832 
ME. F. CEACE-CALVEET AKD ME. E. JOHXSOX OX THE 
one quarter of an hour by a small jet of steam brought into it (for further details see 
the first part of these researches, page 350 of the Philosophical Transactions for 1858), 
when we obtained the following results : — 
Temperature of 
the 75 cub. cent, 
at beginning of 
experiment. 
Temperature of 
75 cub. cent. 
at end of 
15 minutes. 
Conductibility 
found reduced 
to 50 cub. cent, 
water. 
Mean. 
I 
SUver 1000. | 
Mercury vertical... 
14-8 
12-6 
. 16-0 
13*7 
1-801 
1-65 
1-7 
! 
i 
54 1 
We also tilted the apparatus and gave it gradually different angles, and the con- 
ductibility of heat by mercury gradually decreased as the angle increased, showing the 
following results : — 
At a slight angle .... 
13-5 
Silver 1000. 
423 
Angle slightly increased . 
7-3 
229 
Angle still more increased . 
6-9 
216 
Considerable angle 
5-1 
160 
Lastly, the apparatus was used as in our former experiments, the little box being 
placed in a horizontal position, and the results agreed with those ah’eady published; for 
we obtained 
’6 1 Mean. Silver 1000. 
21-8 I 21-63 or 679 
21-6 3 
There cannot therefore be a doubt that the supposed good conductibility of heat by 
mercury arose from not taking into account that mercury being a fluid, its facility to 
conduct heat was owing to currents. The same may be said of water; for we have 
observed, as is already known, that it presents a complete barrier to the conduction of 
heat when the source of heat is applied at the upper part of a column of water. 
Thus in our experiments we have found that the temperature of the water in the 
lower vessel did not rise one-tenth of a degree during the quarter of an hour that the 
water in the upper vessel was maintained at 90° C. 
The bad conductibility of heat by fluids when all currents are prevented in their mass, 
appears to us difficult to explain by the theories of undulation or radiation ; for we cannot 
understand why the imponderable fluid caloric should not travel equally well between 
the molecules in whatever way the source of heat is applied, or why the undulations 
should not be as rapid, nay, more rapid in a fluid than in a solid. All these difficulties 
disappear if we adopt the views of Mr. J. P. Joule, F.E.S., which are, that heat is 
conveyed in bodies by the vibrations of the solid molecules composing them. 
The remarkably low conducting power of mercury presents another point of interest, 
as it establishes a further analogy between heat and electricity. The ratio of conducti- 
bility of these two agents by mercury, as compared with that of silver, shows such close 
