COUNTING THE LIVING BACTERIA IN MILK-APRACTICAL TEST

W. D. FROSTFrom the Laboratories of the Department of Preventive Medicine and Hygiene,

Harvard Medical School, Boston, Massachusetts

The method under test, in the series of analyses reported here,has already been described (1915, 1916a). A comparison ofresults obtained in a series of milk analyses by this and thestandard plate method has also appeared (1916 b). The fol-lowing advantages have been claimed for the method:

1. It is rapid, requiring only about one-eighth of the timeneeded for the standard plate method.

2. The technic is simple and requires little glassware or cul-ture medium.

3. It furnishes a means of keeping a permanent record of eachanalysis.

4. Only the individual bacteria, or groups of the same, thatare alive grow into colonies. This makes it possible to use themethod for counting the bacteria in pasteurized milk, which isnot possible by methods of direct microscopical examinationsuch as those of Slack and Breed.

It seemed desirable to make further tests of the method, andif possible, under practical conditions, such as working in connec-tion with the routine examination of a city milk supply. Throughthe kindness of Dr. M. J. Rosenau of the Department of Pre-ventive Medicine and Hygiene of the Harvard Medical School,arrangements were made for me to carry on a series of tests inthe laboratories of the Boston Board of Health. The workhas been made possible further by the courtesy of Dr. F. H.Slack, director of the laboratories, and Dr. W. M. Campbell,who conducts the bacteriological work on milk.

In this laboratory all samples of milk collected by the milk567

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

44

1-44

0

CSb,C

CS

D-D0fi

gQ.Q

-44

C)

0 0

00

140

z

00 O000CO

- -c-W= e c -oo_ oz0 C) 6 C O' C; 6) C) O' C)..4 ..- ..- ..4 ..4 ..4 ..4 ..4 .-

m c4^ 06 c t:-01 0C

0aoo^o^ oI- xo cl<sa es ab e:n n e4

U: _- C-1 o~U- Co Oo

C C- 0;C C~ 0OQ

0-

oco t$S mo0o_

O o C Oco _ om o o

4

.. i.. . . . *n * . . . . .

co eg e c-O - - - --: -

:OOOOC

cs~~~~~~~~~~~~~~~~~~~~- cq

0q A to o-

x0 4V4

r

4FOI_4 I,'

_J cl

,

9 ctL cses _~~"- cli cl

). CooeCO P N c o0)ecc a_%~t2 g cs es_' coc= _ c

ccoCS mSC co (Mesa

0LVM-;4I +IIIII + + +I++)livoIdooDGaoHIW +

ts tqr X tq t tq tq tq tq t tq t tq tq tq L t tq

0 05050 5 0 5 50 5 0 5 0 0

-4 m m mN' m mb m sPUU P P P4 P4ooo22

O~~~~~~~~~ 1'; *q <D 11<D4$ r'> (A) zVQQQQQQQ

6

co4 or- c;4V r-4 uu4 r- c cq cq

_-- ' _- -_- _- _ -'-4 _ " L0 to t. 0

CPV--

V-4

L6r-

&4Ca

CO

co

CC)14

-4

9-4

Ca:P

568

toE-4

I' ,

04

0,Itc

I-C

44

-41.

9q

.c't

-eI.CIt

IQ

P4cA.q

Iti.c

-4 c

I4pq q

I

9

.1I

I

14

.II4

4I14

%94

50Z3Z)C!1.4k)

NZ

9.1CA)

w.-4eat:l;x0e9C-1-eC)0QQ

i8.ODt-2qD

rt:49t.)

mE-4

4

z

440

0

4)

P0

0S0:

14

V)054)

II

P)0

05

0o)05

4)II

+

0

4),.

4)e05

1.

B

0n)

*i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I

Ir-4 C9 m V 10 = 1- X = (D -4

V-4 r-(

.-esq

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK

It, 2 IVWo- 0U. C. . . -

,: o'4 T'4et b; ;

569

-.. t.q- co CC

... .. ..

