THE PREDICTABILITY OP THE BUTTERFAT T RAN S KITT11'-GAB ILITY OP DAIRY BULLS

By
JAMES PREDERIGK SMITHCORS

A THESIS
Submitted to the School of Graduate Studies of Michigan
State College of Agriculture a nd A pplied science
in partial fulfillment of the requirements
for the degree of
DOCTOR OP PHILOSOPHY

Department of Animal Husbandry

1951

TABLE OP

CONTENTS

INTRODUCTION

1

R EVIEW OF LITERATURE
I.
II.

The heritability of butt; e r f a t

production

The prediction of t r a n s m i t t i n g

III. Environmental influences
production

ability

upon m i l k

II.

6

and butterfat
11

PREDICTION OF BUTTERFAT T R A N S M I T T I N G A B I L I T Y
IN THE MICHIGAN I N S T I T U T I O N H E R D S
I.

4

OF BULLS

Source of data

20

Methods

22

RESULTS
I. Correlation of average d e L x n - d a u g n t e r p r o d u c t i o n
II. Per cent chance of p r e d i c t i n g
certain limits

26

p r o d u c t i o n within
27

III. Correlation coefficients
oetween dam-daughter
production at v a r y i n g l e v e l s o f p r o d u c t i o n

29

IV. Comparison of rankings of* t r a n s m i t t i n g a b i l i t y
w i t h actual daughter a v e r a g e

30

DISCUSSI ON
I.
II.
III.
IV.

Correlation analysis of

production

data

35

Diallel crossing

38

Environmental indices

42

Worksheets

45

SU MMA RY

47

REFERENCES

*9

APPENDIX

61

VITA
Tile au tho r was

born Ma r c h 27, 1920 at Glassboro, New

Jersey an d spent his early life in a rvral background.

he

received the

B.S. degree from Rutgers University in 1941,

ing In dairy

husoandry.

major­

The *:.S. degree was ob tained at

Cornell University in 1943 w h i l e an assistant in animal breeding.
The author then entered the New Y o r k State Veterina ry College
and received

the D.V.M. degree In 1945.

One yea r of this

time

was spent as

a m emb er of the Arm y veterinary training program*

In Nov ember 1945 the author was app oi n t e d Assista nt Pro fessor
of An a t o m y at Mi chi gan State College whe re he entered the Graduate
School in 1945.

He is the author or co- aut hor of two laboratory

textbooks and of four scientific articles in the field of
endocrinology.

He is currently employed as Ass oc i a t e Professor

of A n a t o m y at M i c h i g a n State College.

INTRODUCTION
On e o f tn e m o s t

important

problems

f a c i n g tiie d a i r y m a n

is tie p r e d i c t i o n o f w h i c h o f s e v e r a l p r o s p e c t i v e

sires

is

most l i k e l y to t r a n s m i t tiie ti.igh.est l e v e l o f b u t t e r f a t

produc­

tion*

from

That no a c c u r a t e m e a n s

the m e a g e r p r o g r e s s
pound a year,

in i m p r o v e m e n t

o f the p r o g r e s s

improvement

of p r o d u c t i o n ,

two d e c a d e s or longer.

that has

in environment.

P e w o f t n e s t u d i e s m a d e u p o n the
have

oeen s u c c e s s f u l

tinguishing between hereditary and environmental

dicting

S t i l l f e w e r o f the t e c h n i q u e s
Dutterfat

in d i s ­

factors

In­

in c o m m o n u s e f o r p r e ­

p r o d u c t i o n r e c o g n i z e the

ment other than correcting records
age,

A large po r ­

seen m a d e c a n b e a t t r i b u t e d to

inheritance of producing ability

volved.

about one

as i n d i c a t e d b y h e r d t e s t i n g d a t a w h i c h h w e

b e e n c o m p i l e d i n the last
ti on

is a v a i l a o l e is m a n i f e s t

e f f e c t s of e n v i r o n ­

for s t a n d a r d c o n d i t i o n s o f

times m i l k e d d a i l y a n d l e n g t h o f l a c t a t i o n .
It Is g e n e r a l l y r e c o g n i z e d that a n i n c r e a s e

in the p r o ­

d u c t i o n of a h e r d c a n be a c h i e v e d by a l t e r i n g o n e o r b o t h of
two fa ct o r s ,

i.e., b y i m p r o v i n g

e i t h e r th e h e r d e n v i r o n m e n t

or the h e r e d i t y o f the i n d i v i d u a l a n i m a l s .
former m a y
ments

be i n m a n y cases,

E f f e c t i v e as

the

t h e r e c a n be no d o u b t that i m p r o v -

i n e n v i r o n m e n t a r e t e m p o r a r y a n d do n o t a d d to the g e n e t i c

w o r t h o f the a n i m a l s .

T h e g e n e t i c v/orth o f a n a n i m a l

q u e n t l y o b s c u r e d b y e n v i r o n m e n t a l f act ors .

^ H I R G r e e n B o o k 1949 H o l s t e i n F r i e s i a n A s s o c i a t i o n

is f r e ­

2

Improvement in genetic worth,

other than tne doubtful

possibility of seizing upon worthwhile mutations,
upon selection to change gene frequency.

is dependent

The effectiveness of

any m eth od of selection for a character wnich can oe identified
depends primarily upon the variability present and the selection
differential.

The genetic variability present and the selec­

tion differential.

The genetic variability of males and fe­

males in dairy herds is probably similar,
environment.

but is ooscured oy

The ratio of breeding males to females, already

wide, is oeing increased by the widespread use of artificial
insemination.

Thus while tne greatest use of the selection d i f f e r ­

ential can oe made wi th males,

it i 3 also important that t h e best

possible use of this oe made.

Inter-herd variations in en­

vironment make it essential to d istinguish between genetic and
environmental factors if real progress is to oe made.

Culling

of the lowest producing cows is a useful technique regardless
of the level of herd production even though the selection d i f f e r ­
ential available in most herds makes progress
painfully slow.

by tnis means

Improvement in management to enable a greater

rate of culling of cows would increase the value of tnis technique.
The object of the present study is to consider the merits
of several techniques in use or proposed for predicting the
production of the daughters of a potential sire.

An attempt

will be made to analyze the utility of these methods and to
suggest possible improvements in tneir application.

It is obvious

that the most valuable technique which can be used in the selec­
tion of a breeding animal is that which

na3

the highest predic­

3

tive value and. at tne same time is capaole o f widespread,
application.

E v e n if those methods w h i c h impose the greatest

n umber o f r e s t ri cti ons u p o n the ani mal in question,

in terms

of individu al or fa mily merit, w e r e the mos t accurate,

tneir

utility w o u l d oe lim i t e d to the n u m b e r of animals w h i c h could
meet

the sp ec ifi cat ion s required.

except fr om a n academic
superior animals

In o t n e r words,

standpoint,

It is useless,

to d i s c o v e r a n u m o e r of

if the m e t h o d requires that they oe d e a d before

their w o r t h can be e 3 taolished.

Emphasis in the present w o r k

will oe d i r e c t e d toward tne a p p l i c a b i l i t y of

the several

methods c o n s i d e r e d as w e l l as t h e i r predictability.

RE VIE W OP L ITE RAT URE
I. The he rit abi lit y of outterfat production
The literature concerning p red ictability o f m i l k and b u t t e r ­
fat production can a p p r opr iat ely be div i d e d into two major parts:
one de ali ng primarily w i t h genetic,
factors.

the other w i t h environmental

W h i l e many reports ha ve dealt w i t h the influence o f

various environmental factors u p o n production, m u c h more a t t e n ­
tion has oeen focussed up on genetic factors in actually p r e ­
dicting production.

Ma ny of tne papers dealing w i t h correlations

between the production records of related animals recognize en­
vironmental factors only to tne extent of using the well estab­
lished correcti on factors for age, n umber of times milked,

and

length o f lactation to convert individual records to a mature
equivalent

basis.

In most of these studies the comparison of

the records of individual animals has

d

een used.

These data

form the basis for the present study w n i c h concerns the c orr ela ­
tion of rel ated groups of animals, but because of tne difference in
average aeritabiiity of production of groups as co mpared w i t h
individuals,

these papers will not oe revieived separately.

Rather, at ten tio n is directed to the following authors whose
work can oe considei',ed representative of this po rtion of tne
literature:

Gowen,

Uooeland, Gifford, Lush,

Turner.

Excellent

reviews of this w ork have bee n presented by Johansson and H&nsson (1940), L u s h and Straus
Eldridge

(1948).

(1942),

Shrode and L u s h (1947) and

Depending upon the sample studied the correla-

tion values oetween the production of various relatives ranged
from insignificance to 0.3-0.4 for close relatives w i t h only few
being higher than the latter value.

While many of these values

were significant at the 1% level of pro bab ili ty, they were for
the most part not sufficiently high to do especially useful for
prediction of production under the conditions studied.

In most

of these studies the effect of the common herd environment which
usually exists in the case of dams and their daughters in parti­
cular was not ruled out.

Lush (1945) has pointed out that there

is a correlation of about 0 .2-0.3 between cows in the same herd
regardless of relationship.
The effect of averaging makes the correlation between the
production of groups of related animals considerably higher than
the values determined for individual comparisons.

Thus the dam-

daughter correlation for the daughter average of a sire was
reported to oe on the order of 0.60 for milk

(Rice 1944) and

0.63-0.67 for butterfat (Lush and Schultz 1938, Eldridge 1948).
These values are high enough to have practical significance,
and in addition,

it is these values in which the dairyman is

primarily interested.

That is, the dairy cattle breeder is

more concerned w i t h the average production of a group of
daughters of a sire than wit h that for any individual daughter.
This may not be quite so true for conditions under artificial
insemination where a number of bulls may oe used to inseminate
cows in small herds w ith the result that some nerds ma;/ con­
tain a few single daughters of one or more culls.

This would

6

a p p e a r to be a s e p a r a t e p r o b l e m w o r t h y o f c o n s i d e r a t i o n .

In

one study r e l a t i n g to this p r o b l e m the a v e r a g e p r o d u c t i o n of
a r t i f i c i a l l y s i r e d d a u g h t e r s o f 32 bulls d i d not d i f f e r f r o m
the a v e r a g e o f n a t u r a l s e r v i c e d a u g ht ers o f tne same sires
re gar d l e s s of the le vel of p r o o f o f the sire
II.

(Waahbon 1950).

The p r e d i c t i o n o f t r a n s m i t t i n g a b i l i t y
The o b s e r v e d c o r r e l a t i o n b e t w e e n the p r o d u c t i o n o f r e l a t e d

an i m a l s has le d to the f o r m u l a t i o n of a n u m b e r o f te c h n i q u e s for
p r e d i c t i n g the fu t u r e p r o d u c t i o n of the d a u g h t e r s of a sire
t h r o u g h the f o r m u l a t i o n o f sire indices.
of Hansson, Y a p p ,

Pearl,

Turner,

The p i o n e e r i n g w o r k

Gowen, G iff o r d ,

W r i g h t a n d o t h e r s has o e e n amply r e v i e w e d by L u s h
w o r k has

Heizer, G ood ale ,
(1933).

I'his

l e d to tne w i d e s p r e a d a d o p t i o n of the H e q u a l - p a r e n t ”

Index w h i c h is b a s e d u p o n the s u p p o s i t i o n that the sire a nd d a m
contribute

e q u a l l y to the p r o d u c i n g or t r a n s m i t t i n g a b i l i t y

of t h e i r of f s p r i n g .
Is a d d e d

The d a u g n t e r - d a m d i f f e r e n c e in p r o d u c t i o n

(algebraically)

in d e x o f t he

to the d a u g h t e r a v e r a g e to secure an

bull*s t r a n s m i t t i n g ability.

fu t u r e d a u g h t e r s

The production of

is p r e d i c t e d by a v e r a g i n g the o r i g i n a l pr oof

of the sire a n d the p r o d u c t i o n o f nis mates,

the only c onc ess ion s

to e n v i r o n m e n t b e i n g to u s e the s t a n d a r d c o r r e c t i o n fa ctors to
co nvert re cor ds

to a m a t u r e - e q u i v a l e n t

factors a re s u c h how ev e r ,

basis.

as to inake this

Oth er environ men tal

technique m o s t u s e f u l

on an i n t r a - h e r d basis.
The e x t e n s i v e p r o g r a m o f p rov ing
th eir p roofs

culls and o f re cor d i n g

that has d e v e l o p e d In the past two de cad es has

7

been an important advance in dairy cattle breeding.

