IIIII

I
I

 

THE DEVELOPMENT OF NEW
TECHNICS IN THE MICROSCOPIC
EXAMINATION OF MILK

Thesis for the Degree of M. S.
MICHIGAN STATE COLLEGE

Frank Robert Peabody
1942

 

 

 

THE DEVELOPMENT OF NEW TECHNICS IN THE
MICROSCOPIC EXAMINATION OF MILK

by
FRANK ROBERT PEABODY

A THESIS
Submitted to the Graduate School of Michigan
State College of Agriculture and Applied
Science in partial fulfilment of the
requirements for the degree of

MASTER OF SCIENCE

Department of Bacteriology and Hygiene

191+2

ACKNOWLEDGEMENT

The helpful suggestions and kind assistance
of
Michigan State College
faculty members

Dr. W; L. Mailmann and Dr. C. S. Bryan
of the
Department of Bacteriology and Hygiene
and
Dr. W. D. Baten
of the
Department of mathematics

are greatly appreciated

~ OUITJNE

I. Introduction
A. Accepted methods of milk analysis -------------- 1

B. Purpose of the paper - a study of new developments in
the microscopic examination of milk ------------- 1

II. Historical

A. Original microscopic methods ---------------- 2
l. Centrifuging as carried out by Slack
2. Pipette method of Breed

B. Improvements in the pipette technic ------------- 3
1. Studies made by Brew in l91h
2. Modifications proposed by Breed and Brew in 1916

C. Analytical work comparing the microscopic and plate methods - h
1. Brew and Dotterrer - 1917
2. Bread and Stocking - 1920
3. Robertson - 1921
h. Brew - 1929

III. Experimental Procedure

A. Type of mdlk used ---------------------- 6

B. Comparison of methods of smear preparation --------- 6
1. Pipette procedure
2. Loop technic

C. A study of two methods of staining _ - _ _. ......... 7

1. Breed stain

2. Acid.methy1ene blue stain

D. Effect of a colored light source for microscopic
1. Use of different red lights

2. Use of different yellow lights

examination 8

E. Statistical analysis of the data -------------- 9
Results
A. Data on pipette and loop methods of smearing ------- 10
1. Using a "high count" milk
2. Using a "mediwm count"milk
3. Using a "low count" milk
B. Data on staining technics ----------------- 11
1. Using the Breed stain
2. Using the acid methylene blue stain
C. Effect of varying the color of the light source ------ 12
D. variance in accuracy made by changing the number
of fields counted --------------------- 13
Discussion
A. Advantages of the loop procedure ------------- 15
B. Superiority of the acid stain --------------- 16
C. Use of a yellow light source --------------- 16
D. The number of fields that should be counted for any
desired degree of accuracy ---------------- 18
Summary --------------------------- 19

The Development of New Technics in the
Microscopic Examination of Milk

INTRODUCTION

Of the many and varied tests for checking the sanitary quality
of milk, both before and after pasteurization, the determination of
the bacterial content is one of the most accurate. At the present
time there are two accepted methods of bacterial examination which
are given in the Standard Methods £31; 22 Examination 93 Ram {29;
9393.3 (1). The older method of the two is the standard plate count
in which a suitable dilution of the sample is planted in an agar
plate and incubated for a given time. A little over thirty years
ago, another method of milk analysis was introduced; namely, the
direct microscopic examination. During this period of time the
latter method has been accepted gadually until at the present time
it is in general use. The method is at present almost identical
with the original procedure as outlined by Breed (2).

It is the purpose of this work to evaluate the method of direct
microscopic examination especially from the standpoint of statistical
methods to determine its accuracy and, also, to introduce some new

modifications in the present procedures.

. : s. Luewuflswuaflvur barman... shall-I.

 

 

E’,nll it'll-FIE ‘I‘l-I >I'Iv ilt ll DII'I'I tall I

I'lrrll ‘glllll III

- 2 -

HISTORICAL

The use of the microscope for making counts of bacteria
dates back almost to the origin of the microscope itself.
Although counts had been made, it was not until shortly
after the turn of the century that the idea was applied to
milk. The first method was introduced by Slack (3), who
in 1906 centrifuged a sample of 2 m1. at 2000-3000 revo-
lutions per minute for 10 minutes. ’The entire sediment was
then smeared on a slide and spread over an area of 11 sq. cm.
The preparation was stained with methylene blue and examined
with a 1/12 oil immersion lens. Slack suggested using it as
a "presumptive" test for very high and low quality samples.
Those samples coming between would be analyzed by the plate
method. Be especially noted that the accuracy was poor when
dealing with clean milk.

The next method was introduced by Bread (2) who considered
that centrifuging introduced too many unknown factors to be a
reliable procedure. Breed found that in the microscopic
examination for body cells in milk, the bacteria were also
stained. A standard procedure for the examination of bacteria
microscopically was developed which is almost identical

with the pipette method in use today.

- 3 -

One of the greatest objections to this method was the belief
that dead bacteria would influence the count to a large extent.
Breed discounted this factor because he found that dead bacteria
decompose rapidly and did not retain the stain.