1: 1:_ V

88, 0 517°O$8 "- P- V-

8 °8 °oo °o 8 800 o1 o° o 8

V-4 r-4

C5 §o eD0 _ 00 COOO-O-co Oq O 00 tI-C C1 00

1 m co

_l_,~~~~~~~~~ ~~~~~o;c°l-agI-z>boO g 1;1geX 0

-I

cqV-1 &~~~- P- AO34

343 .434343434 z 343 343.43-434343434 3.4~~~4~~~~4 O4~~~~~~~~~4~~~~404 - 0404 04~~~~~~~~~~~~~~~~~~~0404 0404 0404 04

II - ,.- .. -,. . 18SI *I-~ - - I- +-I- I: *D +I' 4I IQ -4; *I--*

rI.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

MMM c 2 O O U O t2m m m m m m m m m mCS O$} 8 8 } }}oadd440 4 4o4O-C1̂08OtPooo^ -

cO

3-4

A

CO CD 10

o -0

7--

V-- C-

t

0

'---.

0

Co"4

C-t-V-

Ca

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

0

00 LO U1 oO co o-Cq 1U0 e4-t - MCO COCO CO OIC, 1

~10 10O ')O ' O'C

*0o*0 - - CO o0 Cee 0o0Co

........-...........

CO CO CO 4oo

o_fl MoC1010010100 t1- 1i00 coCU01010 10

o o oCO'-'.0 CO 0'. eq 0

S o

E- xo° °°e 8°88oo4o°If 00 R ~ 0 0 000000goo000_0o_ __

o ° G, mxr<o c)to 1, rb Cy: eie cn eD ' t° 1- XOb q ei ci cg X < *n g X >~~m cl N m _ - N to = _ q m ol

00~ 0 00 0 1 0

co'. -41 10eo r-.4CO CO-4 c "1'-4~a) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-4

a e Om C _OC)eqeq,-4 01 rcq C14eqeqe 10,o-.4 Go t- cli-4 W O-'4.

0 o0 a*S.4 e.ED-.4U0c-- 77 X4C'-.4 -.4 CO_-.4ed_O4

X ;; b °: °: °: °::8CC) C0 C):0 c:

C

:c°O°::

1=~~~~~~~~~~~~0°M v-4 r- EfCDc M=x o

(xov-4as)x

1. 4 4 x44o .

8~~~~~~~~~~ ~ ~~~~~~~~~~~~~~-- C,_ _ ,_ s: , _ s a

M~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~N N tA bq tq; Q Q~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Q) a) v w c) ) C) av 4

nW W 3 W W W W > B B W B < <~~~~~~~~~~~~~~~~~~~m;m m mm m m 9Z~~~~~~~~C Cdc' ce Cae e ce cl d ce ce Ca C3 C3 c3 c3 Ca c3 :3 w Ca Ca

00 C)O_o C) "-I 04 ms O RC t XO I= t- 0 = C)4 0CO CO:C COecCO CO. CO tb~ tbb tb -C0

xo CS tOo co co to coco;O---CA _ - 1- t-dz tC 00C6 -_/_ It_ x C_ _ ___- I - W_z ~ ~

CO

7-410

C)4.43e

to

CL)

1-4-4

570

I;s

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK 571

OC8 C= U0 000 0 V0a4 0 C~ec; C4

coC; c; C; CD C; C; coo e-

--- -- * - --- --*- -..