However

the applications of this information that have oeen m ade have
not r esulted in any m a r k e d improvement in production.

Consequent­

ly a n u m b e r o f attempts to fin d an improved technique have been
made.

-*\hese have been di rected toward m aking the records used

in e stablishing the indices more truly representative of the
prooable transmitting aoility.
duction rec ord (phenotype)
her transmitting ability

It is recognized that the p r o ­

of a cow m a y not accurately reflect

(ge not y p e ) .

W i t h this in mind, Lu sh

(1945) has a dvo cated a formula based on the repeataoility
of individual records and the num oer of records

(r)

(N) constituting

the dam's average: prooable transmitting level«>Herd ave. +■
Nr
X- (N - 1) r *
(Cow ave. - Herd ave.) For an average repeatability
2N
3
of 0.4 this formula reduces to 2 N -5 X CoW's ave.
2N-3
Herd ave.

Breed average may be s ubstituted for herd average

wh en the latter is unknown.

Thus for N-l,

reliance on the dam's

record is reduced to 2/5, regressio n toward her d (breed)
accounting for the remainder.

average

The m eth od o f intra-sire regression

of off spr ing o n d a m has oeen advocated by Lush (1940) as the
me t h o d of choice for estimating he rit aui lit y of characteristics
because of the reduction in environmental error.
The phenomen on of regression toward population average
has long oeen estaolished, yet only recently has its use been
ad vocated in sire indexing.

Rice

(1944) has proposed a m o d i f i ­

cation of the equal-parent index in w h i c h the records of both
dama and daughters are averaged w i t h breed average.

Using this

8

regression index, Dickey and Lab art he

(1945) o bta ine d a correlation

of 0*61 bet w e e n actual and predicted product ion of the daughters
of 214 H ols tei n sires.

The simple equal-parent index gave a

correlation of 0*53 for the same data.

The authors found the

regression index to be most useful at h i g h levels and the equalparent index most use f u l at low levels of production.

Lush

(194d) has pointed out that since the regression index is the
equal-parent index regressed h a l f w a y toward breed average,

it

has h a l f the variability but the same accuracy as the equalparent index in d ete rmi nin g transmitting acidity.

Allen

(1944)

nas pr op ose d that twice the d evi ati on of the s i r e ’s daughters
from the expected production''"' be added

(algebraically)

breed a verage to obtain a comparative rating of sires.

to the
This

figure w o u l d represent the probable transmitting ability o f the
sire under average conditions of environment.
the future pro duc tio n of daughters

Por predicting

the in dex is not needed, the

deviation of the s i r e ’s previous daughters

being ad ded to the

expected p rod uct ion of future daughters.
Pr om an analysis of the pedigrees of 207 Ho lsc ein sires
proved in He w Y o r k State DHIA, Eldridge

(1948,1949)

developed

two m ultiple regression equations for predicting the production
of the daughters of a sire. Depending u p o n the information
available,

the equations were:

Ya = 2 9 . S + 0 . 7 5 X 1 + 0 . 0 3 X 2 * 0 . 0 1 X 3 +
Y b =.0.1 + 0.75X!
Ya , Y b ^

0.34X4 - 0.21x5

+ 0Q01X3 + 0.23X4 - 0.22x5 ^ 0.24X6

predicted ave. M.E. prod,

of daughters of the bull

" The expected production is bas ed u p o n experience taoles showing the
average production of daughters of dams of varying levels of p r o ­
duction under u n i f o r m testing conditions.

^ ave.
X 2 a ave.

prod, of the mates of
prod,

of the

the Dull

maternal half sisters of the bull

X3 "

w

,f n

n

dam of the oull

X^ -

11

”

11

”

paternal

X 5 3 11

n

”

’* dams of

_
II
Xg s
1
1

II

H

II
w

II
1
1

half sisters of the uull
4

, 1 ________ U J _______ „

daughters of maternal grandsire of uull

The multiple correlation coefficient between the actual and
predicted production of the daughters of these oulls was 0.70
for Y ^ and 0*69 for Yg.

Deletion of the mates'

production

(X]_)

from the equations resulted in a correlation of 0.37 in either
case.

It was concluded that the most important information

to consider in predicting the average production of the daughters
of a oull is the average of the mates,

and that the sire's index

is the most important consideration in selecting a uull on the
basis of pedigree.
The average production of the da m of tne bull and/or that
of he r daughters contriouted a negligible amount of information
for predicting the transmitting auility of the bull.
Hyatt

Tyler and

(1948^ have snown that one unselected record of a dam may

be as important in predicting production of her future daughters
as the average of her first 2 or 3 daughters in estimating p r o ­
duction of future daughters, but correlation values were of a
low order

(0.18-0.38 for individual comparisons).

Beardsley and coworkers

(1950) have suggested that a

curvilinear regression might account for a greater portion of
the variance in production inheritance studies than would

10

simple linear regression.
significant differ enc e

W h i l e their data d i d not show a

oetween linear and c u r v i lin ear regression

v a l u e s , the latter m e t h o d gave average h e r i t a bil ity figures of
45, 31, 23, an d 17$ respectively for av era ge production levels
of 350, 450, 530 an d 650 pounds of fat.

These compared w i t h

an average figure o f 2 7 $ for simple linear re gre s s i o n for all
data and 28 $ d e t e r m i n e d by Eldridge

(1948)

for nis data.

These re lat ive ly hi gh correlation values, on the order
of 0 .6 - 0 .7, bet w e e n average production of the daughters and
mates of a bull a ppe ar to be val id only for u n s e l e c t e d data.
Turner

(1927b)

and G i f f o r d

(1930)

found that by gro uping bulls

according to m a t e s ’ production, usi ng 100 pound fat intervals,
correlations of 0 . 2 5 and 0.20 respectively wer e o btained for
da m-d aug hte r averages.
20 $ of the

The l atter w ork indicated that about

pro duc tio n of a bull's mates above b r e e d average

was i nhe r i t e d (Holstein data).

Similar data for Jerseys gave

a value o f 15$ (Gifford and 'Turner 1928).
The emphasis placed by Eld ridge

(1948) u p o n the production

of the bull's mates was an t i c i p a t e d by Gowen, who in 1924a stated,
’’The rec ord of the d a m is one of the best pieces o f evidence
on w h i c h to base any estimate o f the probable produc tio n of the
daughter.

In fact the close re lat ion shi p between m i l k yield

of dau ght er and d am makes

it an open q uestion if for the small

breeder it is not better to purchase daughters from relatively
h i g h producing dams than it is to pay too m u c h at tention to
and too m u c h money for a supposedly high pr oducing b u l l . ”
Since m u c h of the c or rel atio n between the product ion of dams

11

and their daughters is undoubtedly due to intra-herd environment,
the term ’’purchase" should probably oe replaced by "select” .
III. Environmental influences upon milk and outterfat production
As mentioned above,

production records used in prediction

of milk or butterfat production are now corrected for the in­
fluence of age, number of times milked and length of lactation,
these influences being well estaolished in the effect they
exert upon lactation.

In spite of correction for these environ­

mental influences, correlations between actual and expected p ro­
duction still lack the accuracy desired.

This lac.-: of corres­

pondence is generally attributed to “ environmental" factors,
but there has been no concerted effort to determine constants
to correct for such errors.
The influence of body size upon milk production is recognized
in the axiom that "within the breed, the big cows excel."

In

a summary of the effect of body size on lactation, Beck and Turk
(1948) 3tated that while tne larger cows of a breed do produce
more mil :, they ruay not necessarily do so more economically than
smaller ones.

That age factors automatically correct in part

for size was shown by Brody and coworkers

(1923).

In over 15,000

Jersey ROM records with age constant there was an increase of
20 pounds of fat per 100 pounds increase in body weight.

If

age was not held constant, there was an increase of 104 pounds
of fat per 100 pounds increase in body weight.
records the increase was 16 pounds
cords— Turner 1929).

In Guernsey AR

( or 14 pounds for DHIA re­

The correlation of body weight and milk

12

production was found to De on the o rde r of 0.6 and was highest
at one m o n t h after calving
cords u n c o r rec ted for age.

(Davis an d coworkers 1943)

^'his value was reduced by one-half

if age cor rec tio n factors w e r e u s e d
coworkers

for r e­

(Gowen 1932^) •

Oaines and

(1940) have gone so far as to advocate correlating

production d i r e c t l y w i t h body weight without regard for age
using the formula PCIvl/body w ei g h t

(PCM- fat corrected milk).

This idea has been vigorously att ack ed by Kleiber and Mead

(1941,

1945) who m a i n t a i n that production is proportional to metabblic
body size ^Kff0.75)
Brody

in de ter min ing relative lactation capacity.

(1945)* poi nted out that Gaines*' formula w o u l d be valid

only f o r cows o f the same weight.
M i sne r

(1939,1941)

w i d t h of hips,

computed a size index for cattle using

l eng th of rump andheight at hips, and found a

correlation o f 0.54 w i t h PCM on 2747 Holsteins.

Measurements

of 100 cows in the Cornell h e r d gave a correlation of 0.49
between size index and the best age corrected record.
mature cows Davis and Wil lett
tween gain in weight,

(1938) found no correlation b e ­

increase in height or chest girth from

birth to two years of age w i t h subsequent production.
and coworkers

Por i m­

Tyler

(1948) found that 30-60;^ of the variation in

mature body size of Holsteins was hereditary and suggested
selection for body size be practiced.
The effect of season of calving upon milk yie ld has been
investigated by Cannon
cowori-ers

(1933),

, Prick and

(1947), Morrow and c©workers

and Sanders

(1927a).

(1943), Edwards

(1938),

Pr odu c t i o n was found to be uniformly

•M-Bioenergetics and Growth

n

y

15

lower for cows freshening in the summer months, but the season
of freshening h a d no effect upon length of lactation.

Pall fresh­

ening resulted In a maximum increase of 12-13$ in milk production
over summer fresnening.
and Jones

A n exception to this was found by Oloufa

(1948) under condition of uniformly mild temperature.

This indicates that temperature is a greater factor than the poor­
er late summer feeding conditions which generally prevail while
cows freshening in the summer are at peak production.

Sanders

(1927a) found the effect of season of calving to range frorn-5
to +7$ of the total yield and proposed that records be corrected
according to the month of calving.
relationship(r =• -0.87)
percent fat.

Bro oks (1931) found an inverse

between environmental temperature and

The same relationsnip existed in four breeds,

was most marked In those with higher tests.

but

The sts-ge of lactation

was less a factor than season.

Similar findings were reported

by Weaver and Matthews

Hays

(1928).

(1926) under controlled ex­

perimental conditions found an increase in fat test of 0.095$
for each 10°P lowering of temperature from 85 to 24°p.

thus

demonstrating that temperature is the chief factor in seasonal
variation in fat test.

Bartlett

(1929) found daily milk yield

to be most variable in hot weather.
The length of the dry period has been demonstrated to affect
milk production.

The optimum length of the dry period was de ­

termined to be 55 days
and coworkers 1945)

(Klein and Woodward 1943) or 65 days

for a 12 months calving interval.

(Morrow

This

interval was found optimum for maximum 305 day production.

The

former workers
day dry period,

suggest correction of records to a standard 55
the fators ranging from 1.403 for 0 d a y 3 to 0.955

for 120 days or longer.

They found h a l f the variation of successive

records o f cows to be explainable by the length o f the dry period.
Tyler and Hyatt

(1950)

found significantly lower produ cti on with

a calving interval of 10-11 months than for 12-13 months.
was due in part to the shorter lactation,

Tnis

but it was sxxggested

that cows w i t h short calving intervals may be p ersistently lower
producers.

Sig nif ic ant ly gre ate r production was not o bta ine d

wi th an increase in calving interval bey ond 13 months,
Gaines and Pal fre y

(1931)

although

found that the calving interval could

be extended to 18 months without seriously a f f e c t i n g total lifer

time yield.
for age,

D ick ers on

(1937)

reported that cor rec tin g records

c alv ing interval and dry period a c c o u n t e d for 35$ of

the intra-herd v ari a t i o n in producti on be twe en cows compared
with 24 $ for u n c o r r e c t e d records.

D i c k er son and C hap man

(1939)

found that increasing the length o f the dry period resulted
in greater percentage increases in the subsequent lactation period
for low p rod ucing than for h i g h producing cows.
of the dr y period was found to increase w i t h age.

The length
Di cke r s o n

(1940)

found that correct ing calving interval to 365 days increased the
repeatability of total yield.

Cows w i t h shorter lactations t e n d ­

ed to have lower records and longer dry periods.