In l91h, Brew (h) discovered that there was little relation-
ship between the plate and microscopic methods. This was especially
true in the low and high quality samples. The differences in the
low-count milk were attributed to the fact that in the samples used
many of the bacteria were of udder origin. These organisms fail to
grow an agar plates at an incubation temperature of 21 degrees C.
and thus would not show up on the plates while they would be counted
by the microscopic method. This gave the advantage of more accuracy
to the microscopic procedure.

Two years later, in 1916, Breed and Brew (5) in continuing this
study, advised a few modifications in technic. When the films were
being dried, it was found advisable to use a slight amount of heat,
(warm table or similar device) because if the slides were dried slow-
ly the increase in numbers would introduce a significant error. They
also found that it was not necessary to use sterile pipettes - thor-
ough rinsing was sufficient to remove all the organisms that would
effect the count to a significant degree. They also made a study
of the accuracy of the loop smear as compared to the pipette method.
It was found that although a loop claibrated to deliver 0.01 ml.
speeded up the procedure, it introduced a variation of 35 per cent

while the pipette technic did not exceed 2 per cent variation. The

.- 1+ -

methods were checked by weighing the amount of milk discharged
rather than by bacterial counts.

It was found by Brew and Dotterrer (6), in.l917, that.more
than 50 per cent of the plate counts of 6H3 samples gave results
intermediate to the microscopic "group" and individual counts. This
showed, as was expected, that the plate count represents fairly well
the number of "groups" present after being broken apart by the
dilution waters; the "group" count gave the number of groups originally
present; and the individual count gave the number of bacteria actually
present in the milk. They also found that the size of the "groups"
increased with the number of bacteria up to a point where the lactic
acid bacteria became predominant. These, of course, formed smaller
"groups" and the size decreased again. Thus, they concluded that
the accuracy of the plate method was affected by two highly variable
factors - the size of the "groups" originally present and the extent
to which they were broken.up in.making the dilutions. For this
reason, the plate method was not considered accurate enough to
grade milk into more than two or three grades.

In 1920, Breed and Stocking (7) published some data showing
that the variations in results of both the plate count and the
direct microscopic count were small enough to Justify the use of
either method for grading milk into two or three classes. Also,
they agreed with the previous conclusions that the methods were not
sufficiently accurate to allow any finer grading. Breed and Stocking

also made some analytical studies to show the large standard deviation

_ 5 -
found in the microscopic methods. These studies showed the variations
among the results of different analysts, some of whom.were inexper-
ienced in the technic, and should not be used as a criterion for

the reliability of the method. Their purpose was to show the need

of experienced technicians in the microscopic as well as the plate
method, if reliable results are to be expected, rather than to
determine the accuracy of the procedures.

It was demonstrated.by Robertson (8) that the ratios between
plate counts and the individual microscopic counts became more uniform
if the incubation periods were lengthened to five days at 21 degrees C.
followed by two days at 37 degrees C. The use of lactose agar also
reduced the number of widely discrepant counts.

Some statistical analysis was done in 1929 by Brew (9), who
found that the standard deviation of a "group" count and the plate
count were in approximate agreement.

It can be seen that in all these studies the aim.has been a
comparison with the plate method. At the same time, several workers
have pointed out the highly variable factores involved in the plating
technic. Thus, it would seem.more reasonable to use only the
microscopic method and obtain a sufficient volume of data so that
statistical anlysis could by employed to determine the results.

By recording the counts of each field, it is possible to determine
the amount of error encountered as well as some data on distribution
of organisms on the smear. This gives a much better and more accurate

picture of the advantages as well as the limitations of a direct

microscopic method.

EXPERIMENTAL PROCEDURE

The milk used in all of these experiments was of good quality.
When high counts were desired, the milk was incubated for a suitable
period of time. By using one grade of milk and varying the bacterial
content by incubation it was possible to have comparable types of
bacteria giving comparable clumps so that any variation that might
enter through change of organisms was largely eliminated. In all
of the studies, comparisons between procedures were made so the
writer believes that the selection of one type milk gave the results
desired.

In the first series of studies, a comparison of methods of
preparing the smear was made. Two procedures were examined; namely,
the pipette and the loop methods.

All smears made by the pipette method were in accordance to

Standard Methods £22 the Examination 2; Dairy Products (1) .

 

The loop method was made according to the procedure recommended
by Bryan (10) for the examination of producer milk. In this procedure,
a standard platinum loop of h mmi outside diameter is used and a
smear of h by 8 mm, is made. ‘Using this method 20 smears were
evenly spaced on one slide.

Smears were made by each.method of three grades of milk so

- 7 -

the data could be obtained for smears showing less than 1 organism
or clump per field, smears showing 10 to 15 organisms or clumps

per field and smears showing numerous organisms or clumps per field.

All slides were stained in.exactly the same manner using the
same stains so that the only variable in the series was the method
of smearing. Two hundred and fity fields were counted for each
smear in the low and medium count milks and one hundred for the
high count milk. In all instances the number of organisms per
field was recorded for each smear.

Examinations were made with a 1.32 mm. oil immersion objective
and a 5 X ocular. This yields a conversion factor of 2h0,000 - one
organims per field represents 2h0,000 organisms per ml. of milk.

The results represent a "clump count" in which the clumps of bacteria,
as well as the individual organisms, are counted as units.

