160 RESEARCHES ON THE VOLATILE HYDROCARBONS. 
bons, viz. that it does not lose its virtue by use, so long as any of it remains; plati- 
num, on the contrary, being liable, especially if the liquid is not quite pure, to become 
after a while slightly coated, and its efficiency thus impaired.* 
Except for low temperatures, the retort rests on a piece of wire gauze laid over the 
ring of an iron lamp-stand, and is heated with a small gas-flame. When operating on 
liquids of low boiling-point, I have observed the liability of the thermometer to be 
considerably affected by the ascending current of hot air striking the sides of the 
retort above the level of the liquid, thus causing an elevation of several degrees of 
temperature. To prevent this, I proceed as follows. For low temperatures, and yet 
above the common temperature, I place upon the gauze on which the retort is to 
stand, a screen of felt or thick woollen paper, which has been provided with a hole in 
the centre about two inches in diameter. This screen extends several inches from the 
sides of the retort, and has been found effectual for the purpose. 
For temperatures below the common temperature, the retort is set in a water-bath 
containing ice-water, the temperature of the bath being gradually raised by means of 
a small gas-flame. 
As is customary, in order to ascertain the temperature by which to calculate the 
correction for the upper column of mercury, a thermometer is attached, by means of 
elastic bands, to the side of the thermometer in the retort; the bulb being placed, 
during ebullition, midway between the centre of the cork and the upper extremity of 
the mercurial column. And, as usual, a paper screen, closely fitting the thermometer, 
is placed across at the top of the cork to shield the upper column of mercury from the 
direct influence of the ascending heat. 
I have observed that it often requires considerable time — variable according to its 
length and the thickness of the glass spindle — for that part of the thermometer 
above the retort to acquire the highest temperature which the boiling liquid can 
communicate to it. During this time the thermometer evidently is not in a fit state 
for an observation. While this gradual change in the condition of the thermometer 
is taking place, it is desirable, for obvious reasons, that no vapors should escape from 
the retort. I therefore proceed as follows. The retort, the neck of which has pre- 
viously been wrapped with a wet cloth, is placed in such a position that the neck 
* For common use in fractioning, when not desirable to use sodium, I have found pieces of coke to be more 
effectual and much more durable than platinum. Not unlikely it would be found equally preferable to plati- 
num for general use in taking the boiling-points of liquids in which sodium could not be employed. It is cer- 
tain that nothing could operate better than coke for the nitro-compounds and alkaloids derived from benzole 
and its homologues. 
