e
air within the bulb and tube expands or contracts. But in general, air
thermometers, however sensible to changes of temperature, are by no
means accurate in their indications.

I can, however, show you an air thermometer of a very peculiar
construction, which is remarkably well adapted for some chemical
experiments, as it is equally delicate and accurate in its indications.

CAROLINE.

It looks like a double thermometer reversed, the tube being bent, and
having a large bulb at each of its extremities. (PLATE II. Fig. 2.)

EMILY.

Why do you call it an air thermometer; the tube contains a coloured
liquid?

MRS. B.

But observe that the bulbs are filled with air, the liquid being
confined to a portion of the tube, and answering only the purpose of
showing, by its motion in the tube, the comparative dilatation or
contraction of the air within the bulbs, which afford an indication of
their relative temperature. Thus if you heat the bulb A, by the warmth
of your hand, the fluid will rise towards the bulb B, and the contrary
will happen if you reverse the experiment.

But if, on the contrary, both tubes are of the same temperature, as is
the case now, the coloured liquid, suffering an equal pressure on each
side, no change of level takes place.

CAROLINE.

This instrument appears, indeed, uncommonly delicate. The fluid is set
in motion by the mere approach of my hand.

MRS. B.

You must observe, however, that this thermometer cannot indicate the
temperature of any particular body, or of the medium in which it is
immersed; it serves only to point out the _difference_ of temperature
between the two bulbs, when placed under different circumstances. For
this reason it has been called _differential_ thermometer. You will see
by-and-bye to what particular purposes this instrument applies.

EMILY.

But do common thermometers indicate the exact quantity of caloric
contained either in the atmosphere, or in any body with which they are
in contact?

MRS. B.

No: first, because there are other modifications of caloric which do not
affect the thermometer; and, secondly, because the temperature of a
body, as indicated by the thermometer, is only relative. When, for
instance, the thermometer remains stationary at the freezing point, we
know that the atmosphere (or medium in which it is placed, whatever it
may be) is as cold as freezing water; and when it stands at the boiling
point, we know that this medium is as h