In the second series a study of two different methods of
staining was made. One method is a modification of "standard methods"
as it is used routinely in this laboratory. The slides are immersed
in xylene to remove the fat, followed by a 95 per cent alcohol
solution. .About one minute is sufficient in each, and the slides
are drained between solutions. The staining bath is prepared by
adding 10 m1. of saturated alcoholic solution of methylene blue
to 90 m1. of 30 per cent alcohol (12). The slides are dipped in the
stain Just long enough for proper staining. They are then rinsed

in water and decolorized in alcohol if necessary followed by

thorough drying.

-8-

The second is a method which is recommended by Mallmann and
Churchill (11) for use in staining egg-meats. The slides are stained
according to the following procedure:

The staining bath is made of

1 gm" methylene blue (Certified for bact. use)
500 ml. 95 per cent ethyl alcohol
5 m1. conc. hydrochloric acid.

The slide should remain in the staining bath from.three to
five minutes. It is then removed and dipped in a tap water bath
only long enough to remove the excess stain. It is important to
decolorize only partially in order to avoid decolorization of the
background to such an extent that it is difficult to find the
location of the various smears when.making the examination. In
addition, excessive washing will soften and loosen the film, The
slide is air dried.

Because this stain contains 95 per cent alcohol, it was found
more satisfactory to omit the alcohol bath before staining, using
only the xylene to remove the fat.

In this series, all of the smears were made by the loop technic.

In the third series, a study was made of the effect of a colored
light source in making bacterial counts. The work was done with
the use of colored solutions - acid fuchsin for the red and potassium
dichromate for the yellow. The concentrations of the dyes used are

as follows:

Yellow Solution #1
"

N N

Red "
I!

II '1

Red.& Yellow"

If n " "

HWMI—‘WN

f0

-9-

12 parts pot. dichromate per 100, 000

30 " " 100, 000
60 " " " " 100, 000
0A " fuchsin " 100,000
1 " " " 100,000
2 " " " 100,000
(2h " pot. dichromate " 100,000
( 0.h " fuchsin ” 100,000

(2h " pot. dichromate " 100,000
( 0.5 " fuchsin ” 100,000

These solutions were made up in quantities of 200 ml. from

stock solutions and were used in a special 250 ml. Florence flask

with a ground glass side which fitted the "Stella" microscope light.

In addition to the solutions, wratten filters were employed which

were fastened to the sub-stage of the microscope.

In much of the work described, the conclusions were reached

by submitting the data to statistical treatment. All conclusions

were made according to the following formulae:

where

"“ a:

u "i
:

 

 

=/E_g__ (ZS/V (1))2

Alt

 

 

 

0" a
t» DI"- ofmean: DI£ {dfmgqng
07:" of mean: /%l 3+ rm: 2

standard deviation of the sample
organisms per field

number of fields

deviation of the mean

test for significance

The data used were obtained from.at least 100 fields and in most

- 10 -
cases 250 or 300 fields were used. The deviations are calculated
from.the original counts of individual fields and not from.a series

of‘averages.

RESULTS

The results of the first series of studies will be found in
Tables I to VII. Before starting a discussion of these data, it
should be understood that the terms "low count", "medium.count",
and "high count” are used as relative terms in the statistical sense
and do not in any way imply that the quality of milk is "high",
"medium", and "low" since this is not the case as can be seen from
the counts per m1.

A study of the data for the high count milk (Tables I and II)
shows that these two methods give very comparable results. The
pipette procedure yielded a slightly higher average. It can.be
seen from.the sub-totals that there is some variation from one
smear to another (652 to loot) as well as from one field to
another when dealing with the pipette method. In the case of the
loop technic, the totals for the ten smears varied only from.667
to 922, or almost 100 less.

In order to determine the variations from.one field to another,
the standard deviation and mean deviation were calculated (Table VII).

However if these values are submitted to the "students" t test, it

pa.nm m.nw o.mm m.Hw o.wm m.mm m.mo e.ooa s.om >.~w n.Hm oneness
san.m mum 0mm mam mow nmm «no sooa pom New sad Hosea

am e» me we we mm mm mas om am
He no mm ooa am on we me mm om
me am am mm me an pea mm mm man
om mm was Hm om om am am mm mm
oaa mm mm as med Hm oma mm am we
no 00H as am am mm was an op .mOH
mm as mm om mma on med om am we
mos om ma om mm me am med mm n»
cos NOH mm mm m» 3» mm am we oHH
am am mm mm so we «ma om flea mm
0H m m p m n a m m H
.oz ssosm

each“ use sflsopesm

ANN .oz oaassmv .safis pesos swan s_so seasons opposes
on». hp. 062 mhdofim 0H Mo dHOHM hon cahovodn. on“... I H OHDGB

mm.am m.mm >.mm a.a~ s.am m.am m.mm m.»~ n.0m H.m~ ~.mm omssesa
omim mum So 1E. :8 m8 wmm mt. new 5. Km Hence

we en mm on as sea we om on mm
mm em mm mm an em mm em mm as
«a mm as so me we mm om oaa om
we we ow on so am we as ooa was
was am no em on we om me we mm
wHH an a» mm mm Hm ooa an mm mm
em an mm mm mm me an em am med
ooa mm mm mm ooa mm an nm No me
mm on mm mm sea an em on nod mm
mad no we mm sad ooa mm moa am we
0H m m a w n a m m a
.oz ssssm

eased sea anaconda

8a .02 3988 an? assoc amen n no cases» aces
can an seas assess ea no eases sea assesses one - HH oases

- 11 _
is found that the result is less than 2 and therefor the variations
between methods is not significant.

