644 
ON THE TESTING OF PETROLEUM SPIRIT. 
the surface of the liquid, the other being only half an inch below the surface, 
a difference of several degrees will be noticed between them at the time the va¬ 
pours will flash ; and, I am happy to say, the following results confirm Mr. 
Tate’s interesting observations :— 
No. Flashed at Flashed at 
4 94° F., inches. 99° F., \ an inch. 
5 94° „ 98° " „ 
6 95° „ 99° 
This curious and unusual fact of a fluid having a much higher temperature 
near the surface than it has an inch below what may be considered the centre 
of the bulk of the fluid, is due, in my opinion, to this,—that petroleum, not 
being a homogeneous liquid, but a mixture of several hydrocarbons, the lightest 
products being first expelled, they rise and carry the heat towards the surface, 
and, in this way, the difference in temperature referred to is produced. It is 
with a view of overcoming this practical difficulty that a series of experiments 
were instituted, in which the operations were conducted in the usual way, with 
this exception, that the liquid was kept in a constant state of agitation (except 
at the time when a flame was passed over to observe the flashing-point) by the 
thermometer, and the results obtained were as follows :— 
No. 1 did not flash at 102° F. 
2 „ „ 99° 
3 „ „ 98° 
5 „ „ 98° 
6 „ „ 104° 
n 
These experiments appear to me to confirm the explanation above given as 
to the cause which produced the difference of the two thermometers, placed at 
unequal depths in the fluids under examination ; for it will be observed that 
the flashing-points of the hydrocarbons are raised several degrees ; in my 
opinion this fact is due to the agitation having facilitated the gradual escape of 
the most volatile products, the flash will not occur until a sufficient quantity of 
the more dense vapours have been volatilized and collected on the surface of the 
liquid to be tested. 
I believe many of the anomalies above-described are principally owing to the 
difficulty, notwithstanding any amount of care that may be bestowed on the 
operation, in raising the temperature of the petroleum spirit from natural tem¬ 
peratures in a certain time, say fifteen, twenty, or thirty minutes, to their 
flashing-points ; and as the true flashing-points of the fluids depend entirely on 
the time employed in raising their temperature, I would propose the following 
method, which will enable every operator, in any part of the United Kingdom, 
to determine the flashing-point of a petroleum with certainty, and feel satisfied 
that another manipulator will obtain identical results. 
The process consists in heating the water of the outer vessel 10 degrees above 
the flashing-point of the spirit (approximately tested) in the apparatus such as 
described in the Act, removing the flames, and then placing the can in the 
water, filling it at once carefully wdth the petroleum spirit. The thermometer 
should then be placed with the bulb half an inch below the surface of the 
spirit, and the flashing, point ascertained in the usual manner. 
The following are the results obtained with the same six samples of petro¬ 
leum employed in the previous experiments :— 
First experiment. 
No. 1 flashed at 
98° F. 
)> 2 ,, 
95° 
„ 3 „ 
96° 
»» 4 ,, 
96° 
»» 5 
96'5° 
Second experiment. 
99° F. 
96° 
97° 
97° 
97° 
