ery particle is successively heated at the
bottom, and cooled at the surface of the liquid; but as the fire
communicates heat more rapidly than the atmosphere cools the succession
of surfaces, the whole of the liquid in time becomes heated.

CAROLINE.

This accounts most ingeniously for the propagation of heat upwards. But
suppose you were to heat the upper surface of a liquid, the particles
being specifically lighter than those below, could not descend: how
therefore would the heat be communicated downwards?

MRS. B.

If there were no agitation to force the heated surface downwards, Count
Rumford assures us that the heat would not descend. In proof of this he
succeeded in making the upper surface of a vessel of water boil and
evaporate, while a cake of ice remained frozen at the bottom.

CAROLINE.

That is very extraordinary indeed!

MRS. B.

It appears so, because we are not accustomed to heat liquids by their
upper surface; but you will understand this theory better if I show you
the internal motion that takes place in liquids when they experience a
change of temperature. The motion of the liquid itself is indeed
invisible from the extreme minuteness of its particles; but if you mix
with it any coloured dust, or powder, of nearly the same specific
gravity as the liquid, you may judge of the internal motion of the
latter by that of the coloured dust it contains. --Do you see the small
pieces of amber moving about in the liquid contained in this phial?

CAROLINE.

Yes, perfectly.

MRS. B.

We shall now immerse the phial in a glass of hot water, and the motion
of the liquid will be shown, by that which it communicates to the amber.

EMILY.

I see two currents, the one rising along the sides of the phial, the
other descending in the centre: but I do not understand the reason of
this.

MRS. B.

The hot water communicates its caloric, through the medium of the phial,
to the particles of the fluid nearest to the glass; these dilate and
ascend laterally to the surface, where, in parting with their heat, they
are condensed, and in descending, form the central current.

CAROLINE.

This is indeed a very clear and satisfactory experiment; but how much
slower the currents now move than they did at first?

MRS. B.

It is because the circulation of particles has nearly produced an
equilibrium of temperature between the liquid in the glass and that in
the phial.

CAROLINE.

But these communicate late