222 MAGNUS ON THE EXPANSIVE FORCE OF STEAM. 
serve the temperature of this space itself more certain. The air — 
thermometer is constructed exactly like the one I employed in 
my investigations on the gases and the comparison between the 
expansion of the air and of mercury. It is figured in Plate III. 
fig. 1, at D E, The tubes, however, which contained the air 
were of a different form and of greater capacity ; their form is 
seen in fig. 2, where they are designated by X Y Z. 
If the thermometer and the steam are to have exactly the same 
temperature, it is requisite that not only both be exposed to one 
and the same temperature, but that this remain constant for 
some time, otherwise there is no certainty that both the steam 
and the thermometer possess the same temperature at the mo- 
ment of the observation, in case, for instance, the two do not 
become heated or cooled with the same rapidity. 1 employed on 
that account for the production of a constant temperature the 
same apparatus which I used in the comparison of the expansion 
of the air and of the mercury at high temperatures*. It is re- 
presented in figs. 1 and 2, designated with S P R, and consists 
of a case constructed of sheet-iron, surrounded by three other 
perfectly similar cases, so that between each two there exists a 
stratum of air 4ths of an inch thick above as well as below and 
on every side. The cases are suspended in one another to avoid 
all metallic connexion in the lower portions. The outer case 
was heated by Argand’s spirit lamps, of which I required only 
two in these experiments. When they burn with a moderate 
flame they constantly yield the same heat, and in this manner 
the air in the inner cases is kept at an invariable temperature. 
M. Regnault has thrown out some suspicions as to the fitness 
of this arrangement+. He compared at the same time with me 
the expansion of air and of mercury at high temperatures, but 
arrived at different results to those which I obtained. He ex- 
plains this difference by the unequal expansion of the glass, 
which he found might exert a very considerable influence {. He 
also throws out suspicions as to the correctness of my experi- 
* Poggendorfl’s Annalen, vol. lvii. p. 177. 
t+ Annales de Chim. et de Phys., 3rd series, vol. vi. p. 3870. 
y I cannot omit to notice in this place a blunder which oceurs in an investi- 
gation which has the same object, viz. the comparison of quicksilver thermome- 
ters made of different glass, and which was executed by M. Pierre under the 
direction of M. Regnault in the laboratory of the Collége de France. M. Pierre 
has made the bulb of his thermometer of a different kind of glass from that of 
the tube (nn. de Chim. et de Phys., 3rd series, vol. v. p. 428), to which atten- 
tion has already been drawn by Prof. Poggendorff in his translation of this 