10 Cq 10 CObS C C- 14uC o 1cli 0

~~~~~~~~~~~~~)co o ccoa7- C css_££_os O

0 +

C~~IC0~~~~~-4~c-I 10C0CO4D t

o oc k

osco NO cro_ooico- o ocq

e> 1O - -o 0D- qO _

H~~~~~~~~" 7. IP-p-_14VD:CS e>co CT'l

ooo o o o o o~~-4 o CD xo o oo o ooo

O6OOmOO O O O O O O O ° O O O O 0 00 0 0 o O

O~~~~~~~~C _, C6 1 (O? 9OZ O 0Huje&o+b scser coesr4744m com c

1 HO -O A4 8CS a OC ) &0 A v A A A0~~~~~~~O ~ -_ ~~~~ ~ ~~~~~- - -I - - - -

10C~~~~~~~~~~~~COCq t 00les

.qco_U.

0

C> tl- 11~O C4

+III+IIIIII+ +++++±++++ IIl+IIIII

00Oo 0 0000 O00 00

oo0 0 0 0 0= 0~ - -

- -~~~~~~~~~~~~~~~~~~~~~~-

cM X CO t 0 C N c )co ) s

14 14r- q qolC

'-4

_ cses t UO co bCa o esbXoo co O_

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

inspector and sent into the laboratory are examined micro-scopically by the Slack method, and rated as above or belowthe legal limit of 500,000 bacteria per cubic centimeter. Thosesamples rated as likely to show more than the permissible numberof bacteria are plated out, as also are all samples of pasteurizedmilk. The presence of streptococci and pus is noted. Themicroscopical estimate is used further as a guide to the dilutionneeded in platimg. Ordinarily duplicate plates of only onedilution are made. This is usually 1: 10,000.At the time the routine plates were poured I made little plates.

From sample 59 to sample 111 (table 1) I also made standardplate cultures.The medium used for the ordinary plates was made according

to the standard methods. It had a reaction of + 1.5 and con-tained one per cent of agar. Ten cubic centimeters were usedfor each plate. The Petri dishes had clay tops. The temperatureof incubation was 37.5°, and the time forty-eight hours. Theair of the incubator was saturated with moisture. The countingwas done either with the naked eye or with the 'aid of a readingglass.The little plates were made in the manner previously described

(Frost, 1916 a) and incubated five to seVen hours. They werethen dried, stained and counted. The same agar was used asthat supplied the laboratory. It would have been better, nodoubt, to have had a medium containing more agar, since somedifficulty was experienced from spreaders, and in highly con-taminated milks the little colonies were not so well individualizedas would have been the case with a stiffer agar.The results obtained from these comparative tests are given

in table 1. The kind of milk is indicated, i.e., whether raw orpasteurized. The microscopical estimate by the Slack method;the plate counts obtained in the routine analyses, together withthe number of colonies actually seen on the plates; the platecounts obtained by the writer; the results obtained from theuse of the little plates; and the ratio which the results obtainedby the different methods of analysis bear to each other, areindicated in the different columns of the table.

572

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK

E--X 1111211 111L11L411WL::::::A0'^ 1 n F 0

573

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

All of the analyses made during the time of the comparativetests are included in the table, although it is apparent thatthedata are too unsatisfactory in a number of cases to be of muchvalue; yet for the sake of completeness all are given.The general results are shown in chart 1.The milks used were both raw and pasteurized, and varied in

bacterial content from very good to highly contaminated milks.One of the first things to note is the fact that because of

spreaders on both plates (which unfortunately were unusuallyprevalent in the laboratory at the time of these tests), no countscould be obtained in 16 out of the 111 tests, and in 16 othersamples one of the duplicate plates was spoiled by spreaders.While spreaders occurred on the little plates under exactly

the same conditions, their presence did not materially interferewith a count of the plates, with the possible exception of one.It would seem then that the little plates could be relied uponto give countable colonies with more certainty than the standardplates.The second thing to notice is that there is considerable varia-

tion between the counts on the duplicate plates of the routineseries. In ten samples the difference amounted to 'over 50 percent (5, 9, 26, 28, 30, 43, 71, 78, 89, and 102) and in eleven othersamples it was over 100 per cent (31, 32, 34, 39, 41, 64, 68, 70, 74,76, and 87).

Again, the number of colonies on the plates in about 40 caseswas below 20, which would seem to be the smallest number ofcolonies that can be relied upon as giving anything like a satis-factory degree of accuracy.

