MAGNUS ON THE EXPANSIVE FORCE OF STEAM. 221 
employed the same or nearly the same method ; they introduced 
some water into the empty space of a barometer-tube, and 
warmed it by surrounding it with water. Watt and Southern 
employed for this purpose a metallic vessel, which was heated 
by a lamp, but this vessel does not always appear to have sur- 
rounded the entire portion of the barometer-tube filled with va- 
pour; Dalton, on the contrary, fixed a wide glass tube by means 
of a couple of corks around the barometer-tube, and filled the 
intervening space with hot water. Biot however has already 
drawn attention to the fact, that in this method of Dalton’s the 
surrounding water can scarcely have possessed the same tempe- 
rature throughout its entire height, and has on that account ad- 
vised inserting several thermometers in it at various depths. 
Ure considered it more advantageous to avoid this evil by the 
employment of a siphon-shaped barometer, the open branch of 
which was tolerably long; by filling it with mercury he suc- 
ceeded in causing the vapour to occupy always the same space, 
even when their expansive force became greater. This space 
only occupied the length of half an inch in the apex of the ba- 
rometer-tube, and he fixed the elongated bulb of his thermome- 
ter close to this space. He mentions, as a peculiar advantage 
which is gained by this method, that the mercury of the baro- 
meter-tube does not become unequally heated. I must however 
confess that I cannot conceive how this mercury, where it is in 
contact with the vapour, should not acquire nearly the same 
temperature which it possesses, and below that of the sur- 
rounding atmosphere. But a correction for the expansion of 
the mercury is not possible as long as the temperature is not the 
same throughout the whole height. Ure indeed appears to have 
felt that this was absolutely requisite ; however, it is impossible 
to understand how he attained it in his method. I have conse- 
quently employed the following apparatus of measuring the ex- 
pansive force of steam. It is, it is true, less simple, but I found 
it impossible to satisfy in any other way the above-mentioned 
requisite condition. 
To measure the temperature, I preferred an air thermometer 
to the quicksilver thermometer, not merely on account of its 
greater sensitiveness, but especially because it admits of the em- 
ployment of any arbitrarily great volume of air. It is thus pos- 
sible to surround the space in which the vapour is contained 
almost entirely with this thermometric substance, and so to pre- 
VOL, IV, PART XIV, R 
