COMPOSITE SAMPLING 63 
served that the sampler did not cover the jars at all! 
Can we wonder why patrons complain so frequently about 
the testing? Where the jars are kept uncovered for sev- 
eral weeks the cream is in a condition in which it can 
not be reincorporated with the milk and the Babcock 
test in this case becomes truly a snare and delusion. 
Should the samples show any dried or churned cream 
on testing day, the sample jars must be placed in water at 
a temperature of 110° F. for five or ten minutes to allow 
the cream or butter to melt. When this is done the 
sample for the test bottle must be taken instantly after 
mixing, as the melted fat separates very quickly. 
Frequency of Testing. It must not be supposed that 
if enough preservative can be added to the sample jars to 
keep the milk sweet for a month or longer that it is just 
as well to make monthly tests as weekly. Far from it. Even 
if the milk does remain sweet, the tendency of the cream 
to churn and become dried and crusty is in itself sufficient 
protest against monthly testing. It is rare, indeed, that 
samples that have been kept for a month or longer can be 
sampled satisfactorily without warming them in a water 
bath, which means a great deal of extra work. 
The best tests are secured when the samples are tested 
weekly- or at most every two weeks. When the tests 
are made weekly it rarely becomes necessary to warm the 
sainples if they have been properly cared for. Then, 
too, if an error is made anywhere in the testing, there 
are three other tests for the month that help to mini- 
mize it. It is not strange at all that a sample jar 
should break occasionally. If the jar should contain a 
whole month’s milk the patron is deprived of his test for 