Sanders

(1927a,

b, 1928a,b) d eve l o p e d correction factors for length of dry period
ranging from -14 to +25 $ (0 interval 40 days)
of service period ranging from -33 to +34$

a nd for length

(0 interval 85 days).

Hammond, and Sanders

(1923) found that variation in records of

individual cows could be reduced 20%,

by correction for age

(15%),

length of dry period (2%) and for month of calving and length
of service period

(3%).

Plum (1935) found that length of the

dry period a cco unted for 1% and calving interval 3% of the total
variance in fat production.
Turner

(1926) has suggested a persistency index for p red ic­

tion of total yield based on the fact that each succeeding months'
production after the peak yield is a constant percentage
with individuals)

of the preceding months production.

(differing

Per­

sistency was found to be largely independent of pregnancy,
temperature, nutrition and other management factors.

Alexander

(1950) in a study of the inheritance of persistency found that
the daughters of inbred sires varied less in average persistency
than those of outbred sires, and the inbred daughters of a sire
varied less than the outbred daughters of the same sire.
differences w e r e found between breeds,

Distinct

strains w ithin breeds and

between the daughters of different sires.

The regression of

daughters on dams was constant, although Turner

(1927a)

re port­

ed the correlation of dam-daughter persistency to oe of a low
order.

Gaines

(1S27A) reported persistency to

the dam and not the sire,

oe inherited through

but Becker and McGill iar d (1929)

found

contributions of both s ire and dam.
Putnam and coworkers

(1944) have reopened the question of

the effect of age u p o n butterfat percentage.

They refer to a

large number of studies w h i c h supports their finding that there

16

is a s tat istically significant d ecr e a s e in Dutterfat test with,
age, the regres sio n of fat test on age

(Ayrshires)

being -0.02509.

1hey presented a set of convers ion factors for c alculating mature
equivalent p rod uct ion in co njunction w i t h those no w used for milk
yield.

A further d isc rep anc y in the a p p l i cat ion o f blanket

conversion factors has

oeen rep orted by C o p e l a n d

(1934) who

showed that the increase in production w i t h 3 time compared wit h
2 time a day m i l ■ing varied w i t h the level of production.

The

increase for Jersey cows w i t h 400-500 p oun d 2 time records was
38$ for subsequent 3 time records, wh i l e

bhat for cows w i t h 2

time records above 750 pounds was only 6$.

The findings

suggest

that a h i g h p rod ucing cow w i l l come closer to m a x i m u m producing
capacity on 2 time m ilk in g than will a poor producer.
limitations

The

imposed by u d d e r size have been suggested as an ex­

planation of this.
slightly higher

Lush and Shrode

(0.833)

(1950) h a v e sugges ted a

correction factor for converting 3 x to

2x mil k i n g aft er the first lactation

(0.80) and h a v e pointed

out m i n o r discrepa nci es in the factors now u s e d for age correction.
The influence of the plane of n u t r i t i o n u p o n production
has long been u n d e r investigation.

Eckles

(1927)

reported that

fat test was h igher w h e n cows fresh ene d in g o o d condition.
ward

(1927)

Wood­

stated that an increase of 50$ in p rod uct ion could

be expected if average h e r d conditions are supplanted by those
w h i c h obtain u n d e r offic ial testing.
by Dawson and Graves
tices alone.

Similar values w e r e obtained

(1936) b y radically altering feeding p r a c ­

A n average productio n of 405 pounds of fat for 46

cows restricted to rou ghage alone was increased to 654 pounds

17

w h e n tiie sa me cows w e r e g r a i n fed in a c c o r d a n c e w i t h a c c e p t e d
px>aci>ice.

Plum

(1935)

f o u n d that 1 8 $ o f the v a r i a n c e in fa t p r o ­

d u c t i o n c o u l d be a c c o u n t e d for by f e e d i n g practices.
Kagadale

(1935) h a v e d e m o n s t r a t e d that the gross

p r o d u c t i o n v a r i e d f r o m 1 5 - 2 5 $ w i t h po or p r o d u c e r s
good p r o d u c e r s ,

the b i o l o g i c a l limit a p p a r e n t l y

Brody a n d

e f f i c i e n c y of
to 3 5 - 4 5 $ w i t h

oeing about 50$.

The i n f l u e n c e o f a n u m o e r o f o t h e r factors u p o n p r o d u c t i o n
has b e e n studied.
been studied.

The c o r r e l a t i o n of t ype w i t h p r o d u c t i o n h as

T he c o r r e l a t i o n of typ e w i t h p r o d u c t i o n h a 3 b e e n

shown to be o f a low order,

0 . 3 0 - 0 . 3 8 or less

land 1938b,

1948b).

T y l e r a n d Hya tt

Gowen

(Gowen 1933b,

suggests

Cope­

that w h i l e

c o n f o r m a t i o n is a p o o r c r i t e r i o n of p r o d u c i n g a b i l i t y it has
the m e r i t of bl o w i n g cu rre nt h e a l t h of the cow.
c o r r e l a t i o n s w e r e f o u n d by L e i g h t o n a n d Gr a v e s
of r u m p an d u d d e r s lop e w i t h product ion .
d e v e l o p m e n t at 4 m o n t h s
highly significant
(Book a n d c o w o r k e r s
on fat pe r c e n t a g e ,
of m i l k

(Jersey)

Non-significant
(1947)

for slope

E v a l u a t i o n of u d d e r

or 6 m o n t h s

(Holstein)

correlations with mature equivalent
1950).

Pregnancy was

had
production

fou nd to h a v e no effect

but r e s u l t e d in a d r a i n o f 400 — 600 pounds

(Gowen 1924b).

L i t t l e o r no r e l a t i o n s h i p w as f o u n d b e ­

t w een p r o d u c t i o n a n d the p r e s e n c e of s u p e r n u m e r a r y teats
1934)

(Gifford

or for t h e age o f the 3 ire a n d d a m u p o n daughter's p r o d u c t i o n

(Gifford a n d E l t i n g 1928).

Washbon and Tyler

in cre ase i n a v e r a g e fat p r o d u c t i o n

(1950)

f o u n d an

( d a m - d a u g h t e r difference)

for the d a u g h t e r s of later p r o v e n sons of a sire c o m p a r e d w i t h
those of e a r l i e r pr o v e d sons,

but this w a s c o n s i d e r e d to be

purely environmental.

Tne uniformity of get of a sire was found

to be unrelated to production of the get (Johnson 1945), nor
was there any correlation between age at first calving and mature
body weight or total production for the first five lactations
(Chapman and Dickerson 1936).

Cows calving early had a higher

total production to seven years of age however.

Bartlett

(1929)

reported Slightly less variability in daily production if m o r n ­
ing yield was added to the subseqtient evening yield rather than
the previous evening yield as is commonly done in testing.

The

number of daily milkings was found to have no effect on persistency
as measured by dope of decline in production (Ludwin 1942).
A significant development in evaluation of environmental
influences upon fat production has oeen the formulation of an
environmental index by Bayley and Heizer (1950).

The index is

based upon deviations of individual lactation records from an
arbitrarily determined base, such deviations being calculated
on the basis of the plus or minus contribution to the mature
equivalent record made by the six factors concerned.

The factors

used, together w ith the level at which the deviations are zero
and the nature of the dev iat i o n s , are:
Factor
Lbs TDN daily/lOOO lb wt.

0 interval

Effect of change

17.0-17.5

varies directly-::-

Nutritive ratio

6.4-6.8

Days with calf while milking

150-203

Number of milking cows in herd
Length of preceding dry period(wks)
Condition at calving

39-48

it

tt

it

tt

directly

6

Fair

inversely-;:-*:?

it

u

-;tPlus deviations result in plus contrioution to index and vice versa
t«t
tt
tt
it
M
M
u
" minus
n
”

19

Tiie index permits the records of* dams and daugiiters, for
example, to toe dompared on the toasis of the effect of the actual
environment upon fat yield.

This value can toe compared w i t h the

actual difference in ma ture equivalent production to determine
the effect of environment thereon.

Thus a large deviation in

actual dam-daughter production would toe considered largely
genetic if accompanied oy a small difference in environmental
index, or largely environmental if the environmental index difference
were large.

A major use of the environmental index might toe to

reflect differences in the environmental levels from one sire
proof to another.

The authors stress however, that the index

measures only six environmental f a c t o r s , and care should toe
exercised in concluding that the remainder of the production
differences are entirely genetic.

PREDICTION OF THE BUTTERFAT TRANSMITTING ABILITY OF BULLS IN THE
MICHIGAN INSTITUTION HERDS
I. Source of data
A preliminary survey of tiie Herd records in several of the
Michigan State Institution Holstein herds indicated that these
herds con sti tut ed a large and possibly more homogeneous group
of animals than might be expected in most populations of this
size.

About two-thirds of the bulls proved in these herds were

sons of her d cow3,
chased animals.

the other third apparently having been p u r ­

For h a l f of the h e r d bred bulls records were

available w h i c h made it possible to calculate prediction indices.
In this group there was one set of 18 30ns of one sire

^Marathon

Bess Burke 32) of w hic h one subgroup of 3 and one of 2 bulls
were full brothers.
ernal first cousins.

In addition,

9 of these 18 bulls w e r e m a t ­

There were two other sets of 2 and one each

of 5 and 3 paternal half brothers, and two sets of 5 and4 maternal
half brothers among other data from these herds.

Considering

the corresponding number of parent-son relationships, this formed
a fairly closely related group whi ch could be compared wit h an
equal sized group of unrelated

(purchased)

bulls.

In addition to the herd records of four institutions;

Tra ­

verse City, Ionia and Pontiac Hospitals and Ionia Reformatory,
an important source of data was the HI R records compiled by
the Holstein Friesian Association of America for these and other
state institution herds.

Most of the latter records were from

the following institutions;

Marquette and Jackson Prisons, and

Newoerry, Kalamazoo, Ypsilanti and Howell Hospitals.

Most of

these data were compiled between the years 1933 and 1947.

Another

bull, King Bessie Ormsby pietertje, with 20 sons proved outside
of the state institution herds was chosen at random to compare
with the 20 sons of Marathon Bess Burke 32 (two of whose sons
were proved in private herds in Michigan).

The DHIA data of

Eldridge (1948) consisting of 207 bulls proved in New York State
were used as a source of comparison with unrelated material.
w h s

believed that tnis represented

Holstein population.

a

It

more random sample of the

The data for the institution herds used

in the present work are compiled and appended to this study.
Eldridge*s data are contained in appendices to his thesis.
Most of the animals included in the analysis were located
in the state Institution herd 3 shown in Graph I.

Over the period

of time represented by most of these data, approximately 15 years,
from 1933-1947, these 10 herds ranged in size from 30-150 milking
cows.

The annual average production (3x-305da.) ranged from 380-

515 pounds butterfat for individual herds; the 15 year average
for all herds, not adjusted for herd size, was 433 pounds.

Ex­

cept for some of the earlier 4x records, practically all of the
cows were milked 3 time 3 daily.

The average length of lactation

for all herds over this period was 314 days.

Graph i:: shows tne

following:
1. Distinct differences existed between the 15 year herd aver­
ages for the several state institution herds.
2. There was a noticeable increase in annual average for
all herds in the latter half of this period.

POUNDS BUTTERFAT
in
C*
o

IONIA HOSP

MARQUETTE PR

3X

•A
o
o

o»
o
o

01

o

c*
oi

o
-L.

H
m

>

CO

777777777 <77777/7/77////.7ZZZ1
777777K 177777777777777/*7\

NEWBERRY HOSP

77777m /77777/A

IONIA REFORM

V777777k 777777777777777,a

K '2 0 0

V77777771 7777/7/7/////A

CD

z 33

777777777 V/////7777&7,77777777777/A

YPSILANTI HOSP

HOSP

305 DAY

“D

H I

C ^

••

H

-.5 2
■oi
W I ?
I m w
3

o ij

^ co o
->i z

.

3s
a i

H 3)

HOWELL S A N IT

77777777k V/////7A

JACKSON PR.

’77777777 V//////////////,7 7 7 7 7 %

> >

™£

TRAVERSE CITY H.

v7//7777a

wo

77777777/7/A

z

(D
PONTIAC HOSP.

7777777, V7777A

T
JO

cn r

r m
JO
m >
O
jo
rn
o
co

22

3.