Excluding those samples in which one-half or more of theplates were rendered useless because of spreaders, and also thosesamples in which both plates had less than twenty colonies, itwill be seen that only 28 of the 111 counts can be considered asthoroughly reliable in indicating the exact number of bacteria inthe sample when the standard plate method was used.The criticisms here made of the routine plates would apply

with equal force to the duplicate Petri dish cultures made bymyself. Of 53 only 10 were entirely satisfactory in that both

574

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK

plates were free from spreaders and had a sufficient number ofcolonies present to be thoroughly reliable.The agreement between the two series of standard plates

varies from a ratio of 1: 0.09 to 1: 8, with an average ratio of1: 1.42. In other words, the plates that I made from one samplecontained slightly less than one-tenth as many colonies (9 percent) as the routine plates, and in another instance my countwas 700 per cent higher than the routine, while my averagefigures were 42 per cent higher than those of Dr. Campbell.The little plates, on the other hand, were all countable. Some

of them were too thickly seeded, and others were overrun tosome extent with spreaders, but neither of these things pre-vented the counting of the plates. It might have been expected,for several reasons, that the counts obtained by this method wouldnot closely approximate the count obtained by the standardplate method, the chief of which is that the medium is somewhatdifferent from the standard in that it is half milk. Again, theshort period of incubation might not be sufficient for the de-velopment of colonies from some of the slowly-growing bacteria;and finally, crowding might be sufficient to inhibit the growthof some colonies. In spite of these differences and possible handi-caps, however, the results obtained in this competitive seriesshow a reasonable agreement.Compared with the routime standard plates (leaving out of

'account three of the samples [39, 44 and 53] which are evidentlynot comparable), the ratio varied from 1: 0.05 to 1: 7.43, with anaverage of 1: 0.96. This appears to be a better showing thanwas obtained in duplicate runs by the standard plate methodalone where the average variation was 1: 1.42. It must bepointed cJut, however, that this comparison gives a false impres-sion, since the counts obtained by the little plates are on thewhole somewhat lower than those obtained by the standardplate method. For example, if we compare the results I ob-tained by the standard plate method with the routine platesmade by Dr. Campbell, we will see that my count fell below theroutine count nineteen times, while-it was higher in 14 samples.That is, the variation occurred either up or down with about

575

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

equal frequency. On the other hand, comparing the resultsobtained from the little plates with the results from the routineplates, it is found that in 48 samples the count on the littleplates fell below the count obtained by the standard plate-method, and above in 26 samples. In other words, the counton the little plates was lower than that on the standard platesin two cases, as against one above.There are marked differences between the results obtained by

the two methods in several samples. Some figures on the littleplates are very much lower than those obtained by the standardplate method. This might be, accounted for by the suppositionthat colonies did not have time to develop in the short period ofincubation, but in some of these cases, at least, a long period ofincubation had little effect in raising the count. What seems

more likely in these instances is that we are dealing with samplesof milks in which the bacteria have a tendency to form groups

Tests comparing the plating of 120 cc. of whole milk with the ordinary dilutionmethods.* (By W. M. Campbell)

1/20 CC. WHOLE mILK DILUTION 1:20 DILUTION 1:100

Number of Count Number o Count Number of Countcolonies

Cut

colonies conoloniesCon

712 14,240 1,109 22,180 401 40,100688 13,760 1,042 20,840 365 36,500181 3,620 257 5,140 98 9,800173 3,460 229 4,580 87 8,700660 13,200 870 17,400 220 22,000720 14,400 680 13,600 200 20,000185 3,700 415 8,300 141 14,100616 12,320 1,030 20,600 331 33,100202 4,040 210 4,200 49 4,900200 4,000 237 4,740 128 * 12,800232 4,640 129 2,580 71 7,100340 6,800 224 22,400449 8,980 217 21,700128 2,560 200 4,000110 2,200 240 4,800118 2,360 128 2,56096 1,920 140 2,800

* American Journaj Public Hygiene, 17, p. 359.

576

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK 577

