“New YorK AGRICULTURAL EXPERIMENT STATION. 83 
In these tests where the strong inrush of air at the close of 
the milking was prevented, the resulting germ contents bear 
no distinct relation to the location of suction filter. Under 
these circumstances the thin layer of cotton in the suction 
filter sufficed for all the duty laid upon it and with the cotton 
in place the results are both very low and quite constant. 
The highest recorded germ content in 34 tests with the cotton 
filters is 16,230 per cc. with an average of 5,840 per cc. Where 
the cotton filters were not used the average of 36 determina- 
tions was 18,060 per cc. with a maximum of 77,500 per ce. 
It is interesting to observe that on June 28 when the teat 
cups were accidentally dropped on the floor the milk contained 
so large a germ content that the colonies on the plates were 
too numerous to count, 
Modification of the filter cwps.— Up to July, 1907, the Bur- 
rell-Lawrence-Kennedy machine was used without any changes 
in its manner of construction. At that date the filter cups were 
replaced by new ones which were much larger, much easier to 
clean and more easily packed with cotton. Plate V, fig 2, 
and Plate VI, fig. 2, show the increased size of these filters 
over the ones first used. The main suction filter on the 
pulsator base between the pail and the nipple for the stanchion 
hose (Plate V, fig 2) is approximately 11 times larger than 
the one originally furnished with the machine (Plate V, fig. 1). 
The relief filters on the pulsator head and those on the teat 
cup connectors as shown in Plate VI, figs. 2 and 3, are approxi- 
mately 9 times larger than those previously tested. 
Our results given in Table V show clearly that the former 
suction filter was inefficient and to obtain milk with a small 
germ content it was desirable to release the vacuum at the close 
of the milking process by allowing the air to flow in slowly 
through the teat cups. This was undesirable in practice be- 
cause of the time lost in the process and also the fact that 
unfiltered air with its accompanying quota of bacteria was 
being admitted to the pail. The efficiency of the new suction 
filter was tested by breaking the vacuum on alternate nights in 
the two different ways. In all of these tests the teat cups and 