There was a tendency for production to oe h i g h e r in

the smaller herds.
The distribution of the 20 sons of 412017 among these herds
was as follows

(some were used in m o r e than one herd):

Traverse

City Hospital-4, Pontiac Hospital-3, Ionia Reformatory-2, Lapeer
(JL
TrAining School-2, Chatham Experiment Station-2, private herds-2,
Ionia Hospital-1, Ypsil a n t i Hospital-1, Kalamazoo Hospital-1,
Howell Sanitarium-1, iaarquette Reformatory-1,

Jackson prison-1,

N e w b e r r y Hospital-1 and Flint School for the Deaf-1.
II. Methods
All records were corrected to standard conditions of age,
times m i l k e d a n d length of lactation in a c c o r d a n c e w i t h accepted
practice in such studies.

The following factors were used to

convert HIR records to 305 days, and to convert 305 day HIR records
to 3x ivi.E. Taasis:
3x

4x

1.25

1.10

1.20

1.05

3

1.15

1.01

0 .95

5k

1.10

0.97

350-359

0.94

4

1.07

0.94

360-364

0.92

4f
‘
£

1.05

0.92

365-

0.90

5

1.02

0.89

6-9

1.00

0.88

10

1.03

0.91

11 -

1.05

0.92

Days

Factor

Age

306-319

0.99

2

320-329

0.97

330-339

0.96

340-349

The data were divided into the sire groups indicated in
Graph II from which the following tendencies can be seen:
1. Distinct differences in group

averages and more so in

distribution of the data within groups existed.
2. Most of the values fell within two standard deviations
on either side of the mean.

In a normal distrioution 95^ of the

values would be expected to fall within this range.
3. The most homogeneous groups were those composed of daughters
of bulls w i thin herds.

The most conspicuously heterogeneous data

were from unrelated bulls used in several herds.

The data from

more or less related bulls used in several herds were intermediate
in their distribution.
4. The standard deviation of neither daughter nor mate
average showed any conspicuous relationship to the size of
the aample.
The averaged M.e. records for the daughters, mates, and
where availaole, the bulls* dam and sire proofs were itemized
(see appendix).

Correlation coefficients

(r= x y / X(x 2 ) (y2 )

Snedecor) ware determined for dam-daughter production in each
sire group.

In addition,

correlation coefficients were ob­

tained between actual daughter production and that predicted
b;/ the several means indicated below.
In reporting the correlations obtained by the several methods,
the symbols used to designate the various groups are as follows:
(subscripts X ^ ..... q ar*e those used by Eldridge 1948).

GRAPH H

CHARACTERISTICS

m a te

IN

a v e r a g e

STATE

OF

DAUGHTER

p r o d u c t io n

I N S T I T U T 'O N

f o r

AND

b u l l s

HERDS

600

5 50 -

>-

<

Q
m
0
rO
1

^
500 -

11

X

rO

MEAN

I—

K

\\

<
Lu

cr
Ll)
1
—

A

V
450

0
CD

A

CO

Q
3
3
O
Q.

\

i
1

£
4 OC

Z
<

Q
z:

<
LU

I S. D.

300

CO

2
<
cr

/ N / W

IAUAVE.

(Y)

D A M AVE

(XO

BULL GROUP

Y

x.
to

o
t r\

cr
—

2

r
O

R*J t
l
Qo
cr &

O to
1.1 o
to or

CD UJ

<t Uj

^

X

T

X

o

OJ

24

Y - average of the daughters of the hull

Xp

”

H

mates

'*

"

"

-

’*

M

u

daughters of the dam of the hull

X3 =

"

w

”

dam of the hull

”

daughters of the sire of the hull

X4 ^
X5 -

"

*'

”

mates

"

"

«

«

»

«

it

it

it
n

daughters of the maternal grandsire of the hull

The use of a subscript with Y denotes a predicted value for
the average production of a bull's daughters as follows:
Y ep ; hy equal parent index*- (2Xa-Xr) + X ^ /2
Y r » hy regression index*
(w- herd ave.)
=

4

Xi/2

(2Xa-X.s4w /2 ) 4 Cfo+w / 2 )/2 4 (Xt4w/ 2 )
----"— — tt~tr- -- — _
-4r— ______ —-1
----

/**

Yet, 3 hy Eldridge equation B*= 0.1 t 0.75Xi -t O.OIX 3 + O. 23X 4 - O. 22X 5
+ 0.24X6
Y 3 , hy equation 5 0.75Xi 4 0.25Xi
YpX c by equation - (Xx -v P)/2
Ypg 3 hy equation =. 0.75Xx * 0*25F
p

X 4 +X3./2 H- Qfr+Xs/g) ■> Xl/g /g

The value P equals the transmitting ability of the bull calculat­
ed hy regressing the average of the daughters of the bull's sire (X4 )
halfvmy toward the average of the mates

(Xx) of the hull, adding to

this the average of the dam (X'g) of the hull corrected for N records
by Lush's formula

(1945)

with r- 0.4 after regressing this value

halfway toward the mates of the hull's sire

(x5)

and further

regressing this latter value nalfway toward the mates of the
bull (Xx), and halving the sum.

Thus each regression utilizes

The theoretical bases for use of thi.se predictions are
given in the review of literature of the prediction of transmitting
ability.

25

the records of cows contemporary to those whose records are being
regressed.
The prediction Y g = 0.75X^ + 0.25 X-^ was based on the same value
found by Eldrldge for regression of daughter average upon dams plus
the supposition that the remaining 25/^ might oe accounted for by
population average (Xx).

The more complicated formulae for Y px

and Y p a represent two of a number of attempts to improve upon Y g
by incorporating data from the pedigrees of the bulls.

RESULTS
I. Co r r e l a t i o n o f ave r a g e d a m - d a u g h t e r prod uct ion
The f inding that one Dull, M a r a t h o n Bess Burke 32, reg. no.
4 1 2 0 1 7 / h a d 20 sons
institution herds

Dred and proved as sires in the several

(two w e r e proved in private herds)

led to e x ­

te nsive u s e of this d a t a for correlating actual w i t h p r e d i c t e d
production.
group

It was thought that this co n s t i t u t e d as large a

(Ms 686 comparisons)

of records of related animals kept

u n d e r r e l a t i v e l y u n i f o r m conditions as might

oe found.

There­

fore it s h ould prove of interest to compare the predic t a o i l i t y
of these w i t h less h o m o g eneous data.

% e

co r r e l a t i o n

oetween

the a v e r a g e records o f the daughters o f these 20 bulls and their
dams was 0.64."

This r value is in Keeping w i t h those found b y

Lush and Schultz

(1938) and E l d r i d g e

of 0 . 6 3 - - 0 . 67 for similar data.
ing w a s

(1948) who reported values

No c o n s i d e r a b l e amount o f inoreed-

evident from e x a m ination of h e r d records, thus it was

assumed that the d a m - d a u g h t e r r e l a t i o n s h i p was not sign i f i c a n t l y
different from that

existing In most herds.

These m e t h o d s gave co r r e l a t i o n values bet w e e n actua l a n d
pr e d i c t e d p r o d u c t i o n as follows:

significant at 1% level

Table I:
Correlations between actual and predicted production
of daughters of 20 sons of warathon Bess Burke 52
Method of
Correlation with
Percent chance of estimating
prediction
actual production
actual production withinC
_________________________________________________50 pounds_______50 pounds
0.64**

55

85

Y ep

0.46*

30

60

Yr

0•5 5**

15

60

*eb

0.58::-*

40

75

Q .64**

40

80

0•63**

50

70

0.63-**

50

80

X1

Ys
V

-^px

Y PS

**significant at 1%, * at 5$ level
t For explanation of symbols see above (page 24).
It will be noted that none of the predicted values exceeds
that for tne correlation of dam-daughter average.

There were

no significant differences between any of these correlation
coefficients.
II.

Per cent chance of predicting production within certain limits.
A factor of greater practical interest than the correlation

between predicted and actual production is the chance of predicting
production within certain limits.

From Table I it can be seen

that the average of tne bulls* mates was the most reliable means
of predicting within 30 or 50 pounds the actual production of his
daughters.

iJot only are these the most available data,

but the

chances of predicting production within the lower limit were Q0fo
better than by using the equal parent index.
Rice ’3 regression index was even poorer.

Predictaoility with

Or it might oe pointed

out that the chances of predicting actual production w i t h i n 30
pounds was n e a r l y twice as great w h e n only the mates*
was u s e d as w h e n the equal parent index was used.
mates'

average

similarly the

a v e r age was ne a r l y four times as reliable as Rice's r e ­

gression index.

The figure of 30 pounds was chosen as re p r e s e n t ­

ing the probable lower limit of accuracy
a 300 d a y l a c t a t i o n ) .

(0.1 pound daily for

The figure of 50 pounds

(about 10# of

a desir a b l e level o f 3x production) was chosen as representing
the o p t i m u m u p p e r limit of error.
other groups

Since the figures for several

(below) w e r e based on 2 x production, only the 30

pound limit was applied to them.
A d d i t i o n a l dam- d a u g h t e r correlations w e r e determ i n e d for
several groups of culls w i t h i n the institution herds and for
a group of 20 sons o f another bull,
reg. no.

hing Bessie Ormsby Pieterje,

52017, not u s e d in these herds.

These data are presented

below (tablejl).
Table II.

Correlations

between mates and daughters of bull groups

Bull Group
H=.no*
rYXi*
_________________ com p a r i s o n s

^predicted
^
w i t h i n 30 lbs.______ r Y Y R_______

Traverse City herd

15

0 .57-*-;:-

Ionia Reform, h e r d

14

0.75

93

0.76

all herds 23

0.52

48

0.54

Purchased, all herds 45

0.67

40

0.67

Herd-bred,

Sons o f MBB32

100

0.58#-#

(Taule 20
0.64
55
0.64
I)
Sons of KBOP
20
0 . 8 1 __________ 7 5 _________________ 0.78_______
all r values significant at 1% level
* rYX-, ^ dam (X-, ) ± daughter (Y) correlation
ttr Y • 7 5X 1 * 725X1

The correlation between equal parent prediction and actual
production determined for the daughters of the herd- b r e d sires
(sous of h e rd cows, not sens of M B B 3 2 ) above was 0.47-::-.
daughters of sons of King Bessie Ormsby Pietertje
the correlation was 0.67-*-*.
are similar to what has
III.

For

(KBOP above)

These and the values in Ta->le II

oeen reported and to those in Taole I.

Correlation coefficients oetween dam-daughter production

at varying levels of production

(Eldridge data).

As m e n tioned above, the correlation coefficients for damdaughter production correspond closely w i t h those of Eldridge
(1948) who found r values of 0.63--0.67 for four sets of damdaughter comparisons and values of 0.69 and 0.70 oetween actual
production and that predicted by his regression formulae.

As

might be expected, the value calculated for r Y Y s (Eldridge data)
was of the same order as that for rYXi heing 0.64 for the data
in his grdup B (3 ee

page 8 ).

reasons for these correlations
data for the two groups

In an attempt to discover the
oeing what they were, E l d r i d g e ’s

(N-= 207) were combined and then partition­

ed into several more homogeneous groups.

The following correlations

were found for dam-daughter averages and for actual-predicted
daughter production

(Taole III).

Table III.

Correlations for bull groups assembled from E l d r i d g e ’s

data.

All factors 400# or more

.
tt
N______ rYXjj_______r
Y
Y
ep
------ =-±---- ------- -----38
0.19
-0.06

Bulls’ mates 450# or more

29

0.18

0.19

more
Mates 350# or less, other factors 400# or/ 22

0.19

0.04

Mates 350-399#, other factors 400# or more 28

0.33

0.40

All factors less than 400#

0.77*-::-

0.77*#

Bull group____________

16

significant at '1 % level
t rYXp « dam(X^)-daughter (Y) correlation
tt Y Qp = equal parent pr#dtf.etion
The use of Eldridge's regression formulae on the first group
gave an r value of 0.19, the same as for rYX^> while rYYs ~ 0.22.
Eldridge*s formulae for prediction with the factor for trie m a t e s ’
average deleted gave an r value of 0.002.

It is evident that

when the data are partitioned in such a manner as to destroy the
usual h e r d relationships, predictability, for practical purposes,
i 3 lost.

While the last group represents a selected group as

far as these data are concerned,

it is suggested that it is also

a group in whihch, in practice, no discrimination had oeen exercised.
IV. Comparison of rankings of transmitting ability of bulls with
actual daughter average rankings.
Another approach to the problem was to see which of several
methods of prediction would come closest to placing various
daughter groups in their actual rank of production.

