xxiv Historical Sketch of the Physical Sciences, 1818. 
but we are afraid from the momentary nature of the temperature 
thus evolved, and the insignificant weight of air when compared 
with that of other bodies, that it would produce but very little 
effect on even the most delicate thermometer.—(Ann. de Chim, 
et Phys. ix. 305.) 
7. Melting Point of Bismuth, Tin, and Lead.—The melting 
point of every particular substance has been hitherto considered 
as perfectly fixed, when other circumstances remain the same. 
Sufficient attention has not been paid to a fact which has been 
very o.ten observed in water ;~but which appears not to be pecu- 
liar to that liquid. Water may be cooled down a good many 
degrees below the freezing point without congealing. I have 
sunk it nearly to zero by cooling it in thermometer tubes ; but 
the instant it begins to congeal, it starts up to the temperature 
of 32°. In like manner, as we learn from the experiments of Mr. 
Crichton, bismuth always sinks 8° below its fusing point before 
it begins to congeal; but the instant the congelation begins, it 
starts up to its true fusing point.* Tin always sinks 4° below 
its congealing point, and starts up to it again the instant it 
begins to congeal. At the congealing point the thermometer 
remains long stationary, indicating that the congelation is going 
on slowly and regularly. Lead, on the other hand, does not sink 
at all sensibly below its point of congelation (Annals of Philoso- 
phy, xii. 224). I consider this curious circumstance to depend 
upon the latent heat of these bodies. The subject is entitled 
to a much fuller investigation than it has yet received. 
8. Bowling Point of Liquids—It has been long known that 
when water is heated in a glass vessel it boils much less equably 
than it does in a metallic vessel. The temperature is raised a 
degree or two above the regular boiling point; then a torrent of 
steam rushes up through it, and the temperature sinks a little. 
This is repeated during the whole continuance of the process, 
and the temperature continues always vibrating between two 
points, distant two or three degrees from each other. If a few 
slips of platinum wire, or indeed of any other metal, are put into 
the glass vessel, these vibrations are prevented, and the water 
boils regularly, as it does in a metallic vessel when it has reached 
the boilmg pomt. The cause of this difference has not yet been 
accounted for in a satisfactory manner. We owe some late inge- 
nious speculations on the subject to M. Gay-Lussac, which I 
think it unnecessary to repeat here, as.they have been inserted 
in the Annals of Philosophy, xii. 131, to which the reader is 
referred. 
9. Lamp without Flame.—The only other set of facts con 
* Chemists are not agreed about the true melting point of this metal. Mr 
Crichton, of Glasgow, fixes it at 476°, which I am disposed to adopt from the 
known precision of this excellent artist. Berzelius makes the melting point 4513°; 
Thenard, 492°: and Gay-Lussac, 541° (Ann, de Chim. et Phys. ix, 308). This last 
Sumber must be very erroneous.”~ °° 