In studying the data on a medium count milk, Tables III and IV,
it is found that the results obtained were almost identical. This
time the variations of the smears are very close - 207 for the pipette
against 197 for the loop technic. However, the deviation of the
fields show more striking results. Table VII shows that the standard
deviation is 6.93 for the pipette and 6.50 for the loop; while the
deviation of the mean was .h38 and .All respectively. In this case,
the value of "t" exceeds the limit of 2 and therefor the variations
have considerable significance.

In considering the results of Tables V and VI one should
realize that in dealing with so few organisms on each smear, small
variations will give relatively high percentage errors. In this
reapect it is wll to point out that the difference in the two
totals would amount to only 15,000 organisms per m1. In practical
use, this amount is not too important. For this reason as well as
the low value of "students" t, the difference encountered here
can be considered as insignificant even though they show slightly
larger variations with the loop technic.

In making the series of studies on staining, counts were taken
on two duplicate sets of smears of varying quality. In examining
the smears made by the Breed procedure of staining, the usual
difficulties were encountered; i. e., the background of milk and

any debris which was on the slide retained the stain almost as well

mw.HH
Ham.m

as.
as

am
ma
mm
o

ma
m

0H
ma
ma
ma
Hm
ma
om
oH
m

d

as
ea
ea
ea
m

mm
as

O
OH

OH

:.OH
HmN
m

m.m a.w
new mom
a ma
Hm mm
m as
m n
ma ma
ma m
a as
a m
w s
m 0H
m ea
m «a
m m
as o
Ha m
as m
m o
a a
ma m
a a
ma a
m w
m a
NH m
ma ma
m a

m.OH
OON

O
OH

O
OH
NH
OH
m
m
:H
HH
:H
OH
OH
MH
mH
O
NH
OH
OH
mH
w
:H
m
m
mH

O

.03 Adosm

eases son sasoposm

w.wH

mH:
mu
m
mm
om
OH
m
NH
O
h
HH
N
m
mH
OH
NN
hm
mm
mm
NH
OH
mH
:H

an
em

n

0.:H

can

00
HO]

\Ob-NmH-d'd) NCO

“mm .oz oHaasmv .maas pasoo_ssaeos s so ceases» opposes

can he sees assess oH so eased sea assesses one - HHH sagas

n.m

PMN

HH
0

ma
ad
ma

m O\t‘-\Ol‘-O\\Ol{\\0

>.:H

mmm

mH
PH
P

NN
PH
NH

m.0H

saw

OH
mH
OH
O

OH

O
m
m
mH
m
OH
m

m
w
mH
w
HH
NH
m
n
NH
m
m
m
nH

H

omeho><

H6908

3;:
mHa.m

m.NH
mHm

w

mH
ON
NH
HH
PH
OH
w

m

m

mH
m.
OH
nH
wH
wH
HH
:H
NH
nH
NH
:H
m

:H
mH

OH

:.NH :.NH
HHm OHM
NH ON
NH :H
HH NH
n :H
m nH
NN wH
wH m
m m
MN m
OH HH
NH OH
nH w
nH n
P v
mH #H
: MH
PH NH
: NH
OH HN
n NH
NH :H
mH HN
OH HH
NH O
nH mH
m w

AmN .oz oHasemv .MHHs pnsoo_s:Heos e no OHonoop m00H

O.nH
Own

mN
PH
NH
NN
mN
HN
mH
0N
:H
>

OH
HN
m

PH
OH
m

OH
n

m

OH
w

NH
OH
>

nN

P

N.wH

o.nH
mos mam
mH om
HH am
a s
0H am
mH mH
n a
aH mm
am mH
a an
mm mm
H: m
mH 3H
m e
MH oH
mH mH
NH mH
om mH
pH «H
mm «H
om »
mH m
Hm sH
«H s
0H m
aH nH
o n
.oz sense

eHoHM hon eHhopoem

m.:H
mmm

m

sH
Hm
HH
nH
om
mm
mm
eH
em
n

m

«H
om
pH
on
ma
mH
PH
0H
m

n

P

m

mH

a

m.ON
pom

0.:H

:Om

OH
NH
O

NH
>

PH
O

NH
wH
OH
:H
mH
PH
:H
OH
O

OH
ON
:H
PH
mH
NN
HN
0N
mH

N

any hp eons unease OH «0 eHon non eHnopoep one a >H oHpnB

H.nH omwuob<
mpm Hopes

NH
mH
HH
om
oH
oH
Ha
m
mm
m
on
m
MH
«H
0
HH.
Ha
om
Hm
oH
oH
«H
Hm
om

OH
H

Table V - The bacteria per field of 10 smears made by the

pipette technic on a low count milk. (Sample No. 2h)

Bacteria per field

Smear No.

