386 MILK AND MILK PRODUCTS 
public health demand that the control of milk supplies, both as to production 
and distribution, should include regular laboratory examinations of milk by 
bacteriological methods. They stated by resolution that— 
Among present available routine laboratory methods for determining the 
sanitary quality of milk, the bacterial count occupies first place. and that bac- 
terial standards should be a factor in classifying milk of different degrees of 
excellence. 
The adoption and enforcement of bacterial standards will be more effective 
than any other one thing in improving the sanitary character of public milk 
supplies. The enforcement of these standards can be carried out only by the 
regular and frequent laboratory examinations of milk for the numbers of bacteria 
it may contain. 
It is of the utmost importance that standard methods should be adopted by 
all laboratories for comparing the bacterial character of milks, since by this 
means only is it possible to grade and classify milks and properly enforce bac- 
terial standards. 
Concerning the methods which should be used by milk laboratories for 
determining the numbers of bacteria the commission unanimously resolved: 
That there be adopted as standards for making the bacterial count the stand- 
ard methods of the American Public Health Association Laboratory Section. 
One of the chief objections raised against pasteurization is the claim that it is 
frequently employed to cover filthy methods, the milk producer using less care 
in his methods if he knows that the milk is to be subsequently pasteurized. To 
meet this objection the commission believes there should be bacterial standards 
for raw milk as well as bacterial standards for pasteurized milk. In the case of 
pasteurized milk, standards should be required of the milk before pasteurization 
as well as after pasteurization. 
Reltability of Bacterial Tests. The commission has considered the numerous 
criticisms that have been raised as to the unreliability of bacteriological analyses, 
and has made extensive inquiry as to the force of these criticisms. An opinion 
concerning the reliability of laboratory tests for numbers of bacteria has been 
reached based on voluminous statistics secured for the most part by groups of 
observers working together, as well as by individuals. One of these researches 
alone carried out by members of the commission in coopération with others 
included the testing of over 20,000 samples of milk. In other instances repeat- 
edly the same sample of milk was tested one hundred times. Some variations 
in the analysis of duplicate samples are inevitable, due to the fact that the bac- 
teria are not in solution, but are floating in the milk more or less clustered to- 
gether in clumps, each of which will count only as a single colony. Under such 
conditions only an approximate agreement can be expected. 
The results of extensive study justify the commission in the conclusion that 
the analysis of duplicate samples of milk made by routine methods in different 
laboratories may be expected to show an average variation of about 28 per cent, 
with occasional samples of wider variation. In some good laboratories the varia- 
tion may not be greater than 10 per cent. Variations in results diminish with 