that are not easily shaken apart except as they axe diluted withwater. This fact has been brought out especially well by aseries of tests given in the Reports of the Boston Board ofHealth (1907).

This grouping of the bacteria no doubt occurs to some extentin all milks, but appears to be more noticeable in the bettergrades of milk.The little plates occasionally give much higher counts thanthe

standard plate method; see, for example, nos. 39, 44, and 53(table 1). In these samples. the little plates showed a largenumber of small colonies of the lactic acid type, and it is quiteconceivable that the count obtained by the little plates is moreaccurate than that obtained by the standard plate method.

TA&BLE 2

Number of bacteria in plain and lactose agars-from Sherman 1916

NUMBER OF BACTERIA PER CUBIC CENTIMETERSAMPLE NUMBER*

Plain agar Lactose agar Ratio

1 817 1,017 1:1.252 1,340 2,070 1:1.553 1,630 2,000 1:1.234 3,230 3,530 1:1.09(3) 6,000 69,000 1:10.50(2) 6,500 42,500 1:6.545 7,330 7,330 1:1(1) 7,600 25,300 1:3.336 8,000 11,300 1:1.417 8,900 12,100 1:1.368 9,030 8,970 1:0.999 11,800 11,170 1:0.95(4) 18,100 43,600 1:2.41(5) 23,800 58,000 1:2.44(6) 72,000 177,000 1:2.4610 186,000 610,000 1:3.2811 192,000 278,000 1:1.4512 260,000 361,000 1:1.3913 369,000 463,000 1:1.2614 53,700,000 109,000,000 1:2.03

* The numbers of the samples are the same as those used by Sherman in histable 1. The numbers in parenthesis are from his table 2. The last four analyses(nos. 15-18) are not given, because of the difficulty of charting such high numbers.See chart 2.

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

578 w. D. FROST

c~qE0V4

a

tf

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK

Sherman (1916) has recently shown that sugar agar gives highercounts than plain agar on the same milks. I have combined histables 1 and 2 in my table 2. These figures are representedgraphically in chart 2.The large amount of milk in the little plates makes this a

sugar medium, and it is possible that the discrepancies underdiscussion can be at least partially explained on this basis.As far as the accuracy of the little plate method is concerned,

when judged by the standard plate method, this study wouldseem to indicate that on the average the results are in fairlyclose agreement, with the probability that in about half of thesamples the count will be somewhat low, and that other sampleswill show a much higher count than that obtained in the usualway.

It is interesting, in this connection, to compare the resultsobtained by different workers using the standard plate methodin analyzing milk, and for this purpose I have brought togethersome of the results obtained by Conn (1915), and incorporatedthem in table 3. The figures were obtained by taking alter-nate numbers in Conn's table 2 and arranging them in an ascend-ing scale. The column headed "S. Method and S. Media" wastaken, but any other would apparently have given similaryariations. The same figures are represented in chart 2.

TABLE 3ata from "Standards for Determining the Purity of Milk". H. W. Conn. Table ,p. S866, Public Health Reports, August 18, 1915. Averages obtained by labora-tories A. B. and C. Rearranged in one series using alternate figures.

REISULTS OBTAINED FROM THE SAME MILKS BYRATIO

Laboratory A Laboratory B Laborator C

760 2,960 2,725 1:3.9:3.5800 610 6,900 1:0.7:8.6950 1,350 2,780 1:1.4:2.9

1,000 1,010 3,000 1:1.0:3.01,000 3,100 3,000 1:3.1:3.0

1,010 5,330 8,130 1:5.2:8.11,150 2,770 2,250 1:2.4:2.01,200 500 6,700 1:0.4:5.61,250 1,586 9,650 1:1.3:7.71,340 1,840 2,760 1:1.3:2.0

579

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

TABLE 3-Continued

RESULTS OBTAINED FROM SAME MILKS BY

Laboratory B

3,3101,1601,5305,5001,100

7002,5601,9752,9602,320

9002,1751,2002,9502,100

2,2301,9758,000760

57,000

22,50013,30053,000198,00030,500

29,00055,00024,500135,00027,400

158,000244,000930,000

2,700,0001,600,000

7,000,000

Laboratory C

5,3812,5003,5007,8007,450

7,7102,8807,3752,7253,560

6,9307,5805,7505,5107,210

14,3005,750

24,00012,30055,000

99,000138,000113,000

1,730,000145,000

75,000121,000124,000523,000146,000

1,700,0001,600,000

22,000,0008,190,0002,700,000

164,000,000

580

Laboratory A

1,5001,8101,8301,8301,840

1,8502,0502,1002,1302,440

2,6602,7003,3253,4503,900

4,1005,05011,56017,20025,100

29,00035,00039,20040,00042,300

49,00052,60054,60059,100130,000

337,000533,000

2,350,0003,500,0006,720,000

14,600,000

RATIO

1:2.2:3.61:0.6:1.31:0.8:1.91:3.0:4.21:0.6:4.0

1:0.3:4.11:1.2:1.41:0.9:3.51:1.3:1.21:0.9:1.4

1:0.3:2.61:0.8:2.81:0.3:1.71:0.7:1.61:0.5:1.8

1:0.5:3.51:0.4:1.11:0.7:2.11:0.04:0.61:2.2:2.1

1:0.7:3.41:0.3:4.01:1.3:3.01:4.9:43.01:0.7:3.4

