and the Mode of its Communication. 107 
body, in order that the water with which it is occasionally filled 
may not run out of it, when the body of the vessel is laid down 
in a horizontal position, in the manner represented in the above- 
mentioned figure. 
A thermometer, with a cylindrical bulb 4 inches in length, 
being inserted into the body of this vessel, through its neck, 
shows the temperature of the contained water. 
Care is necessary, in constructing a thermoscope, to choose a 
tube of a proper diameter : if its bore be too small, it will be 
found very difficult to keep the spirit of wine in one mass ; and, 
if it be too large, the little horizontal column it forms, (which 
I have called a bubble,) will be ill defined at its two ends, 
which will render it difficult to ascertain its precise situation. 
After a number of trials, I have found, that a tube, the bore of 
which is of such a size that 1 inch of it in length contains about 
15 or 18 grains Troy of mercu^, answers best. For a tube of 
that size, the balls may be about \\ inch in diameter; and they 
should both be painted black, with Indian ink, which renders 
the instrument more sensible. 
I have an instrument of this kind, the tube of which is quite 
filled with spirit of wine, excepting only the space occupied by 
a small bubble of air, which is introduced into the middle of the 
horizontal part of the tube ; but it does not answer so well as 
those which contain only a very small quantity of that liquid, 
sufficient to form a small bubble. 
But, without enlarging any farther, at present, on the con- 
struction of these instruments, I now proceed to give an account 
of the experiments for which they were contrived. 
Having found abundant reason to conclude, from the results 
of the experiments of which an account has already been given, 
P 2 
