Page images

due to the fact that such a small quantity of milk was examined in proportion to the number of bacteria present that the few present did not appear in the microscopic fields examined. On the other hand the distribution of the 32 counts of this type among the higher count milk indicates that some of the errors arose from unrecognized contaminations on the plates.

The second class contains 123 samples (17 per ct.) in which the plate counts were less than the “group

group ” counts. In this class, it is also evident that there must have been errors of considerable size. As these occurred evenly distributed throughout the series of counts, it is evident that the cause of them must be something which affects all classes of milk. The most evident explanation and apparently the most probable is that in many of these cases bacteria were present which did not grow or grew poorly on ordinary agar. In some cases, the plate counts were probably low because of the presence of dead bacteria. The plate count in all of these cases gave a mistaken idea of the quality of the milk.

The third class contains 345 samples (54 per ct.) in which the plate counts are greater than the "group" count and less than the individual count. If however the 200 counts which were made from samples containing less than 30,000 bacteria per cubic centimeter are excluded, the number of instances in which the plate count is intermediate between the microscopic counts form 71 per ct. of the 443 counts which remain. Since comparative counts would naturally be expected to show this relationship, there is reason for thinking that these counts are not subject to the gross errors which were noted in the case of classes I and II. The plate count probably represents fairly well the number of “ groups” of bacteria in the milk after they are broken apart by the dilution waters; the “group” count represents the number of “groups” of bacteria originally present in the milk; and the individual count represents the number of bacteria actually present. Under these conditions the variations in the size and compactness of the "groups" of bacteria originally present in the milk have the greatest influence in determining whether the plate count will be about the same as the “ group” count or whether it will be more nearly equal to the individual count.

Market milk containing very few bacteria, the majority of which were presumably derived from udders which were not infected with streptococci, were found to show "groups” of bacteria of very

[ocr errors]

the "

small average size (2.2 individuals per group ”). There appears to be a marked increase in the average size of the

groups market milks containing larger numbers of bacteria, some of them long chain streptococci, others masses of bacteria washed off of the surfaces of utensils and like sources. The largest average size of groups”

was found in market milks containing less than 1,000,000 bacteria per c. C. In milks containing larger numbers of bacteria there was a gradual decrease in the average size of the groups

as the number of bacteria increased. The apparent reason for this decrease in the average size of the “groups” in high count milks is that there is a gradual change in the types of bacteria present. The longer the milk stands and the more growth there is, the more the bacteria originally present are overgrown by lactic acid types which grow normally in small " groups."

Variations in the size of the “groups " of bacteria affect the accuracy of the microscopic count because of the difficulty in counting the individuals in large, dense clumps and also because a few large clumps increase the error due to unevenness of distribution. The accuracy of the plate count is affected both because of the variations in the size of the “groups" and because the clumps do not entirely separate into their component individuals during the plating process. The groups

" of bacteria ini milk break apart on the average into 2.6 smaller “groups " in the process of preparing dilution waters. That is, under usual conditions the plate count will be increased from 2 to 3 times as a result of this factor alone. In individual cases, however, it may be increased more than this, the highest increase noted being 16.5 times.

The accuracy of the plate count is affected both because of the variation in the size of the “groups" and also because the clumps do not entirely separate into their component individuals during the plating process. The actual size of any plate count is largely the result of the interaction of these two highly variable conditions. Because no way exists in which the errors thus introduced can be discounted except by microscopic examination of the milk, the use of plate counts may result in great injustice being done where an attempt is made to make fine distinctions in the bacterial quality of milk.

