SOME FACTORS ASSOCIATED WITH THE RATE OF MILK FLOW
IN DAIRY CATTLE

By
Eugene Wesley Hupp

AN ABSTRACT

Submitted to the School for Advanced Graduate Studies
of Michigan State University of Agriculture and
Applied Science in partial fulfillment of the
requirements for the degree of

DOCTOR OF PHILOSOPHY

Department of Dairy
1953

Approved

EUGENE WESLEY HUPP

ABSTRACT

This study was undertakento investigate the effect of
several

factors involving milk let down and rate of removal

of milk from the udder of the dairy cow.
In the first experiment, the effect of addition of
water to the concentrate mix during milking was investigated.
Water was added at the rate of V/z pounds of water per pound
of concentrate mix.

This change reduced the eating time 30

percent

for the 16 cows used. The maximum rate of milk flow

was not

changed significantly, but the average rate of flow

decreased 12 percent.

However, this statistically signifi­

cant decrease was due primarily to increased time spent by
the machine operator during stripping.

This cannot be at­

tributed to the effect of addition of water to the concen­
trate mix upon the milk flow characteristics of the cow.
In the second experiment, intravenous injections of
20 International Units of Oxytocin were given immediately
before milking to assure a complete let down of milk.

These

injections had no effect on rate of flow in trial 1 involv­
ing seven slow milking cows, and only a slight effect on
rate of flow in trial 2 in which seven average and fast
milking cows were used.

This effect on the faster milking

cows may have been due to a greater increase in intra-mammary
pressure following oxytocin injections.

These changes in

2
ABSTRACT

EUGENE LESLEY HUPP

rate of flow were small compared to differences in rate of
flow between cows.

This indicates that let down failure was

not responsible for the slow rate of flow of all but one of
the cows studied.

One cow did exhibit a faulty let down

which was quite different from the normal let down pattern
for that cow.
The effect of varying the interval between stimulation
of let down by preparation of the udder by a warm water wash
and use of the strip cup and attachment of the teat cups was
also investigated.
and 20 minutes.

The intervals used were 0, 4-, 8 , 12, 16,

Ten cows were milked at three milkings fol­

lowing each of these intervals.

The maximum rate of flow

was significantly faster when the teat cups were attached
immediately after preparation.

This change in rate of flow

was quite small, but could be responsible for some milking
to milking variation.

Milk production and butterfat percent­

age decreased slightly but not signficantly as the interval
between preparation and milking increased.

These results in­

dicate that the shorter intervals between stimulation of let
down and attachment of the teat cups did result in a more
complete emptying of the udder.
Data were presented to establish the precision of the
measuring devices employed in these studies to determine the

3
ABSTRACT
rate of milk flow.

EUGENE WESLEY HUPP
The Milk-O-Meter Had a gross error of

2.2 percent and a net error or bias of minus 0.5 percent
in tests in tHe barn involving 54-5 actual milkings.

The

spring scales used Had an error of 0.4-3 percent wHen weigHing weights of known mass and 1.4-3 percent under milking
conditions.

SOME EACTOES ASSOCIATED WITH THE EATE OE MILK ELOW
IE DAIEY CATTLE

By
Eugene Wesley Hupp

A THESIS

Submitted to the School for Advanced Graduate Studies
of Michigan State University of Agriculture and
Applied Science in partial fulfillment of the
requirements for the degree of

DOCTOE OE PHILOSOPHY

Department of Dairy
1958

ProQuest Number: 10008539

All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

uest
ProQuest 10008539
Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author.
All rights reserved.
This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106- 1346

TABLE OE CONTENTS
Page
INTRODUCTION .....................................

1

REVIEW OE LITERATURE .............................

2

Development of the Milking Machine

..........

2

....

6

Repeatability of Observations ................

8

Differences Between Individual Cows ..........

9

Methods of Measuring Rate of Milk Elow

The Role of Let Down in Controlling Rate of
E l o w ...................................

13

Methods of Stimulating Milk Let D o w n ........

26

The Effect of Adding Water to the Concentrate
Mix Eed During M i l k i n g ..................

29

The Effect of Variation of the Interval Between
Preparation and Milking ................

31

The Effect of "Managed Milking" Routines

...

37

The Role of Teat Morphology..................

41

Other Eactors Associated with Rate of Elow

46

. .

OBJECTIVES.......................................

57

SECTION A
EXPERIMENTAL PROCEDURE...........................

58

Part I. The Effect of Adding Water to the Con­
centrate M i x ............................

58

Part II.

62

The Effect of Oxytocin Injections . .

Part III. Variation in Interval from Prepara­
tion to M i l k i n g ........................

66

TABLE OE CONTENTS
(continued)
Page
RESULTS AND DISCUSSION . . . .

..................

72

Part I. The Effect of Adding Water to the Con­
centrate M i x ........................
72
Part II.

The Effect of OxytocinInjections

. .

78

Part III. Variation in Interval from Prepara­
tion to M i l k i n g
.

95

SECTION B
DETERMINATION OE THE ACCURACY OE MILK MEASURING DE­
VICES .......................................

108

Part I.

The Milk-O-Meter................

108

Part II.

Spring S c a l e s ..................

129

SUMMARY AND C O N C L U S I O N S .........

132

LITERATURE C I T E D .......................

135

LIST OF TABLES
TABLE

PAGE

1

The effect of feeding a watered concentrate
mix on the maximum rate of milk flow, aver­
age rate of milk flow and average eating
r a t e ................................
73

2

The effect of feeding a watered concentrate
mix on average milk production, average
milking time and average rate of milk flow
for the three components of the milking
75
c u r v e ..............................

3

Analysis of variance. Sources of variation
in maximum rate of milk flow intrial 1
. .

80

Milkings arranged in ascending order by
maximum rate of flow for the two oxytocin
treatment trials..... ......................

81

4-

5

Milkings, with and without oxytocin treat­
ment, arranged in ascending order by average
rate of f l o w ........................
84-

6

Milkings, with and without oxytocin treat­
ment, arranged in ascending order by level
of p r o d u c t i o n ......................
85

7

Milkings arranged in ascending order of the
number of seconds before a high rate of flow
was reached, with and without oxytocin
treatment............................
87

8

Milkings arranged in ascending order of the
number of seconds a high rate of flow per­
sisted, with andwrithout oxytocin
treatment.

88

Milkings arranged in ascending order of the
number of seconds between the end of a high
rate of flow and the time stripping started
with and withoutoxytocin treatment . . . .

90

Milkings arranged in ascending order of
stripping time in seconds, with and without
oxytocin treatments ............ . . . . .

91

9

10

LIST OF TABLES
(Continued)
TABLE
11

12

13

14

13

16

17

18

19
20

PAGE
Analysis of Variance. Sources of variation
in maximum rate of flow when the interval
from preparation to milking was varied. . .
The influence of the time elapsing between
preparation of the udder and the attachment
of the teat cups upon the milk flow. Milk­
ings arranged in ascending order of maximum
rate of flow.............................

96

97

The average rate of flow and the mean pro­
duction for each of the six intervals be­
tween preparation and milking ............

98

The mean number of seconds in each of the
four components of the milk flow curve for
each of the six intervals between prepara­
tion and milking
..................

100

The mean number of pounds of milk produced
during each of the four components of the
milk flow curve for each of the six inter­
vals between preparation and milking. .. .

102

The average rate of flow during each of the
four components of the milk flow curve for
each of the six intervals between prepara­
tion and milking.........................

103

The percentage of butterfat, total solids,
solids non-fat and protein of milk follow­
ing six different intervals between prepa­
ration and milking.......................

104

The gross and net error of four types of
Milk-O-Meters in estimating the production
at a single mi l k i n g........ * ...........

116

The effect of rate of flow on gross and net
error in estimating daily production. .. .

119

The effect of level of production upon
gross and net e r r o r .....................

121

LIST OF TABLES
(Continued)
TABLE

PACE

21

The gross and net error of different meters.

124

22

The amount of milk removed by different
combinations of meter, sampler, and gaskets.

124

23

24

The percentage error of the three milk-ometers used in part II at four rates of flowbefore and after adjustment.................
The percentage error of the recording milko-meters before and after the experiments in
part III, and with the pin-type baffle tray.

126

128

LIST 01 FIGURES
Figure
1

Page
The Milk-O-Meter in operation in the tiestall main dairy b a r n ....................

64-

The recording Milk-O-Meter with the butterfat sampler attached and the recorder . . .

68

Inside detail of the recorder used with the
recordingMilk-O-Meter.....................

69

4-

Two

92

5

The

6

The flow apparatus for testing the Milk-OMeter ...................................

2
3

milking curves for cow No. 3006.

...

Milk-O-Meter, showing component parts .

109
113

ACKNOWLEDGEMENTS
The author wishes to express his sincere gratitude
and thanks to Dr. R. C. Lewis, Prof. W. W. Snyder, and Dr.
E. P. Reineke for guidance and help in conducting this
study.
The investigator appreciated the financial support
provided by the National Science Foundation through a fel­
lowship granted during a portion of this study; Michigan
State University through a Graduate Research Assistanceship;
and the taxpayers of America through benefits provided under
Public Law 550.
The author is indebted to Dr. N. P. Ralston, Chairman
of the Department of Dairy, for making facilities and ani­
mals available, and to Mr. E. S. Smiley, herdsman, and his
assistants in the barn whose cooperation implemented the
collection of the data.
Finally, I am deeply indebted to my wife whose assist­
ance and encouragement has been invaluable.

INTRODUCTION
Modern milking machines have greatly increased the
efficiency of milking cows.

However, the physiological

mechanisms that control the rate of milk removal are still
not well understood.
Two groups of conditions may be considered to act in
controlling the rate of milk flow from the udder.

The

first may be referred to as let down and involves the trans­
fer of milk from the alveoli to the teat and gland cisterns
and the pressure which it exerts.

The second is the ten­

sion of the teat sphincter which surrounds the streak canal.
This controls the pressure required to open the teat sphinc­
ter as well as the degree to which it will open under a
given set of conditions.
The object of this investigation has been to ascertain
the effect of several factors involving milk let down and
rate of removal of milk from the udder.

-

1

-

REVIEW OF LITERATURE
Development of the Milking Machine
The perfection of a machine that would rapidly and ef­
ficiently remove the milk from the cow's udder has long been
the aim of inventors and investigators.

Although the milk­

ing machines in operation today are all of the intermittent
suction type, the methods of achieving the desired amount
of intermittent suction and transporting the milk from the
teat cups to the storage receptacle are as varied as Ameri­
can inventive ingenuity can make them.

In spite of these

developments, research in machine modifications is still
being conducted.
Erf (1906) stated that American inventors attempted
to perfect milking machines as early as 1819.

He described

many types of milking tubes and pressure devices designed
to remove milk from the udder.

Woll and Humphrey (1909)

reported that the first American patent on a continuous
suction machine was obtained in 1860.
Matthews et a.1 , (1928) credited Scottish inventors
with the development, in 1902, of a milking machine which
successfully used intermittent suction.

William Lawrence

and Robert Kennedy patented the machine which included a
pulsator on the cover of the milk receptacle and a vacuum
-

2

-

3
regulating device*

D. H. Burrell of Little Falls, New York,

made some improvements on this machine and in 1903 marketed
the first modern American milking machine as the BurrellLawrehce-Kennedy cow milker.

The machines in operation

today are all improvements of this intermittent suction
type milker.
Recently, several types of adjustments in operation of
the milking machine have "been investigated.

Several groups

of workers, Smith and Petersen (194-6), Baxter et al. (1950),
G-regoire et al. (1954-), Caroulo et al. (1955)j and Stewart
and Schultz (1956 & 1958) have all reported increased milk­
ing rate with increased level of vacuum.
Smith and Petersen (194-4-) and Baxter et al. (1950) re­
ported that the increase in rate of flow was greater than
might be expected due to change in vacuum level alone, and
postulated that increased vacuum levels also cause the
sphincter to open wider.
However, there are limits beyond which the vacuum
level should not be increased, since such increases may
cause udder injury.
(1955)•

(For a discussion see Mochrie et al.

Petersen (194-4-a) also reported that increased

vacuum levels increase the time required for stripping due
to a greater tendency for the teat cups to crawl upward as
the rate of flow decreases near the end of milking.

4
Although it is generally agreed that increased vacuum
increases the rate of flow of milk from the udder, there
are reports in opposition to this view.

Whittlestone and

Verrall (1947) found no significant difference in milking
rate when cows were milked at a vacuum of 10 and 19 inches
of mercury.

Only in the case of an extremely hard milker

was there a noticeable change, and the difference here was
not significant.

Dahlherg (1943) noted that the amount of

suction, within reason, did not materially vary the rate
of milking.
Smith and Petersen (1946) also reported changing the
pulsator ratio so that vacuum was applied in the ratio of
2:1 or 3 si of vacuum to release, rather than the 1:1 of con­
ventional milkers.

Phis resulted in an increase in rate of

flow comparable to that obtained with a two inch increase
in vacuum level.

Phere was a lesser increase in rate of

flow with the wider pulsator ratio’s as higher levels of
vacuum were used.

It was postulated that at the higher

vacuum levels, maximum drainage rates from the udder had
been reached.

Clough and Dodd (1956) also reported increased

rates of flow with 2:1, 3:1, and 4:1 ratio's of vacuum to
release, when compared to the normal 1 :1.
Phe effect of varying the pulsator rate has also been
investigated.

Using a standard milking machine which

5
normally operated at 20 pulsations per minute, Clough et al.
(1953)

found that decreasing the pulsator rate to 20 pulsa­

tions per minute decreased the peak rate of flow 18.9 per­
cent.

Increasing the rate to 80 pulsations per minute in­

creased the peak flow 9.7 percent.

They point out that the

milking machine must allow a complete evacuation of the
pulsation system during each cycle if the higher pulsation
rate is to he effective.

In a later publication, Clough

and Dodd (1956) reported increased rate of flow with pulsa­
tor rates of 110, 14-0, and 170 pulsations per minute.
Stewart and Schultz (1956, 1958) reported the effect
of variations in pulsation rate on the rate of milking in
slow milking cows.

Decreasing the pulsation rate from 50

to 20 pulsations per minute resulted in a 0.5 pound per
minute decrease in maximum rate of flow.

Increasing the

pulsation rate from 50 to 80 pulsations per minute resulted
in only a very slight increase in maximum rate of flow.
However, loot (1951a) reported very little difference
sind Whittlestone and Yerall (194-7) found no difference in
milking rate which could be attributed to pulsation rate.
As pointed out by Clough et al. (1953) these conflicting
reports may have been due to the failure to completely
evacuate the pulsation system during each cycle.

6
Methods of Measuring Rate of Milk Flow
The evaluation of the effect of machine modifications
on the rate of milk flow and the recognition of the differ­
ences in rate of flow "between cows has created the need for
methods of measuring rate of flow.

Early investigators

such as Woll and Humphrey (1909) and Smith and Harding
(1912) determined the rate of milking "by the time required
to milk the entire herd "by machine.

Dahlherg (1935) and

Elting and La Master (1930) used the average milking time
required per cow to test different milking procedures.
Dalton et al. (1953) also reported the total milking time
per individual cow.
Matthews et al. (1928) were the first to report meas­
urements taken during the course of milking.

The milker

pail was placed on a scale during the milking process.
The amount of milk accumulated in the pail at one minute
intervals was recorded.

Many other workers, Foot (1935),

Petersen (194-3), Smith and Petersen (1944), Korkman (1948),
Dodd and Foot (1947), Baxter et al. (1950), Gregoire et al.
(1954) and Stewart and Schultz (1950) have measured the ac­
cumulated amount of milk as indicated "by scale readings at
various intervals throughout the milking process.

The

length of interval used by different workers has varied
from 10 to 60 seconds.

7
Hupp and Plum (1957) reported the use of a system in
which the maximum rate of flow was determined by measuring
the time required to produce the five pounds of milk in the
interval between five and ten pounds of milk produced*

This

method avoids the difficulty of attempting to read the scale
at an exact instant.
Several groups of workers have used continuous flow
kymographs to obtain a graphic record of the amount of milk
as it accumulated in the milking receptacle*

The use of

such a device was first reported by Gaines (1927)*

Beck et

al. (1951a & 1951b), Roark et al. (1952), Thoele (1954-), and
Ace (1957) have also reported the use of a continuous flow
kymograph.
Matthews et al. (194-1) used a milking machine which
transferred the milk from separate quarters to four cylin­
drical containers fitted with a graduated scale.

Hall (1951)

reported the use of a similar device designed to give instant
transfer of milk to the cylinders.
Smith and Petersen (194-6) used two vacuum flasks joined
to the milking machine pail by a vacuum hose.

With this de­

vice they could transfer milk from one teat to the vacuum
flask with one hose while the milk from the other three quar­
ters was transferred to the milker bucket by separate hoses.
Through the use of a ’T 1 tube the milk was deflected into one

flask until full flow was obtained, then shunted into the
second flask for exactly one minute so that the amount of
milk produced by an individual quarter in one minute of
maximum flow under various experimental conditions could
be measured.
Knoop and Monroe (1950) used a special Surge milker
with two latex bags within the milk receptacle.

By means

of a valve the milk could be deflected into either of the
bags or into the milker receptacle proper.
Whittlestone and Phillips (1953) described an automa­
tic apparatus which they designed for study of milk ejection
curves of individual animals.

An ordinary set of teat cups

was used and the milk was transferred from the teat cups to
a special receptacle.

Through a series of electrical cir­

cuits which plotted the curve automatically this apparatus
recorded the accumulated milk in relation to elapsed time.
It also determined the end point of milking as the time at
which rate of flow dropped to a pre-determined level.

The

apparatus was completely controlled by the milking machine
operator, so that an additional observer was not required
during the milking procedure.
Repeatability of Observations
Through the use of measuring methods just discussed,
various workers have shown that the repeatability of

9
observations on individual cows, when milked under the same
conditions, is high.

Beck et al. (1991a.) reported that in

three successive afternoon milkings of 26 cows, the coeffi­
cient of intra-class correlation for maximum rate of flow
was 0.96.

The coefficient of correlation, based upon 12

cows, for observations taken once weekly from the fifth to
the tenth week was 0.89»

The correlation coefficient, based

upon 25 cows, measured at the same stage of lactation in two
successive lactations was 0.94.

Thoele (1954) reported the

coefficient of intra-class correlation of 1118 milkings of
58 cows to be 0.95.

Ace (1957) reported repeatabilities of

0.904 on morning and evening milkings, and 0.955 considering
morning and evening milkings separately.
Whittlestone (1946) and Beck et al. (1951t>) have shown
that each cow has a rather characteristic flow curve that
changes little in appearance from day to day.
Differences Between Individual Cows
While the repeatability of rate of flow within indivi­
dual cows is high, it has long been recognized that indivi­
dual cows differ greatly in their response to any method of
milking that may be employed.

Gaines (1927) noted differences

between individual cows in the rate at which they milked when
milked under the same conditions.

Matthews et al. (1928)

10
reported that individuals producing at the same level differ
markedly in the rate at which they can he milked.

The high­

est producing cow observed milked very rapidly while other
high producing cows milked very slowly.
Elting and LaMaster (1936) reported that when grouped
by level of production the average total milking time of
each group of cows was almost equal.

For individual cows,

large differences were found in total time required per cow.
The shortest time required was 3.37 minutes, while the long­
est time required was 8.28 minutes.

Seven percent of the

cows milked out in four minutes or less, 81 percent milked
out in 4 to 6.5 minutes, and 12 percent required 6.5 minutes
or more.
Whittiestone (1946) discussed in detail the milk flow
curves of 12 cows which were recorded over the entire lacta­
tion.

He showed, graphically, individual differences between

cows and variations in response during the lactation.
Beck et al. (1951^) studied the response of individual
cows to a standard method of machine milking.

The studies

were conducted on 102 animals of various ages and represent­
ing four breeds.

The average milking time was 3«5 minutes

per cow, with 79 percent of the cows milking out in four
minutes or less, but only 39 percent of the cows milked out
in three minutes or less.

11
Dodd (1953) also reported large differences in milking
rate within the herd studied,

The total time required

varied from 4.42 minutes to 21.30 minutes.

Only 2.8 per­

cent milked out in five minutes or less, while 8.3 percent
of the cows observed required over 14 minutes to be milked
out.

The peak rate of flow ranged from one pound to 10.4

pounds per minute.

Only 3.4 percent of the cows had a peak

flow of less than two pounds per minute while only 1.5 per­
cent had a peak flow greater than nine pounds per minute.
Sixty percent of those observed had a peak flow of from 2
to 4.9 pounds per minute.

The coefficient of variation be­

tween cows for the various measures of rate of milking were
in the range of 30 to 40 percent while the within cow vari­
ations were on the order of 8 to 11 percent.
Petersen (194*3) also found great variation in the rate
of milking between cows, while Thoele (1934-) reported dif­
ferences in the rate of milking between cows to be signifi­
cant at the one percent level of probability.
Whittlestone (1954) presents four types of milk flow
curves which he describes as representing most of the cows
which will be encountered.

The first and most desirable is

the cow that starts immediately and milks out rapidly with
little machine stripping.

The second starts immediately,

but milks out slowly due to the tightness of the teat

12
sphincter#

In the third a large amount of machine stripping

is required to remove the last milk from the udder.

In the

fourth type, the cow is described as having a faulty let­
down response.

The milk is not rapidly transferred to the

gland cistern; then a second letdown or release of more
letdown hormone occurs upon stimulation of machine strip­
ping.

Whittle stone attributes much of the difficulty in

slow milking to this

type of response.

Several workers

have investigated

the contributionof

the individual quarters to the individuality of the milking
response of each cow.

Matthews et al. (1941) found that in

most cows there were wide differences in the performance of
the separate quarters of the udder.