10

M00

OOoH

0000

85

11

ll

10

Total

0.32 0.3M

0.h0 0.28 0.36 0.hh 0.hh 0.28 0.28 0.28 0.32

Average

Table VI - The bacteria per field of 10 smears made by the

loop technic on a low count milk. (Sample No. 2h)

Bacteria per field

10

Smear No.
6

5

ricn<>

Or-Ir-i

OH

OH

101

ll

11

ll

10 ll

13

Total

0.h0h

0.h0 0.hh 0.32 0.36 0.32 0.hh 0.hh 0.36 0.hh

0.52

Average

Table VII - A Statistical Summary of Tables I through VI

Deviations Between the Two Methods

Arithmetic Standard Deviation

Method Sample mean deviation of the mean
Pipette 22 85.77 16.66 1.67

" 23 11.88 6.93 .h38

" 2h .3h0 .A78 .032
Loop 22 81.36 16.h8 1.65

" 23 1h.85 6.50 .hll

" 2h .hoh .6Ah .OAY

Significance of Deviations Between the Two Methods

Sample No. value of "Students" t. Conclusion
22 1.88 Not Significant
23 h.9 Significant

2h 1.12 Not Significant

- 12 -
as the organisms themselves. This showed that many of the bacteria
were obscured by the other stained material, and good results could
be obtained only in the hands of a trained technician, if at all.
On the other hand, in examining the smears made from.duplicate samples
which had been stained with the acid stain, the background was found
clear of debris and stained very faintly. This slight amount of stain
was retained during the washing procedure in order that there would
be something to focus on while making the examination. The organisms
retained the stain well and appeared very distinct from the rest of
the smear. This afforded much more ease and speed in counting.

Another factor to consider is the ease of preparing the staining
bath and the method of staining. It can be seen from.the technics
presented in the preceding section that the acid stain is even more
simplified than the Breed technic. The preparation is about the same
except that only two baths are required and there is no need for a
separate decolorization in using the acid stain.

Of course in comparing the relative value of two staining
technics, the important thing to consider is which will give the
higher counts; i. e., which stains the larger number of bacteria.

In this respect, the data in Table VIII and the first two columns
of Table IX show that the acid stain gave consistently higher results.
Thus, the acid stain was the better of the two in all respects.

The data on the studies on the effect of a colored light

source are found in Tables IX to XI. Those presented in Table IX

are of a preliminary nature. In several cases it was found upon

Table VIII - The Bacteria per m1. on Two Duplicate
Sets of Smears using Different
Methods of Staining.

Sample Type of Staining
N0. Breed technic Acid stain
25 5,000 5,000
26 5,000 9,000
27 lhh,000 115,000
28 15,000 38,000
29 53,000 100,000
30 53,000 200,000
31 lh,000 65,000
32 9,000 5,000
33 38,000 57,000
3h 115,000 300,000
35 5,000 H.000
36 3h,000 h8,000
37 9.000 9,000
38 1h,000 19,000
39 96,000 lhh,000
to 72,000 75,000
he 192,000 2h5,000
is 53,000 67,000
Ah 19,000 25,000

. HHS 83%.. was sou HHHsewsso one agenda: .3 eons seeds 1.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oom.s oom.n oos.m omo.H oom.H oow.m oos.m mHn.s osm.m ow
me as am omH mo me ssH mm mm on mH
oom.H osm.H omH.m oom.~ oos.m oom.m ooo.H ooa.m ons.a oom.m oom.H omo.m omn.m wH
oww.a owH.H om>.m oso.m ooH.m owe.m oom.H omo.m omw.m coo.m own.m oeo.m oem.w AH
mam mom mom ems omm ooo.m asH . oH
6mm mmm mmm mmm 8N mm on we nH
osm.m oos.m oes.m ooo.m om~.H ooe.H ooH.m «Hm AH
08 5H 08 5H 5H :3 96 m5 MH
oom.H omm.H oom.H new one osH.H mom «H
n n n m HH
n m nH m 6H

m m nH m m

emu moH mnH cam .‘moH meH meH ooH m
m 4H m n a

m n 3 4H m

Hm me on S n

mm m... 3 0H s

«so sHo aHo ems mum one «Hm now 6mm 4mm own new m
n n m n m n m m n N

omH on mm mm mm m: on mm me an em as H

sHspm eHo< “wwwm
mm _ nH s mm m H H m _ m _ H m _ m a H
nsneHHm season: .aoHHoH s_eom eom _soHHo» seem osHm .oz ‘
oonsom p . HA . 0H5 .
. moonsom psmHH one madam “3833.0 was: 3303. ASH 05 hp eds:
9386 gang mo Amended—on» sH couuohmkov HHHE M0 .Ha .HoHH erases one u NH 3909

 

 

 

Table I - A Study of Five Different Filters
of Sample No. 21 (Smear A)

Number of Bacteria per Field

Type of Filter

Blue wratten Pot. Fuchsin Fuchsin.&
22 dichromate Sol. 2 dichromate
S01. 2 801. 2

