MICROSCOPIC METHOD OF MILI ANALYSIS 395 
hours rather than twenty-four hours because on an average the count 
is doubled or trebled. In this connection he advises meat infusion 
agar in place of meat extract agar. Sherman (1916) recommended the 
addition of dextrose or lactose to agar used in the examination of 
milk. A much higher count was obtained and the colonies grew 
much larger. 
Incubation. Only one period of incubation and one temperature 
is regarded as standard, forty-eight hours at 37.5° C. In crowded 
incubators ventilation shall be provided. 
Counting. If among the different dilutions, there are plates contain- 
ing from 30 to 300 colonies these should be counted, and the number 
multiplied by the dilution, be reported as the final count. All colonies 
on such plates should be counted and the numbers averaged. If there 
are no plates within these limits the one that comes nearest to 300 
is to be counted, unless it happens that there are no other plates with a 
larger number of colonies, or, unless the numbers in the plates check 
with other dilutions. ° If the number of colonies on the plate is over 300, 
a part of the plate may be counted and the whole plate averaged. 
Counting shall be done with a lens magnifying 2.5 diameters (or what 
opticians call a 2.5 lens. Near-sighted persons shall use their glasses 
in counting while far-sighted persons shall remove them. In case it is 
doubtful whether certain objects are colonies or dirt specks they should 
be examined under a compound microscope. 
Reports. In making reports it must be borne in mind that with 
high numbers obtained by routine methods, only an approximation to 
accuracy can be obtained. Only the left-hand figures of the final 
numbers are significant. It is best, therefore, to report only the two 
left-hand figures of the results, in order to avoid an unwarranted im- 
pression of accuracy. For example, where the numbers are in the mil- 
lions no figures smaller than millions have any significance in the routine 
analysis of milk. In making the report raise the number to the next 
highest round number, but never lower it. In no case shall the count 
of a single plate be regarded as sufficient for the purpose of grading milk. 
If a single sample of milk only is to be tested, there should be at least 
three plates counted before a report is made. 
Microscopic Method of Analysis. This method has been brought 
to such a stage by Breed and his colleagues that it is now included 
in the Standard Methods of Bacteriological Milk Analysis of the Labora- 
tory section of the American Public Health Association, 1917. Brew 
has summed up the advantages and disadvantages of the plate and 
direct microscopic method as follows: 