There were wide vari­

ations between individuals in milking rate and in the extent
of the differences between quarters.

Of the 94 cows observed,

48.9 percent showed a difference of 20 percent or more be­
tween the quarter with the lowest percentage of its total
yield milked out following the first 2.5 minutes of milking.
A difference of 0.6 pounds per minute or more between the
quarter with the highest peak flow and the quarter with the
lowest peak flow was encountered in 40.4 percent of the cows.
Smith and Petersen (1956) also presented data showing
considerable variation in
within the same cow.

rates of the

individual quarters

The reaction of the individual

13
quarters to increased levels of vacuum also varied.

In one

cow an increase of five inches of mercury in the vacuum
level resulted in a 111.4- percent increase in mean rate
of flow in the quarter showing the greatest response and
only 60.8 percent in the quarter showing the least response.
The Role of Let Down in Controlling Rate of Flow
The term let down will he used in this thesis even though
a considerable degree of controversy exists as to the proper
term to he used to describe the phenomena generally known to
the dairyman as let down.

This word itself has been written

"let down," "letdown,11 let-down," and "letting down," and all
have been written both with and without quotes.

A group of

British workers (Cowie et al., 1951) proposed that milk
ejection, a term used as early as 1915 by Gaines, be adopted
for uniform use.

However, Turner (1954), Whittlestone et al.

(1952) and others have continued to use let down.

Other

terms have also been used; "draught" is a term which has
been frequently used in human medicine; Espe and Smith (1952)
prefer to use the term expulsion, while Kay (1955) and others
have used pertrusion to describe this phenomena.

Thus there

seems to be no common agreement on the proper term to describe
this phenomena.
The various aspects of milk let down have been the sub­
ject of a great deal of investigation.

Although the effect

14
of one of tJie hormones of the posterior pituitary was
described as early as 1895 by Oliver and Schafer, and the
structure of both oxytocin and vasopressin have been eluci­
dated, as reported by du Vigneaud et al. (1953 and 1954)
there is much to be learned concerning the mechanism of re­
lease and the action of these hormones.
In contrast to the other areas covered in this review,
the mechanisms involved in milk let down have been considered
in several review articles.

Various phases of the problem

have been covered by Petersen (1944c), Volley (1947), Cross
(1955), Kay (1955) Van Dyke et al. (1955), Benson and Cowie
(1957), and Hammond (1957)•

Since these reviews are readily

available, only the more significant material necessary for
an understanding of the neurohormonal mechanism of milk let
down will be considered here.
Elucidation of the let down mechanism.
Gaines (1915) was the first to demonstrate that the
suckling stimulus caused milk let down.

He reported that

suckling of one gland of a goat by its kid caused a copious
flow from the other (cannulated) gland about 50 to 60
seconds after suckling commenced.

He stated that practi­

cally all of the milk obtained was present as such in the
gland at the beginning of milking and was "actively ejected
by a reflex contracture of the gland musculature under the

15
stimulus of milking."

Gaines also investigated the effect

on the mammary gland of injection of posterior pituitary
extract.

(The effect of these extracts upon the mammary

gland had been reported by Ott and Scott in 1911.)

Although

Gaines concluded that the extract caused a constriction of
the milk ducts and alveoli, with a subsequent expression of
milk, he did not implicate the pituitary in milk ejection,
but considered milk let down to be a purely reflex act.
Turner and Slaughter (1950) suggested that the let down
reflex might act through the posterior pituitary.

However,

they concluded that it was still an open question whether the
posterior pituitary played an active part in the normal dis­
charge of milk during milking.
In 1940 and 1941 Ely and Petersen reported the results
of experiments which demonstrated the neuro-hormonal nature
of milk let down.

They stated that various stimuli such as

manipulation of the udder, sight of the milker or calf, or
sounds associated with milking, stimulate let down.

The re­

flex nature of the stimulation was substantiated by demonstra­
tion that an interruption of the normal milking routine or
injection of adrenalin inhibited let down.

The hormonal

nature of the efferent side of the reflex was demonstrated
by showing that let down occurred in the half of the udder
in which the efferent nerves had been interrupted through

16
sympathectomy at the inguinal ring.

The hormonal nature

was further demonstrated by showing that blood which had
been collected from cows after stimulation of let down, when
perfused through an isolated cow's udder, produced let down
in that udder.
Demonstration of the afferent mechanisms.
Although the report of Ely and Petersen (194-1) that
sight and sound as well as tactile stimuli can initiate let
down in the cow has been well substantiated, several inves­
tigators have attempted to elucidate the afferent nerve path­
ways from the mammary glands.

The work of Peeters et al.

(1952), Denamur and Martinet (1954-), Tsakhaev (1954-), and
Eayrs and Baddeley (1956) offer conclusive evidence that
the afferent stimuli from the mammary glands enter the cen­
tral nervous system via the dorsal roots of nerves supplying
the gland or glands stimulated and are conveyed up the spinal
cord deep in the lateral funiculus of the same side.
Grosvenor and Turner (1957), through injection of vari­
ous drugs which inhibit cholinergic or adrenergic responses,
concluded that both cholinergic and adrenergic links exist
in the neuro-hormonal arc which is responsible for the dis­
charge of oxytocin by the posterior pituitary.
The effect of stimulation of the reproductive tract
upon the let down of milk has been reported by several workers.

17
Hays and Van Demark (1955) cited the review of literature of
Nuesch (1904-) who reported a publication in 1727 in which
the Hottentots brought about milk let down in the absence of
the calf by blowing air into the vagina.
Piana (1950) and Andersson (1951) reported milk let down
following massage of the cervix and uterus per rectum.

Hays

and Van DeMark (1955) reported increased intramammary pres­
sure in 89 of 99 trials following artificial insemination
techniques.

Harris and Pickles (1955) presented evidence

of let down in lactating women following coitus.

The exact

relationship responsible for this phenomena is not clear;
however Hays et al. (1955) and other workers have suggested
that oxytocin may play a role in sperm transport in the fe­
male following copulation.

Thus the same hormone may play

a role in both the mammary and reproductive systems, and a
stimulus in either system which brings about a release of
oxytocin evokes a response in both systems.
Establishment of the role of the pituitary.
Although the action of exogenous oxytocin in causing
let down has long been known, and Ely and Petersen (194-0)
have clearly demonstrated the hormonal nature of the efferent
portion of the reflex, much additional work has been done in
an attempt to demonstrate the role of the pituitary in normal
milk let down.

Several approaches to the problem have been

18
made.

The Gomori chrome Alum hematoxylin stain appears to

he specific for demonstrating the neurol lobe hormones,
Brightman (1955) reported that there was less of the neuro­
secretory material (as demonstrated by the above method)
in the pituitaries of mice killed immediately after nursing
their young.

However, Drager and Rennels (1955) and Rennels

et al. (1955) have been unable to demonstrate a relation­
ship between the Gomori-positive material and conditions
which should cause depletion of oxytocin in the pituitary.
Several workers have reported the results of electri­
cal stimulation of the pituitary and hypothalmus.

Cross and

Harris (1950) reported that an electrical stimulation of the
pituitary stalk caused a flow of milk from a cannulated duct
in an anesthetized rabbit which was similar to that obtained
by injection of 0,2 IU. of posterior pituitary extract.

In

1952 the same workers reported similar results on stimulation
of the supraoptical-hypophysial tract, which supplies nerves
to the posterior pituitary.

Andersson (1955) reported milk

let down in sheep following electrical stimulation of the
anterior hypothalmus.
Cross and Harris (1952) also showed that electrolytic
lesions in the supraoptico-hypophysial tracts of lactating
rabbits caused a marked diminution in the quantity of milk
obtained by their litters.

19
In the light of the above reported evidence, and other
data, there can be little doubt that the release of oxyto­
cin from the posterior pituitary takes place as a result of
nervous stimuli.
In connection with the role of the pituitary, it should
be mentioned

that recent work has established that the hor­

mones of the

neurohypophysis are actually produced in the

supraoptic and paraventricular nuclei of the hypothalmus and
that the pituitary only serves as a source of storage until
the hormones

are released.

This work has been reviewed by

Scharrer and

Scharrer (1954-) •

The nature of the effector tissue.
The elucidation of the nature of the contractile tis­
sue responsible for the phenomena of let down has been diffi­
cult.

Hammond (1915) presented evidence that the pituitary

extracts acted directly upon the glandular epithelium.

How­

ever, Schafer (1914-) proposed that contractile tissue which
he called "plain muscle" in the walls of the alveoli was re­
sponsible for let down.
Schafer's view persisted for some time; Turner and
Slaughter (1950) attributed the effects of oxytocin to action
on smooth muscle surrounding the alveoli.

Swanson and Turner

(194-1) reported the presence of smooth muscle cells in the
udder of the cow to which they attributed the action of
squeezing out the milk.

20
Richardson (194-9), using a silver staining technique,
was the first to clearly demonstrate that an abundance of
myoepithelial tissue existed in the udder.

It was shown to

be arranged in such a manner that it could actively squeeze
the milk from the alveoli.

Linzell (1952) confirmed the

results of Richardson, and also demonstrated that there was
not enough smooth muscle present in the udder to cause con­
traction of the alveoli.

He also reported that the myoepi­

thelial cells are not innervated, therefore they can respond
only to direct stimuli such as that produced by hormones.
Silver (1954-) criticized silver impregnation techniques as
being rather capricious, but confirmed the earlier work of
Richardson (194-9) and Linzell (1952) by demonstrating the
presence of myoepithelial cells with an alkaline phosphatase
stain.
The above conclusions have been further substantiated
by Linzell (1955 & 1955) through direct examination of the
mammary gland of living mice.

He observed contraction of

the alveoli following topical application of oxytocin or
vasopressin.

He also reported that the smaller ducts,

around which the myoepithelium is arranged longitudinally,
shortened and widened in response to oxytocin.

He presented

evidence that this shortening and widening is not due to
pressure of the milk alone.

21
Cb.aB.ges in intramammary pressure.
Turner (1935) cited the work of Tgetgel (1926) wbo
found that the intramammary pressure in the cow before let
down ranged from 25 to 35 mm. of mercury and after let down
the intramammary pressure ranged from 40 to 60 mm. of mer­
cury.

Whittlestone (1955) reported similar pressures, and

reported only a slight drop in pressure after 22 minutes if
the cow's udder was unmilked.

He stated that this is in

contrast to the sow (which he has studied extensively) in
which a rapid rise and fall in pressure is noted.

He sug­

gested some type of non-return mechanism which is responsi­
ble for the sustained pressure after the effects of oxyto­
cin are presumed to be dissipated.

However, such a mechanism

has not yet been demonstrated anatomically.
Baxter et al. (1950) proposed that changes in internal
pressure in the gland could be responsible for small differ­
ences in rate of flow from milking to milking.
Duration of let down.
The time during which let down persists following stimu­
lation has not been definitely established.

Ely and Petersen

(1941) reported that blood from the donor cow was not effective
in eliciting a let down in the perfused udder if more than
eight minutes elapsed between stimulation and perfusion.
Miller and Petersen (1941) and Ward and Smith (1948) reported

22
decreased production if more than eight minutes elapsed "be­
tween stimulation and the time milking began.

They pre­

sented this as evidence for the rapid dissipation of oxy­
tocin; however, the validity of conclusions based upon
these experiments will be discussed in greater detail in a
later section.

Petersen (194-4-b) stated that not more than

seven minutes should elapse between preparation and comple­
tion of milking if all of the milk is to be obtained.
Additional information is required to reconcile these
results with those of Tgetgel (1926) and Whittlestone (1955)*
Tgetgel reported that the pressure in the gland cistern had
not fallen to the expected pre-let down pressure level
within 10 to 60 minutes, while Whittlestone reported that
the pressure had fallen but little within 22 minutes fol­
lowing stimulation of let down.

The possibility that the

continued stimulation of milking causes a constant release,
or stimulates several releases of the hormone sufficient to
maintain the contraction of the myoepithelium should not be
overlooked.
Frequency of let down.
The frequency with which a let down may occur has been
the subject of some controversy.

Smith (194-7) reported that

milking cows at 2-hour intervals resulted in extreme milkingto-milking variation which he attributed to the inability of

23
the cows to respond to the normal stimulus with a let down
of milk.

However, Whittlestone et al. (1952b) demonstrated

that if the normal milking routine was maintained and the
regular milker was present, cows did respond with a normal
let down when milked at 2-hour intervals.
In this connection, a comparison of the amounts of oxy­
tocin stored in the pituitary and the possible amount re­
quired to elicit a let down is of interest*

Whittlestone

et al. (1952a) reported that the oxytocic activity of the
whole pituitary of mature cows ranged from 23 to 178 I.U.
Donker et al. (1954-) reported let down on injection of 1.5
I.U., but doses smaller than that were not tried.

Evidence

presented by Denamur and Martinet (1954-) in which 10”^ I.U.
produced a response, and 10

_p

I.U. produced a maximal response

in the goat indicates that amounts smaller than those reported
above may be effective, when released by the pituitary of the
cow.

In a later publication, Donker (1958) reported let down

in the cow with as little as 0.06 I.U. of oxytocin and nor­
mal responses with 0.12 and 0.25 I.U. of oxytocin.

Thus the

pituitary apparently stores enough hormone to elicit many
let downs.
Several workers have reported a “second let down” of
milk which results in an increased rate of flow after a
period of low rate of flow following the first let down.
Whittlestone (194-9) first reported a second let down which

24
was elicited “by the additional stimulus of stripping.

He

attributed this phenomenon to a faulty original let down.
Whittlestone concludes that too much stripping may cause
some cows to become slow milkers by producing the second let
down while weakening the response to the stimuli which caused
the first let down.
Whittlestone (1951) reported that a second let down could
be obtained by milking two quarters at a time, if the second
let down was stimulated by washing the udder before the
second two quarters were milked.

In 1955 Whittlestone re­

ported that one cow observed consistently responded to
stripping with a second let down following a milking pro­
cedure using cannulas which required 16 to 18 minutes milk­
ing time before stripping began.

He reported that the nor­

mal amount of milk was obtained in this way and concluded
that procedures which produced a second let down were just
as efficient in removing all the milk as faster milking pro­
cedures involving a single let down.
Thoele (1954) reported that one cow exhibited a spon­
taneous second let down on several occasions which were pro­
duced without any manipulation of the udder on the part of
the milking machine operator.

Ace (1957) reported the ob­

servation of spontaneous second let downs in three different
cows.

These reports would indicate that some let down hormone

25
was released in response to the original stimuli, then a
later stimulus caused a second release of the hormone.
Effect on milk composition.
Injection of oxytocin has been used extensively as a
tool to get complete evacuation of the udder.

The effect

of oxytocin injections on the level of production has also
been investigated extensively.

For a review of this work

see Benson and Cowie (1957)•
Many workers, Shaw (1942), Knodt and Petersen (194-2,
194-4), Johansson (194-9), Swanson and Hinton (1951), Adams
and Allen (1952), Johansson (1952a, 1952b), Donker et al.
(1952, 1954-), and others, have shown that the “residual
milk" obtained by oxytocin injections following normal milk­
ing is higher in butterfat.
It has also been amply demonstrated by Johansson (194-9,
1952a), Whittlestone (1955), and others that the fat content
of the milk increases as milking progresses.

Whittlestone

attributed the increase in fat percentage in the cow to the
clustering of the fat globules.

This gives the fat globules

a higher viscosity and hence a slower rate of flow through
the small ducts, resulting in an increased fat percentage
in the last milk drawn.

He further stated that if the let

down does not persist long enough to enable the cow to be

26
milked completely at a normal milking, a greater amount of
high fat milk will be obtained following the injection of
oxytocin.
Joh.an.sson (1949) attributed milking-to-milking varia­
tions in the fat content of the milk to variations in the
degree of completeness of milking.

At an incomplete milking

the fat content is lower than usual, but at the subsequent
normal milking it is higher than usual.

This report would

indicate that the fat percentage of the milk could be used
as an indication of completeness of milking under varying
experimental conditions.
Methods of Stimulating Milk Let Down
Various methods of obtaining a satisfactory let down
of milk have been reported by several workers.

Gaines

(1915) demonstrated that the natural stimulus of a kid
suckling its mother caused a let down of milk in the un­
suckled gland as well.
The effects of various pre-milking preparations have
been extensively investigated.

Dodd and Toot (1947) re­

ported upon the effect of washing the udder with warm water
(115 to 120 degrees T.) and using the strip cup one minute
prior to the attachment of the teat cups.

The routine

previous to this had been to hose all of the animals with

27
cold water and employ the strip cup to the entire herd prior
to the commencing of milking.

In this, the usual manner,

from 15 minutes to over an hour elapsed between preparation
and attachment of the teat cups.

They used the experi­

mental regime for five weeks and found very little difference
in the rate of flow which they could attribute to the method
of preparation.

However, it was noted that whereas in the

initial control period, the cows failed to let down their
milk after washing, most of them did let down their milk
following washing in the first few days of the post experi­
mental period.

Several days were required before the cows

ceased to let down their milk following the cold water wash.
This indicates that these cows had been conditioned to wash­
ing as a stimulus for let down.
Smith and Petersen (1948) investigated the effect of
stimulation of let down by a 10 to 15 second udder wash fol­
lowed by udder massage two minutes before application of the
teat cups.

They found that this procedure resulted in higher

average rate of removal of milk and shorter milking time
than a group which received no stimulation prior to milking.
Knoop and Monroe (1950) also used several methods of
pre-milking preparation.

The response to washing with cold,

warm, or hot water was the same.

Variations in amount of

water used, ranging from massage with a damp cloth to bathing

28
■the entire udder in hot water gave the same response.
Brushing the udder, followed hy the use of the strip cup,
was not sufficient stimulation to cause milk let down.
When brushing the udder was used to promote let down, the
milking time was prolonged approximately one minute in
comparison to the other methods of pre-milking preparation.
They concluded that some type of preparation in addition
to use of the strip cup is necessary to promote the initial
let down of milk.
Beck and Clayton (1951) found that damp paper towels,
damp cloths, or 15 to 20 seconds massage with hare hands,
in addition to the use of the strip cup in all cases, were
equally effective in the stimulation of let down.
Roark et al. (1952) found that udder massage combined
with the use of the strip cup was more effective in stimu­
lating let down of milk than either of the preparations
alone.

Stimulation by the combined method shortened milk­

ing time as much as 27 percent, decreased stripping time and
increased yields slightly.
A report by Witt and Andreae (1955) should also be
mentioned.

They found that rate of milk flow was slightly

higher and the duration of milking slightly less when udders
were massaged with dry hands in contrast to the usual rou­
tine of washing the udder with a damp cloth.

They concluded

29
that warm water has an inhibitory effect on milk production
and does not offer any advantage under farm conditions.
These reports indicate that various types of stimuli
may be used to stimulate the let down of milk, but that the
response to a specific method may vary from herd to herd.
However, from the results of the vast amount of research
that has been conducted along these lines, it may be con­
cluded that an adequate stimulation of let down is the key
to successful removal of milk from the udder.
The Effect of Adding Water to the Concentrate Mix
Red During Milking
The possible effect of adding water to the concentrate
mix upon the let down of milk has also been investigated.
Dalton (1952) and Dalton et al. (1955) reported the effect
of feeding concentrates with different degrees of fineness
and water content on the eating and milking times of dairy
cattle when they were fed in the parlor during milking.
They reported the results of two trials.

The first involved

12 Brown Swiss cows and was conducted during the barn feed­
ing season; the second involved 11 Brown Swiss and two Hol­
stein cows and was conducted during the change-over to pas­
ture in the spring.

Reed ground to three different degrees

of fineness was fed; fine, medium, and coarse ground with
one-sixteenth, one-eighth, and five-sixteenths of an inch

30
screens respectively.

Each of the above concentrates was

fed for a three-day period, first as air dry and after that
water was added at the rate of 0 .5» 1 .0 , or 1.5 pounds of
water per pound of feed.

One fineness of ground concentrate

was fed with each of the three water contents, then another
grind was used.

A one day adjustment period intervened be­

tween each of the 12 three-day periods.
Consumption of the grain was the most rapid with the
coarsely ground feed, both wet and dry, although the differ­
ences in eating rate due to fineness of grind were not great.
With all three degrees of fineness of grind, eating rate in­
creased as the water content of the grain increased.

The

most rapid eating rate was attained when 1.5 pounds of water
was added to each pound of the coarsely ground feed, fed
during the barn feeding season.

The cows then required an

average of 0.61 minutes to consume a pound of grain, which
was a decrease of 64- percent from the 1,70 minutes required
to eat the air dry grain.
The change in average milking rate closely paralleled
the change in eating rate although the percentage change was
not as great.

The greatest increase in average milking rate

was attained when 1.5 pounds of water were added per pound
of coarsely ground feed during the pasture season, when 0.26
minutes were required per pound of milk produced.

This was

a decrease of 37 percent from the 0.4-1 minutes required to

31
produce a pound of milk when the cows were eating air dry
grain.

These workers postulated that the increased milking

rate may have been due to a faster let down produced by a
conditioned reflex or an increased feeling of satisfaction
associated with.the ease of eating the watered grain, which
resulted in a faster let down.
Thoele (1954-) also investigated the effect of adding
water to the concentrate mix.

The eating rates and milking

rates when air dry grain and grain to which water had been
added at the rate of 1.5 pounds of water per pound of grain
were compared in a double reversal trial involving 15 cows.
He confirmed the increased eating rate and average rate of
milk flow as reported by Dalton et al. (1953) although the
differences were not as marked.

However, Thoele found no

significant differences in maximum rate of flow or maximum
rate of milking time.

He advanced no explanation for the

increased average rate of flow which was obtained without an
increase in the maximum rate of flow.
The Effect of Variation of the Interval
Between Preparation and Milking
Several workers have studied the effect of varying the
interval between stimulation of let down and the time milking
began.