2 2 h 3 2

2 2 h l 1

3 2 2 2 3

2 2 2 0 h

2 1 l 0 6

2 2 0 0 h

1 h 1 0 2

2 h 3 l 0

h 2 1 0 1

h h 2 1 1

3 3 1 3 2

h h 2 2 5

3 l l 1 2

3 1 1 0 l

2 1 1 0 3

1 3 0 h 2

3 2 0 1 2

l 1 1 2 2

2 2 l 3 2

2 0 3 3 1

2 h 1 1 2

2 1 2 l 1

h 5 3 2 1

2 h 1 0 3

h 1 l 1 2
Total 62 58 39 32 55
Average 2.5 2.3 1.6 1.3 2.2
Deviation

of the mean .19 .26 .22 .2h .27

Table XI - A Study of Five Different Filters
of Sample No. 21 (Smear B)

Number of Bacteria per Field

Type of filter

Blue Wratten Pot. Fuchsin. Fuchsin.&
22 dichromate Sol. 2 dichromate
Sol. 2 Sol. 2

2 1 0 3 2

h 3 0 2 0

I. 3 2 l 3

l 3 h 1 2

l l l l 2

0 2 1 1 3

1 h 3 o 2

3 1 2 0 2

2 6 1 3 h

1 0 1 h 2

3 3 2 2 2

3 2 1 2 1

h 2 0 1 2

6 2 2 1 1

2 l 3 3 2

2 1 1 2 3

1 2 1 3 1

h h 1 2 3

3 h 3 1 2

2 2 2 2 3

2 1 1 2 2

1 1 1 3 0

2 1 l 2 1

1 l 2 2 3

1 2 2 0 3
Total 56 53 38 Ah 51
Average 2.2 2.1 1.5 1.8 2.0
Deviation

of the mean .27 .27 .19 .21 .19

- 13 -
examining only a few slides that the filter used would be unsatis-
factory and it was discarded. Some were too weak, some destroyed
contrast by too great an intensity, and others were hard on the
eyes. Those samples containing 15,000 organisms per m1. and less
are not considered in these results because they represent less than
four organisms per fifty fields and variations would be largely due
to selectivity rather than variation of the light source.

Since it is rather hard to get a true picture from.the averages
as presented here, five of the lights which showed up the best were
checked more carefully. The counts per field were recorded from.25
fields on two duplicateasmears. The data are recorded in Tables X
and XI. They show almost no difference between the light blue color
and the yellow wratten filter No. 22. In close agreement to these
is the mixture of red and yellow (Solution 2) which is a comparable
substitute for the yellow‘Wratten filter.

To complete the analysis of the procedures used in.milk
examination, it was thought advisable to make some study concerning
the relative accuracy of the counts made by using a varying number
offields. To do this satisfactorily, the counts were recorded for
300 fields on one smear (Table XII) and 200 fields on a second smear
(Table XIII). It can reasonably be assumed that this amount of data
would give results close to a "true mean". The data are given in
Table XIV as Samples 3 and 17 respectively. In addition to these
samples four more were selected to show the results for higher and

lower figures. Although counts such as found in Samples 22 and 23

Table XII - The Bacteria per Field on

One Smear of Sample No. 3

In. 5533‘"... 53uh667222362h2251h031hh 52234.45142 511 2 5133753

an. en. 3666 35514 3131 321252.4275hhh00016u6135736h33022h.5

011733532 31 5.4141357I46 7101632 52635321355355hh 561276.42

683.452222215376310h66672275hh83056578262214 5.419.714.1111

214 15632h155hh683388h‘42hh202h3h. 505503.4162333222 95h». 3

1 1

2351311351261463h.3567.15213h145h. 7h667756h2312 021.4 3500

Average - 3.66
Deviation of the mean - 0.117

Total - 1098

Table XIII - The Bacteria per field on

1h
10
17
10
11
12
23
1h
10
12
13
10

9
1h
11
13
10

9
1h
11
10
15
16

9
2O
12
ll
15

8
19
1h
15
13
10
2O
11
11
15
1h

13

One Smear of Sample No. 17.

15
19
1h

7
15
l6
15
17
1h
11

9
10
13
2O
13
10

6

7
1h
ll
10
12
16
10

5
11
1h

Total - 2,6h3

11
2O
ll
12
1h

6
11
15
15
13
15
17
12
19
23
15
1h
15
18
11
15
13
1h
13
17
11
15
17
1%
1h
12
13
21
2O
19

9
12
12
14

19

Average ~ 13.22

Deviation of the mean -

1h
15
12
21
13
13
12
13

8
12
10
18
13
17

6
12
2O
10
13
15
13
ll
13
1h
11
11
19

9
l6

5
15
ll

6
1h
15

9
ll
12

8
22

0.28

10
15
17

18

l8
l7

17

l6
15

1h
12

10
22
12

16
2o
15

16
10

19

13
11
10
1h
16
10
10
10
17
1h

15

Table XIV - To Show the Number of Fields Necessary
to Obtain a Desired Accuracy

Sample Arithmetic Standard Allowable Number

mean deviation per cent of Fields
error
2h .hoh .6hh 25 A7
15 113
10 25k
21 2.12 1.h8 25 8
15 22
10 #8
3 3 .66 2.02 25 5
15 11
10 31
17 13.22 3.96 25 2
15 h
10 9
23 1h.85 6.5 25 3
. 15 9
10 19
22 85.77 16.6 25 1
15 2
10 h

- 1h -
would not be encountered in actual practice, they were included here

to show the trend of the results.
DISCUSSION

When the microscopic technic was proposed by Breed(2), it
was standardized by using 0.01 ml. of milk spread over an area of
1 sq. cm. This procedure gave a sufficiently uniform.thickness to
the smear that consistent results were possible. However, the method
is somewhat cumbersome and time-consuming'because it is necessary to
accurately measure the amount of milk in the pipette, and two
Operations are necessary to make the smear - depositing the milk and
spreading it over the required area. In addition, it is advisable
to have some sort of guide to indicate the proper area.