If this method

could be shown to have the advantage of placing bulls in proper
rank according to transmitting ability,

it might obviate the

51

dependence upon the production level of their potential mates.
Accordingly the actual daughter averages for three oull groups
were ranked by production level and the production rank predicted
b; the EP a n d regression indices compared w i t h the average rank
of the respective d a m groups.

The average deviation of predicted

production rank for these three methods ranged from 3.7 to 4.5
places out of 20 or 25 from perfect correspondence

(zero

deviation w ith actual ranx of daughter average (Tables IV, V,
and VI).
Taole IV.
Actual
.. ave
581
546
523
521
492
490
487
487
478
476
457
449
410
406
405
400
400
390
378
364

Actual and predicted average daughter production and rank.
Sons of Marathon Bess Burke 32.
Rank

Daughter average and rank predicted by:
Rank
Reg. index Rank
EP index
D a m ave . Rank

1
2

3
4
5
6

7
7
9
10
11
12

13
14
15
16
16
18"
19
20

srage deviat ion'*'

573
442
499
529
458
444
461
420
500
449
427
423
440
442
402
439
516
424
382
406

1
10

5
2

7
9
6

17
4
8

14
16
12
10

19
13
3
15
20

18

512
445
473
519
469
458
450
425
485
439
437
449
471
441
445
454
474
448
422
402

3.7*3.2

•^Deviation of predicted from actual rank

2

13
5
1

7
8
10

18
3
16
17
11
6

15
13
9
4
12

19
20

5.4*3.5
(column 2)

477
445
459
480
456
450
447
433
457
441
442
445
456
466
444
448
458
445
432
430

2
12

4
1

7
9
11

18
6

17
16
12

7
3
15
10

5
12

19
20

4.5*3.8

32

In each, case the standard deviation nearly equalled its statistic, hence
there was no significant difference between any of the methods.
However in 3 out of 4 comparisons of dam average rank with that pre­
dicted oy either EP or regression index, the average deviation of
dam average rank from daughter average rank was the lesser of the
two figures.
Taole V.

Actual and predicted daughter average production and rank.
Herd-bred bulls not sons of MBB 32.

Actual
Rank
.ave.
596
579
575
550
540
504
498
487
486
481
473
470
459
448
442
442
436
431
429
425
409
403
400

1
2

3
.4
5
6

7
8

9
10
11
12

13
14
15
s 15
17
18
19
20
21
22

23

Average deviation

Daughter average and r
Rani
Dam ave.
471
'503
529
490
483
507
443
545
453
497
408
450
441
464
532
427
466
497
419
436
391
443
388

10

5
3
8

9
4
15
1

13
6
21

14
17
12
2

19
11
6
20

18
22

15
25
'4'. 5fc3’3

EP index
471
551
539
521
574
464
506
581
496
47 5
468
476
496
514
508
421
482
424
441
504
408
440
402

Rank
15
3
4
5
2

17
8
1
10

14
16
13
10
6

7
21
12
20

18
9
22

19
23
4 . 3 I3.4

33

Table VI. Actual and predicted average daughter production and rank.
Sons of King Bessie Ormsby pietertje
Actual
Dau.ave.
461
435
424
418
411
393
392
390
389
389
382
381
379
372
371
359
352
341
323
316

Rank
1
2

3
4
5
6

7
8

9
9
11
12

13
14
15
16
17
18
19
20

Average deviation

Daughter average and rank predicted by:
D a m ave.
Rank
Rank
EP index
452
419
462
410
374
374
351
373
390
394
361
381
376
383
316
332
397
367
324
336

2

439
444
448
444
403
404
423
414
396
422
382
408
404
426
390
383
409
404
399
402

3
1

4
11
11

16
13
7
6

15
9
10
8
20

18
5
14
19
17

4
2
1
2

14
12
6

9
17
7
20
11
12

5
18
19
10
12

16
15
4'.3 +3. 4

3 .813.3

Several methods which did not utilize the dam average were
also u s e d to rank bulls according to transmitting ability.

In

every case the average deviation of rank of transmitting level
from actual daughter average rank was about twice that when the
mates average was used, the figures ranging from 7.1 to 9.2
with correspondingly high standard

deviations.

Again no significant

difference was found between any of the methods of prediction.

A

more crucial test was made by selecting groups of 7 to 10 bull 3
from each sire group who were mated to cows whose average produc­
tion fell w ithin a range of 30 pounds.
above, no significant differences

Applying the same technique

Detween the rankings of trans-

mi t t i n g a b i l i t y of the individual bulls on the basis of pedigree
were found*

N o r w e r e any differences found w h e n the mates'

age was included*

aver­

In b o t h instances the average deviations w e r e

of the -iame o r d e r as found for the entire group.
It is apparent in all of these analyses of transmitti n g
ability that w h e n the only corrections of production records
made are the u s u a l ones

for age, times m i l k e d and length of

lactation, m et h o d s of prediction based in part on pedigree
are no m o r e reliable than the use of the d a m average alone in
the p r e d i c t i o n of d a u g h t e r average production.

M o r e o v e r in no

case w e r e p redictions based on t r a n s mitting a b i l i t y alone as
accurate as those w h i c h included the average of the mates to w h i c h
the bulls w e r e bred.

DISCUSSION
I. Correlation analysis of production data.
In each, of the groups of unselected data studied, the cor­
relation o f production of dams and daughters was as high or
h i g h e r than any other value determined.

It seems doubtful that

the data could nave Dean ma n i p u l a t e d in any other logical m anner
to produce hi g h e r values.

In most of the literature dealing

w i t h methods of predicting production the existing dam-daughter
correlation has not been adequately pointed out.

In light of

this, and the apparent fact that at desirably high levels of
production predictability is essentially lacking,

it would 3eem

that present methods are not adequate for predicting production
from pedigree.

This does not m e a n that tnere are no differences

in transm i t ting abi l i t y of bulls,

but that such differences as

do exist are obscured by environmental factors.
emphasis placed u p o n the mates'

Nor does the

records in this and Eldridge*s

v/ork n e c e s s arily refute the theory of Mendelian inheritance.
Rather the greater weight given to the female side of the pedigree
is apparently only a reflection of the similarity of environment
w hich exists for dam3 and their daughters.
It seems logical to assume that m u c h of the unpredictability
w h i c h attends the use of a bull with h i g h pedigree promise on
cows of lesser demonstrated ability stems from differences in
management likely to be placed upon the bull as a herd Improver.
Often, w i t h

the purchase of a good bull, the owner maices improve­

ments in h e r d environment commensurate w i t h the interest w h i c h

prompted, purchase of* the hull.

Others m a y place primary reliance

up o n the bull alone to do the job.

While some bulls appear to

have oeen h e r d improvers regardless of* h e r d level, there seems to
be no doubt that genetic and environmental factors are so inter­
m i n g l e d as to confuse cause and effect.

The present study,

like

others that have preceded it, leaves the vital question u n a n s w e r ­
ed, namely, what is to oe the source of h i g h producing cows?
the a v erage breeder at least,

For

the answer w o u l d appear to lie in

G-owen»s early admonition to pay less attention to getting a bull
of h i g h pedigree promise and more to securing more calves from
his h i g h producing females.

A logical follow-up of tils advice

w o u l d be to pay attention to supplying the o p t i m u m h e r d environ­
ment .
A n o t h e r factor to be considered in the application of
techniques to predict future production is the number of restrictions
to be p l a c e d u p o n the individuals concerned.

Two or three gen­

eration pedigrees w i t h complete production records are relatively
rare.

E l d r idge found only 207 bulls proved in New Y o r e DHIA nerds

w h i c h met his specifications.

W h e n the restriction that they

be u n u s u a l l y good prospects as nerd improvers,
the entire breed oecomes infinitesimally small.

the number in
Was h o o n

(1948)

lists o n l y 22 Holstein bulls and 4 of two other breeds the sons
and grandsons of which, according to his criteria,
pended u p o n to increase production.

can oe d e ­

Further, W a s h b o n admits

that even these may not stand the test of experience.
The use of already proved sires is limited by the fact

that most are dead oy the time they are proved (Lush 1945),
and very few of those living are proved in a second herd (Beardsley
1969)»

The time factor involved makes dependence upon proved

bulls a less useful procedure than would be the case otherwise.
As far as availaoility of data is concerned, methods of prediction
based principally upon the records of the bull»s mates would be
most useful, and according to this study, a 3 reliaole as any
other method tested.
With t hi 3 in mind it seems logical that the average breeder
should pay more attention to factors of management and selection
within his own herd than to pedigree promise.

Experience does

not indicate however, that the problem of herd improvement can
be resolved 3imply by improved environment and selection of dams.
The former does not make any permanent contribution, and progress
by the latter method i 3 painfully slow in practice.

The increase

in production due to the amount of selection for production now
possible in most herds is low, being 011 the order of one pound
per year regardless of herd level (Seath 1940).

Chance and

Mather (1949) concluded that cow families were not sufficiently
differentiated to receive much consideration in selection.

This

pound a year improvement is about what is being accomplished by
the use of all methods in common practice(DHIA).

More could oe

achieved by increasing the selection differential by better
management-- larger calf crops and fewer replacements due to
disease, injury, etc.

Nelson and Lush (1950) found that by

selecting bulls from the better cows it was possible to raise
genetic ability 40 pounds in 12 years in the face of an inbreed-

ing program which resulted in an overall decrease of production
at the rate of -4.5 pounds of fat per one percent of inbreeding.
II.

Diallel crossing
Because of the poor predictability with common pedigree

methods, the Traverse City records were examined to see whether
data were present to afford analysis by the method of diallel
crossing.

This method has the advantage of reducing the source

of error inherent in the mates inasmuch as the same mates are
used to prove two males.

In dairy cattle breeding such a method

has obvious limitations.however, because of the relatively small
numbers of cows that nave tested daughters by more than one bull.
In the Traverse City herd only 35 such cows were found despite
the fact that a potential of 1500 lactations was represented.
Prom 12 bulls mated to these 35 cows there were 15 sets of 2
or more common mates for 2 bulls, of which 8 had sets of 3 or
more, and 2 had 9 mates in common.

Because of the limited amount

of data available, the following is reported here rather than
under experimental results.
Since each of these bulls had already been proved on a large
number of comparisons, an attempt was made to see whether the
method of diallel crossing could be used to more accurately
predict the final index of a bull on the basis of a smaller number
of comparisons.

This method was limited to the four bulls which

h a d comparisons based on 6 or more common mates.

It was assumed

that one bull had oeen proved on the basis of the total number
of comparisons available.

His proof on the basis of the mates

in common with the other bull was corrected by adding (algebraically)

to the dam's average the amount required to m a k e the two indices
equal.

The proof of tne ’’unproved** bull was determined on the

basis of the mates in common w i t h the proved bull, and the same
correction was then applied to give the corrected proof.
ing example will illustrate this technique.

The follow­

The nine common

mates for bulls 412017 and 353211 averaged 440 pounds.

The

daughters of 412017 averaged 438 giving an EP index o f 538 p o u n d s .
On the oasis of
509 pounds.

26 comparisons the EP index of the latter bull was

To m a k e the index on the

basis of the 9 comparisons

equal 509 pounds, 29 pounds must be added to the production of
the dams.

That is,

if the 509 figure is more correct because it

is based on more data, the situation is as if the daughters in
the 9 comparisons were from dams w h i c h transmitted at the level
of 469 r ather than 440 p o u n d s .
If this value of 469 pounds for t.ie 9 dams is
corrected value

used as a

to determine an index for the bull 412017, the

c o r r ected index w i l l be 407 rather than the 436 pound figure
using tae actual dam average.

The actual proof of this uull

on 67 comparisons was 418 pounds.
tne dams'

Thus the corrected value for

average gave a more conservative index, and one which

was 7 pounds closer to the final index of the bull than did
the u n c o r r e cted figure.
the dams'

Conversely if the corrected value for

average is used to predict the index for the other

bull, 353211, 18 pounds must be added to the actual dam average
(440) to m a ke the index for 412017 on the basis of the 9 dams
equal his final index of 418 pounds.

Using the corrected value

for the dams 1 average(458) to determine an index for 353211 gives
an index of 520 pounds compared with 538 pounds for the uncorrected
index.

The final index based on 26 comparisons being 509, the

corrected provisional index came 18 pounds closer to the final figure
than the uncorrected. These data are presented below in tabular
form (Table VII).
Table VII, The use of diallel crossing to predict transmitting
ability of bulls.
Bull
412017
353211
1 . Actual ave. prod.