Dodd and Root (194-7), Smith and Petersen (1948),

Dodd et al. (194-9), Knoop and Monroe (1950), and Roark et al.

32
(1952) have shown that application of teat cups without
preparation or immediately following preparation delayed
attaining peak flow and resulted in a decreased average
rate of flow.

However, an effect on peak flow, once it was

attained, or on the level of production has not been demon­
strated.
The results which different groups of workers have ob­
tained concerning the effect of longer intervals between
preparation and milking on the production at that milking
have varied*

Little has been reported concerning the effect

of the longer intervals between preparation and milking
upon the rate of flow.

Miller and Petersen (1941), Peter­

sen (194-5), and Ward and Smith (1948, 194-9) have reported
decreased production following longer intervals.

Dodd et al.

(1949) and Roark et al. (1952) have reported slight decreases
in production following the longer intervals.

However,

Knodt et al. (194-7, 1949), using a large number of animals,
failed to find decreased production due to regular delays
in milking.

Dodd et al. and Roark et al. reported slower

rates of flow following the longer intervals.
The results of an experiment involving 19 cows were re­
ported by Miller and Petersen (1941).

They found that

manipulation and washing of the udder, a procedure which
stimulated let down, 20 minutes before, stripping 15 minutes

33

"before, and massage 10 minutes before milking brought an
average decrease in fat production of 8.39 percent.
was great variation in response between cows.

There

Production

decreased in 14 of the 19 cows studied, while the changes
in production ranged from plus 4*3 percent to minus 33*7
percent.

The mean decrease in fat test was 0.13 percentage

units.
The effect of milking one quarter at a time has been
reported by Miller and Petersen (1941).

Tkey found that pro­

duction decreased with each successive quarter milked.

This

corroborates the work of Bitting (1901) and is very similar
to the results of Babcock (1889), Emery (1895)* and Beach
(1904) who found the greatest production in the second quar­
ter, followed by the first, third, and fourth.

A decrease

in butterfat test was also reported by Miller and Petersen.
The tests for the first, second, third and fourth quarters,
respectively, were 4.76, 4.52, 4.34, and 3*64 percent butterfat.

Injections of pituitrin yielded amounts of milk and

fat so that the addition of the amount obtained after in­
jection to that obtained before resulted in almost equal
amounts of milk and fat for each of the four quarters.

They

postulated that the reason for lowered production after a
period of time following stimulation is due to dissipation
of the oxytocic principle or principles in the blood.

34

Ward and Smith. (1948, 1949) reported on an experiment
involving five cows in which the control half of the udder
was milked two minutes after let down was stimulated by
washing the udder and use of the strip cup, and the other
half was not milked until intervals of 4 , 8 , 12, 16, and
20 minutes following stimulation.

The differences between

the amount produced in the control half milked two minutes
after stimulation and the other half milked four and eight
minutes after stimulation were not significant.
slight decrease was noted after eight minutes.

However, a
When the

experimental half of the udder was milked 12, 16, or 20
minutes following stimulation, significantly less milk was
produced than in the control half of the udder.
Dodd et §6L. (1949) reported the results of two experi­
ments involving varied intervals between preparation and
milking.

In the first experiment, cows accustomed to prepar­

ation one minute before milking showed no change in rate of
milking when the interval for a single milking was three
minutes and a slight but not statistically significant de­
crease in the rate of flow and production when the interval
was six minutes.

In a second experiment, the same workers

milked four cows each at one minute or 20 minutes following
preparation.

No significant differences in rate of milking

or milk production were noted.

However, it was suggested

35
that milking immediately after preparation was the more effi­
cient routine, although the difference was slight.

Dodd and

Clough (1957) in a popular review article cite the above work
to substantiate the need for milking within three minutes fol­
lowing preparation, in order to develop a strong conditioned
reflex.
Roark et al. (1952) reported varied responses resulting
from 3 , 5 , and 8 minute delays following stimulation of let
down.

Eighteen cows were used in 18-day experiments, each

cow being subjected to each delay for three 2-day periods
during the trial.

The eight-minute delay resulted in five

percent less production in one trial, but in another trial,
no significant decrease in production was shown.

Delaying

milking for eigjht minutes consistently increased total ma­
chine time while a delay of five minutes caused a signifi­
cant decrease in milking time in one trial.

However, in a

longer trial in which six cows were subjected to the same
treatment for a six-week period, a significant increase in
machine time was noted following the five-minute delay.

A

figure is presented which shows some evidence of a second
let down when an eight-minute interval elapsed between pre­
paration and attachment of the teat cups.
Knodt et al. (1947, 1949) reported the results of a
series of six trials, involving 66 cows in all, in which

36

regular delays up to 20 minutes were made from the beginning
of washing of the cows udders until attachment of the teat
cups.

Each cow was milked at the same interval following

preparation for the entire test period, some of which were
as long as 84 days.

No significant differences in production

due to variation in interval were noted either in the first
week in which the routine was changed, or over the entire
period.
In one of these trials the effect of milking the entire
udder either 2 or 12 minutes following preparation was compared
to the effect of milking half of the udder two minutes after
preparation and the other half twelve minutes after preparation.
While there was no decline in production if the whole udder
was milked 12 minutes after preparation, there was a highly
significant decline in production in the half of the udder
milked 12 minutes after preparation if the other half was
milked two minutes after preparation.

This may explain the

discrepancy between these results and those which indicate
a decline in production in parts of the udder milked at
various intervals following preparation.

These results

cast doubt upon the validity of extrapolation of data based
upon halves or quarters of the udder milked at different
intervals following preparation, to the effect of milking
the entire udder at the same interval following preparation.

37
(The Effect of "Managed Milking11 Routines
Several workers, Dahlberg (194-1, 194-3), Zehner (1943),
and Hopson (1944) have advocated restricting the milking
time for each cow to a certain length of time in a strict
"managed milking" regime.

They maintain that the time re­

quired to milk a cow can he materially reduced.
Dahlberg (1941) reported the effect of gradually re­
ducing the milking time in a herd in which the initial milk­
ing time was 9 to 10 minutes.

At the end of a year, the

machine was left on the cow for only five minutes.

Approxi­

mately one-third of the cows required hand stripping, but
the amount of milk obtained from most of these was slight.
First lactation cows adapted very easily to the shorter
milking period when they were milked under this system for
calving.

Monthly production of butterfat and persistency

were equal to that obtained by hand milking and were better
than that of the previous period in which machine milking
time was longer.
In a later publication, Dahlberg (1943) restated his
previous conclusions and emphasized that hard milkers can
be trained to become easy milkers.

He stated that most slow

milkers have been developed by habit and chances are that a
slow milker may have been a fast milker as a heifer.

He ad­

vocated installing a timer on the milker, set to sound at

38
four and five minute intervals following the attachment of
the teat cups, the two intervals being provided to account
for the individual differences between cows.
Zehner (194-3) advocated an even greater reduction in
milking time.

He recommended milking each cow for three

minutes and suggested the use of a three minute egg timer
or similar instrument to measure the proper interval for
milking.

He stated that two men handling three units can

milk 50 cows per hour in a stanchion barn.

Zehner reported

that most cows rapidly became accustomed to the routine and
less strippings were obtained than previously.

He stated

that 5 to 10 percent of the cows do not respond to the
method and should be moved to the end of the line so that
the routine of rapid milking will not be interrupted.
Hopson (1944) advocated the use of a routine similar to
that presented by Zehner above.
In contrast to the above results, Dodd and his associates
have been unable to increase the rate of milk flow through
reducing the milking time.

In 194-7 Dodd and Foot reported

the results of an experiment in which they reduced the
milking time to 60 percent of normal as determined in an
initial control period.

The milking time was gradually re­

duced in increments of 10 percent per week until the 60 per­
cent level was reached.

This milking time was maintained

39
for four weeks.

They were unahle to produce an increased

rate of flow, in spite of the fact that in the shorter
milking times, considerable milk was left in the udder when
milking was completed.

They postulated that the discrep­

ancy between these results and those published by other
workers may be in that the reduced time gained by other
workers was only an elimination of the common fault of
leaving the teat cups on the cow too long after the milk­
ing process had actually been completed.
Dodd and Foot (194-9) reported the results of an experi­
ment in which a group of 16 animals were subjected to 60
percent of the normal milking time, as determined by a pre­
liminary control period, without any gradual period of re­
duction.

In this experiment the cows on the 60 percent

controlled time showed an increased maximum rate of flow
during the first week.

These workers suggested that this

increase in rate probably was due to an increased intramammary pressure caused by the accumulated milk within the
gland.

After the first week a statistically significant

decline in the maximum rate of flow was noted.

This decline

was probably associated with the decline in production which
was brought about by the strict limitation of milking time.
An experiment in which cows in their first lactation
were subjected to a rigid control of milking for the entire

40
lactation was reported, by Dodd et al. (1950)*

Forty cows

were paired by calving date as they freshened, and assigned
at random to either a four or an eight minute controlled
milking period for the entire lactation.

No statistically

significant differences in the peak rate of flow, or mean
yield in the first three minutes were found between the two
groups.

During the early part of the lactation, the dura­

tion of peak flow was shorter in the cows milked for four
minutes because the machine was removed during peak flow in
some cases.

For about one-half of the cows in this group,

four minutes was not sufficient time to complete the milk­
ing during the early part of the lactation.

However, in

spite of the incomplete milking of some of the cows in this
group, there were no significant differences between the
$05 day lactation yields or in the persistency in the two
groups.
Thus the above reported work of Dodd and his associates
indicates that the milking characteristics of the cow can­
not be changed by reducing the milking time and suggests
that the results reported by the earlier workers may have
only reflected improved milking practices which eliminated
idle machine time.

41
The Role of Teat Morphology
Although let down of milk has received a great deal of
attention, many workers maintain that the morphology of the
teat plays a very important role in determining the rate at
which a cow can he milked.
The transition of the teat cistern from the gland cis­
tern is sharply distinguished by a circular fold of the
membrane lining the teat.

According to Petersen (1942),

this fold may be so extensive as to interfere with the free
passage of milk from the gland cistern to the teat cistern.
Petersen (1943, 1944a) reported that the morphology of this
area affects the ease with which the passage between the
gland and the teat cistern may be occluded if the teat cups
are allowed to crawl upward when the rate of flow decreases
near the end of milking.
The shape and size of the teat cistern vary greatly
between cows, and Petersen et al. (1944) state that these
variations affect the rate of machine milking.
Turner (1939, 1932) and Espe and Cannon (1942) have
described the teat sphincter.

The streak canal through the

teat sphincter usually varies from 8 to 12 mm. in length
and is held closed by the involuntary sphincter papillae
muscle.

This canal is designed to retain the milk in the

udder against the pressure developed in the storage system

42
during the interval between milkings.

Along the axis of the

streak canal there extend folds of the mucosa which lie so
close together that under ordinary circumstances the streak
canal has no free lumen as the 3, 4, or 5 convex epithelial
projections touch each other in such a manner as to leave
a star-shaped slit, which as a rule is not open.
Espe and Cannon (1942) stated that upon application of
pressure to the teat, the length of the canal through the
sphincter is reduced from about 12 to 8 mm. in length.

At

the same time the cistern tends to balloon out as milk in
the cistern is put under pressure until the folds in the
mucosa disappear, causing the opening of the canal.
The histological structure of the teat sphincter has
been described by Turner (1939, 1932), who cited the work of
Christ (1903) and Kamm (1923), and by Venzke (1940).

In the

region of the streak canal under the epithelium of the
papillary eminences there is found an inner longitudinal
smooth muscle layer.

Next there is a circular smooth muscle

layer which forms the muscular sphincter papillae.

The inner

longitudinal muscles are found in small bundles; in females
in the prime of life these bundles are abundant and may form
a closed circular section, while in young or very old animals
only small isolated bundles may be found.

Thus the constric­

tion of the canal may be attained by the close elastic network
of the tunica propria or by the sphincter muscles.

43
Many workers, Ace (1957), Baxter et al. (1950), Dodd
(1952), Dodd and Foot (1953), Dodd and Kneave (1951), Bspe
and Cannon (1942), Foot (195H), Knoop (1948), Knoop and
Monroe (1950), Petersen (1942), Smith and Petersen (1946),
Stewart and Schultz (1958), Turner (1939, 1952), and others
have attributed to the teat sphincter the major role in
determining the rate of milk flow under a given set of con­
ditions.
Baxter et al. (1950) have shown that eliminating the
effect of the teat sphincter eliminated most of the vari­
ation in the rate of flow between cows.

When the two hind

quarters of 4 cows were milked a number of times with ordi­
nary teat cups the range in peak rate of flow for individual
quarters was found to be from 1.1 to 2.9 pounds per minute.
The same quarters were then milked at a similar vacuum level
through a simple cannula so that the size of the teat ori­
fice was standardized for all quarters.

Under these condi­

tions, the rate of milking for all quarters fell within the
range 2.6 to 2.9 pounds per minute.
However, to date, attempts to make direct measures of
the teat sphincter which can be related to the rate of flow
through the sphincter have been unsuccessful.

Espe and Can­

non (1942) used a standard Tycos sphygmomanometer with a
small rubber bladder and sleeve to fit around the teat.

The

44
teat was then closed off at the "base with a pair of rubber
covered clamps just sufficient to prevent the milk from
returning to the gland cistern.

Readings were made when

pressure on the bladder caused a very small stream of milk
to be forced from the teat.

They obtained only fair results

after considerable practice and considered the method gen­
erally unsatisfactory.
Murphy (1944) attempted to estimate teat "patency" by
repeatedly applying lateral pressure to the middle of the
teat with the thumb and first finger.

Care was taken to see

that the teat was full of milk to start and vertical motion
was avoided as much as possible.

No attempt was made to

estimate degree of patency, teats were classed as either pa­
tent or non-patent.

Results by this method were not very

repeatable, as frequently a cow would be classed as patent
one day and non-patent on another.

Although it is not men­

tioned by the author, let down of milk may have been a com­
plicating factor in determining patency.
Ace (1957) measured the length of the teat, thickness
of the teat, and length and width of the teat sphincter and
was unable to correlate any of these measures with the maxi­
mum rate of milk flow.
Although the teat sphincter plays a major role in de­
termining the rate of flow, apparently little can be done to

45

alter the sphincter of a slow milking cow.

Turner (1959)

cited the work of Kraft (1927) who described a technique
which could be used to remedy the condition of slow milking.
A special knife called a teat bistourie was inserted into
the teat; the knife was then unsheathed and carefully drawn
downward, making an incision through part of the streak
canal and cutting part of the sphincter lying below.

For

obvious reasons, this technique has not gained wide accept­
ance.
Another question yet unanswered concerning the teat
sphincter is the one posed by Tan Dyke et al. (1955)*

Could

oxytocin relax the teat sphincter while contracting the
myoepithelim of the udder and the smooth muscle of the
uterus?

They concluded that sphincter relaxation seems to

be primarily under the control of the nervous system, rather
than that of oxytocin.

However, it is possible for the

muscle fibers to be arranged so that contraction of these
fibers, or parts of these fibers would open the sphincter,
although such an arrangement has not been demonstrated.

It

is also possible that the leaking of milk observed in many
cows is simply a pressure phenomena due to the increased
intramammary pressure obtained upon let down, and relaxation
of the teat sphincter is not involved.

46
Other Factors Associated with Hate of Flow
There are several other factors associated with rate
of milk flow which need to he considered in evaluating the
results of experiments involving the rate of milk flow.
Heredity.
Since the milking response of the individual cow seems
to he a characteristic feature which is highly repeatable,
several workers, as indicated helow, have concluded that the
rate of milking is an inherited characteristic.

However,

the experimental data in support of this conclusion is not
abundant•
Several workers, Matthews et ad. (1941), Beck et al.
(1951b), Dodd and Boot (1953), and Caroulo et al. (1955)
showed that statistically significant differences in rate
of milking existed between breeds.

Beck et ad. and Thoele

(1954) cited these breed differences as evidence in support
of inheritance of the rate of milking characteristic.
Several groups of workers have presented other evidence
that the rate of milk flow is an inherited characteristic.
Beck et al. (1951b) found statistically significant differ­
ences in the rate of milking that could be attributed to
sires within a breed.
51 sires.

These were based on 102 daughters of

However only 14 of these had more than one daughter.

47
Therefore the data are actually based on 85 daughters of
14 sires.

Dodd and Boot (1953) reported significant cor­

relations of 0.57 on 40 dam-daughter pairs and 0.37 on 35
pairs of sibs and half-sibs.

They also reported one sire

with 8 daughters, all of which were slower milkers than their
dams; and another sire with 8 daughters, whose rate of milk­
ing was very similar to their dams.

Both of these reports

were cited as preliminary observations, but neither group
has presented supplementary evidence.
Hansen (1952 to 1956) has presented data showing a
large range in rates of flow of progeny groups of different
sires tested in the Danish cow testing associations; however
no statistical analysis of the data have been made.
Thoele (1954-) presented data on 33 daughters of four
sires, the mean maximum rate of flow for the daughters of
each sire were 8.2, 7*8, 7-6, and 6.2 pounds per minute.
Ho statistical analysis was presented.

However these data

are reported to substantiate the differences between sires
as shown by Hansen.

Brumby (1956) reported data from iden­

tical and fraternal twins that showed the maximum rate and
the average rate of flow to be strongly inherited, with a
great deal of genetic variation existing between twin sets.
However, evidence has also been presented that rate of
milk flow is not a characteristic, in which the role of
heredity has been clearly elucidated.

48
Hupp (1956) reported statistically significant differ­
ences in the maximum rate of flow in one herd, while the
differences between sire groups were not significant in two
other herds.

In the latter, the variance was greater within

sire groups than between sire groups.

In the herd where

the means of sire groups were significantly different, there
was a great variation within sire groups.
Ace (1957) reported that only the mean maximum rate of
milk flow of Holsteins differed significantly from the other
major dairy breeds in the herd studied.

He reported signi­

ficant differences between sire means, based upon 117 daugh­
ters of 28 sires with two or more daughters each.

However,

the intra-class correlation, or index of repeatability,
among daughters of sires was 0.174- which is not statistically
significantly different from zero.

This indicates that

large variation exists among individuals in the subsample.
Ace also reported an intraclass correlation of 0.589 for
22 daughter-darn pairs.

While this is not statistically

significant, he concludes that a correlation coefficient
of this magnitude indicates that the dam-daughter relation­
ship may be greater than that of the sire to his daughters.
In view of the many factors which can contribute to the
rate of machine milking, it is not surprising that heritability of rate of flow per se is hard to demonstrate, but

4-9
heritability of a single factor which might in some cases
limit rate of flow could be high.
Stage of Lactation.
Somewhat conflicting views have been presented as to
the effect of stage of lactation upon rate of milking.

In

many cases the effect of stage of lactation upon rate of
milking has been confounded with the effect of the decline
in level of production associated with progressing stage of
lactation.
Matthews et al. (194-1) found that the mean maximum rate
of flow for 1.5 minutes at peak flow in the sixth month of
lactation was only 62.6 percent of the rate when measured
in the third month of lactation.

The mean maximum rate of

flow and production was rather low in both cases; 4-.58
pounds of milk per minute and 50.6 pounds per day, respec­
tively, in the third month and 2.74- pounds per minute and
17.5 pounds per day, respectively, in the sixth month of
lactation.
Beck et al. (1951b) found that the differences in
average and maximum rate of milking between early, middle,
p-nH late stages of lactation were highly significant.

The

differences in rate between the early and middle part of the
lactation were less pronounced than between the middle and
late part of lactation.

However, the mean yields of milk

50
per milking for the three periods were 17.0 , 11.5, and 6.1
pounds, respectively.

Thus a large part of the differences

reported by these two groups could he due to low levels of
production at the later milking periods.

At the earlier

period, rate of flow could have been primarily dependent
upon the teat sphincter, while at the later periods, rate
of drainage of the milk into the teat cistern may have been
the controlling factor.
Brumby (1956) and Stewart et al. (1957) also reported
decreased average rates of flow with progressive stages of
lactation.

However, Stewart et al. reported that the average

rate of flow decreased less rapidly than level of production,
so total milking time decreased.
Dodd (1955)? through analysis of data by covariance
techniques, concluded that lactation decline in milking rate
is related both to change in milk yields and to interval
since calving.
On the other hand, Gregoire et al. (1954-) found that
in cows in their first lactation, the rate of flow increased
slightly, but not significantly as the lactation progressed
when milked at a vacuum level of 15 and 17 inches of mercury.
There was no increase in rate of those milked at a vacuum
level of 10 inches of mercury.

51
Caroulo et al. (1955) found that the linear change in
rate of milking from the fourth to the forty-fourth week of
lactation was: not appreciable.

This was based upon 4-8 cows,

24 of which were in the first lactation and 24 of which
were in the second or subsequent lactation.
Hupp and Plum (1957) reported that the first lactation
cows increased in maximum rate of flow in a four-month period
during the portion of the lactation in which production re­
mained over 30 pounds per day.

Cows in a second or sub­

sequent lactation did not show a significant change in rate
of flow in the same period.
Level of production.
The work of Matthews et al. (1941), Harshbarger (1950),
Beek et al. (1951b), Brumby (1956), and Sandvik (1957)? and
others indicates that a decreased rate of flow is associated
with the decrease in level of production as the lactation
progresses.

Hupp (1956) reported that in the period while

the level of production remained above 15 pounds per milk­
ing, at the milking measured, there was no significant change
in rate of flow that could be associated with level of pro­
duction.
Hupp (1956) also reported correlation coefficients of
0.131 and 0.147, between the maximum rate of milk flow and
the level of production at that milking, when production was

52
above 15 pounds per milking.

Neither of these correlation

coefficients are statistically significant.

