
are oe. 
IMPORTANT NEW EXPERIMENTS 143 
Some little practice is required to do this properly 
—that is, on the one hand to seal the vessel while 
se a there 
~ a is at 
= gentle outpouring 
of steam, and, on 
the other hand, to 
do it in such a way that there 
is no inbending of glass at the 
sealed extremity. Even a small 
amount of such inbending is 
very apt to lead to a minute 
crack at the next stage. After 
allowing an interval of }? of a 
minute for the sealed tip to cool 


Fic. 6. a little, the retort is inverted, 
Sealed’ Reiort containing | and im this position is at ence 
liquor-potassze tube. immersed in a can of boiling 
water, prepared and ready to hand for this purpose. 
Here the experimental vessel is left for 8 minutes 
or more. 
Three purposes are served by this double process 
of heating. In the first place, it simplifies the ex- 
perimental conditions to get rid of the air by boiling ; 
secondly, the speedy closure of the vessel and the 
prolongation of the heating in a can of boiling water 
reduces the loss of fluid by boiling to a minimum; 
and thirdly, and principally, the inversion of the 
experimental vessel during the second period of 
heating brings those upper portions of the internal 
surface, as well as the outer surface, of the liquor- 
potassee tube (which, during the boiling over the 
