976 On the Power of Fluids to conduct Heat. 
immersed it into the cold mixture, when it rose, 
at first slowly, but increasing in velocity, it passed 
60°, 70°, and was going up towards 80°, when 
I took it out to see if there was any ice in the 
bulb, but it remained perfectly transparent: I 
immersed it again and raised it to 75° per scales 
when in an instant it darted up to 128°, and that 
moment taking it out, the bulb appeared 
white and opake, the water within being frozen: 
Fortunately it was not burst; and the liquid 
which was raised thus to the top of the scale was 
not thrown out, though the tube was unsealed. 
Upon applying the hand, the ice was melted and 
the liquid resumed its station. This experiment 
was repeated and varied, at the expence of several 
thermometer bulbs, and it appeared that water 
may be cooled down in such circumstances, not 
only to 21°, but to 5° or 6°, without freezing, 
and that the law of expansion abovementioned 
obtains in every part of the scale from 42° to 
10° or below ; so that the density of water at 10* 
is equal to the density at 75°. But as the dis 
covery of this curious, and I believe hitherto un- 
noticed property, has little to do with the object 
before us, I shall say no more of it at present. 
Count Rumford’s principal experiments are 
those in which a cake of ice was confined on the 
bottom of a cylindrical glass jar, of 4.7 inches 
in diameter, and 14 high, and water poured upon 
