162 RESEARCHES ON THE VOLATILE HYDROCARBONS. 
constant temperature, the end of the mercurial column being in a state of motion. 
He states that a boiling-point taken in this manner may lie several degrees above 
that found with the thermometer bulb in the vapor. As bearing on this point, I pro- 
pose, a little farther on, to give the results of a few experiments and observations, 
which, with others of similar character, have induced me to depart from the now more 
common custom of taking boiling-points with the thermometer bulb in the vapor. 
Under normal conditions, the temperature of the boiling liquid and that of the 
vapor evolved should be the same. The only disturbing influence which appears to 
have been specially dwelt upon as likely to alter these conditions in the taking of 
boiling-points, is the liability of some liquids to adhere to the surface of the glass in 
such a manner as to produce abnormal elevation of temperature, generally attended 
with irregular ebullition, and consequent fluctuation of the thermometer. ‘To remedy 
this it is usual to introduce pieces of platinum ; iron filings, coal, etc., have also been 
employed. As above remarked, pieces of coke — or, when admissible, sodium — are 
found to be more surely effectual with hydrocarbons than platinum. Indeed, during 
more than three years of experience and careful observation upon a large number of 
hydrocarbons, I have not yet met with a single instance in which irregular ebullition 
and its consequent disturbing influence upon the boiling-point might not be completely 
prevented by these means. Although I cannot, of course, go so far as to say that 
equally satisfactory results would be obtained with other liquids by the use of coke, 
it is nevertheless my belief that in a majority of instances such would be the case. 
I have dwelt upon this point for the reason that the objections to the custom of 
taking boiling-points with the bulb in the vapor, appear to be even greater than those 
which Kopp has raised against the opposite course of placing the bulb in the liquid, 
as I shall proceed to show. It therefore becomes a matter of some importance that 
the objections to one or the other custom should be removed; and I think it will be 
found easier to overcome the objections to placing the bulb in the liquid, as I have 
done in the case of many hydrocarbons, even if coke shall not be found equally effi- 
cient with most other liquids. 
My experience has shown that, when irregular ebullition is effectually prevented, 
the temperature of the vapor from a boiling liquid is more liable to lead to an erro- 
neous determination of the boiling-point, than the temperature of the liquid itself. 
The reasons for this are, first, that the vapor is liable to become superheated by the 
hot air from the flame coming in contact with the sides of the retort above the 
surface of the liquid; second, that, with the bulb in the vapor, the thermometer is 
more liable to sudden depression from currents of cool air passing over the retort. 