However, if

they were the true correlation coefficients, variations in
level of production would account for only about two per­
cent of the variation in rate of flow.
Several workers have also investigated the relationship
between milking rate and the yield of milk in a standard
lactation.

An analysis of the data on rate of milking col­

lected in the Danish cow testing work (Hansen, 1954 to 1956)
shows a high correlation between the average production of
the daughters of a sire and the average maximum rate of
milking of these daughters.

If the amount of milk produced

in the second minute is used to represent the maximum rate,
the coefficient of correlation is found to be 0 .72, based
upon the daughters of 142 sires measured in the years 1952
to 1955*

This statistic is significant at the one percent

level of probability.
Clough and Dodd (1957) report a highly significant re­
gression of standard lactation yield on peak rate of flow,
based upon a total of 352 cows.

Records of 135 animals with

yields between 15*0 and 17*9 pounds per milking gave a re­
gression value almost identical with the one based on all
the cows.

This indicates that any differences in yield due

to level of production at the time the recording was made

53
did not bias the results to any marked degree.

These data

support preliminary results reported earlier by Dodd and
Foot (1953).

They attribute the differences in standard

lactation yield to the failure to remove all the milk from
the slow milking cow during the time oxytocin is effective
in expelling milk from the alveoli.
Two workers have reported results in conflict with the
above.

Brumby (1956) was unable to demonstrate any relation­

ship between milking rate and milk yield on a between animal
basis.

He postulated that this may have been due to good

milking techniques, so that all of the milk was removed from
the slow milking cows in this study.

Sandvik (1957) reported

that when the effect of milk yield at the recorded milkings
on the observed characteristics was eliminated, through the
use of a correction factor, the milking character had no
significant effect on the lactation yield, the maximum daily
yield, or the persistency.

The same results were obtained

when only the data from observations on the milking charac­
teristics in which the yield was between 6.1 and 7 Kg. were
used.

From this he concluded that slow milking cows are not

necessarily poor producers.
Age of the cow.
Matthews et al. (1941) found that in cows which were
studied in two successive lactations, the milking rate for

54
1.5 minutes at peak flow was 8.29 percent lower in the later
lactation.

The average age of the cows at the earlier lac­

tation was 4.6 years, and at the subsequent lactation, 6.2
years.
A tendency for the maximum rate of milk flow to decrease
with increasing age of the cow was reported by Korkman (1948).
The decrease was significant at the six percent level of
probability.
However, the maximum rate of milk flow in the first
lactation has been reported to be significantly less than in
the second lactation (Dodd, 1953).

The cows milked slightly

faster with each subsequent lactation, but the differences
between these subsequent lactations were not significant.
Hupp (1956) reported that the mean maximum rate of milk
flow of cows in their first lactation was slightly greater
than the mean maximum rate of cows which were in their
second through ninth lactations.

With the relatively small

number of animals involved, the differences were not sta­
tistically significant.
Differences between morning and evening milking.
Matthews et_ al. (1941), Dodd and Foot (1947), and
Brumby (1956) have shown that if the differences in interval
between milking are great, such as 9 and 15 hour intervals
or 8 and 16 hour intervals, differences in the rate of flow

55
between the two milkings occur.

However, Thoele (1954-) has

shown that differences in rate of flow between morning and
evening milkings were slight and not statistically signifi­
cant when 12 hours intervened between milkings.

Hupp (1956)

reported that no significant differences occurred in the
rate of flow when 12 hours intervened between milkings, or
when 11 and 13 hours intervened between milkings.
Differences due to length of milking time.
In cows in their first lactation, in which half of the
udder was milked for twice normal periods for 4-0 weeks,
Gregoire et al. (1954-) found a significantly faster average
rate of flow in the half which was milked the twice normal
period.

One possible explanation was that the prolonged

milking over the 4-0 week period stretched the teat sphincter,
causing a larger opening and a greater rate of flow.

Thus

conceivably, a herd in which the machine is consistently
left on the cow too long might be capable of a higher rate
of flow due to the prolonged milking.
Relationship to mastitis.
Dodd and Kneave (1951) reported a high correlation be­
tween milking rate and incidence of clinical and sub-clinical
mastitis in cows in their first lactation.

They postulated

that bacteria can gain entrance to the udder more easily

56
through, the slackened teat sphincter of the more rapid milk­
ing cows.

They suggested that a portion of the increased

incidence of mastitis in the fast milking cows could be due
to udder injury resulting from leaving the machine on the
cows too long after they were milked out.

In contrast to

the above report, Caroulo et al. (1955) found that in 24cows in their first lactation and 24- cows in their second
or subsequent lactation, cows with a previous history of
mastitis milked at a slightly lower rate.

Thus the exact

relationship between mastitis and rate of milk flow has not
been demonstrated.

OBJECTIVES
1. To investigate the effect that the addition of
water to the concentrate mix which is fed during
milking may have upon the milking rate and eat­
ing rate.
2. To investigate the effect of injected oxytocin
upon the rate of milk flow.
3. To investigate the effect of variation in the
interval between preparation of the udder and
attachment of the teat cups upon the rate of
milk flow, level of production, and milk compo­
sition.
4. To determine the accuracy of the milk measuring
and weighing devices used in these studies.
The procedures and results of these studies will be
presented in two sections.

Section A includes the study

of the effects of several variables as outlined in objec­
tives 1 to 3 .

Section B comprises a validation of the

milk measuring and weighing devices used in this research
(objective 4).

- 57 -

S E C T I O N

A

EXPERIMENTAL PROCEDURE
The methods used in this section will he presented in
three parts*

Part I describes the methods used to investi­

gate the effect of addition of water to the concentrate
mix fed during milking.

Part II describes the procedures

used to investigate the effect of oxytocin injections,
while Part III describes the procedures used to investigate
the effects of varying the interval between preparation of
the udder and attachment of the teat cups.
Part I.

The Effect of Adding Water to the Concentrate Mix

In this experiment, 16 registered Brown Swiss cows, a
part of the Michigan State University Dairy herd, were used.
They were housed and milked in a loose housing steel barn.
The milking room used was a four stall side opening U-type.
This parlor was equipped with a two unit Chore-Boy pipeline
milker which operated at a vacuum level of 10.6 inches of
mercury.

Ranges in vacuum level of from 10.4- to 10.8 inches

of mercury were noted during the course of the experiment.
Each milking unit included a weigh pail suspended from
a spring scale so that the amount of accumulated milk could
be readily noted.

An evaluation of the accuracy of the milk

weights obtained by this method will be presented in Section B.
-

58

-

59
A concentrate mix consisting of 60 parts shelled corn,
15 parts soybean oil meal, 15 parts oats, 10 parts barley,
1 part salt and 1 part di-calcium phosphate was fed during
milking.

The grain had been ground to a coarse consistency

using a B. J. Humdinger hammer mill with a 3/16 inch screen.
The concentrate mix was stored in an overhead automatic
feeder, and was measured into each individual feed trough
with a crank that delivered two pounds with each turn.

When

water was added to the concentrate mix, the water was poured
into the trough; then the concentrate mix was added.

About

30 seconds were required for the water and grain to become
thoroughly mixed.

Preliminary experiments indicated that

hand mixing did not hasten the mixing, so no hand mixing
was done.
In the normal milking routine, the cows entered the
parlor from a holding pen, usually two at a time.

Then the

udders were washed with warm water and the strip cup was
used to check for abnormal milk.
pared at a time.

Two cows were usually pre­

The average interval between preparation

and attachment of the teat cups was 2.65 - 1.20 minutes
during the period when dry grain was fed and 2.81 + 1.20
minutes during the period in which water was added to the
grain.

The standard deviation is quite large because the

first cow prepared usually had the machine stripping time
of one cow between preparation of the udder and attachment

60
of the teat cups, while the second cow prepared had the ad­
ditional interval while another cow was machine stripped
before the teat cups were attached.
The following observations were taken for each cow
milked:
1. The time the udder was washed.
2. The time of application of the first teat cup.

The

concentrate mix was fed by the person recording the
data, to coincide with the application of the teat
cups.
3* The time at which each pound of milk accumulated in
the weigh bucket.
4-. The time concentrate eating was completed.
5. The time stripping started and the pounds of milk in
the weigh bucket at that time.
6. The time at completion of milking.
7. The total pounds of milk weighed with the claw hang­
ing in the same position at which the scale was ad­
justed to zero with the bucket empty.
These observations yielded data for each cow from which
the average eating rate, average milking rate, and a milk
flow curve could be computed.

The average eating rate was

obtained by dividing the pounds of grain eaten by the total
eating time.

The average milking rate was determined by

61
dividing total pounds of milk produced "by total milking
time.

The milk flow curve was obtained by computing the

time required to produce each successive pound of milk.
Although most other workers who have made measurements
of milking rate during the course of milking by other than
graphic means have determined flow rates from measurements
of pounds of milk produced in a fixed period of time, these
data were obtained as the time required to produce each
pound of milk.

There is less error in reading the scale in

this way than when the accumulated amount of milk must be
noted at an exact instant.

In that case, the swinging dial

of the scale must be read instantly, while in the method
used in this experiment, by observing the swinging dial, it
can be determined quite accurately when each pound of milk
is reached.

At that instant the watch, the hand of which is

moving at a constant rate, can be easily read.

Flow rates,

computed as seconds required per pound of milk produced, were
converted to pounds per minute before presenting the results
so that they can be more readily compared with the other
data in the literature.
This experiment was conducted in November, 1957*

Six

control observations, three of which were morning and three
evening milkings were made during which the normal milking
routine prevailed.

Following the control observations, water

62
was added to the concentrate mix at the rate of V/z pounds
of water per pound of concentrate mix for a three day ad­
justment period, during which no observations were taken.
Then observations were taken for six more milkings.

The

procedure was similar to the control observations, except
for the addition of water to the feed trough just before
the concentrate mix was added.

Following this observation

period, watering the concentrate mix was stopped.

After a

four-day adjustment period, two additional control observa­
tions were made.
Part II.

The Effect of Oxytocin Injections

Registered Holstein-Friesian and Brown Swiss cows, com­
prising a portion of the Michigan State University Dairy
herd, housed and milked in the main tie-stall barn on the
Michigan State University campus, were used in this study.
In the normal milking procedure, the cows were milked in
the stalls by one man operating two units of a DeLaval pipe­
line milker.

The udders were washed and the strip cup em­

ployed to check for abnormal milk approximately two minutes
before the teat cups were attached.

Since two cows were

usually prepared at once, this interval varied considerably.
Milk-O-Meters were used with both milker units.

The meters

were hung on mounts which were located below each inlet in

63
the pipeline, one of which was located between each two
cows.

The milk from the cow passed through the meter and

directly into the pipeline.

A Milk-O-Meter in operation is

shown in Figure 1.
In Trial 1, seven Holstein-Friesian cows were used.
They were selected on the basis of their slow milking char­
acteristics, as determined by preliminary observations.
Meter No. 1-1460 was used for the pre-treatment observa­
tions, except for three observations in which meter No. 1-1454
was used.

Meter No. 1-1555 was used for the treatment and

post-treatment observations.
Two control observations were taken, in which the nor­
mal milking procedure was followed, except that all cows in
the experiment were milked with one milker unit to facili­
tate taking the observations.

The following data were taken

for each cow which was included in the experiment:
1. The time the udder was washed.
2. The time the last teat cup was applied— this is start­
ing time.
3. The time each pound of milk accumulated, as indicated
by the Milk-O-Meter dial.
4. The time stripping started and the pounds of milk
produced up to that time.
5. The time the teat cups were removed and the total pounds
of milk as indicated by the meter dial.

64

Figure 1.

(The Milk-O-Meter in operation in the Tie-stall
main dairy barn.

65
Following the control observations, the seven cows were
given injections of 20 I.U. of oxytocin (Parke-Davis Pitocin)
preceding each milking for three milkings.

The cows were

tied in the stanchions at the south end of the barn and
given the injections of oxytocin intrajugularly.

Each cow

was returned to her tie-stall, immediately after injection,
the udder washed, strip cup employed, and the teat cups at­
tached.

The interval between the

the jugular vein

time the oxytocin entered

and the time the last teat cup was attached

was timed which averaged 99*5 - 9.1 seconds.

The same data

were taken as in the control observations, except that the
time the udder was prepared was not noted.
Three days later two more control observations were
made.

The same procedures were used as in the control ob­

servations made before the treatment milkings.
The procedures used in Trial 2 were similar to those
used in Trial 1.

Seven Holstein and one Brown Swiss cow

were selected to cover the range in rate of milk flow en­
countered within the herd.

Meter Ho. 1-1460 was used

throughout, but was equipped with the modified baffle plate
rather than the pin-type baffle plate used in Trial 1.

The

time between injection and attachment of the last teat cup
was 110.4 - 20.6seconds, as compared to 99.5 in Trial 1.

The

9*1 seconds

other procedures followed were the same as

66

in Trial 1*

However, in the three day interval between the

injection period and the post-injection control observations,
difficulties in the operation of the milker necessitated
repairs.

These repairs produced a greater vacuum potential

without changing the vacuum level and may have affected the
pulsation pattern as well.
Part III. Variation in Interval from
Preparation to Milking
The same Holstein-Priesian herd used in the experiments
described in Part II was used in this experiment.
milking procedure was also the same.

The normal

The data obtained in

this experiment were taken with the recording Milk-O-Meter.
The operating parts of this meter are the same as in the
regular meter, except that two electrical contacts have been
added so that each time the weigh tray trips an electrical
impulse is sent to the recorder.

This impulse from the meter

is carried by a wire which is plugged into a jack.

One jack

is located behind each Milk-O-Meter mount in the barn.

These

jacks are connected to a low voltage wire which goes all the
way around the barn and leads to the recorder.

Thus the

recording meter can. be used for any cow milked in the main
barn while the recorder is located in one place.

Each time

the recorder receives an impulse from the meter, it makes a
mark with an indelible marker on a tape which is driven past
the marker at a constant speed.

67
Figure 2 shows the recording Milk-O-Meter with the
sampler attached, plugged directly into the recorder.

In

the barn an additional length of* wire would separate the
meter and recorder.

Figure 3 shows the inside detail of

the recorder.
In addition to the graphic record made of each milking
by the recorder, a stop watch synchronized with the recorder,
was used to obtain additional data to complete the flow
curve.

The following observations were made:

1. The time the udder was washed.
2. The time the last teat cup was applied— this is the
starting time.
3. The time stripping started.
4. The time the teat cups were removed and the total
pounds of milk as indicated by the meter dial.
A total of ten cows were selected for this experiment
on the basis of rather high milk production and freedom from
recent clinical cases of mastitis.

They covered the range

in rate of flow of the cows in the herd which had high cur­
rent levels of production.

Each of these cows was milked

at 0, 4, 8, 12, 16, and 20 minutes following preparation of
the udder by washing the udder and using the strip cup.

In

the milking designated as the zero interval, 0.58 minutes
actually elapsed between the time preparation of the udder

68

Figure 2.

The recording Milk-O-Meter with the butterfat
sampler attached, and the Eecorder.

In actual

operation, both pieces are mounted in the barn
and the wire leading from the meter to the Re­
corder is much longer.

69

Figure 3.

Inside detail of the Recorder used with the
Recording Milk-O-Meter.

70
began and the ‘time the last teat cup was attached.

This

mean time had a standard deviation of 0.40 minutes.

The

means and standard deviations for the other intervals are
as follows:

4.37 + 0.32 minutes; 8.22 + 0.44 minutes;

12.15 + 0.41 minutes; 16.32 + 0.58 minutes, and 20.43 + 0.78
minutes.
The order of the treatments was randomized for each cow
for each replication, using a table of random numbers.

There

were three replications, so that in the entire experiment,
each cow was milked following each interval three times.

All

observations were made at the evening milking so that any
residual effect of the treatments would be obtained at the
morning milking.

The first, second, and third replications

were taken May 26 to 31? June 1 to 11, and June 22 to 30,
1958, respectively.
The only change in the regular milking routine was that
the recording meter was substituted for one of the regular
meters, and the cows used in the experiment were milked
with the milking unit which had this meter.

The cows were

milked in the regular order as nearly as possible.

However,

at times when the longer intervals between preparation of
the udder and attachment of the teat cups were used, the cow
was prepared quite a distance up the milking line so that
the desired interval would have elapsed by the time her turn
came in the line.

71
In the first replication, the meter was used without
the sampler; in replications two and three, the sampler was
used with the meter, and the entire volume withdrawn by the
sampler was removed for use in composition analysis.
The percent butterfat in these samples was determined
by the Babcock method.

The percent protein was determined

by Hichardson's modification of the Eormol titration
(Richardson et ad., 1953)*

Total solids were determined by

the small Watson Lactometer, the accuracy of which has been
reported by Madden and Brunner (1958).

RESULTS AND DISCUSSION
The results obtained in these experiments and a dis­
cussion thereof will be presented in three parts.

In Part

I the effect of addition of water to the concentrate mix
will be presented.

In Part II, the effects of injected oxy­

tocin upon the rate of milk flow and level of production
will be presented, while Part III presents the results ob­
tained when the interval between preparation of the udder
and attachment of the teat cups was varied.
Part I. The Effect of Adding Water
to the Concentrate Mix
The object of this experiment was to investigate the
effect that addition of water to the concentrate mix fed at
milking time might have on the rate of milk flow and also on
eating rate.
The effect on average and maximum rate of milking and eating
rate.
In Table 1 data are presented showing the effect of
addition of water to the concentrate mix on the average rate
of flow, maximum rate of milk flow and the average eating
rate.

The average rate of milk flow and average eating rate

were computed as described in the procedures.
- 72 -

The maximum

73
[TABLE 1
THE EFFECT OF FEEDING A WATERED CONCENTRATE MIX
ON THE MAXIMUM RATE OF MILE FLOW, AVERAGE RATE
OF MILK FLOW AND AVERAGE EATING RATE

Maximum
Rate of Flow
(lbs./min.)

Average
Rate of Flow
(lbs./min.)

Eating
Rate
(lbs./min.)

Dry Grain

3-73

2.19

0.63

Watered Grain

3.60

1.92

0.90

-0.13

-0.27**

Difference

+0.27**

j|(

Significant at the one percent level of probability, as
analyzed by the "T" test for paired observations.
rate of flow was determined from the flow curve.

It is the

rate of flow in which from two to five pounds of milk were
produced at the fastest rate during that milking.

In most

cases the period encompassed the time required to produce
five pounds of milk.

However in some cases, due to low

levels of production or the particular flow curve of the
cow, the maximum rate of flow did not persist that long.
In that case, the time required to produce two or three
pounds of milk was used.

However, the period measured en­

compassed the same number of pounds for all milkings of each
individual cow.
The increased eating rate with the watered grain, as
reported by Dalton et, &1. (1933) and Thoele (19540 was

74
-

confirmed (Table 1).

However, the increase in eating rate

was not as marked as that reported by Dalton et al.

The

time required in this experiment to eat a pound of grain to
which water had been added was only 70 percent of that re­
quired to eat the dry grain.

This difference is significant

at the one percent level of probability.
The maximum rate of milk flow decreased slightly with
the watered grain.
significant.

This decrease was not statistically

In nine of the 16 cows, the maximum rate was

slower and in seven it was faster.

In contrast to the re­

sults of Dalton et al. (1953)» who reported an increased
average milking rate, the average rate of milk flow actu­
ally decreased 12 percent with the watered grain.

This

change was significant at the one percent level of probabil­
ity.
The effect on the components of the milk flow curve.
In order to further investigate the decreased average
rate of flow, the milk flow curve was broken into three com­
ponent parts.

These components are as follows:

Part 1 is

from the time the last teat cup was applied to the time that
the rate of flow decreased to half of the maximum rate of
flow at that milking.

Part 2 is the interval beginning at

the end of Part 1 and continuing until machine stripping be­
gins.

(This includes the idle time before stripping began.)

75
Part 3 is the interval from the time machine stripping be­
gan until the teat cups were removed.
The average production, average milking time, and
average rate of flow for each of these three components and
the entire curve are reported in Table 2.

This table shows

TABLE 2
THE EFFECT 0F FEEDING A WATERED CONCENTRATE MIX ON
AVERAGE MILK PRODUCTION, AVERAGE MILKING TIME AND AVERAGE
RATE OE MILK PLOW EOR THE THREE COMPONENT PARTS
OF THE MILKING CURVE

Average
Production
(lbs.)

Average milk­
ing Time
(min.)

Average Rate
of Flow
(lbs./min.)

Part 1
Watered Grain
Dry Grain
Difference

9.46
9.96
-0.50

3.26
3.31
-0.05

3.42
3-54
-0.12

Part 2
Watered Grain
Dry Grain
Difference

2.02
2.22
-0.20

2.72
2.49
+0.23

0.74
0.87
-0.13

Part 3
Watered Grain
Dry Grain
Difference

3.06
3.07
-0.01

1.91
1.69
+0.22

1.50
1.80
-0.30**

Entire Curve
Watered Grain
Dry Grain
Difference

14.54
15.25
-0.71**

7.89
7.49
+0.40

1.92
2.19
-0.27**

%*
Significant at the one percent level of probability, as
analyzed by the "T" test for paired observations.

76
that the production was slightly less with the watered grain
while the total milking time increased an average of 0.4
minutes per cow.

This resulted in the significant decrease

in average rate of flow.
In Part 1 of the milk flow curve, production with the
watered concentrate mix was 0.5 pound lower while the milk­
ing time decreased only slightly.

The resulting decrease

in average rate of flow was not significant.

Since less

milk was produced in Part 1 with only a slight change in
milking time when watered concentrate mix was fed, this
could indicate a slight effect on let down.

Therefore the

time required to reach a high rate of flow was determined.
Since the maximum rate of flow persisted for only a short
period of time, another criterion, the high rate of flow,
was established.