9 common mates

440

440

2 . Dau. ave. 9 mates

(1)

438

489

3. Provisional EP index (9 pairs)

436

538

4. Final EP index (67 pairs)
(26 pairs)

418

5. Discrepancy: final minus provisional
6 . Corrected m a t e s ’ ave.

(1 ♦ 5)

7. Corrected provis ional'^EP index (9pr.)
8 . Discrepancy:

final minus corr. provis

9. Improvement using corr. provis. index'

509
18

29

458

469

407

520

11

18

7

18

ttUsing corrected m a t e s ’ average (s transmitting ability) from
proof of bull A with daughter average of bull B to determine
corrected provisional EP index for bull B and vice versa.
Coi*rected provisional index came 7 and 18 pounds closer
respectively than uncorrected provisional index in pre­
dicting final EP index.
For three of the four bulls the provisional index on the
basis of the corrected d a m s ’ average was closer to the final
index than that using the uncorrected average.

For the four

bulls the provisional corrected index averaged 428 pounds, a
deviation of 4 pounds from the average final index; the average
uncorrected provisional index was 440 pounds, a deviation of
16 pounds from the final index.
It was decided to see whether such a procedure might be
extended to the use of individual corrected comparisons rather
than averages of common mates.

In this case the only requirement

was that each mate of the bull to oe proved have another daughter
by a proved bull.

The deviation of the daughter by the proved

bull from EP expectation was noted and the average of these devia­
tions added

(algebraically) to the average of the dams.

This

method has the advantage that it does not require that the mates
of the bull to be proved be mated to a single other bull, thus
more comparisons will usually be available.
For 16 mates of the bull 412017, the average deviation o f
their daughters by other (proved) bulls was 15 pounds more than
EP expectation.

Adding this to the mates average (441) gives

a corrected value of 456 pounds.

The daughters of 412017 from these

16 mates averaged 437 pounds, thus an index of 420 was obtained
with the corrected d a m s 1 average compared with 433 for the u n ­
corrected average.

The final index for the bull was 418 pounds.

For the bull 353211 the correction was less than one pound.
The dam average was 448, the daughter average 478 for 11 comparisons.
This gave an index of 508 compared with 509 for 26 comparisons.
Thus the smaller number of comparisons was apparently a good sample
of the total, no correction oeing needed.

It is not suggested, that any conclusions be drawn from this
limited group of data, but the matter would seem to warrant fur­
ther investigation.

The procedure might be termed one of proving

the dams w h ich are used to prove a bull.

If the method can be

shown to increase the accuracy of prediction, it would be useful
to prove a bull on dams w h i c h have one or more daughters by
already proven bulls.
Such a technique might be u s e d as follows:

Assume there

is in a h e r d a bull whose first daughters had been outstanding
and who therefore was bred to a large number of cows over a short
period of time following his early proving. As these cows freshen­
ed small groups w o u l d be bred to young unproved bull 3 .

Rebreeding

of a number of the cows would be necessary to secure heifer
calves from each cow w h i c h had produced a daughter by the proved
bull.

This could be largely accomplished by the time the daughters

of the proved bull had completed their first records.

The records

of the mates of the young bulls could then be corrected on the
basis of their deviation from the figure required to give the
latest index of the proved bull.

The corrected figure for the

average of t h e mates of the young bulls would be u s e d to predict
what their final proofs w o u l d be.

By such a technique it should

be possible to discover a group of cows which were apparently r e ­
latively prepotent for transmitting at a certain level of pro­
duction.

As they were found they should be maintained as a

group of tester brood cows,

i.e., they should no longer be pushed

to their limit but be managed so as to encourage regular breeding
and longevity.

III. Environmental indices
Rather than simply suggest that primary attention be directed
toward improvement of environment,

it seems desirable that methods

be developed to measure tne effects of environment in order to
more accurately assess the genetic worth of animals.
of this is

A corollary

b a s e d on the fact that some of these 11environment a lw

factors m a y be inherited and thus selected for independently
but in conjunction with production per se.

of

For example, body

size and persistency are not merely environmental factors which
surround the producing machine, but are in part, inherited com­
ponents of

the machinery for

Bayley and

Heizer

butterfat production.

(1950) appears to be

effort in this direction.

The work of

the first large scale

It would seem that the state institution

herds might be an ideal proving ground for establishing environ­
mental indices because of the large number of cows kept under
somewhat u n i f o r m and better than average conditions.

The possibility

of keeping extensive records may also be somewhat greater than
exists in average herd3.
It is suggested therefore that this step be taken by the
state institution herds of Michigan,

and that tne records so

accumulated be made availaole to a central agency, probably one
connected w i t h the College of Agriculture.

It would be necessary

to determine prediction values for each factor under the condi­
tions w h i c h exist In these herds before their general applicability
could be estaolished.

Some factors,

such as climate, must

necessarily vary from one region to another regardless of other

43

conditions of environment.

Certain discrepancies in the literature

must be resolved, notably the question of whether body weight or
metabolic body size more accurately reflects production, and whether
a separate set of age correction factors for fat test should be
us e d .
The environmental factors discussed in the review of litera­
ture should provide an adequate array to initiate such a program.
Others will undoubtedly be discovered and snould be added as soon
as their influence is known or suspected.

A work sheet similar

to that used by Bayley and Heizer could be used to accumulate
data.

It is suggested that such a work sheet would be more use­

ful for large herds if the data for herd practices were recorded
on a separate sheet, and the factors measured be arrayed in columns
and the cows in rows.

The following ( pages **',*<,) is suggested

as a form of work sheet that might be used.

This sheet differs

from that used by Bayley and Heizer In that only original herd
data are called for, all calculations, as for length of dry
period and days with calf while milking, would be made at the
central agency.

This would compensate in part for the time re­

quired to make the additional observations called for and elimi­
nate one source of error.
In addition to the worksheet data the monthly cow summaries
of production should be available for calculation of persistency.
Evaluation of udder development as calve 3 should be made.

Some

’’extension work” might be necessary to impress herd managers
w i t h the potential value of such data for th© project should

toe worth whatever extra time would toe required to compile the data.
If time and the expense of testing oecome a factor in tne numtoer
of cows that can be tested, some simplifications of the present
testing procedure might be worth investigating.

Tyler and Chap­

man (1944) have suggested a simplified method of estimating 305
day production by multiplying the sum of the first 10 test periods
toy 30.5 to save time and avoid some error of computation.

A

correlation of 0.993 was obtained between this technique and that
in DHIA use.

Alexander and Yapp (1949) have suggested a further

simplification involving testing only three times during the
lactation period.

They found that testing during the second,

sixth and tenth months was sufficiently accurate to merit adoption
as a means of lowering the cost of testing and increasing the
number of cows tested.
It is suggested that as the yearly data sheets are completed
for each herd they oe placed upon punch cards at the central
agency to facilitate handling.

The data for monthly and total

production, sire and dam numbers, and calculated values for
length of dry period, days with calf, daily TDN, persistency,
etc, could toe added at the same time.

4S

Owner

Address

Year

H e r d Pracrdcea
1.

Size

of mil d n g nerd____

2.

G r a i n ra tio n— amounts and ua sis on w!iicn. fed to fresh
d r y cows

Centents:

5.

Rou gha ge - - a m o u n t s and kind fed to fresh. and dry cows,
s e a s o n a l variation
Silage

Hay
I
Other
4.

I n c i d e n c e of disease

5.

C l i m a t i c conditions
Average

oam

temperature while stabled

U n u s u a l conditions

S.

(drougnt,

excessive 1

Dates
or cold,

nemar-Ks
H o w were measurements mate
Oth. er

(actual or estimated)

etu.)

Com J k u

Body weight/
month, lactation
1st,J3rd oth lOtii

prev,
Measure. 1st mo.
30
iy
jthrs
Heart
dry
girth Ingih h t . date

hate
frsb.

Cond.
when
frsh

Ho.
se rv

Date
conceiv.

Dace
dry

I
Remarks

-

SUMMARY

Correl ation analysis of the butterfat production records of
the d a u g hters and mates of several groups of bulls in the state
instit u t i o n herds of Michigan showed the following:
1* The correlation between daughter and m a t e average was
on the o r d e r of 0 . 6 0 — 0.65 for most groups.
2. N o n e of the accepted or experimental methods of predicting
d a u ghter average gave higher values.
but not significantly,

Several w e r e distinctly,

below the value for dam-daughter correlation.

3. T h e re were no significant differences

between the values

o b t a i n e d w i t h any of the methods used.
4. The institution data showed either no increase in h o m o ­
g e neity over data f r o m m o r e diverse sources, or no increase in
p r e d i c t a o i l ity due to w h a t e v e r increase in h o m o g eneity that might
h a v e existed.
5. The percent chance of predicting daughter average pro­
d u ction w i t h i n certain limits was somewhat greater u s i n g only
the d a m a v e rage than
6 . There were no

any m e t h o d involving pedigree prediction.
significant differences between the rankings

of bulls according to transmitting ability w h e n compared to the
rankings o f daug h t e r average production.

The most accurate

rankings w e r e those w h i c h jncluded the averages of the mates of
the bulls.
7. P a r titioning of Eldridge's data into more homogeneous
groups on the basis of dams'

production and/or pedigree promise

resulted in a significant loss of predictability of daughter
production.
It was concluded that environmental differences were of
such magnitude as to obscure any genetic interpretations that
might be placed upon such data.
A proposed worksheet for accumulating data to be used in
the formulation of an environmental index was constructed.

49

REfERE'.CE 3
Alexander, R. h.
1950 a study of cue inheritance of oer dstency in mil • oroduct ion. Jour, dairy 3ci. 33:
575 (abst.).
Alexander, a. H. and Yapp, ’7.
19-i9 Comparison of methods of esti 1mating mil r and fab pro­
duction in dairy cov/s. Jour. Dairy 3 ei. 32:
621-629.
i
Allen, a. 1!.
1944 A standard for evaluation of dairy sires proved in
•dairy herd improvement ass Delations. Jour. Dairy 6 ci. 27:
35-847.
Arnold, P. f. Dix and Bee-or, R. 3 .
1036 Influence of preceding dry period and of mineral sup­
plement on lac bation.
Jour. Dairy 3 o i . 19:
257-263.
Ba r11 e11, 31 eph en
1929 formal day to day variaoility of yield of mile an i
. fat
of intivi luul cons.
Jour. Ap;r. 3ci. 19: mo3-451.
z e r , 3. E.
•the effects o f certain environ'venta1 i
dairy C: 1113. Jour. Dfliry 3c i. 55: 375
(a o st . )

o pra nod supsi 0 r.Q'.lt.
1 oa .Lie,ijLi.ay , Li. j •
tton, R . '7. an .

lo 50

t-ie enrv i 1in en r i ty of ncorit ability of on tt erf a t produc­
tion.

tour. Dairy 3ci. 35:

95-97.

Been, G. H. and firm,
. L.
19-.B Size and its influence o . nil.' production.
oy tne fure bred Dairy Gar,tie association.
Boeder, A. 3., an :t ncGilliard, P.G.
192$ Inheritance of persistency of lactation.
An in. Prod., P r o c . 29-53.

Dictriouted

Arner. Soc.

Berry, J. G.
1945
Reliability of averages of different nuasaro of lactation
records for com.pfl.riny lairy cov/s. Jour. Dairy 3ci. 28:
uOO“00'J •

1939

and Lush, Jay L.
High records contrasted with unselected records and
with average records as a basis for selecting cows.
Jour. Dairy Sci. 22:
607-617.

Book, J. H . , Swett, W. W. and Matthews, C. A.
1950
•^valuation of mammary development of heifer calves.
Jour. Dairy Sci. 33:
383 (abst).
Brody, S ■ and Proctor, R. C.
1935
Growth and development XXXV energetic efficiency of
milk production and the influence of body weight there­
on.
Mo. Agr. Exp. Sta. Res. Bull. 222 . (c <?/
1935

, and Ragsdale, A. C.
Evaluating the efficiency of dairy cattle.
Sta. Bull* 351. 'Columbia.

Mo. Agr. Exp.

, Ragsdale, A. C. and Turner, C. W.
The rate of growth of the dairy cow.
Ill The relation
between growth in weight and increases in milk secretion
with age.
Jour. Gen. Phy 3iol.
6 : 21-30.
Bonnier, G.
Is the shape of the lactation curve genetically deter­
1935
mined?
Hereditas 20:
199-213.
Brook 3 , I. J.
The influence of environmental temperature on the per­
1931
centage of butterfat in c o w ’s milk.
Jour. Dairy Sci. 14:
483-493.
Gannon, 3. Y.
Seasonal effect on yield of dairy cows.
1933
16:
11-15.