The solution appeared to be the use of a loop for depositing the
0.01 ml. of milk. However, Breed and Brew (5) found that this
introduced an extremely large error, and the method was discarded in
favor of the pipette procedure.

The purpose of the procedure is to obtain a smear of constant
thickness. Since the area used is much larger than is necessary for
the examination, it would be possible to reduce the area and the
amount of milk used with a resultant smear of the same thickness as
obtained by the original method. If this smaller amount of milk were
used, it was thought that it might be possible to calibrate a loop

that would have greater accuracy. This method was recently proposed

- 15 -
by Bryan (10) and was used in this study. The results showed the
loop to be superior in every way. To further demonstrate its relia-
bility, several distribution curves have been prepared from.the fore-
going data. Figure l was prepared from Tables III and IV in which
25 fields were counted for each of 10 smears. The pipette method
shows two peaks indicating two different predominating sized fields.
The wider distribution of the counts indicates more variability
among the slides. The loop technic, on the other hand, shows a
rather compact graph without too large a distribution.

Figure 2 was prepared from Tables XII and XIII in which 300 and
200 fields were examined on one smear for each sample respectively.
These curves give a very good idea as to the small amount of variation
found when using the 100p technic.

In addition to the greater reliability of the results as
previously shown, the loop method is much.more rapid. By using
three 100ps, it is possible to be flaming one and cooling the second
while the smear is made with the third. In this way there is no
waiting andclean, sterile equipment is always immediately available.
The smears of h by 8 mm“ are easily made by placing the h mm. loop
on the slide and drawing the smear out to twice its width. This
requires only one operation as the sample is deposited and spread
over the required area in one motion.

Another consideration is the saving of equipment. The smaller
size of the smear allows about four times as many samples to be placed

on one slide, and the replacement and care of the pipettes are

 

lyre.

xenx

NE u..I§h-.\0

I41 1‘.

1"“

 

I . h; ~.I-nl|.7.‘! 1.-.

 

 

\o\um\ «0K nEgtumIKO

bu. 06. .o\ O\ b 0‘ Q e V d. O

 

- 16 -
entirely eliminated.

One of the biggest difficulties in.microscopic procedure has been
the staining of debris and background. This same problemiwas encount-
ered in working with eggs to such a extent that microscopic examination
was not possible. Recently, Mallmann and Churchill (11) proposed a
staining technic for eggs which overcomes this difficulty. They found
that when an acid solution of methylene blue is used there is a
resultant shift in isoelectric point which causes the organisms to
stain blue while the protein of the egg-meat does not retain the stain.

Since the protein material of milk is somewhat the same as that
encountered in.egg-meats and the problem was the same, it was decided
to try this new technic on milk. Upon examination of the film.the
advantage was at once noticed. In place of the heavy blue clouded
background of the Breed smear, there is a clear field with the bacteria
showing very plainly and almost no debris stained so that it is visible.

The desirability of the method is easily apparent and its relia-
bility has already been demonstrated. It gives much.more speed and
ease in counting as well as the higher results as shown inthe data.

.Mallmann.and Churchill also found that by the use of a red.Wratten
filter No. 2h the ease of counting was greatly increased when dealing
with eggs. The principle behing this is a matter of simple colorimetry.
In order to secure a maximum.contrast, two complimentary colors should
be used. In using a red filter, all the blue light is removed and the

bacteria appear black against a red background. Their main reason for

- 17 -
discarding the blue light was the fact that it would filter out many
of the lightly stained organisms so that they were not counted. 'By
the use of a red filter, these bacteria were easily seen. To avoid
this error in making milk examinations, a study of light filters was
made.

In studying the data obtained in this series, one might be led
to think that the blue or daylight filter should be retained.
However, there are several factors which will not show up in a
tabulation of data.

First, is the fact that the results show that for these parti-
cular samples there was little difference between the blue and the
wratten.No. 22 filters. As in working with eggs, one occasionally
encounters some of those organisms which for some reason will not
retain the stain as well as the others. In this case it is easy
to see that the blue background would "hide" many of these bacteria
so they would not be counted. ‘Unfortunately, no organisms of this
type were found while this study was being made so that it was not
possible to tell exactly what difference it would make.

In addition, there must be considered factors such as ease and
speed of counting which do not show up in the tables of data. It
was found that where the yellow'Wratten filter No. 22 is used, the
bacteria could be seen distinctly and more easily than with a blue
background. This filter has Just enough red to make the organisms

appear dark against the light background of yellow, thus providing

- 18 -

a very good contrast. When the red wratten filter No. 2% was used,
it was found to cause irritation of the eyes, while the yellow
presented a restful, pleasing condition. If work was done
continuously for an extended period, the eyes became fatigued

much sooner with the red than with the yellow light source.