It is defined as the interval during

which a pound of milk is produced in up to two seconds more
than the time required to produce a pound at the maximum
rate.

This latter period should correspond closely to the

period of maximum rate of flow reported by other workers.
During the period in which the dry concentrate mix was fed,
an average of 60.5 seconds elapsed before a high rate of
flow was reached.

When water was added to the concentrate

mix, an average of 59.7 seconds elapsed before a high rate
of flow was reached.

These times are almost identical and

77
the slight difference is not statistically significant.
Therefore no indication of an effect on let down is noted
in this part of the curve.
In Part 2 of the milk flow curve, the idle time before
stripping began, production decreased 0.2 pounds while
milking time increased 0.22 minutes.

Due to wide variation

between cows, the resulting decrease in average rate of flow
was not statistically significant.

The increased idle time

in Part 2 reflects the increased stripping time which was
responsible for the highly significant decrease in average
rate of flow in Part 3«
Discussion.
The above results indicate that the highly significant
decrease in the average rate of flow was due primarily to
increased time spent by the operator during stripping.
Apparently this was in an attempt to overcome the slight
decrease in production which was encountered.

This increased

milking time is a function of the machine operator and cannot
be attributed to the effect of addition of water to the con­
centrate mix upon the milk flow characteristics of the in­
dividual cow.

The fact that the maximum rate of flow, aver­

age rate of flow before the machine stripping should have
begun, and the time required to reach a maximum rate of flow
were not changed significantly by the addition of water to

78
the concentrate mix shows that this procedure did not have
an effect on the rate of milk flow as controlled by the
cow.
Part II.

The Effect of Oxytocin Injections

Although the primary role in control of rate of milk
flow has been attributed to the teat sphincter, several
workers (Whittlestone, 1954; Turner, 1954; Dodd and Clough,
1957; and others) have stated that some cows may be slow
milkers because of an insufficient let down of milk.

In

order to investigate the relative importance of these two
factors, in controlling rate of milk flow, injections of
oxytocin were used to assure an adequate level of oxytocin
in the blood.
Maximum rate of flow.
Differences between cows.

An analysis of the sources

of variation in maximum rate of flow in Trial 1 is presented
in Table 3.

Each milking is considered separately.

A simi­

lar analysis was carried out for each characteristic inves­
tigated.

In Trial 2 the data from only six cows were used

in the analysis.

There were highly significant differences

between cows in maximum rate of flow.

It has been shown

repeatedly by other workers that differences in maximum rate
exist between cows, but these data show that these differences

79
can be demonstrated even in a group selected for their
slow milking characteristics.

Significant, or highly sig­

nificant differences between cows were also noted for
average rate of milk flow, pounds of milk produced, seconds
before a high rate of flow began and seconds that a high
rate of flow persisted in Trials 1 and 2 and stripping time
in Trial 2.

The differences between cows were not signifi­

cant for the seconds between the end of a high rate of flow and
the time stripping started in Trials 1 and 2, and stripping
time in Trial 1.

A lack of significant differences for

these characteristics indicates that they are primarily
under the control of the machine operator, rather than the
cow.
Differences between treatment and control.

The data

in Table 3 also show that there were no significant dif­
ferences in the maximum rate of flow between the seven
milkings in Trial 1.

Thus oxytocin injections did not

change the maximum rate of flow significantly for these
seven slow milking cows.

The mean maximum rates of flow,

and their order, from the slowest to the fastest rate are
presented in Table 4.
Bight cows were used in Trial 2; however, one cow de­
veloped mastitis following the treatment period.

Thus post­

treatment control observations could not be made, so the

80
TABLE 3
ANALYSIS OF VARIANCE
SOURCES OF VARIATION IN MAXIMUM RATE
OF MILK FLOW IN TRIAL 1
Source of
Variation

Degrees of
Freedom

Mean
Square

"F"
—

48

15.00

Cows

6

90.20

Treatment

6

7.95

2.18

36

3.64

—

Total

Error

14.09**

**Statistically significant at the one percent level of
probability.
data from the other five milkings on this cow were eliminated.
The one Brown Swiss cow showed great variation in maximum
rate of flow between milkings.

This resulted in within-cow

variation that was not compatible with that of the other
cows.

The data from this cow were eliminated from the group

comparisons, but will be presented separately at the end of
this section.

Thus the group comparisons in Trial 2 are

based upon observations bn six cows.
Analysis of variance showed highly significant differ­
ences in maximum rate of milk flow between milkings in
Trial 2.

The means were tested by Duncan's multiple range

test (Duncan, 1935)

see which differed significantly.

The mean rates of flow ranked from slowest to fastest are

81
TABLE 4
MILKINGS ARRANGED IK ASCENDING ORDER BY
MAXIMUM RATE OB BLOW EOR THE TWO OXYTOCIN TREATMENT TRIALS
Order:

1

2

3

4

5

6

7

A-l

0-2

0-3

B-l

A-2

0-1

B-2

3.16

3.4?

3.51

3.64

3.66

3.75

3.64

B-2

0-2

B-l

0-1

0-3

A-l

A-2

Mean Rate*3

4.81

5.08

5.14

5.68

5.84

6.76

6.84

Rt crni fi

----

Trial 1:
MiIkingsa
Mean Rate*5

(7 cows)

Significance0
Trial 2:
Milkingsa

(6 cows)

a. A = alter treatment; ± = tne u r s u miiong in m e penoa,
etc.
B = Before treatment; 1 = the first milking in the per­
iod, etc.
0 = oxytocin treatment; 1 = the first milking in the
period, etc.
b. The mean of the maximum rate of flow is expressed in
pounds per minute.
c. Those means underscored by the same line are not signifi­
cantly different.
presented in Table 4.

Those means underscored by the same

line are not significantly different.

(All other character­

istics in which significant differences were shown between
milkings by analysis of variance were analyzed in a similar
manner.) In the tables presented, those means shown to be
not significantly different by the analysis of variance were

82
not tested by Duncan's multiple range test.

They are under­

scored by a single line to indicate that they are not signi­
ficantly different.
The results in Table 4- show that the maximum rate of
flow in the two post-treatment observations is significantly
faster than the other milkings.

This is due to the repairs

made on the milking machine in the period between the treat­
ment and post-treatment controls.

The rate of flow in the

third oxytocin treatment is significantly faster than the
two pre-treatment controls and the second oxytocin treatment.
The first oxytocin treatment is significantly faster than
the second pre-treatment control and the second oxytocin
treatment.

The second oxytocin treatment is not signifi­

cantly different from either of the pre-treatment controls.
These results indicate that the oxytocin had a slight effect
on these cows, some of which milked quite rapidly.

This

slight effect of rate of flow could be due to a slightly
greater increase in pressure within the gland and the teat
cistern due to a more rapid let down caused by the high
level of oxytocin injected.

This increased pressure was

indicated by the fact that more of the cows leaked milk from
the teats following injections of oxytocin, than following
let down induced in the normal manner.

If there was an in­

creased pressure, the higher pressure could be reflected
more fully in the faster milking cows in which there is a

83
larger teat orifice and also possibly less tension in the
sphincter.

Thus the teat would open to a larger degree

under the increased pressure.

This could explain the

slightly increased rate of flow obtained in Trial 2 and not
in Trial 1 in which slower milking cows were used.
Average rate of flow.
The average rates of milk flow for each of the milkings
in Trials 1 and 2 are presented in Table 5*

There were no

significant differences in average rate of flow between
milkings in Trial 1, but there were significant differences,
as indicated by the means not underscored by the same line,
in Trial 2.

As was shown with the maximum rate of flow, the

two post-experimental control observations were significantly
faster.

The other milkings are in much the same order as

the maximum rate of flow, but the differences are not as
marked.

The average rate of flow also indicates that the

oxytocin injections had a slight effect on the rate of flow.
However, the effect is not as large, since average rate
of flow is partially a function of the machine operator
and does not reflect changes in the rate of flow of the
cow quite so sharply.
Level of Production.
The average production at each of the milkings in
Trials 1 and 2 is presented in Table 6.

In Trial 1,

84
TABLE 5
MILKINGS WITH AND WITHOUT OXYTOCIN TREATMENT
ARRANGED IN ASCENDING ORDER BY AVERAGE RATE OR PLOW
Order:

1

2

3

4

5

6

7

0-2

A-l

A-2

B-l

B-2

0-3

0-1

2.46

2.56

2.66

2.67

2.69

2.86

3.00

B-2

B-l

0-2

0-5

0-1

A-l

A-2

Mean Rateb

3-37

3.46

3.58

3.86

4.21

4.24

4.46

Significance

----

Trial 1:
Milking3,
Mean Rate^

(7 cows)

Significance0
Trial 2:
Milking3

(6 cows)

a. A = after treatment; 1 = the first milking in the period
etc.
B = Before treatment; 1 = the first milking in the peri­
od, etc.
0 = Oxytocin Treatment; 1 = the first milking in the per­
iod, etc.
b. The mean of the average rate of flow is expressed in
pounds per minute.
c. Those means underscored by the same line are not signifi­
cantly different.

production at the two pre—treatment milkings was highest.
However, it was significantly higher than only the produc­
tion at the second and third treatment milkings.

In Trial

2, the production at the first treatment milking was con­
siderably higher than the other milkings.
■yygpg

However, there

significant differences between the production at

the different milkings in the trial.

85
TABLE 6
MILKINGS WITH AND WITHOUT OXYTOCIN TREATMENT
ARRANGED IN ASCENDING ORDER BY LEVEL OF PRODUCTION
Order:
Trial 1:

1

2

3

4

5

6

7

0-2

0-3

A-l

A-2

0-1

B-2

B-l

(7 cows)

Milking3Mean Production

17.75 17.96 18.29 18.32 19.00 19.93 20.18

Significance0

-----

Milking3

(6 cows)
B-l

0-2

hCN
00

i
—1

Mean Production13 18.54- 18.79

B-2

0-3

A-l

A-2

0
1
j—1

Trial 2:

20.20 20.33 20.88 21.67

Significance0
etc.
B = before treatment; 1 = the first milking in the period,
etc.
0 = Oxytocin Treatment; 1 = the first milking in the per­
iod, etc.
b. The mean production is expressed in pounds per milking.
c. Those means underscored by the same line are not signi­
ficantly different.
The components of the milking curve.
The milking curve was divided into four component
parts, in order to see if the injection of oxytocin had an
effect on any of these components.
follows:

The components are as

86
Part

1 - The time required to reach a high rate of flow.
(The high rate of flow, as defined earlier, is
the interval during which a pound of milk is
produced in up to two seconds more than the
time required to produce a pound of milk at
the maximum rate.)

Part

2 - The time a high rate of flow persisted.

Part

3 - The interval from the completion of the high
rate of flow to the time machine stripping
began.

Part 4 - The machine stripping time.
The above parts of the milking curve do not correspond
directly to those listed in Part 1 of the thesis.

In Part

1 of the thesis, the three components were designed pri­
marily to eliminate the time in which the machine was idle,
while in this section, the curve is broken into 4 parts by
flow characteristics.
Table 7 shows the number of seconds which elapsed fol­
lowing attachment of the last teat cup, before a high rate
of flow was reached.

Although a high rate of flow was

reached sooner, following injection of oxytocin in both
trials, the differences were not statistically significant
in Trial 1, and not significantly different from all of the
control observations in Trial 2.

The greatest and most

consistent effect of the oxytocin injections is shown in

87
TABLE 7
MILKINGS ARRANGED IN ASCENDING ORDER OF THE
NUMBER OF SECONDS BEFORE A HIGH RATE OF FLOW WAS REACHED,
WITH AND WITHOUT OXYTOCIN TREATMENT
Order:
Trial 1:

2

3

4

5

6

7

0-2

0-3

0-1

A-2

A-l

B-l

B-2

(7 cows)

Milkinga
Mean Time

1

"K

37-14- 42.27 42.37 51-14 60.71 73-86 81.86

Significance0
Trial 2:

(6 cows)

Milkinga

0-1

A-2

0-2

B-l

A-l

B-2

19-33 23-67 30.50 31-67 41.33 4-2.33 54.00

Mean Time13
Significance

0-3

£

etc.
B = before treatment; 1 = the first milking in the period,
etc.
0 = Oxytocin Treatment; 1 = the first milking in the
period, etc.
b. The mean time is expressed in seconds.
c. Those means underscored by the same line are not signifi­
cantly different.
this portion of the curve.

This indicates that the milk

was let down more completely at the start of milking fol­
lowing the oxytocin injections, than at the control milkings,
Thus the high rate of flow started sooner.
The mean number of seconds during which a high rate of
flow persisted is shown in Table 8.

One milking in each

88
TABLE 8
MILKINGS ARRANGED IN ASCENDING ORDER OE THE
NUMBER OE SECONDS A HIGH RATE OE FLOW PERSISTED,
WITH AND WITHOUT OXYTOCIN TREATMENT
Order:

1

Trial 1:

2

5

4

-

5

6

7

(7cows)

Milkinga

A-2

Mean Timeb154.57

0-2

B-2

0-1

0-5

B-l

A-l

166.57 189.86 195.00 198,86 204.86 264.29

Significance0 _________________________ ______________
Trial 2:

(6cows)

Mi Ikinga
Mean Time13
Significance

Q

A-l

A-2

B-l

B-2

115-5

126.0

148.2

149-5

— -------------

0-2

-

0-5

0-1

155-8 157-7

195-2

-___ — ______

a. A = after treatment; 1 = the first milking in the period,
etc.
B = Before treatment; 1 = the first milking in the period,
etc.
0 = Oxytocin treatment; 1 = the first milking in the
period, etc.
B. The mean time is expressed in seconds.
c. Those means underscored By the same line are not signifi­
cantly different.
trial is significantly different from the others.

The high

rate of flow persisted significantly longer in the first
post—treatment control oBservation in Trial 1, that milking
had the slowest maximum rate of flow in that trial; this
indicates that machine factors may have Been responsiBle for
the slow rate of flow and consequent longer duration of the

89
high, rate of flow.

At the first milking at which oxytocin

was injected in Trial 2, the high rate of flow persisted
significantly longer than in the two post-treatment con­
trols.

A portion of the difference is due to the increased

rate of flow in the post-treatment control period result­
ing from the machine repairs.

However, the average time

during which the high rate of flow persisted in this milk­
ing is considerably, although not significantly, longer
than in all the other milkings.

A large portion of this

difference is due to one cow whose production was very high
at that milking.

Thus the high rate of flow persisted

longer.
Tables 9 and 10 present the mean number of seconds
between the high rate of flow and the time stripping
started, and the stripping time, respectively.

Although

large differences are shown, due to the large within milk­
ing variances, there are no significant differences be­
tween the means.

However, as the analysis of individual

cow differences shows, these two parts of the curve are
primarily under the control of the machine operator, thus
a lack of significant differences here does not give strong
evidence for a lack of an effect of oxytocin injections
upon this portion of the flow curve.

90
TABLE 9
MILKINGS ARRANGED IN ASCENDING ORDER OE THE
NUMBER OE SECONDS BETWEEN THE END OF A HIGH RATE OF FLOW AND
THE TIME STRIPPING STARTED WITH AND WITHOUT OXYTOCIN
Order:
Trial 1:

1

2

5

4

5

6

0-1

B-2

B-l

A-l

65.0 68.0

70.4

82.4

7

(7 cows)

Milking9,

0-5

Mean Timeb

50.0

A-l

0-2

99*9 100.5

Significance0_____________________________________________
Trial 2:

(6 cows)

Milking9
Mean Time13

0-1
-7-7

A-2

0-5

B-l

B-2

31-7 32.5

41.8

54.2

A-l

0-2

56*7 67.5

Significance0_____________________________________________
a. A = after treatment; 1 = the first milking in the period,
etc.
B « before treatment; 1 = the first milking in the period,
etc.
0 = Oxytocin treatment; 1 = the first milking in tbe
period, etc.
b. The mean time is expressed in seconds.
c. Tbose means underscored by tbe same line are not signi­
ficantly different.

An example of a faulty let doWn.
Figure 4 presents two milking curves for tbe Brown
Swiss cow No. 3006.

Number 1 is a normal control milking,

taken at tbe second post—treatment control observation.
Number 2 is a control milking at wbicb tbe cow failed to
let down ber milk normally.

Tbis was at tbe first post-

treatment control observation.

Altbougb tbis cow was a

91
TABLE 10
MILKINGS ARRANGED IN ASCENDING ORDER OF
STRIPPING TIME IN SECONDS WITH AND WITHOUT OZYTOCIN TREATMENT
Order

1

2

3

4

5

6

7

A-2

B-l

0-1

A-l

B-2

0-3

0-2

61.1

75*9

76.3

76.3

0-2

A-2

0-3

0-1

62.3

66.3

68.6

98.7

Trial 1: (7 cows)
Milkinga
Mean Time13

87*3 101.3 101.4

Significance0
Trial 2: (6 cows)
Milkinga
Mean Time^

B-l

A-l

B-2

99*0 106.0 150.3

Significance
etc.
B = "before treatment; 1 = the first milking in the period,
etc.
0 = oxytocin treatment; 1 = the first milking in the
period, etc.
h. The mean time is expressed in seconds.
c. Those means underscored "by the same line are not signifi­
cantly different.
slow milker at even the fastest milking, at which the maxi­
mum rate of flow was 4.5 pounds per minute, the maximum
rate of flow at the milking at which let down was faulty
was only 3*0 pounds per minute.

This rate was not attained

until late in the milking, during machine stripping.
Curve number 2 illustrates the type of curve which is
obtained when let down is faulty.

It is readily seen that

22

POUNDS OF MILK

PRODUCED

20

12

10

l

2

5

k

5

6

7

8

MINUTES AFTER ATTACHMENT OF THE TEAT CUPS

Figure

k.

Two making curves for cow No. 5006. Curve No. 1 repre­
sents a normal milking. Curve No. 2 illustrates the
type of response obtained when a faulty let down occurs.
The arrow indicates the time stripping began.

the rate of flow is slowed "by the let down failure.

The

operator began machine stripping when only six pounds of
milk had been removed.

The highest rate of flow was at­

tained during stripping, which substantiates the report of
Whittlestone (1954-) that in the case of faulty let down a
release of more oxytocin may be induced by manipulation of
the udder.

This type of response should not be confused

with one in which let down is normal, but rate of flow is
confused with one in which let down is normal, but rate of
flow is slow due to the morphology of the teat.

Oxytocin

injections did increase the rate of flow of this cow over
that of the faulty let down but did not cause an increase
over that when let down was normal.

The response of this

cow to oxytocin injections was also variable.

This raises

the question whether the udder may have been somewhat re­
fractory to oxytocin, rather than an insufficient let down
by the cow.

However, variations in sensitivity of the mam­

mary gland from milking to milking would be difficult to
explain.
Discussion.
The question of what effect the struggling of the cow
during injection may have should be posed at this time.
Cross (1955) reported that the primary effect of nervous
inhibition of let down is through a block of the reflex arc

94

-

since this arc is not needed when oxytocin is injected, no
effect of a block of this arc would be noted.

Adrenalin,

which may be released at this time, can also affect the
mammary gland.

However, the work of Cross (1953) and a

series of reports reviewed by Benson and Cowie (1957) in­
dicate that adrenalin affects let down by vaso-constriction,
reducing the amount of blood to the udder, rather than a
direct effect on the gland.

The doses injected in this ex­

periment should have been high enough to insure a blood level
of oxytocin high enough to overcome the effect of any decrease
in the amount of blood reaching the udder.

The work of

Turner and Cooper (194-1) and Donker et al. (1954-a) indi­
cates that if the animal is excited, larger doses of oxy­
tocin are needed to obtain a response, but the response is
the same as if smaller doses are given without struggling
on the part of the animal.

Thus, very probably, the strug­

gling of the cows had no effect on the let down caused by
the oxytocin injected.

It might be profitable to investi­

gate further the effect of injected oxytocin in unexcited
animals through the use of a continuous intravenous arrange­
ment, so that the cow would not be disturbed at the time
the oxytocin was given.
The major portion of the evidence from this experiment
indicates that let down failure was not responsible for the

95
slow rate of £low of all but one of the cows studied.

The

curves for the Brown Swiss cow illustrate a case where let
down failure caused an extremely slow rate of flow in a cow
that had a slow rate of flow following a normal let down.
Part III. Variation in Interval from
Preparation to Milking
In this experiment, the interval between the time let
down of milk was stimulated by preparation of the udder and
the time milking began was varied.

The purpose of the ex­

periment was to ascertain the effect which this variation
in interval might have on the rate of milk flow, level of
production, and milk composition.
Maximum rate of flow.
An analysis of variance of the effect of variation in
interval from preparation to milking upon maximum rate of
flow is presented in Table 11.

The data are based upon

180 milkings, 18 on each of 10 cows.

Each interval— 0, 4,

8, 12, 16, and 20 minutes— was used for each cow three
times, once in each of three replications.

The order of

the observations within each replication was randomized
for each cow.

(A similar analysis was also carried out for

average rate of flow, production and the first two of the
four components of the milk flow curve as described in

96
TABLE 11
ANALYSIS OP VARIANCE
SOURCES OF VARIATION IN MAXIMUM RATE OF FLOW
WHEN THE INTERVAL FROM PREPARATION TO MILKING WAS VARIED
Source of
Variation

Degrees of
freedom

Total

Mean
Square

"F"
---

179

4-.7^

Cows

9

84.03

186.7**

Interval

5

1.68

4.2**

45

0.41

0.91

Replication

2

0.78

0.70

Rep. x Cows

18

1.12

2.49*

Rep. x Interval

10

0.25

0.56

Residual (error)

90

0.45

---

Cow x Interval

♦♦Significant at the one percent level of probability.
Part II.