Jour. Dairy Sci.

Cash, J, G. and Yapp, W. W.
A study of the effect of two-and three-times-a-day m i l k ­
1950
ing upon milk yield.
Jour. Dairy Sci. 33: 382 (abst.)*
Chance,
1949

3. M. and Mather, R. E.
Relation of cow families to milk production, reproductive
efficiency and longevity.
Jour. An. Sci. 8 : 603-4 (abst.),

Chapman, A. B. and Dickerson, G. E.
The relation of age at first calving to butterfat pro­
1936
duction in the first five lactations.
Amer. Soc. An.
Prod. Proc. 52-55.
Copeland , Lynn
The contribution of the dam in inheritance of milk and
1931
butterfat.
Jour. Dairy Sci. 14:
379-393.

1934

Pedigree analysis a 3 a basis of selecting bull calves.
Jour. Dairy Sci. 17:
93-102.

51
1954b

'‘Mill- raid butterfat yields 01 Jersey cows as affected
by freouency of mil-ing.
Jour. Dairy Sci. 17:
815-82.L.

lUStT

Effect of breeding efficiency and culling on herd pro­
duction.
Jour. Dairy Sci. 18: 450 (abst.).

1936

persistency of production in Jersey cows and its prac­
tical application.
Jour. Dairy Sci. 20:
151-158.

1938a

Tie use of records in evaluating tie inheritance of
cows and in the proving of uulls. Jour, -dairy Sci. 21;
651-660•

1938b

The old story of tyoe and production.
21:
295-303.

"1939
Davis, H.
1943

1938

“

A study of factors concerning
production.
Jour.
Dairy Sci.

Jour. Dairy Sci.

her i management and herd
22: 229-239.

P., Morgan, R. P. and Gaines, W.
L.
Live weight and mil : energy yield in the Psoras a
station herd.
Jour. Dairy Sci. 26: 625-641.
and Willett, E. L.
Relation between rate of growth and m i l • and fat pro­
duction.
Jour. Dairy Sci. 21:
637-642.

Dawson, J. R. and Graves, R. R.
1936
Milk and outterfat yields of Holstein cows on rations
restricted to roughage.
Jour. Dairy Sci. 19: 464-465.
Dickerson, G.E.
"1957
’ Evaluation of different measures of inherited producing
ability in dairy cattle.
Jour. Dairy Sci. 20:
457-a4v.
Dickerson,
1940

G. E.
Estiraatea of producing ability in lairy cittle.
Agr. Res. 61!
561-586.

jour.

and Chapman, A. B.
The effect of age and dry period on production at
different levels of producing auility. Amer. Soc. An.
Prod. Proc. 73-76.
1942

, and Hazel. L.H.
Effect of progeny testing on progress from sol;c fclon^
within closed herds. (Abstract) Sour. Anim. Sci. 1: 34^.

i

58
Dickey, H. C., and Labarthe, pedro
1945
Predicting the transmitting aoility of young dairy sires
for milk production, outterfat test, and outterfat pro­
duction.
Jour. Dairy Sci. 28:
893-900.
Eckies, C. H.
1927
Influence of fatness of cow at parturition on percentap;e
of fat in the mi la.
do. Agr. Exp. Sta. Bull. l O u . Xoluntria).
_

____ , and Reed, 0. E.
T912
A study of the cause of wide variation in mil
production
by dairy cows.
do. Agr. E x p t . Sta. Res. Bui 1,2 . p° ium'b'ifcu

Edwa rda , Jo s e ph.
1932
The orogeny test as a method of evaluating the dairy sire.
Jour. Agr. Sci. 22:
811-837.
1938

Effects of climatic
prod. Proc. 48-53.

factors on lives

sock.

Amer. soc. An.

Eldridge, F. E.
1948
Evaluation of dairy culls by multiple regression and a
study of variability of mil and butterfat r ec o r d s . Thesis
borne 11 Univ. Xtlfaca.)
1949

, and dal is .jury, G. IV.
The relation >f pedigree promise to performance of proved
Holstein-Friesian culls.
Jour. Dairy Sci. 32:
841-848.

Frio:, G. E., Harm, a . I. and Johnson, S.
1847
The relation of season of freshening to milk production.
Jour. Dairy Sci. 30:
o31-G40.
Ga in e s, VJ. D .
1922
The inheritance of fat content of milk in dairy cattle,
li'ner. Soc. Anim. Prod., proc. 29-32.
1927a

Heasures of persistency of lactation.
373-583.

Jour. Agr. Res. o-*:

T9271T' inheritance of persistency of lactation.
prod., proc. 37-41.

Artier. Soc. Anim.

1927c"

film yield in relation to recurrence of conception.
Dairy Sci. 10:
117-12 5.

1951'"'

Size of cow and efficiency of rail;: production.
Dairy Sci. 14:
14-25.

jour

Jour

53
1935

Correction ft.ctors ani gorw Jlasin in ■l&iry cattle breed­
ing.
Araer. Soc. Anim. 1’rod., Proc. 50-53.

""1039

Live Weight and mil,-:-energy yield in tlie -Viscons in pairy
Cow Competition.
J o u r . dairy Sci. 22;
49-53.

~ 19-xO

1946

i..-ive weight and mil.-.- energy yield in ^o latein cows .
Dairy Sci. 23: 259-265.

Jon

"Live Weight versus Metabolic 3ody Size in dairy cows
and goats.
Jour. Dairy Sci. 29: 259-272.

1931

, and Palfrey, J. K.
Length of calving interval and average rail
Jour. Dairy Sci. la; 2 94-306.

1940

,Rhode, C. S., and Oasn, G. J.
Age, live w e i g a t , and mil
energy yield in Illinois
c ow s . Jon r . Da iry Sci. 23:
1031-1043.

yield.

Gifford, w.
1950
Mode of inheritance of yearly oubterfat production; an
analysis of the orogeny perfor janoe of Holstein-Priesian
sires.
ho. Agr. tbt.pt. Sta. Res. Su lU 14 vh’aoiujnbiaM
’_93a

The outterfat records of cows possessing suyernu.merarie ;
co;.ipared with cows waving the normal number of teats.
Jour. Dairy Sci. 17:
571-573.

_________ ,an d
t in g , L . C .
1528
The effect of ages of the sire and darn on the average
butterfat production of offspring in dairy cattle.
5’ou r . Da i ry Sci. 11:
1-18.
1928

and Turner, 6 . yi;.
The mode of inoeritance of yearly butterfat production.
An analysis of the progeny perfoinance of Ayrshire sires
and darns . Mo . Agr. E x p t . Sta. Ri s. Bull* 120. (loLunibia.'v,

i•
dowen, John V*
Studies in milt secretion. V.
On the variations and
1920a
correlations of mi lie secretion with age.
Genetics 5:
111-188.
1920P

Studies in mil.: secretion.
VI.
On the variations and
correlations of outterfat percentage with age in Jersey
cattle.
Genetics 5:
249-324.

54
Gowen, Jo on W.
1920c
Studies in milk secretion.
VIII.
On the influence
of age on mil;: yield and outterfat pe rcentage as d e ­
termined from the 3S5 day records of Holstein-Priesian
cattle.
Annual Report Maine Agr. Expt. Sta. BuiI, 295:
185-196. 'CTronok
1922

The inheritance of milk yield and some of its practical
applications.
Amer. Soc. Anim, prod., Proc. 102-104.

1925a

Studies in milk secretion.
X.
The relation between
the mil.: yield of one lactation and the mil ■
- yield of
a subsequent lactation in Guernsey Advanced Registry
cattle.
Jour. Dairy Sci. 6 : 102-121.

1923b

Studies in milk secretion.
XI.
The relation Detween
the butterfat percentage of one lactation and the but­
terfat percentage of a subsequent lactation in Guernsey
Advanced Registry cattle.
Jour. Dairy Sci. 6 : 339-346.

*1924a

Interpretation of dairy pedigrees.
Annual Report Maine
A g r . E x p t . Sta., Bull. 318:
53-80. .'Orono'j

1924P

Intrauterine development of the bovine fetus in relation
to milk yield in Guernsey cattle.
jour. Dairy Sci. 7:
311-317.

1925a

Studies in mil.: secretion.
XV.
Guernsey sires' progeny
performance for mile yield, outterfat percentage, and
butterfat . Annual Report moire Agr. E x p t . Sta., Bull.
324:
37-124. COronO'.;.

"1925b ‘ Studies in mil:-; secretion.
XVI.
progeny performance ox
Guernsev sires' sons.
Annual Report maine Agr. Expt. 3 (-a.
Bull. 327:
197-252. Cronoi
1925c

Studies in mil ' secretion.
XVII.
Transmitting qualities
of Guernsey sires for mi lie yield, outterfat percentage,
and outterfat.
Annual Report Maine Agr. Expt. Sta., Bud.
32 9:
1 - ~8 . Urona.;,

1925d

Genetics of breeding better dairy stoc :.
Sci. 9:
153-170.

Jour. Dairy

55

*

Gowen, John W.
1927
Mil:-, secretion as influenced 07/ inheritance•
Quarterly Rev. of Biol. 2;
516-551.

The

1953a

On the genetic constitution of Jersey cattle, as influenc
by inheritance and environment.
Genetics 18:
415-440.

19535

Conformation of the cow as related to milk secretion.
Jersey Register of Merit.
Jour. Agr. Sci. 23:
485-513.

1934

The influence of inheritance and. environment on the
milk product ion and butterfat percentage in Jersey cattle
Jour. Agr. Res. 49: 433-465.

____________, and Covell, Mildred R.
.'..921a Studies in mil : secretion.
IX.
On the performance of
the progeny of Holstein-Eries-an sires.
Annual Report
Maine Agr. Expt. Sta., Bull. 300: 121-252. brono. '
1921b

Studies in mil
secretion.
XII.
Transmitting qualities
of Holstein-Friesian sires for uil'' yield, bunterfat per
cento j-e and outterfat.
iuaine Agr. Expt. Sta. Bulf. 301:
Orono. 5
Graves, R. R.
1926
Transmitting aoility of tventgg-three Rolstein-Friesian
sires.
U.S.D.A. Bull. 1372. fyhshin^ton.
IU47

Best records vs. the average of all records for the
evaluation of a sire’s inheritance for level of pro­
duction . Jour, dairy Sci. 30:
21-24,

TlammonO., Jorm, and Sanders, H. G.
1925
Some factors affecting milk yield.
74-119.

Jour. Agr. Sci. 13:

Hays, W. P.
1926
The effect of environmental temperature on the percent­
age of fat in cows mile.
Jour. Dairy Sci. 9: 219-55.
l-I'eizer, E. E.
1932
The inheritance of milk production and ou reterf at per­
centage In a nerd of pureored Ayrshire cattle. Artier.
Soc. Anim. prod., Proc. 25:
273-277.
Johansson, I., and Hans son, A.
1940
Causes of variation in mi Jo-: and outterfat yield of
d-iry cows. Jour, of one Royal Swedish Aca. of Agr. 79:
65-84. (Cited oj Eldridge-1948),

>56
Johnson, neslie E.
19 -5
The importance of uniformity of get in evaluating a sire.
Jour. Dairy Sci. 28:
109-120.
Kay, R. R. and wcGandlis i, Andrew C.
1929
Factors affecting tas yield and quality of milk.
age of tne cow.
Jour. Agr. Sci. 29:
342-372.
Kleioar, wax and Lead, S. W.
1941
Body size and m i l 1: production.
127-134.
1945

Body size and lactation rate.
49-55.

I. The

Jour. Dairy Sci. 24:

Jour. Dairy Sci. 28:

Klein, John W . , and Woodward, f. E.
Influence of length of dry peri od upon the quantity
1943
of mil': produced in th e subsequent lactation.
Jour.
Dairy Sci. 26: 705-713.
Leighton, R. E., and Graves, R. R.
The relation of inclir at ion of rump to in.cl ination
1947
of udder, pro-duetion aeility and oreeding efficiency.
Jour. Dairy Sci. 30:
25-40.
Ludwin, I.
1942

Tne effect of numoer of •daily mil .ings upon persistency
of rid i " production.
Jour. Anim. Sci. J:
300-308.

Lush, J'ay L.
The number of daughters necessary to prove a sire.
1931
Jour. Dair'y Sci. 14:
209-220.
T933

The bull inde ■: proolem in the light of modern genetics.
Jour. Dairy Sci. 16:
501-522.