When these factors are considered it can be seen that the yellow
filter has a distinct advantage over the blue, or daylight, and red
filters.

In making an analytical study of the data obtained, one of the
important factors which can be determined is the number of fields
that must be counted to give any desired degree of accuracy. By
taking the formula for determining the deviation of the mean and
then calculating this value from.the allowable per cent error, it
is possible to find the number of fields which must be counted to
obtain the desired accuracy. Referring again to Table XIV, it may
be seen that for poor quality mdlk, samples 3 and 21, when only 10
fields are counted, the error might be as high as 25 per cent. At
the same time if 50 fields were examined, there would occur an
error of only 10 per cent.

In considering the high quality milk (sample 21), comparable
accuracy was obtained by counting #7 fields for 50 per cent, and 25h
to remain within 10 per cent. In routine work the examination of
this number of fields is not possible for each sample, and so one

might be inclined to think that the inaccuracies are so great that

- 19 -

the method is of little practical value. However, in turning to the
practical side there is another consideration which must not be
overlooked. This is the fact that an error as large as 50 per cent,
when dealing with such low counts, is not too significant. By
counting 50 fields, the amount of error can be kept within the limits
of practical usage. In dealing with samples of higher counts, it is
possible to reduce the number of fields counted to 25 and still
maintain reasonable accuracy. Thus, if a sample had an average of
three or four organisms per field one would be safe in counting only
25 fields. With lower counts the minimum of 50 fields should be
counted.

Although this number is considered sufficiently accurate for
routine work, in case a finer grading of mdlk is desired which
requires greater accuracy, the number of fields will have to be

varied accordingly.

SUMMNRY

The results of these studies show that all three of the proposed
modifications in the procedures used in examining milk gave very
satisfactory results. The loop method of preparing the smear gives
a preparation that can be interchanged with the pipette method and
still obtain the same results. In addition, it is much easier and
faster. The series on the acid stain showed that this method of

staining the slides is much superior to the method now in use.

- 20 -

It has also been shown that the use of a yellow wratten filter No.
22 greatly increased the ease and speed of counting.

On the basis of the results obtained, the following recommendations
are proposed for the Standard Methods of Milk Analysis:

1. The method of preparation of the smear on the slide be changed
to the 100p method of Bryan in order to increase the speed of the
examinations and also to obtain greater accuracy of results.

2. The use of the acid methylene blue stain be adopted in place
of the Breed stain now in use. This stain gives a much cleaner field
to work with and a resultant higher count is obtained.

3. When making the examdnations of the smear, the light source
containing a yellow wratten filter No. 22 or its equivalent be adopted
for standard use.

h. It is further recommended that the number of fields counted
follow the data presented in the foregoing section. For routine work
this would be 50 fields for those samples where one organism.or less
appears on each microscopic field. In those instances where there are
more than two or three organisms per field, it would be necessary to
examine only 25 fields.

These changes in procedure will give a greater accuracy in the
results as well as speeding up the method of examination and making
it possible for one technician to handle more samples in a shorter

time with greater accuracies.

- 21 ..
LITERATURE CITED

(1) Standard Methods for the Examination of Dairy Products, Amer.
Pub. Health Aes‘n, 7th Ed. (1939).

(2) Breed, R. 8., The determination of the number of bacteria in milk
by direct microscopic examination. Centbl. fur Bakt.,
Abt. II 30: 337 - 31+0 (1911).

(3) Slack, F. H. , Methods of bacteriological examination of milk.
Jour. Inf. Diseases, Sup. 2, 21h - 222 (1906).

(h) Brew, James D. , A comparison of the microscopical method and the
plate method of counting bacteria in milk. N. Y. Agr. Exp.
Sta., Bull. 373 (1911:).

(5) Breed, R. S. , and J. D. Brew, Counting bacteria by means of the
microscope. N. Y. Agr. Exp. Sta., Bull. 149 (1916).

(6) Brew, J. D. and W. D. Dotterrer, The number of bacteria in milk.
N. v. Agr. Exp. Sta., Bull. l+39 (1917).

(7) Breed R. S. and W. A. Stocking, Jr., The accuracy of bacterial
counts from milk samples. N. Y. Agr. Exp. Sta., Tech.
Bull. 75 (1920).

(8) Robertson, A. H., The relation beWeen bacterial counts from
milk as obtained by microscopic and plate methods.
N. Y. Agr. Exp. Sta., Tech. Bull. 86 (1921).

(9) Drew, James D. , The comparative accuracy of the direct microscopic
and agar plate methods in determining numbers of bacteria in
milk. Jour. Dairy Sci., XII: 1+, 30h - 319 (1929).

(10) Bryan, C. S., G. J. Turney, W. K. Fox, L. H. Begeman, X. A. Miles,
J. S. Bryan, The microscope in the production of high quality
milk. Jour. Milk Tech. I: 5, 26-31!- (1938).

(ll) Mallmann, W. L. and Elbert Churchill, The detection of bacteria in
liquid egg-meats during processing. Ice and Refrigeration,
May (1919).

(12) Bryan, C. S., G. J. Turney, W. K. Fox, L. H. Begeman, X. A. Miller,
J. S. Bryan, Microscopic examination of producer's milk
samples. Mimeographed material from Department of Bacteriology,
Michigan State College (1938).