This analysis was also carried out for the percent

hutterfat and total solids, except that samples were taken
for composition analysis only in the second and third repli­
cations.

Thus only 120 samples were available for analysis.)

The highly significant differences between cows for
maximum rate of flow, as shown in Part II, were also shown
in this experiment.

All other criteria analyzed, including

milk composition data also, showed significant differences
between cows.

97
There were highly significant differences in maximum
rate of flow "between the different intervals "between pre­
paration and milking.

Table 12 shows the means, arranged

in order of increasing maximum rate of flow.

This table

shows that, in general, the maximum rate of flow decreased
TABLE 12
THE INFLUENCE OF THE TIME ELAPSING BETWEEN PREPARATION OF
THE UDDER AND THE ATTACHMENT OF THE TEAT CUPS UPON THE MILK FLOW.
MILKINGS ARRANGED IN ASCENDING ORDER OF MAXIMUM RATE OF FLOW.
1

2

3

4

5

6

Interval3"

20

16

8

12

4

0

Mean Rate13

6.32

6.42

6.53

6.59

Order:

Significance

6.45

6.80

Q

a. The figure listed is the time in minutes which elapsed
between preparation of the udder and attachment of the
teat cups.
b. The mean of the maximum rate of flow for the three ob­
servations on each of ten cows is expressed in pounds
per minute.
c. Those means underscored by the same line are not signifi­
cantly different.
with increasing interval from preparation of the udder to
milking.

The rate of flow when the udder was prepared and

the teat cups attached immediately afterward was significantly
faster than all of the other intervals except the fourminute interval.

The rate of flow with the four-minute

interval was significantly faster than when 20 minuses in­
tervened.

98
There was also a significant replication by cow inter­
action for maximum rate of flow.

This indicates that the

maximum rate of flow for some cows changed significantly
from one replication to the other.

This significant inter­

action is probably due to a rather marked decrease in pro­
duction for two cows between replications two and three.
Thus their rate of flow may also have decreased.

With the

sharp decrease in production, rate of drainage into the teat
and gland cistern, rather than the teat sphincter may have
become the controlling factor.
Average rate of flow.
There were no significant differences in average rate
of flow between the different intervals between preparation
and milking.

However, an inspection of Table 13 shows that

the means are in approximately the same order as the means
TABLE 13
THE AVERAGE RATE OE FLOW AND THE MEAN PRODUCTION FOR
EACH OF THE SIX INTERVALS BETWEEN PREPARATION AND MILKING

0
Ave. Rate (lbs/min)
Production (lbs.)

Interval expressed in minutes
8
12
16
4
20

4.69a
27.3

^.82
26.0

4.76
23.6

4.61
26.1

4.58
25.4

4.60
26.2

a. Each figure represents the mean of 3 observations on each
of 10 cows.

99
of the maximum rate of flow, except for the zero interval,
in which the teat cups were attached immediately after
preparation of the udder.

The average rate of flow for the

milking following this interval was less, due to a delay in
the start of a high rate of flow, as will be shown later.
The smaller magnitude of the other differences can probably
be attributed to the fact that average rate of flow is par­
tially controlled by the machine operator, thus any effect
upon the cow is reflected less sharply in the average rate
of flow.
There was a significant difference between replications
in average rate of flow.

The average rate of flow for the

third replication was significantly slower; this may be
due in part to a decrease in production on the part of all
of the cows, although it was more marked in some than
others.

The decreased production results in a decreased

average rate of flow, while it may not result in a decrease
in maximum rate of flow, because a smaller proportion of
the total production is produced at a high rate of flow.
Level of production.
Table 13 presents the mean production of the cows fol­
lowing each of the six intervals between preparation.

There

were no significant differences in the production between

100

TABLE 14
THE MEAN NUMBER OF SECONDS IN EACH OF THE FOUR COMPONENTS
OF THE MILK FLOW CURVE FOR EACH OF THE SIX INTERVALS
BETWEEN PREPARATION AND MILKING

Component

0
Sec.

Part 1

33.9a

Interval Expressed in minutes
4
16
8
12
Sec.
Sec.
Sec.
Sec.

20
Sec.

24.4

23.9

29.4

29.0

30.6

Part 2

162.8

, 157.5

152.0

145.6

143.6

157.2

Part 3

91.4

73.8

92.0

81.4

76.8

97.3

Part 4

60.3

71.0

70.8

78.4

75.2

82.2

a. Each, figure represents the mean of 3 observations on each
of 10 cows.

the various intervals between preparation and milking, al­
though the production when milking began immediately after
preparation was considerably higher.
There were significant differences between replications;
the mean production was significantly less in replication
three and was especially lowered for two cows.

The lowered

production for these two cows contributed to the signifi­
cant replication by cow interaction which was also shown.
The decreased production contributed to the decreased aver­
age rate of flow in replication three and the greater decrease
for "the two cows probably contributed to the replication by
cow interaction for maximum rate of flow.

101

The milk flow curve.
The milk flow curve for each milking was broken into
four component parts as described in Part II of this thesis.
The time between attachment of the teat cups and the start
of high rate of flow, and the time during which high rate
of flow persisted were analyzed in detail as described
earlier.

Since the time from the end of a high rate of

flow to the time stripping started, and the stripping time
are less closely associated with the individuality of the
cow, a less detailed analysis, using the means of the values
obtained for the three replications, was carried out for
these criteria.
The mean number of seconds in each of the four compo­
nents of the milk flow curve for each of the six intervals
are presented in Table 14.

There are no statistically

significant differences in the number of seconds between
treatments for any of these components.

However, the time

before a high rate of flow was reached following the 0,
12, 16, and 20 minute interval was considerably greater
than the 4 and 8 minute intervals.

The time for the zero

interval would be expected to be longer, because the let
down reflex may not be completed until after the teat cups
have been applisd*

The fact that the time for the 12, 16,

and 20 minute intervals is nearly as long may indicate that

102

the effect of the original let down, may have been dissi­
pated.

The high rate of flow persisted somewhat longer

with the zero interval than the other intervals.
In Tahle 15 the mean number of pounds produced during
each of the four components of the milk flow curve is pre­
sented.

There were no statistically significant differences
TABLE 15

THE MEAN NUMBER OR POUNDS OR MILK PRODUCED DURING EACH OR
THE ROUR COMPONENTS OR THE MILK PLOW CURVE ROR EACH
OR THE SIX INTERVALS BETWEEN PREPARATION AND MILKING

Component

Part 1

Interval expressed in minutes
4
16
8
12
lbs.
lbs.
lbs.
lbs.

0
lbs.
2.1a

20
lbs.

1.5

1.5

1.9

1.9

2.0

Part 2

17.5

16.7

16.7

15*6

15.4

15.4

Part 5

5.1

4.6

4.3

5.0

4.5

5.3

Part 4

2.6

3.2

3.1

3.6

3.6

3.5

a. Each figure represents the mean of three observations on
each of 10 cows.
in the pounds produced between treatments for any of these
components.

However, the amount produced at a high rate

of flow following the zero interval between preparation
and attachment of the teat cups was considerably higher than
any of the others.

The "R" value obtained in analysis of

variance approaches significance.

The differences between

103
the means of the other components were slight.
The average rates of flow for each of the four compo­
nents of the milk flow curve for each of the six intervals
are presented in Table 16.

There are no significant dif­

ferences in the average rates of flow between treatments
TABLE 16
THE AVERAGE RATE OE FLOW DURING EACH OF THE FOUR COMPONENTS
OF THE MILK FLOW CURVE FOR EACH OF THE SIX INTERVALS BETWEEN
PREPARATION AND MILKING

Component

Interval expressed in minutes
0
16
20
4
8
12
H d/ihilh« lb/min. lb/min. lb/min. lb/min. lb/min.

Part 1

3-75a

3.66

3.75

3.88

3.96

3.92

Part 2

6.83

6.56

6.63

6.68

6.66

6.13

Part 3

3-36

3.74

2.81

3.68

3.51

2.66

Part 4

2.57

- 2.70

2.63

2.75

2.88

2.55

a. Each figure represents the mean of three observations on
each of 10 cows.
for any of the components, although the average rate of flow
following the zero interval in Part 2 of the curve is
slightly higher.

This reflects the significantly higher

maximum rate of flow for the milkings following this inter­
val between preparation of the udder and attachment of the
teat cups.

104
Composition analysis.
In replications two and. three, samples of milk were
taken from each, cow for each milking to see if changes in
milk composition occurred, which could "be related to the
rate of milk flow or production.
The mean percentage of hutterfat, total solids, solids
non-fat, and protein are presented in Tahle 17.

An analysis

of variance was made for percentage of "butterfat and total
TABLE 17
THE PERCENTAGE BUTTERPAT, TOTAL SOLIDS, SOLIDS NON-FAT AND
PROTEIN OF MILK-FOLLOWING SIX DIFFERENT INTERVALS
BETWEEN PREPARATION AND MILKING

0

Constituent

Interval expressed in minutes
4
16
8
12
20

%

%

%

%

°/o

%

3.8

3.6

3.6

3.7

3.3

3.5

12.5

12.3

12.3

12.4

12.1

12.1

Solids Non-fat

8.7

8.6

8.7

8.6

8.7

8.6

Protein

3.4

3.4

3.4

3.3

3.4

3.4

Butterfat
Total Solids

solids, as already described.

No significant differences

between the means of the samples from the milk: obtained for
these two components.

Since protein percentage is closely

correlated with total solids, and solids non-fat is related
to total solids and hutterfat, these two components were

103
not analyzed, statistically.

However, an inspection of Table

17 shows that there is little difference between treatment
means.

Although the difference was not statistically sig­

nificant, the percentage butterfat and total solids was
higher following the zero interval between preparation and
milking.

Since solids non-fat changed but little, the dif­

ference in total solids was due to the difference in butterfat percent.

This higher butterfat percentage with the zero

interval indicates that more of the residual milk was ob­
tained following the shorter interval.

This interval which

yielded the highest butterfat percentage also yielded the
highest production, and although differences between the
means are not statistically significant, they indicate that
more of the residual milk may have been removed at the
milking after the shorter interval following preparation.
The normal amount of residual milk has been reported to be
approximately 20 percent by Swanson and Hinton (1951) and
Espe and Smith (1954).

Thus changes in the amount of re­

sidual milk would be quite small in relation to the total
amount produced.

The many other variables that affect milk

production at each milking make these rather small changes
difficult to demonstrate statistically.

106
Discussion.
The results of this experiment indicate that a slightly
faster maximum rate of flow is obtained when the teat cups
are attached immediately after preparation.
is little difference in average rate of flow.

However, there
In the case

of attachment of the teat cups immediately after prepara­
tion, the lack of a change in average rate of flow is ac­
counted for by the increased time before the maximum rate
of flow is reached.

These results indicate that let down

may have a slight effect upon the rate of flow, but again,
as pointed out in the section on oxytocin injections, these
differences are slight in contrast to the differences be­
tween cows.

However, the results of these two experiments

indicate that differences in completeness of let down may
have some effect on day to day variation in rate of flow.
Although the differences are not statistically signi­
ficant, the production and butterfat tests indicate that the
shorter interval between preparation of the udder and attach­
ment of the teat cups did result in a more complete evacu­
ation of the udder.

The differences are small and not as

marked as those reported by Miller and Petersen (1941) and
Ward and Smith (1948, 1949) but appear to be comparable to
the results reported by Dodd et al. (1949) and Roark et al.
(1952).

Thus it would seem that increasing the interval

107

between preparation and milking beyond tbe time required
for let down to occur has a slight deleterious effect on
milk production and fat test.

Therefore, the most desirable

milking routine would include attachment of the teat cups
to the udder as soon after let down has occurred as possible.
It is also possible that if the interval between pre­
paration and milking is too long, or is too irregular, the
cow may fail to let down her milk upon stimulation by pre­
paration of the udder.

This could be responsible for the

faulty let down illustrated in Part II.

This possibility

is another reason that the teat cups should be attached as
soon after let down has occurred as possible.

S E C T I O N

B

DETERMINATION OF THE ACCURACY OF MILK MEASURING DEVICES1
Part I.

The Milk-O-Meter

The Milk—O-Meter was used to measure rate of flow and
milk production in the experiments which were conducted in
the main dairy barn on the Michigan State University campus.
Prior to this experiment, consideration was given to install­
ing weigh buckets suspended from spring scales.

However,

this installation did not appear to be practical, and the
availability of the Milk-O-Meter prompted its testing.
Description of construction and operation of the Milk-O-Meter.
The Milk-O-Meter is constructed of Plexiglas, and the
entire device weighs about five pounds.

The shell of the

meter is designed in two pieces held together by vacuum
when the meter is in operation.

The weighing receptacle is

a two-compartment tray which operates on a rocker arm like
a "teeter-totter.11 The rocker arm extends outside the shell
to the dial portion of the meter which contains two magnets
and the mechanism that turns the pointer.

The meter and

its component parts are illustrated in Figure 5*

1A portion of these data have already been published.
See Hupp et. al. (1957).
- 108 -

109

Figure 5*

Th.© Milk-O-Meter, showing component parts.

The

upper shell and haffle plate are on the left;
the lower shell with rocker arm and weigh tray
are on the right.

110

Milk entering the top of the meter flows into one com­
partment of the tray.

When the tray contains slightly less

than 4 ounces of milk, the weight of the milk pulls against
the tension of one magnet, allowing it to rock downward and
discharge 4 ounces of milk while the other end of the tray
is moved into position to catch the next 4 ounces.
side of the tray fills, the other side drains.

As one

The downward

movement of the milk tray on the rocker arm moves the dial
to indicate each quarter-pound increment.

Milk weights are

read directly from the dial on the front of the meter.

Be­

cause the system is under vacuum, the milk goes directly
into the pipeline after leaving the weigh tray.
The sampler, which can he attached directly to the
lower shell of the meter, removes a small amount of milk
from each half-pound that passes through the meter.

At the

end of the milking, the drain at the bottom of the sampler
is opened while the unit is under vacuum.

Most of the milk

is drawn into the pipeline, leaving about one ounce of
milk in the sampler.

Then the vacuum is turned off, and

the remaining amount of milk is drained into a sample bottle.
In this way, a representative sample of uniform size is ob­
tained for each cow.
Several modifications of the meter were tested, and
although only two of the types were used to obtain data on
rate of milking characteristics, the results of tests on all

Ill
types will be presented, since they indicate the general
accuracy of the meter.
The meter which will be referred to as Type A was de­
signed for zero error at a rate of flow of six pounds per
minute.

Type B meter, designed for zero error at a rate of

flow of three pounds per minute, was available commercially
at the time it was tested.

Meters of Types 0 and I) were

both designed for zero error at a rate of flow of four
pounds per minute.

Meter D was a modified model which was

available commercially at the time of its testing.

It had

an improved type of baffle plate designed to reduce the
error due to differences in rates of milk flow between cows.
Method of testing under operating conditions.
In the earlier trials, conducted in the main dairy
barn, which is equipped with a DeLaval pipeline milker, the
milk from the cow was passed through the meter and into a
milker bucket so that it could be weighed on a Toledo plat­
form balance, the accuracy of which had been verified.

The

weights then were compared with the meter reading.
The maximum rate of milk flow of the individual cows
was determined in each trial by noting the time required to
produce the 5 pounds of milk in the interval between 5 and
10 pounds.

The sampling device was tested by taking a two-

milking composite sample from the sampling device on the

112

meter.

A duplicate sample was dipped from the bucket in

the usual manner.

Routine Babcock tests were run on these

samples and the results were compared.
After the tests in the barn had been completed, a flow
device was constructed which delivered water to the meter at
constant rates of flow so that further tests could be made
under more optimum conditions.
The test apparatus consists of a reservoir for the
water, a series of glass tubings with openings of various
sizes which are changed to regulate the rate of flow, an
air inlet into the hose leading from the reservoir to the
meter, a mounting bracket for the meter, a milker bucket
to receive the water from the meter, and a source of vacuum.
The test
The

apparatus is shown in Figure 6.
reservoir has a valve at the bottom which is used

to turn the water on and off at the beginning and end of
each test.

The air inlet into the hose leading from the

reservoir to the meter is provided so that the water goes
into themeter in a manner simulating
the cow.

the flow of milk from

The four rates of flow used and their standard

deviations were 2.9 - 0.19; 5*2 - 0.20; 7*2 - 0.27; and
11.9 - 0.99 pounds per minute.

The variation in rate of

flow was primarily due to the fluctuations in vacuum level.
It was not possible to control the vacuum level adequately.

1X3

Figure 6.

The flow apparatus for testing the Milk-O-Meter.
From left to right are the 10 gallon reservoir
with the shut off valve, the removable glass
tubing which controls rate of flow, the hose
leading to the meter, the Milk-O-Meter, the
bucket which receives water from the meter and
a hose leading to the source of vacuum.

114
Vacuum levels ranged from 13 to 18 inches of mercury.

How­

ever, the effect on rate of flow was not large, as indicated
by the relatively small standard deviations about each rate
of flow.
The meters were tested by running water through the
meter for a fixed period of time for each rate of flow.
Then the meter reading was recorded.

Next, the weigh pan

was tripped and the amount of water in the weigh pan was
measured in the calibrated hose leading to the milker
bucket.

This amount was added to the meter reading to ob­

tain the total amount of water as indicated by the meter.
The amount of water that had passed through the meter and
into the milker bucket was weighed on an accurate Dayton
platform scale, or in some cases a spring scale, the accu­
racy of which had been checked repeatedly by comparing the
weights on it to those on the Dayton scale.

The percent

error for each run was calculated by subtracting the scale
weight from the meter weight and dividing the difference
by the scale weight.
The factory recommended method of testing*
The factory recommended method for testing the MilkO-Meter was also used to check its accuracy during the course
of the experiments.

The method is as follows:

place the

Milk-O-Meter on the pipe clamp bracket and make sure that

113
I'fr Is level, then install the lower shell and. weigh tray,
leaving the upper shell and. baffle system removed..

Pour

water in both sides of the weigh tray and trip to drain
naturally.

Using a standard volumetric pipette graduated

to deliver 125 mis. or a 100 ml. volumetric pipette plus a
25 ml. graduated pipette, fill one side of the weigh tray
and note the exact milliliter at which the rocker tripped;
repeat for the other side.

The combined capacity of the

two sides in tripping should be between 218 and 250 mis.
If the total required is outside these limits, the meter
will not yield results of acceptable accuracy and should be
recalibrated.

In these tests a 100 ml. volumetric flask and

a 25 ml. graduated pipette were used.
Results of tests conducted under operating conditions.
A total of 545 observations were made in the main tiestall barn in which milk from the cow was passed through the
meter and into a milker bucket so that it could be weighed
on a platform balance.

These observations were made in a

series of seven trials on four types of meters.

The results

of these observations are presented in Table 18.
The gross error as expressed in the table is the mean
of the errors encountered in each individual milking.

The

net error is obtained by subtracting the negative deviations
from the positive deviations and dividing by the total number

116
TABLE 18
THE GROSS AND NET ERROR OE FOUR TYPES OF MIIK-O-METERS
IN ESTIMATING THE PRODUCTION AT A SINGLE MILKING

Trial no.

Metera
Type

No.

Average
production
per milking
lb.

Observa­
tions

Gross
error

Net
error

%

%

1

A

54-

21.3

1.37

-0.74-

2

A

40

14.4

1.46

-0.47

3

A

26

24.9

2.07

-1.62

4

B

36

14.7

1.52

-0.31

5

C

108

15.9

2.89

-0.53

6

D

62

16.1

2.36

-0.06

7

D

219

15.4-

2.28

-0.52

545

16.5

2.20

-0.53

All .
a. Type A
pounds
Type B
pounds
Type C
pounds
Type D
pounds

was
per
was
per
was
per
was
per

designed for zero error at a flow
minute.
designed for zero error at a flow
minute.
designed for zero error at a flow
minute.
designed for zero error at a flow
minute and Has a modified type of

of observations witHin tHe trial*

rate of 6
rate of 3
rate of 4
rate of 4
Uaffle plate.

THe net error represents

the amount the meter was in error in estimating the total
amount of milk produced in a trial.
error, or bias, was negative.

In all cases the net

117
With, the Type A meter, the smallest amount of error was
found in Trials 1 and 2.

Trial 1 was conducted with a

group of 10 cows with random rates of flow.

In Trial 2, a

group of 10 cows was selected for their slow rate of milk
flow.

In Trial 3, using the same meter as for Trial 2 on

cows selected for rapid rate of flow, the error was consid­
erably larger.

In Trial 4, in which a meter designed for

zero error at a flow rate of three pounds per minute was
used on cows selected for a slow rate of flow, the error
was also quite low; however, preliminary trials had shown
this meter to underestimate the production of the faster
milking cows.
The mean gross error in Trials 5 through 7» using the
meter designed for zero error at four pounds per minute,
was somewhat greater than that encountered in the above
trials.

However, the net error was not increased, indicat­

ing that the increased gross error was primarily due to
random variation.

A portion of the increased error in Trial

5 may have been due to a changed procedure in which more
milkers who were not as familiar with the procedure were
involved.
When samples were taken for butterfat testing in Trial
5, correcting for the amount of milk removed may have intro­
duced some error.

Some of the increased error in Trials 6

118
and 7 m&y have been due to the effect of continuous opera­
tion of the meter over a period of several months.
Factors contributing to the error of the meters.
Several factors may contribute to the error of the meters.
In preliminary observations and during Trial 1, it was noted
that the error was greater for the fast milking cows.

There­

fore the cows in Trials 2, 3> and 4 were selected on the
basis of rate of milk flow.

The same 10 slow milking cows

were used in Trials 2 and 4.

The mean of the maximum rate

of flow of these cows was 4 pounds per minute.

In Trial 2,

they were milked with the meter designed for zero error at
a rate of flow of 6 pounds per minute.