1:0.5:1.51: 1.4:2.31:0.4:2.21:2.2:8.81:0.2:1.1

1:0.3:5.01:0.4:3.01:0.4:9.01:0.8:2.41:0.2:0.4

1:0.5:11.2-L

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING IMVING BACTERIA IN MILK 581

The variations, it- will be noticed, are similar to those foundin comparing my method with the standard plate method,, and Ibelieve I am warranted in saying that variations very similarin magnitude to those obtained by using the standard platemethod and the little plate method may occur when two or moreworkers analyze milk by the standard plate method alone.

In considering the accuracy of this method, one of the thingsthat must be determined is the uniformity of the distribution ofthe colonies on the little plates. Experience in these tests isthat the distribution of colonies in the various fields in anylittle plate is quite uniform. To show this, the counts of twentyfields are given here by the mechanical selection of these fieldson ten different plates.

NO.1LOW

1376161010.71716141811101516101161010

NO. 5LOW

10910981011131114132115914148131419

NO. 10OIL IM.

811106343

* 109116437976643

NO. 15

13131215888556663958111286

NO. 20LOW

4253414034242426222725453529332632252235

NO. 30LOW

6250555768544744505839405043546050395347

NO. 40OIL IM.

87111217192099171113131313101114912

NO. 50LOW

6262585862666246433843345753595655706868

NO. 60LOW

01322202234410143221

NO. 70LOW

106127.788681311791211964912

20)233 20)245 20)130 20)167 20)640 20)1020 20)248 20)1120 20)39 20)175

12 12 6.5 8 32 51 12 56 2 9

THE JOURNAL OF BACTERIOLOGY, VOL. II, NO. 5

Another matter of importance is counting at different magni-fications. In a former communication (1916 b) some emphasis

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

W. D. FROST

was laid on the discrepancies obtained with different magni-fications used for counting the little plates. The data presentedwere for counts where the period of incubation was only four orfive hours. If the period of incubation is seven or eight hoursthere seems to be little difference, and that magnification may beused which will give the most satisfactory number to count.

In comparing the two methods other considerations need briefmention.

First, the method is rapid, requiring not over seven or eighthours.

Second, the little plate method is reliable in that it isnot likely to fail entirely because of spreaders, too great a dilu-tion, etc.

Third, it permits a study of a relatively large amount of themilk, usually 1/20th of a cubic centimeter.

Fourth, it furnishes a permanent record of the bacterial con-tent of the milk.

Fifth, it requires less material in the way of glassware andmedia. Microscopical slides, sterilized in the flame just beforeuse, are substituted for Petri dishes, and the amount of mediumis not over 1/20th and may be as little as 1/200th part of thatused by the standard plate method.

Sixth, it is less time-consuming to make the little plates thanit is to make the standard plates, and the only previous prepa-ration necessary is the sterilization of pipettes and test tubes foriixing the samples. The counting of the little plates requires

more time if 20 fields are carefully gone over. It seems proba-ble, however, that the extensive counting may not be requiredin routine work. If the little plate is gone over quickly togain an idea of the distribution of the colonies, the counting offour or five representative fields will be sufficient. Tests of thiskind made during the progress of this work gave very satis-factory results. By varying the magjification, fields with com-paratively few colonies in them can always be obtained.The method may seem complicated to some, but in reality it

is not, and if one will take the time to master the principlesinvolved, the method is much simpler than the standard platemethod.

582

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COUNTING LIVING BACTERIA IN MILK 583

REFERENCES

CONN, H. W. 1915 Standards for determining the purity of milk. Pub. HealthReports, 30, 33.

FROST, W. D. 1915 Rapid method of counting bacteria in milk. Science,N. W. 42, 255-256.

FROST, W. D. 1916 a A rapid method of counting living bacteria in milk andother richly-seeded materials. Jour. Am. Med. Assoc. 66, 889-890.

FROST, W. D. 1916 b Comparison of rapid method of counting bacteria inmilk with standard plate method. Jour. Inf. Dis., 19, 273-287.

HEALTH DEPARTMENT, Boston, Mass. 1907 36th Annual Report, p. 76.SHERMAN, J. M. 1916 The advantages of a carbohydrate medium in the rout-

ine bacterial examination of milk. Jour. Bact., 1, 481-488.

on March 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from