520 The Trials of Cream Separators at tfoncastert 
recorded so far iu an exhaustive trial. The worst result waS 
that from the " Windsor " machine, but it is only fair to add that 
at the time of trial there was a leak in this machine, and this 
may have had to do with the separation not being so perfect. 
As between the six other machines there was practically little to 
choose. It is noticeable that the new arrangement of discs 
employed in the case of the " Alpha Baby " and " Al^jlia Windsor " 
machines, while neai-ly doubling the outturn, did not in any way 
reduce the efficiency of separation ; also that, taking these two 
machines as representatives of the horizontal and vertical 
patterns, the separation was equally good in either case. The 
facility for regulating the thickness of the cream was observed 
in separate trials. In this respect the Danish machine, No. 
4294 (Burmeister & Wain) was decidedly the best, then came 
the " Victoria " and the two " Windsor " machines. The two 
" Baby " machines were not so good, and the " Alexandra " 
decidedly the worst. The only machine that produced cream of 
excessive frothiness was the " Victoria." 
It may be interesting to append, for the purpose of com- 
parison, the results, as regards efficiency of separation, obtained 
at some previous trials : — 
Percentage of Butter-fat left 
iu Skim-milk 
R.A.S.E., 
Bath and 
Bath and 
Windsor, 
West, Dor- 
West, New- 
1889 
chester, 1887 
port, 1888 
per cent. 
per cent. 
per cent. 
" Victoria " hand machine . . 
•17 
" Star " liand machine .... 
•15 
" Laval" horizontal hand separator 
•17 
" Laval " vertical hand separator . 
•14 
Shallow pan S3'stem (hand skimming) . 
•8G 
■io 
Cooley deep system (hand skinlming) . 
•50 
The results now obtained, it will be observed, are very con- 
siderably superior. » 
Power Maclt ines. — The trials took place on June 18 and June 
20. Much larger quantities of milk being recjuired, it was 
impossible to obtain sufficient of one quality to deal it out alike 
to each machine. Consequently, se]')arate analyses were made of 
the whole-milk in each case, as well as of the skim-milk. The 
time of running was from half-an-hour to three-quarters. The 
loss, as in the case of the hand machines, includes what was left 
in the drum or bowl, though, owing to the longer time of run-- 
ning, the results when stated per hour are more nearly the real 
ones than with the hand machines. There was a leak in No. 