In Trial 4, they

were milked with the meter designed for zero error in Trial
2 was 1.26 percent; in Trial 4, it was 0.95 percent.

In

both cases the errors were small, indicating that the meter
was quite accurate for the slower milking cows.
Five cows selected for rapid rate of flow were used in
Trial 3.

The average of their maximum rate of flow was 9*6

pounds per minute.

In this case, the gross error in esti—

mating the daily production was 1.87 percent——quite high in
comparison with that obtained on the slow milking cows.
Note that the error in estimating the production of the
rapid milking cows was almost all in one direction; the net
error or bias was minus 1.63 percent.

These results show

119
that the meter with the original type of "baffle plate was
less accurate and underestimated the production of the
faster milking cows.
In Trial 7> using Type D meters which had a modified
"baffle plate designed to reduce the "bias due to rate of
flow, one-half the herd was measured three times, the other
half only once.

The cows which were measured three times

have "been divided into groups according to rate of flow
for comparison purposes.

The results of these comparisons

are presented in Table 19.
TABLE 19
THE EFFECT OF RATE OF FLO# OH GROSS AND HET ERROR
IH ESTIMATING DAILY PRODUCTION IN TRIAL 7
Range in
rate
lb./min.

Cows
Ho.

Average daily
production
lb.

Gross
error

Het
error

%

%

2-4

7

40.8

1.13

+0.47

4-6

9

39.4

1.92

-1.37

Over 6

2

34.1

2.63

-1.44

Although the new type of baffle plate apparently reduced
the error due to rate of flow, it was not eliminated.

The

gross and net error both increased as the rate of flow in­
creased.

The meter was biased against faster milking cows

120

and tended to underestimate their production.

The fast

milking cows measured earlier in Trial 3 were nearing the
end of their lactation at the time of Trial 7, so the rapid
rates of flow were not attained.

It is possible that

greater errors would have been met had these rapid milking
cows been available for measurement.
The tendency to overestimate the production of slow
milking cows and to underestimate that of the faster milking
cows is due to differences in the amount of milk which
drains into the weigh tray after it begins to trip.

The

tray is designed to start falling before it contains four
ounces.

Less than four ounces in the case of slow milking

cows, and more than four ounces with fast milking cows, are
in the tray before milk begins to flow into the other side.
This results in an overestimation or underestimation of the
production.
The new baffle plate apparently reduced the error by
shortening the effective time required for the weigh tray
to fall, but it has not solved the basic problem responsible
for the bias due to rate of flow.
The cows which were measured for three days in Trial
7 were divided into three groups according to level of pro­
duction.

Table 20 shows the mean gross and net errors in

estimating the daily production of these cows.

As the level

121

TABLE 20
THE EFFECT OF LEVEL OF PRODUCTIOH
UPON GROSS AHD HET ERROR IN TRIAL 7
Level of Obserproduction vations
No.

Production per
Day
Mean
Range
lbs.
lbs.

Lowest V3

25

16.7

2.9-27.0

0.38

Middle V3

26

32.4

27.4-36.8

Highest V-3

28

48.1

37.4-61.5

Gross error
lb.

error
Net i
lb.

>

2.25

+0.06

+0.36

0.57

1.75

-0.19

-0.58

0.79

1 •66

-0.35

-0.73

%

of production increased, Loth the gross and net error in
pounds of milk increased.

While the percent gross error

decreased with increased level of production, the percent
net error or bias increased.
During Trial 7, there appeared to be a difference in
error between meters.
used in this trial.

Three meters, all of Type D, were
The errors of the individual observa­

tions were analyzed by meter.
are presented in Table 21.

The results of this analysis

Two of these meters, No's 1-1460

and 1-1553 were used in the experiments reported in Part II.
These results indicate that there is a difference in error
between meters; both meters used in Part II had a negative
bias.

122

The accuracy of the sampling device.
Since milk samples were taken for composition analysis
in the trial in which the effect of the variation in the
interval between preparation of the udder and milking, the
results of tests on the accuracy of the sampling device
which was used with the meter will also be presented.

Two

milking composite samples were taken from 54 cows for butterfat testing.

The mean test of all the samples taken from

the meter was 3.82 percent butterfat, while the samples
dipped from the buckets in the usual manner averaged 3*79
percent butterfat.

This difference of 0.03 percent butter-

fat was 0.79 percent of the total amount.
The greatest difference between tests was 0.3 percent
which was encountered three times, 15 of the samples dif­
fered by 0.2 percent; 28 differed by 0.1 percent, and 8
were identical.

The test of the sample removed by the

meter was higher in 28 cases and lower in 18 cases.
The above data indicate that the sampling device re­
moves a very representative sample of the milk.

Since each

half pound produced is sampled under the same conditions,
this sample may be more accurate than one obtained in the
usual manner.
It was observed that different combinations of the
meter, sampler, and gaskets and weigh pan removed different

123
amounts of milk for each pound of milk produced.

The amount

of milk removed by several different combinations of meter,
sampler, and gaskets and weigh pan is presented in Table
22*

Since the capacity of the sampler receptacle is about

280 milliliters (ml.), the combination removing the largest
amount of milk from each pound would fill the receptacle
when 30 pounds of milk were produced.

A cow producing more

milk than this probably would be underestimated in fat test
since the last milk would not be sampled.

If the sampler

removed the average amount, 53 pounds would have to be pro­
duced before the sampler would be filled.

This level of

production is seldom attained.
On the other hand, a minimum of about 20 ml. are required
for the Babcock test.

On several occasions, less than 2.9

ml. of milk were removed per pound of milk.

In this case,

a minimum production of 7 pounds per day would be required
for a butterfat test.

If the sampler removed the average

amount, daily production of 4 pounds would be sufficient to
obtain a butterfat test.

However, in the trials in which

the sampler was used, production was quite high, so that the
latter would not be a problem.

The sample was drained into

a 240 ml. container which in no case ran over, so the capacity
of the 280 ml. sampler was not exceeded.
Although variations exist in the amount of milk removed,
a representative sample of milk which was suitable fop compo­
sition testing was obtained.

124
TABLE 21
THE GROSS AND NET ERROR OF TWO DIFFERENT MILK-O-METERS
Observa­
tions
No.

Average prod,
per milking
lb.

Gross
error

1-1460

58

13.4

2.09

-0.81

1-1553

110

15.5

2.34

-1.39

Meter no.

Net
error

%

TABLE 22
THE AMOUNT OF MILK REMOVED BY DIFFERENT COMBINATIONS
OF METER, SAMPLER AND GASKETS FOR THE MILK-O-METER SAMPLER

Cows

Average prod,
per milking

No.

lb.

A

4

25.8

9.4

7.0-10.5

B

5

17.6

5.8

4.2-6.4

C

9

13.5

7.2

6.4-8.0

D

6

. 16.5

4.3

4.0-6.4

E

14

17.3

5.8

4.3-7.6

F

12

18.7

4.5

2.9-6.5

G

15

17.0

4.8

2 .9-6.5

H

16

15.9

4.1

2.3-6.4

All

81

17.1

5.3

2.3-10.5

Combina­
tion

Milk removed/lb. 0:
milk
Range
Mean
ml.
ml.

125
On the basis of the results of Trials 1 through 5j the
meter was adopted for use as the measuring device for deter­
mining the rate of flow and production in the experiments
reported in Part II.

Meter trial 6 was conducted following

Trial 1 in Part II and Meter Trial 7 was conducted following
Trial 2 in Part II.

The results of these trials are included

to give an indication of the accuracy of the meter at the
time it was used in these experiments.
Results of tests in the laboratory.
Although the experiments reported in Part II were con­
ducted in August and September 1957} and- the flow apparatus
was not completed until March 1958*

the meters used were

tested on the flow apparatus at this time.

The results of

tests on the accuracy of the meters used are presented in
Table 23.

It must be realized that the meters could have

become less accurate due to constant use during the inter­
vening period.

Two of these three meters required re-adjustment,

as indicated by these tests.

The accuracy of the meters after

re-adjustment is also presented in Table 23.
In evaluating the percentage errors reported in Table
23, it should be pointed out that these errors do not cor­
respond directly to the error which might be ejected in
measuring production at a given milking.

These values are

obtained at a constant rate of flow, while that of the

126
TABLE 23
THE PERCENTAGE ERROR OF THE THREE MILK-O-METERS
USED IN PART II AT FOUR RATES OF FLOW,
BEFORE AND AFTER ADJUSTMENT
Meter
No.

Percent Error at a Rate of Flow of:
(Expressed in pounds per minute)
2.9
5.2
7-2
11.5

Mean
Error

Mean
Bias

A - Before Adjustment:
l-1553a■ -0.10

+0.64

-1.04

-3.58

1.34

-1.02

1-14-60

+2.72

+1.66

-0.28

+0.82

1.37

-0.98

1-14-54-

+3.02

+0.42

-0.51

+1.62

1.39

+1.14

B - After Adjustment:
l-1460b

+2.59

+2.04

-0.56

-1.08

1.57

+0.75

1-14-54-

+2.02

+1.95

+1.72

-1.60

1.82

+1.02

a. Meter No. 1-1553 was not adjusted.
b. THe pointer on Meter No. 1-14-60 may Have been slipping
following adjustment, yielding a positive error.

individual cow varies.

A period early in tHe milking and

during stripping will be slow wHile tHe middle period will
be at a faster rate.

THus in meters wHicH Have a positive

error at tHe slow rates of flow, and a negative error at
tHe faster rates, tHere will be some compensation wHicH will
reduce net errors in estimating total production.

However,

tHe errors presented in tHe table sHould correspond closely
to tHe error at a given rate of flow for any cow.

These

values were obtained under optimum conditions and do not

127

include errors which may result from vibration, improper
leveling of the meter, bursts of air, and other factors.
Thus errors in herd operation could be slightly higher.
It should also be mentioned that no cows tested had an
average rate of flow that even approached 11.5 pounds per
minute at the maximum rate.

Thus errors in excess of 3 per­

cent at the 11.5 pound per minute flow rate should not be
considered as serious as errors of that magnitude at other
rates of flow.
The recording meter used in the experiments reported in
Part III of the thesis was tested on the flow apparatus
before and after the experiments were conducted.
of these tests are presented in Table 24.

The results

The accuracy of

the pin-type distribution tray, used in Trial I of Part II
was also investigated at this time.

The results of this

investigation are also presented in Table 24.

This table

shows that the accuracy of the recording meter changed but
little during the course of the experiment, but the results
indicate that some negative bias probably was present in all
observations during the trial.

In this particular laboratory

test, the pin-type baffle tray gave greater accuracy than
the modified baffle tray that was used in the meter when re­
cordings were made.

However, other trials indicated that

the overall characteristics of the pin-type baffle tray were

128
TABLE 24
THE PERCENTAGE ERROR OF THE RECORDING MILK-O-METER
BEFORE AND AFTER THE EXPERIMENTS REPORTED IN PART III, AND
WITH THE PIN-TYPE BAFFLE TRAY
Combin.
No. Percent Error at a Rate of' Flow of:
Obser.
No.
(Expressed in pounds per minute)
2o9
5.2
7.2
11.5
fo
%
%
%

Mean
Error

Mean
Bias
%

%

1

20

+0.75

-1.56

-2.48

-3.30

2.02

-1.65

2

24

-0.45

-1.30

-2.70

-3.99

2.11

-2.11

0.72 -0.28
-2.00
+0.02
+0.64
+0.22
8
3
Combination 1 is the Meter with the LMB distribution tray before
experiments.
Combination 2 is the Meter with the LMB distribution tray after
experiments.
Combination 3 is the Meter with the pin-type distribution tray.

not as desirable.

These results show that there snould oe

only slight differences in error between Trial 1 in Part II
in which the pin-type baffle tray was used and Trial 2 in
Part II in which the modified baffle tray was used.
Discussion.
All of the experiments that were conducted to determine
the accuracy of the Milk-O-Meter indicate that it is quite
accurate and that a portion of the error is random and tends
to cancel with repeated observations.

The amount of negative

bias increases with increased flow rate, but the increase is
not rapid, nor does it appear to be strictly linear, so that

129
it would, bo difficult to attempt to correct for this bias due
to rate of flow.

The meter also tends to become inaccurate

with continued use.
Tests conducted on the meters immediately after use in
the trials reported in Part II indicate that the meters were
quite accurate at the time they were used in these trials.
The accuracy of the recording meter changed little during the
course of the experiments in which it was used.

However, it

would have been desirable to have had the meter adjusted for
zero error at a slightly higher rate of flow, to avoid the
negative error which was encountered.
Part IIo

Spring Scales

Spring scales were used for measuring production and
rate of flow in the experiments reported in Part I.

Although

spring scales have been used extensively by several groups of
workers, and Mackay et al. (1958) reported errors ranging from
zero to 1.89 percent with an overall average of 0.40% it was
considered desirable to make limited tests on the accuracy
of the scales which were to be used in this experiment.

Three

scales were tested, two were Chatillon 60 pound capacity spring
scales graduated in tenths of a pound.

One was new, the other

had seen limited use in production testing.

The third scale

was a 40 pound capacity scale, also graduated in tenths of a
pound.

130
The three scales were tested by weighing test weights
of known mass*
used.

Weights ranging from 5 to 30 pounds were

The weights were placed in a bucket, and the gross

weight on the scale noted.

At the end of the tests, the

tare weight of the bucket was determined and subtracted
from the net weight as indicated by the scale to determine
the error of the scales.
The accuracy of the scales in determining the amount of
milk in the milker bucket with the hoses attached was then
investigated.

The weigh bucket with the hoses attached was

suspended from the spring scale.

The amount of milk at the

end of milking as indicated by the scale was noted then.
The milk was carefully drained into a pail and weighed on
the 40 pound capacity scale, which had been shown to have
a high degree of accuracy.

The weight as indicated by the

check scale was subtracted from the weight as indicated by
the scale from which the weigh bucket was suspended, to
determine the error of the latter.
Results of tests.
In testing the accuracy of the spring scales in weigh­
ing test weights of known mass, the greatest error en­
countered with either scale of 60 pounds capacity was 0.10
pound.

The greatest error encountered with the 40 pound

capacity scale was 0.05 pound.

All errors with all three

131
scales were positive.

The mean error of the new 60 pound

capacity scale was 0.39 percent.

The used 60 pound capacity

scale had an error of 0.47 percent, while the 40 pound capa­
city scale

had an error of only 0.07 percent.

The greatest

difference between weight of the milk as

determined

in the milker bucket with the hosesattached and

the weight

on the check scale was 0.7 pound, this was an

error of 2.94- percent of the total amount of milk weighed.
The mean error of the scales in all tests was 1.43 percent.
This indicates that the weight of the hoses and the weight
of the milk in the hoses accounts for a considerable degree
of error in estimating the weight of the milk.

It was noted

that at the start of the milking, milk accumulated in the
milk hose for several seconds before any weight was indi­
cated on the scales.

Thus part of the inaccuracy may be

reflected in the first pound produced, and less of the
error is reflected in each additional pound.

These errors

are small enough that they should not materially affect the
results of the rate of flow determinations based upon these
weights.

SUMMARY AND CONCLUSIONS
Addition of water to the concentrate mix fed at milk­
ing time reduced the eating time by 30 percent.

The maxi­

mum rate of milk flow was not changed significantly by the
addition of water to the concentrate mix, but the average
rate of flow decreased 12 percent.

However, this statis­

tically significant decrease was due primarily to increased
time spent by the machine operator during stripping and can­
not be attributed to the effect of addition of water to the
concentrate mix upon the milk flow characteristics of the
individual cow.
Oxytocin injections had no effect on rate of flow in
one trial involving slow milking cows and only a slight
effect on rate of flow in a trial involving average and fast
milking cows.

This effect on faster milking cows may have

been due to a greater increase in intra-mammary pressure
following oxytocin injections.

These changes in rate of

flow are small compared to differences in rate of flow be­
tween cows, and indicate that let down failure was not re­
sponsible for the slow rate of flow of all but one of the
cows studied.

One cow did exhibit a faulty let down which

was quite different from the normal let down pattern for
that cow.
- 132 -

133
Variation in the interval between stimulation of let
down by preparation of the udder, and attachment of the
teat cups resulted in slight changes in rate of flow, level
of production and butterfat percentage.

The maximum rate

of flow was significantly faster when the teat cups were
attached immediately after preparation.

This change in

rate of flow is small when compared to the differences be­
tween cows, but could be responsible for some milking to
milking variation.

Production and butterfat test decreased

slightly but not significantly as the interval between
preparation and milking increased.

However, these results

indicate that the shorter intervals between preparation of
the udder and attachment of the teat cups did result in a
more complete evacuation of the udder.
Data were presented to establish the precision of the
Milk-O-Meter and milk scales employed in these studies to
determine the rate of milk flow.
shown to be quite accurate.

The Milk-O-Meter was

A portion of the error is ran­

dom and tends to cancel with repeated weighings.

The amount

of negative bias increased with increased flow rate, but
the increase was not rapid.

The meters also tended to be­

come less accurate with continued use.

Tests conducted on

the meters immediately after use in the trials reported in
the thesis indicated they were quite accurate at the time

134they were used in these trials.

Errors encountered when

using the milk scales were small enough that they did not
materially affect the results of these studies.

l i t e r a t u r e ci te d

Ace, D. L.
flow.
Univ.

Some factors affecting the maximum rate of milk
Unpublished Masters Thesis, Pennsylvania State
1957.

Adams, H. P., and IT. IT. Allen. The effect of removal of
residual milk by use of oxytocin upon the yield and fat
content of subsequent milkings. J. Dairy Sci. 35:1121.
Andersson, B. Some observations on the neuro-hormonal regu­
lation of milk ejection. Acta. Physiol. Scand. 23:1.
1951.
Babcock, S. M. Variations in the yield and quality of milk.
Wise. Agr. Exp. Sta. Ann. Rept. 6:4-2. 1889.
Baxter, E. S., P. M. Clarke, P. H. Dodd, and A. S. Root.
Factors affecting the rate of milking. J. Dairy Res.
17:117. 1950.
Beach, C. L. The milk yield of quarters on the same side
of the udder. Conn. (Storrs) Agr. Exp. Sta. Rept.
p. 135. 1904-.
Beck, G. H. and T. J. Clayton. Relative effectiveness of
paper towels, cloth towels, and dry bare hands in
cleaning the udder and in stimulating milk let-down.
J. Dairy Sci. 34-:593. 1951.
Beck, G. H., H. C. Pryer, and D. B. Roark. Use and inter­
pretation of milk flow curves in measuring variations
in the response of cows to machine milking. J . Dairy
Sci. 34-:58. 1951a.
Beck, G. H., H. C. Fryer, and F. W. Atkeson. Variations
in the response of dairy cows to a standard method of
milking cows. J. Dairy Sci. 34-:598. 1951b.
Benson, G. E. , and A. T. Cowie. Reviews of the progress of
dairy science, Section A. Physiology of dairy_cattle.
Hormones in reproduction and lactation. J. Dairy Res.
24-:252. 1957.

- 135 -

136
Bitting, A. W. The physiology of milk secretion.
Agr. Exp. Sta. Ann. Eept. 14:51. 1901.

Indiana

Brightman, M. M. Neurosecretion and milk ejection in the
mouse. Anat. Rec. 121 (2):268. 1955.
Brumhy, P. J. The milking characteristics of identical
twin cattle. Proc. N. Z. Soc. An. Prod. 16:89. 1956.
Caroulo, E. V. , C. M. Dembicjak, R. B. Mochrie, R. E. Johnson,
H. D. Eaton, A. A. Spielman, and Geoffrey Beall. Ef­
fects of vacuum level and milking duration on Guernseys
and Holsteins differing with respect to lactation number
and status of udder health. II Milk yield, milking
time, and rate of milk flow. J. Dairy Sci. 58:1283.
1955.
Christ, P. Untersuchen uber die Muskulature und des elastische
Gewebe in der Milchdruse der Haussaugetiere. Inaug. Diss.
Giessen. 1905. Cited from Turner, C. W. The Compara­
tive Anatomy of the Mammary Gland. Univ. Co-op. Store,
Columbia, Mo. 1959*
Clough, P. A., P. H. Dodd and E. W. Hughes. Pulsation and
milking rate. J. Dairy Res. 20:375* 1955.
Clough, P. A., and P. H. Dodd. Effect of pulsation rate and
ratio on milking rate. Agriculture, Lond. 63:354. 1956.
Clough, P. A., and P. H. Dodd. Relationship between milk
secretion and the rate of milking by machine. J. Dairy
Res. 24:152. 1957Cowie, A. T. Pregnancy Diagnosis Tests, A Review. Common­
wealth Agric. Bureaux Joint Publ. Ho. 13. 194-8.
Cowie, A. T., S. J. Polley, B. A. Cross, G. W. Harris, Dora
Jacobsohn, and K. C. Richardson. Terminology for use
in lactational physiology. Nature, Lond. 168:421.
1951*
Cross, B. A. Sympathetico—adrenal inhibition of the neurohypophysial milk ejection mechanism. J. nndo. 9*7*
1955.
Cross, B. A. The posterior pituitary gland in relation to
reproduction and lactation. Brit. Med. Bull. 11:(2) 151.
1955*

137
Cross, B. A., and. G. W. Harris. Milk ejection following
electrical stimulation of the pituitary stalk in
rabbits. Nature, Lond. 166:994. 1930.
Cross, B. A., and G. W. Harris. The role of the neurohypophysis in the milk-ejection reflex. J. Endo., 8:148.
1932.
Dahlberg, A. C. The influence of machine milking.
(Geneva) Agr. Exp. Sta. Bull. 654. 1935*

N. Y.

Dahlberg, A. C. The influence on milk production of a defi­
nite time interval for milking cows by machine. New
York (Geneva) Agr. Exp. Sta. Bull. 697* 1941.
Dahlberg, A. C. Increased milk production by improved machine
milking. Cornell Agr. Exp. Sta. Ext. Bull. 606o 1943.
Dalton, H. L. The effect of feeding concentrates with dif­
ferent moduli and water contents on the eating time
and milking time in dairy cattle. Unpub1. Masters
Thesis. Mich. State Univ. 1952.
Dalton, H. L., C. F. Huffman, and N. P. Ralston. The effect
of feeding concentrates with different degrees of fine­
ness and water contents on the eating and milking times
in dairy cattle. J. Dairy Sci. 36:1279* 1953*
Denamur, R. and J. Martinet. Sensibilite' de la glande
mammaire de la chevre aux hormones posthypophysaires.
C. R. Soc. Biol., Paris. 147:1217* 1953* Cited from
Dairy Sci. Abstr. 17:132. 1955*
Denamur, R. and J. Martinet. C. R. Soc. Biol., Paris. 148:
833. 1954• Cited from Benson and Cowie, J. Dairy
Res. 24:252. 1957*
Dodd, F. H. Fast milkers yield more.
Breeder, Oct. 21, p. 45, 1952.