Inasmuch as there are as yet no other more satisfactory methods in use for distinguishing between the bacterial quality of different samples of milk, it is fortunate that either of these methods (and this is more especially true of the microscopic method) can be used in a simpler way than is generally being done at the present time. The fact that accurate counts cannot be made does not prevent the use of either method as a means of distinguishing with a satisfactory degree of accuracy between milks containing few bacteria, a medium number, and a large number. The fact that this is all that is necessary in commercial or in control work has been shown in the inspection work which has been done for the City of Geneva during the past three years, a description of which will be given in a forthcoming bulletin.13

It is evident from all of this work that the current and popular ideas of the number of bacteria in milk based as they are upon counts made by the plating method must be revised. The results obtained by the microscopic method of counting bacteria in milk show that plate counts are not counts of the number of bacteria in milk as usually stated; but are counts of such groups of bacteria as grow upon agar at the incubation temperature used. Figures commonly given as showing the number of bacteria in milk are therefore ordinarily very much less than the actual number of bacteria present.

13 Breed, Robert S., and Brew, James D. N. Y. Agr. Exp. Sta., Bull. 443. 1917.





1. A preliminary series of tests was made in a commercial bacteriological laboratory at Hobart, N. Y., in which the microscopic method of grading milk according to its bacterial quality was compared with the more generally used agar-plate method. The results secured from the agar-plate method at this laboratory were used as a basis of payment to dairymen, a premium being given for milk containing less than a specified number of bacteria.

2. The results of the comparative analyses showed that out of 1504 samples graded during the months of February, March and April, 1914, 1280 were graded A by the plate method and 1270 were graded A by the microscopic method. In 1339 instances there was an exact agreement in the grades given. In the case of 426 samples taken during July, 234 were graded A by the plate method and 212 were graded A by the microscopic method with an exact agreement in grade in 362 instances.

3. It was found to be impossible to get an exact agreement in results by the two methods as the plate count is not a count of the individual bacteria, but rather a count of the number of groups of bacteria as they exist after the groups are more or less broken apart by the shaking of samples and dilutions in the process of plating. The microscopic count, on the other hand, is either a count of the groups of bacteria as they originally existed in the milk, or is a count of the individual bacteria.

4. By a cooperative agreement with two milk companies in Geneva, the microscopic method was tried out during a period of twenty-six months for the purpose of grading the milk delivered into three grades. The results were used as a basis for payment to the dairymen.

* Reprint of Bulletin No. 443, December, 1917.

5. No unforeseen difficulties were found in using the microscopic technique in this way. In all, 11,851 cans of milk were examined and classified in three grades. The entire work was done by one man who found it quite possible to handle from fifty to seventy samples each working day. Special precautions were taken to get accurate results by both the plate and the microscopic methods in a series of 643 comparative analyses. The results showed an exact agreement in grade in 587 instances (91.29 per ct.). This figure probably represents as high a percentage of agreement as can be obtained even under the very best conditions.

6. The results secured from microscopic examination of the milk were found to be more useful than plate counts in interpreting the cause or causes of excessively high counts in the milk.

Thus a record of the samples in which the predominant flora consisted of the long-chain streptococci frequent in udder infections, showed that one-fifth of all of the milk delivered at these milk stations which contained bacteria in excess of 1,000,000 per c. C. contained large numbers of streptococci. So far as was determined, all of the milk of this type was originally infected from the udders of some one or more cows in a herd. The conditions observed made it probable that most of the remaining high counts (except where the milk was not properly cooled) were caused by the improper care given cans and milking machines.

7. The simplicity and relative inexpensiveness of the microscopic examination of milk as well as the fact that the microscopic preparations are permanent, gives this technique many advantages over the plating method for commercial work.


Three years ago the Station published accounts 1 of the results obtained by using a microscopic method of counting the bacteria and cells in milk. At the same time the counts of bacteria were compared with those obtained by the more commonly used agarplate method. Soon after those bulletins were published plans were made to test the value of the microscopic technique under practical

1 Brew, James D. A comparison of the microscopical and the plate method of counting bacteria in milk. N. Y. Agr. Exp. Sta., Bul. 373. 1914.

Breed, Robert S. Cells in milk derived from the udder. N. Y. Agr. Exp. Sta., Bul. 380. 1914.

« PreviousContinue »