.

d. Fraction of tested daughters resulting from those
first services; and/or the fraction of first ser-
vices resulting in production tested daughters.

Usage rate of sires available?

a. Number of sires in the permanent stud.
b. Number of services possible per sire.

Mortality rate?

a. How many of the sires in the permanent stud must
be replaced each year?

b. How many of the young bulls will remain four to
five years after their test matings.

How many (what fraction) of the original young bulls
to be tested will survive the testing period? Survive
the waiting period?

Age of young sires at the end of the testing period?

Age of young sires saved at end of waiting period?

66

 

II.

III.

IV.

COMPUTATIONAL FORMULAS

Minimum tested bulls needed.

N x A x P

 

Tm = S where:

Young bulls needed.

Y1 = T x D where:
m

Young bulls to test.

Y1
x I where:
s

Progeny testing possible.

T = N x A x To x IS

p S
y where:

 

67

20') "U>I-]

Cit-3%

min. tested bulls need
fraction AI

fraction first services
to proven sires
services per sire
population size

young bulls needed

see above

fraction of permanent
stud replaced annually

young bulls to test
young bulls which
survive waiting
fraction of bulls
saved of those tested

testing possible
population size
fraction AI

fraction on official
test

fraction identified
by sire

services per young
sire

 

 

l