Farmer and Stock

Dodd, F. H. Normal variation in the rate of machine milking.
J. Dairy Res. 20:301* 1953*
Dodd

F. H. and A. S. Foot. Experiments on milking techni­
que. 1. Effect of washing the udder with hot water.
2. Effect of reducing milking time. J. Dairy Res.
15:1. 194-7.

138
Dodd, F. H. and A. S. Fool:. Experiments on milking techni­
ques. 3 . Combined effect of reducing the milking time
and washing the udder with hot water. 4. Effect of
increasing milking time. J. Dairy Ees. 16:14. 1949.
Dodd, F. H . , A. S. Foot, and E. Henriques. Experiments on
milking techniques. 5* Effect of temporary changes
in the interval between washing and milking. 6 . Com­
parison of established washing and milking routines.
J. Dairy Ees. 16:301. 1949.
Dodd, F. H., A. S. Foot, E. Henriques, and F. K. Heave. Ex­
periments on milking techniques. 7 . The effect of
subjecting dairy cows for a complete lactation, to a
rigid control of the duration of milking. J. Dairy
Ees. 17:107. 1950.
Dodd, F. H. and F. K. Heave. Machine milking rate and mas­
titis. J. Dairy Ees. 18:240. 1951*
Dodd, F. H., and A. S. Foot. The importance of machine
milking rate in dairy cow management and breeding.
Dairy Ees. 20:138. 1953Dodd, F. H. and P. A. Clough. Machine milking.
Agriculture. 1:131. 1937*

J.

Outlook in

Drager, G. H. and E. G. Eennels. The independent behavior
of Gomori-positive neurosecretion and the oxytocic
hormone in the rat neurohypophysis. Anat. Eec. 121
(2) 287. 1955.
Donker, J. D. Lactation studies. I. Effects upon milk
ejection in the bovine of various injection treatments
using oxytocin and relaxin. J. Dairy Sci. 41:537.
1958.
Donker, J. D., W. E. Petersen, M. C. Hervey, and J. Koshi.
The relationship of milk letdown factors to milk pro­
duction. J. Animal Sci. 11:791. 1952.
Donker, J. D., J. H. Eoshi, and W. E. Petersen. The influ­
ence of oxytocin induced udder evacuation on milk and
butterfat production in a complete lactation. J. Dairy
Sci. 37:299. 1954*

139
Duncan, D. B. Multiple range and multiple range F tests.
Biometrics 11:1 . 1933.
Dyke, van, H. B., K. Adamson, and S. L. Engel. Aspects of
the "biochemistry and physiology of the neurohypophyseal
hormones. Rec. Prog. Hormone Res. 11:1. 1955.
Eayrs, J. T. and R. M. Baddeley. Neural pathways in lacta­
tion. J. Anat., London. 90:161. 1956.
Elting, E. C., and J. P. La Master. The response of the
individual cow to the milking machine. S. C. Agr. Exp.
Sta. Ann. Rept. 4-9:54-. 1936.
Ely, P. and W. E. Petersen.
mammary gland. (abs.)

The ejection of milk from the
J. Dairy Sci. 23:536. 194-0.

Ely, P., and W. E. Petersen. Factors involved in the
ejection of milk. J. Dairy Sci. 24-:211. 194-1.
Emery, P. E. Milk records and tests.
Bull. 116. 1895.
Erf, Oscar. Milking machines.
14-0. 1906.

N. C. Agr. Exp.

Sta.

Kansas Agr. Exp. Sta. Bull.

Espe, D. and C. Y. Cannon. The anatomy and physiology of
the teat sphincter. J. Dairy Sci. 25:155. 194-2.
Espe, D. and V. R. Smith. Secretion of Milk.
Collegiate Press Inc. Ames. 1952.

4-th ed.

Polley, S. J. The nervous system and lactation.
Bull. 5:1102. 194-7.
Foot, A. S.
6:313*

The rate of milking by machine.
1935.

Foot, A. S. How fast does a cow milk?
Breeder. April 3} P* 102, 1951a.

Brit. Med.

J. Dairy Res.

Farmer and Stock

Foot, A. S. Management and milking rate.
Breeder. 65:100. 1951D.

Farmer and Stock

Gaines, W. L. A contribution to the physiology of lactation.
Amer. J. Physiol. 38:285. 1915*

140
Gaines, i
ff. L. Differences measured in the rate at which cows
milk. 111. Agr. Exp . Sta. Ann. Rept. 40:158. 1927.
Gregoire, A. I., R. D. Mochrie, P. I. Elliott, H. D. Eaton,
A. A. SpMman and Geoffrey Beall. Effects of vacuum
level and milking duration on milk production, milking
time and rate of milk flow in mastitis-free first
calf heifers. J. Dairy Sci. 37:276. 1954.
Grosvenor, C. E., and C. V/. Turner. Evidence of adrenergic
and cholinergic components in milk let-down reflex in
lactating rat. Proc. Soc. Exp. Biol. 95:719. 1957.
Hall, H. S. Equipment for individual-quarter machine milk­
ing. J. Dairy Res. 18:151. 1951.
Hammond, John. The effect of pituitary extract on the se­
cretion of milk. Quart. J. Exptl. Physiol. 6:311.
1913.
Hammond, John. (Editor) Progress in the physiology of
farm animals. Vol. 3* Butterworth Scientific Publi­
cations. London, Eng. 1957*
Hansen, K. Afkomspr^ver med. tyre, beretning fra forsogslaboratoriet. Vol. VI to VIII* 1952 to 195^*
Harris, G. W. and Pickles, V. R. Reflex stimulation of the
neurohypophysis (posterior pituitary gland) and the
nature of the posterior pituitary hormone (s). Dature
172:1049. 1953.
Harshbarger, K. E. Observations of the rate of milk removal,
(abstr.) J. Dairy Sci. 33:382. 1950.
Hays

R. L., H. L. Van Demark and E. E. Ormiston. Effect of
’oxytocin and epinephrine on conception rates of cows.
J. Dairy Sci. 38:587* 1953*

Hays, R. L., and H. L. Van Demark. Effect of stimulation of
the reproductive organs of the cow on the release of an
oxytocin—like substance. Endo . 52:634. 1953*
Hopson, G. H.
1944.

Hew techniques in milking.

Arg. Eng. 25»295»

141
Hupp, E.^W. A study of the influence of heredity upon the
maximum rate of milk flow in Holstein—Eriesian dairy
cattle. Unpub1. Masters thesis. Univ. of Nebr. 1956.
Hupp, E. W. and M. M. Plum. Normal variations in the maxi­
mum rate of milk flow in Holstein—Priesian dairy cattle.
(Ahstr.) J. Dairy Sci. 50:624. 1957Hupp, E. W. , J. P. Dietrich, W. W. Snyder, and R. C. Lewis.
The accuracy of a milk metering device. Michigan Quart.
Bull. 40:565. 1957.
Hupp, E. W. and R. C. Lewis. Effect of adding water to the
concentrate mix on maximum milking rate, average milk­
ing rate and eating rate. (Ahstr.) J. Dairy Sci.
51:724. 1958.
Johansson, I. Variations in the fat content of milk caused
hy varying degrees of udder evacuation during milking.
Proc. Xllth Int. Dairy Cong. Sec. 1., p. 171. 1949.
Johansson, I., N. Korkman, and N. J. Nelson. Studies on
udder evacuation in dairy cows. I. The rise in fat
percentage during milking. Acta. Agric. Scand. 2:(1)
43. 1952.
Johansson, I. Studies on udder evacuation in dairy cows.
II The amount of milk after normal milkings. Acta.
Agric. Scand. 2 (1) 82. 1952.
Kamm, M. Uber den histologischen San der Rinder und
Jeigenzitze. Inaug. Diss. Bern. Cited from Turner,
C. ¥., The Comparative Anatomy of the Mammary Gland.
Univ. Co-op Store. Columbia, Mo. 1939Kay, H. D. (Compiler) Milk Liberation.
Science. 12:45* 1955*

Advancement of

Knodt, C. B. and Petersen, W. E. The effect of continuous
injections of pitocin upon milk and milk fat production.
(Abstr.) J. Dairy Sci. 25:709* 1942.
Knodt, C. B. and W. E. Petersen. The effect of complete
evacuation of the mammary gland by pitocin upon milk
and fat production. J. Dairy Sci. 27:449. 1944.

142
Knodt, C. B., J. J. Reid, P. S. Williams, and E. M. Kessler.
The effect of interval between washing of the udder and
attachment of milking machines upon bacterial flora and
milk production of dairy cows. (Abstr.) J. Dairy
Sci. 30:580. 1947.
Knodt, C. B., J. J. Reid, P. S. Williams, and E. M. Kessler.
The effect of the interval between attachment of milk­
ing machines upon the milk production of dairy cattle.
J. Dairy Sci. 32:637. 1949.
Knoop, C. E. Relation of management to let-down of milk.
(Abstr.) J. Dairy Sci. 31:694. 1948.
Knoop, 0. E. and C. P. Monroe. Influence of pre-milking
preparations of cows udders upon the let-down of
milk. J. Dairy Sci. 33:623. 1950.
Korkman, Uils. Om faktorer som paverka Juvrets tomning hos
mjolkor. Sartryck Ur Kungl. Lantbruksakademiens
Tidskrift. 87:161. 1948.
Kraft, K. Study of the technique of widening the streakcanal of cattle by a bloody operation. Inaug. Diss.
Hanover. Cited from Turner, C. W., The comparative
Anatomy of the Mammary Gland. Univ. Co-op Store,
Columbia, Mo. 1939*
Linzell, J. L. The silver staining of myoepithelial cells,
particularly in the mammary gland, and their relation
to the ejection of milk. J. Anatomy 86:49. 1952.
Linzell, J. L. The contractility of the alveoli of the
mammary gland. J. Physiol. 123:32. 1953.
Linzell, J. L. Some observations on the contractile tissue
of the mammary glands. J. Physiol. 130:257. 1955.
Mackay, D. R. , J. G. Hartsock, and J. H. Book. Preliminary
report on tests of the accuracy of milk scales. J.
Dairy Sci. 41:316. 1958.
Madden, D. E., and Brunner, J. R. Comparative accuracy of
small lactometers for determining the total solids of
individual
cow's milk. J. DairySci. 41:783*
1958.

143
Matthews , C. A., J. M. Shaw and. E. Weaver. The economy
and efficiency of a milking mackine. Ia. St. Coll.
Agric. Bull. 24-8. 1928.
Matthews, C. A., W. W. Swettand E. E. Graves. Milk yields
and milking rates of the individual quarters
ofthe
dairy cow udder. USDA. Tech. Bull. 827. 1941.
Miller, Kenneth, and W. E. Petersen. Some factors involved
in efficient milking. J. Dairy Sci. 24:225. 1941.
Mochrie, E. D., H. H. Hale, C. M. Dembiczak, H. D. Eaton,
W. U. Plastridge, E. E. Johnson and Geoffrey
Beall.
Effects of vacuum level and milking duration onGuern­
seys and Holsteins differing with respect to lactation
number and status of udder health. I Udder health.
J. Dairy Sci. 38:1272. 1955.
Murphy, J. M.
fection.

The relationship of teat patency to udder in­
Cornell Vet. 34:64. 1944.

Ueusch* A. Inaug. Diss. Zurich. 1904.
Van Demark, Endo. 52:634. 1955*

Cited from Hays and

Ott, I. and J. C. Scott. The action of Infundibulin upon
the mammary secretion. Proc. Soc. Exp. Biol. 8:48.
1911.
Peeters, G., J. H. Bouchaert, and W. Oyaert. The influence
of unilateral lumbar sympathectomy on the udder of
the sheep. Arch. Int. Pharmacodyn. 89:197- 1952.
Petersen, W. E.
Bull. 361.

The cows udder.
1942.

Minn. Agr. Exp . Sta.

Petersen, W. E. The action of the milking machine in rela­
tion to milking and udder injury. J. Dairy Sci. 26:752,
1943.
Petersen, W. E. The action of the mechanicla milker to com­
pleteness of milking and udder injury. J. Dairy Sci.
27:433. 1944a.
Petersen W. E. Physiological factors in lactation related
to milking and mastitis. Cornell Vet. 34:38. 1944b.
Petersen, W. E.

Lactation.

Physiol. Eev. 24.340.

1944c.

144
Petersen, W« E., Knodt, C. B. and. W. L. Boyd. Morphology of
the teat in relation to milking and trauma. (Abstr.)
J. Dairy Sci. 2?:64-6. 194-4.
Piana, G. Postipofisi ed emissione del latte. Boll. Soc.
Ital. Biol. Sper. 26:14-4-3. 1950• Cited from Dairy
Sci. Abstr. 15:44-. 1953.
Bennels, E. G., G. V.
son of pituitary
amount of Gomori
and experimental
555. 1955.

Bussell and G. A. Drager. A compari­
content of oxytocic hormone and the
positive neurosecretion under normal
conditions. Anat. Bee. 121: (2)

Richardson, G. A., J. 0. Young, S. H. Dalai, S. J. Pearce
and P. H. Horula. The use of the protein test (formol
titration) to estimate the solids not fat content of
milk, ProQ. Ann. Meet. Western Div. Amer. Dairy Sci.
Assn. 34:70. 1953*
Bichardson, K. C. Contractile tissues in the mammary gland,
with special reference to myoepithelium in the goat.
Proc. Boyal Soc. (London) B 136:30. 1949.
Eoark, D. B., G. H. Beck, and H. C. Pryer. Differences in
milking response under prescribed variations in methods
employed to stimulate milk let-down. J. Dairy Sci.
35:1017. 1952.
Sandvik, Oystein. The relationship between milking rate and
milk yield. J. Dairy Res. 24:(3) 316. 1957.
Schafer, E. A. Bote on the paper by Simpson, S. and B. L.
Hill: The mode of action of pituitary extract on the
mammary gland. Quart. Jour. Exper. Physiol. 8:379.

1914.

Scharrer, E. and Scharrer, Berta. Hormones produced by
neurosecretory cells. Recent Prog. Hormone Res. 10:183.
1954..
Shaw, J. C. The effect of oxytocin on milk and milk fat
secretion. J. Dairy Sci. 25:1051* 1942.
Silver, I. A. Myoepithelial cells in the mammary and parotid
glands. J. Physiol. 125:8. 1954.

145
Smith, G. A. and. H. A. Harding. Effect of machine method of
milking upon the flow of milk. N. I. (Geneva) Agr.
Exp. Sta. Bull. 353. 1912.
Smith, V. R. The effect of milking at short intervals with
and without injections of oxytocin. J. Dairy Sci.
30:703. 1947.
Smith, V. R. and W. E. Petersen. The effect of level of
negative pressure on the rate of milking. (Ahstr.)
J. Dairy Sci. 27:636. 1944.
Smith, V. R. and W. E. Petersen. The effect of increasing
the negative pressure and widening the vacuum release
ratio on the rate of removal of milk from the udder.
J. Dairy Sci. 29:45. 1946.
Smith, V. R. and W. E. Petersen. The effect of preparation
of the cow on rate of milking. J. Dairy Sci. 31:589*
1948.
Snedecor, G. W. Statistical Methods.
College Press, Ames. 1956.

4th Ed.

Iowa State

Stewart, W. E., and L. H. Schultz* Effect of vacuum and
pulsation rate on rate of milking. J. Dairy Sci. 39:
926. 1956.
Stewart, W. E., L. H. Schultz, and S. P. Coker.
the rate of machine milking of dairy cows.
variations. J. Dairy Sci. 40:258. 1957*

Studies on
I. Normal

Stewart, W. E. and L. H. Schultz.
The rate of machine milk­
ing of dairy cows. II. Effect of vacuum and pulsa­
tion rate. J. Dairy Sci. 41:849. 1958.
Swanson, E. W. and C. W. Turner. Evidence for the presence
of smooth muscle elements surrounding the alveoli of
the mammary gland. J. Dairy Sci. 24:635* 1941.
Swanson, E. C. and S. A. Hinton.Residual milk from oxytocin
injections throughout the lactation. J. Dairy Sci.
34:419. 1951.
Tgetgel, B. Schweiz. Arch. Tierheilk 68:335- 1926.
from Hays and Van Demark Endo. 52:634. 1953.

Cited

146
Thoele, H. W. Factors affecting the milking curve of the
bovine. Unpubl. Doctoral Diss. Univ. of Minn. 1954.
Tsakhaev, G. A. 0 prirode afferenykl putei refleksax molokootdachi. Dakl. Akad. Uauk. SSSR. 95: (5) 941. 1953.
Cited from Dairy Sci. Abstr. 16:384. 1954.
Turner, C. W., and I. S. Slaughter. The physiological effect
of pituitary extract (posterior lobe) on the lactating
mammary gland. J. Dairy Sci. 13:8. 1930.
Turner, C. W. The secretion of milk and the milking pro­
cess. Mo. Agr. Exp. Sta. Bull. 346. 1935*
Turner, C. W. The Comparative Anatomy of the Mammary Gland.
Univ. Co-op Store. Colombia, Mo. 1939.
Turner, C. W. The Mammary Gland.
Mo. 1952.

Lucas Bros.

Turner, C. W. Harvesting your milk crop.
Chicago, 111. 1954.

Columbia,

Babson Bros. Co.

Turner, C. W. and W. 0. Cooper. Assay of posterior pituitary
factors which contract the lactating mammary gland.
Endo. 29:320. 1941.
Venzke, C. E. A histological study of the teat and gland
cistern of the bovine mammary gland. J. Am. Vet. Med.
Assn. 96:170. 1940.
Vigneaud, du V., H. C. Lawler, and E. A. Popence. Enzymatic
cleavage of glycineamide from Vasopressin and a proposed
structure for this pressor-antidiuretic hormone of
the posterior pituitary. J. Am. Chem. Soc. 75:4480.

1955.
Vigneaud, du V., C. Ressler, J. M. Swan, C. W. Roberts and
P. G. Katsoyannis. The synthesis of oxytocin. J. Am.
Chem. Soc. 76:3115* 1954.
Ward, G. M., and V. R. Smith. The effect of time of milking
after milk excretion on total milk production.
(Abstr.)
J. Dairy Sci. 31:694. 1948.
Ward, G. M. and V. R. Smith. Total milk production as affected
by time of milking after application of a conditioned
stimulus. J. Dairy Sci. 32:17. 1949.

147
Werle, E. and L. Maier. Uber den Abbau von Oxytocin and
Vasopressin durch Milchdrusenextrakte der Kuh.
Naturwissenschaften 41:380. 1954.
Whittlestone, W. G. The characteristics of the milk ejection
curve of normal dairy cows under standard milking con­
ditions. N. Z. J. Sci. Technol. A., 28:188. 1946.
Whittlestone, W. G. How the cow releases her milk.
J. Agric. 1. 1949.

N. Z.

Whittlestone, W. G. Studies on milk ejection in the dairy
cow. The effect of stimulus on the release of the
"milk ejection" hormone. TS. Z. J. Sci. and Technol. A.,
.32:1. 1951.
Whittlestone, W. G. Variations in the fat content of milk
throughout the milking process. J. Hairy Res. 20:146.

1955.
Whittlestone, W. G. The control of the milking process.
Soc. Hairy Techol. 7:95* 1954-.

J.

Whittlestone, W. G. Intramammary pressure changes in the
lactating cow. I Changes during the milking process.
J. Hairy Res. 22:290. 1955.
Whittlestone, W. G. and S. A. Verall. The influence of
certain milking machine adjustments on the rate of ma­
chine milkings. N. Z. J. Sci. Technol. A., 28:406. 1947.
Whittlestone, W. G. , E. G. Bassett and C. W. Turner. Factors
influencing the amount of milk "Let-down" hormone in
the posterior lobe of Jersey cattle. J. Hairy Sci.
35:889. 1952a.
Whittlestone, W. G. , H. R. Perrin, R. H. Parkinson, and C.
W. Turner. Frequency of discharge of pituitary milk
"let-down" hormone. J. Hairy Sci. 35:894. 1952b.
Whittlestone, W. G., and H. M. S. Phillips. Automatic ap­
paratus for drawing the milk ejection curves of dairy
cows under controlled milking conditions. J. Hairy
Res. 20:319* 1955*
Witt, M. and U. Andreae. Milchfluss und Melkdauer nach dem
Aurusten des Kuheuter mit warmen Wasser. Milchwissenschaft 10:386. 1955*

148
Wo 11, E. M. and G. C. Humphrey. Milking machine experiments.
Wise. Agr. Exp. Sta. Ees. Bull. 3. 1909.
Zehner, C. E. Three Minute Milking.
41: (8) 19. 194-3.

Successful Farming.