1935

Progeny test; and Individual performance as indicators
of an animal s oreeding value.
Jour. Dairy Sci.
18: 1-

1940

Intra-sire correlations or regressions of offspring
on dam as a method of estimating aeritability of
characteristics.
Amer. Soc. Anim. prod., Proc. 293-301.

1944

The optimim emphasis on d a m s ’ records wren proving
dairy"sires.
Jour. Dairy Sci. 27:
937-951.

1915

Animal Breeding plans, Iowa State College press, Arnes,
Io wa.

J

5

?

Lusn, 9ay L. , and Arnold, Floyd
1937 Differences between records, real productivity, and
oreeding values of dairy cows. Jour, dairy Sci. 20:
440-441.
____________ , Norton, III, H. V/., and Arnold, Floyd
1941 "affects which selection of dams may have on sire indexes.
Jour, dairy Sci. 24:
695-721.
____________, and Scnrode, R. R.
1950 Changes in milk production with age end mil "ing frequency.
Jour. hairy Sci. 33:
338-357.
____________ , and Scnultz, Earl K.
1936 Heritaoility of outterfat percentage and butterfat produc­
tion in tne data with wnich sires have Deen proved in
Iowa. Jour. Dairy Sci. 19:
429-430.

1938

Pedigree promise and orogeny tea,: among sires proved in
Iowa cow testing associations.
Jour, dairy Sci. 21: 421-432.

Lush, 9. L., and Straus, F. 3.
1942 The neritaoility of outterfat production in dairy cattle.
Jour, dairy Sci. 25:
975-982.
liisner, E. G.
1939 Relation of size of cows to production and cost of product ion of mil . N. Y. (Cornell) Agr. Expt. Sta. Bull. 719..
1941

Measurements and weights of one hundred cows in the Cornell
dairy herd.
N . Y. (Cornell) Agr. Expt. Sta. Bull.771.

hatson, 9.
1929 Tne effect on lactation of tne preceding calving interval
end its relation to railing capacit?/, to age, and to other
factors of influence.
Jour. Agr. Sci. 19:
553-562.
k!1Candlish, Andrew C •
1920
Environment and oreeding as
factors influencingmil ' oroduction.
Jour. Hered. 11:
204-214.
Morrow, A. S., Keener, H. A. and Hall C. p.
1945 Analysis of certain factors involv .d in dairy nerd management
in ew Hampshire. N . H» Agr. Exp. Sta. Tec^. null. 8 6 .v|)iu±ianuh
Nelson, R. H.
1943 Measuring the amount of genetic change in a nerd average.
Jour. An. Sci. 2: 358 (a.jst.)t
j

& xicl

« J • Xj •

rosO-The effects"of mild inoreeding on a nerd of Holstein F’
riesian cattle.
Jour, dairy Sci. 33:
186-193.

56
Oloufa, Li. Ivl. a n d Jones, I. R.
1948
The relation oetween the m o n t h of calving and yearly
butterfat production.
Jour. Dairy Sci. 51:
1029-1051.
Parker," J. B. an I. M a t t h e w s , 0. a .
1950
Inheritance of outterfat test in the Beltsville Holstein
herd.
Jour. Dairy Sci. 53: 3 7 6 (abst.).
Plum. m .
1935

Oauses of differences in outterfat production of cows
in Iowa cow testing associations.
Jour. Dairy Sci. 18:
811-825.

Putnam, D. N., Bowling, 0. A. and Con d i n , 0. T.
1944
The recognition of the influence of age on outterfat
percentage when calculating mature equivalents.
Jour. Dairy Sci. 27:
905-907.
Rice, V. A.
19 j;4
A new method for indexing dairy culls.
26:
921-936.
Sanders, ii. 0.
1925
The shape of the lactation curve.
169-179.

Jo r . Dairy Sci.

Jour. Agr.

Sci. 13:

T9*27a

The variations in mi in: yields caused uy season of the
year, service, age, and dry period, and their elimination.
J.Season of the year.
Jour. Agr. Sci. 17:
389-379.

1927b

The variations in mil • yields caused oy season of t :.e
year, service, a g e , and dry period, and their elimination.
II. Service, Jour. Agr. Sci. 17:
502-523.

'1927c

The length of tne Interval
Sci. 17:
21-32.

oetween calving.

Jour. Agr.

1928a

The variations in mil-.' yields caused b; season of the
year* service, age, and dry period and tneir elimination.
111. Age. Jour. Agr*. Sci. 18: 46-67.

1928b

The variations in mi ,k yields caus mi by season of the
year, service, age, and dry period and their elimination,
IV Dry/ period, and standardization of yields.
Jour. Agr.
Sci. 18:
209-251.

$9
death* Dwight lvi.
19-iO
The intensity and ind of selection actually prac­
ticed in dairy herds.
Jour. Dairy Sci. 23:' 931-951.
Shrode, Robert R. , and Mish, ^ay L.
1947
1'he genetics of cuttle.
263.

Advances in genetics 1:

Turner, 0 . V/•
1925
A comparison of Guernsey ares.
Res. Bu 1L 79. fc'ofumbla.) -

210-

yo. Agr. Expt. 3ta.

1926

A quantitative form of expressing persistency of mil
or fab secretion.
Jour. Dairy Sci. 9; 203-214.

1927a

i-i. comparison of Guernsey sires III Based up m t ie
average persistency of fat secretion during the lactation
of the daughters.
Jour. Dairy Sci. 10:
479-500.

1927o

The mode of inheritance of yearly outterfat production.
An analysis of tie progeny performance of Jeriey sires
and dams.
ho. Agr. Expt. Sta. Res. Bu It 1 1 2 .(id!limb fa ••^■

_929

The relabion oetween weight and fat production if Guern­
sey cattle. Jour, hiiry Sci.
12: 60-73.

Tyler,
W. J., and Chapman, A. B.
1944
A simplified metnod of estimating 305-daylaeta cion
product 1 m . Jour. Dairy Sci. 27: 463-469.
1948a"

and Hyatt, G
Tme relative merits of a cow's own recor i. and her
progeny tesb for pro tic ting tie butterfat production
of ier future •daugnt era. Jour. Dairy Sci. 31: 657 (aost.).

T 9 48 b

The heritauility of official type ratings an 1 b he
correlation oetween tyoe ratings and ouster '-t pro­
duction of Ayrshire cows.
Jour. Dairy Sci. 31:
63-70.

T950

’Some' of' t ie effects of calving interv-T on mil • and
butterfat promuction of Ayrshire cattle.
Jour. Dairy
Sci. 33:
57 5-376 (aoso.),

'1948

, Chapman, A. 3. and .jickerson, G. E.
T h e meritsoility of u ode size of hoIscein Friesian
and AvrS: lIuo cu ^ ^>le . J o ur. A-*• oci • f• 516 (abot.

4

Wa shoo n,
1947
1948
1950

1950

E.
dairy sire directory and line oreeding guide Vol. I
Vol. II
dairy Sire directory, publisned oy the aut'.i5r , Salamanca
1!. 7.

Vv.

an d ■
j?y1 e r, ^ . J .
The nunioer of proved sons necess- ry to evaluate the
transmitting ability of a dairy sire.
Jour, dairy
Sci. 33:
293-298.

Vf e a v e r , B . tin d lv*a 1 1 h e w s , 0 . A .
1928 I'he influence of tera .era ture an i certain ot her factors
on tiie pe r cent age of fat i • jail . Amer. Soc. An. Prod.
j?r o c . 145-1 a9 .
■Vood’.vard, T. E.
1927
Influence of fro planes of feeling and cnre noon milk
prO'.iuction.
Jour, dairy Sci. 10:
285-291.
'iriyit, Sew all
1931
On t.ie valuati )n of dairy
Proc. 71-78.

sires.

Amer. Soc. An in. prod.

yapp, y. v.
1919 a study of tne relative reliaoility of official tests of
lair;/ cattle. Ill Agr. Expt. Sta. Bult 215. (Hfbana* )•

161

Ao )endlx Pauls I
Bur:-- 32 reg. i o. 43.2017
Data I'or Sons of iviara fcii >n Bet
O'vn daughters ave. 448, 'antes 479 ox - 305 da. 13.S.

Bull Reg. to.

Dau. prod.
1

o*0046
609774
540339
565365
565366
504402
620920
650015
553353
548515
640043
528469
522685
531686
588188
675184
504401
650025
675185
537726

523
457
546
.-87
47 6
581
564
492
410
406
378
400
487
478
390
490
400
449
521
405
*57

pates prod.
-^1

-1-99
41.7
442
420

449
573
406
458
440
402
382
516
461
500
424
444
439
423
529
442
45?

Doaas prod.
X5
477
477
477
441
441
483
376
542
586
559
506
447
461
469
526
525
520
532
598
661
505

prod.
of
462
462
462
430
430
430
430
430
430
430
430
430
455
467
656
490
467
417
464
456
457

dau.

62
Appendix iable II
2X production of daughters and mates of culls proved in
fie Traverse City and Ionia Reformatory lords.
Travers e C i ty Ru11 s
Reg. Wo.
815094
553553
609774
700278
412017
486040
566774
729194
659862
353211
721464
480572
65002 5
522685
720535

Dau. Prod. pates Pro i
506
335
341
329
330
347
347
546
359
343
335
536
540
335
363
555
346
345
345
355
541
343
349
548
356
5 58
352
358
543
559
342
345

Ionia Refonnatory Bulls
g. Wo.
3‘,3285
808309
675183
629478
480572
774578
401108
576509
87 3672
545551
o22685
748700
694844
5*,2oo5

Dau. Prod.
367
424
375
37 6
365
599
398
4 .8
400
389
399
408
413
320
389

gates proc
354
424
391
405
383
413
348
408
419
406
372
390
407
500
387

6$

Appendix 'Cable III
3X 1Production data for Institution iierd-ored uulla other than
dons of viaratiion Bess Bur -e 32
Prod. of Sir
Dau. prod.
M ates •f’rod.
Bull Reg. ri0 •
B a m s ’ Prod.
Dau.
Mates
Y
X4
X3
X5
442
532
5-::5551
430
-±67
573
430587
497
431
307
467
430
409
425812
391
466
375
366
7 56888
540
48 3
739
581
573
504507
660470
353
581
573 ■
487
728591
545
644
573
581
6017 52
498
443
546
581
573
634064
550
490
532
581
573
653204
429
419
562
400
439
436
466
694843
505
515
540
72.1464
403
443
495
410
440
663129
408
473
579
598
503
694844
529
575
598
515
540
464
448
462108
598
480
430
551021
436
654
425
515
540
441
459
532
500
720335
430
450
532
700278
470
491
512
480
596
471
602293
407
425
442
4-27
650022
408
431
443
497
591
701944
481
473
434
453
566
473
486
434
670896
529
579
503
598
568009
463
388
465
401
400
436
714235
m
465

Appendix Taole lv
3X production of daughters and nates of purchased
bulls proved in institution herds
Bull Reg. ’
ho.
609133
678668
629478
659862
671583
744578
729194
600154
432090
573627
477989
344502
696989
341217
700526
602205
700089
708507
574194
621262
738394
708582
742591

Dau. Prod.
417
531
515
500
587
596
520
528
499
456
473
453
538
499
581
612
617
600
554
624
478
488
5o 9

Mates’ Prod.
459
463
540
430
581
534
497
493
546
448
434
451
519
406
486
652
532
504
645
577
490
425
558

Bull Reg. ho. Dau
prod.
486040
682582
493589
651543
725465
58 5068
713093
67804 5
646346
566744
697436
480572
556586
519074
290298
353211
401108 :
746901
618734
669545
669566
708580

458
471
498
480
419
469
503
446
393
431
401
491
455
474
436
467
529
446
543
410
576
560
500

Mate
Prod
439
428
450
460
415
427
489
517
420
443
436
512
444
569
364
431
463
507
573
442
518
501
484

65
Aooendix 1'aule V
2 X Production data for sons of -^ing Bessie Ormsoy Pietertje
R e . Wo, 520107, Dau. prod. 444 , m a t ’;s 504
Bull Reg. Ho
717602
752532
723901
727778
763855
688260
644439
593854
702034
695524
738394
708578
652274
669545
669547
669546

688266
711322
676466
763855

Dau. Prod,
Y
461
381
371
382
435
424
359
389
390
352
341
389
392
323
316
393
379
372
418
411
384

Hates Prod.
45^
381
316
361
419
462
332
390
373
397
367
394
351
324
336
374
376
383
410
374
37^

D a m s 1 prod.
X3
465
484
541
419
552
483
483
419
525
458
500
513
605
561
552
485
552
552
570
471

J