446 REporT OF INSPECTION WoRK OF THB 
takes his milk to the factory and it contains, by test, a certain 
amount of fat. On this basis he receives a given price per hun- 
dred for his milk, this price being fixed by the returns from the 
butter sold. When he:meets his neighbor who patronizes an 
adjoining factory and whose milk tests the same as his and whose 
butter is sold at the same price, but who gets more per hundred 
for the milk, he condemns the test; when the trouble is not in 
the method but in the way it is handled. A competent, honest 
man with clean, correctly graduated glassware, will give uniform 
results and we must have that combination to make the Babcock 
fat test uniformly acceptable. 
Some states require the operators of the Babcock test to pass 
an examination to determine whether they have sufficient knowl- 
edge of its workings to make a correct test. This is a proper 
safeguard but it lacks in one particular, that it does not tell 
whether the applicant for a position is an honest man, which is 
quite as necessary as that he be intelligent enough to operate the 
machine. In order to have the work of the Babcock test per- 
fectly satisfactory it may be necessary for the State to have 
careful inspection made at factories and creameries to know that 
the work is done in an honest, careful way. 
The method followed at the Station in testing the bottles is 
as follows: A graduated burette, which has been carefully tested 
beforehand to insure its accuracy and uniformity at all points of 
the scale, is filled with cleaned, dried mercury. If the bottle to 
be tested has been used it is first thoroughly cleansed and dried; 
but this is omitted with new, clean bottles. The bottle is then 
placed under the burette and filled with mercury, first rapidly to 
the o mark, then slowly, with repeated comparison with the 
burette scale, to the top of the scale on the bottle. If the filling 
does not show any irregularity in the neck of the bottle, and if 
the variation is not over ;, of one per ct. in the length of the 
10 per ct. graduation of the bottle, it is passed as correct, as 
the variation in the ordinary sample of milk would be so small 
that it would be imposible to detect it. If the variation is 
of 1 per ct. or over, the bottle is rejected and destroyed. ‘The 
