remove some of the atmospheric pressure, the water should boil with a
less heat than will cause the mercury in the thermometer to rise to 100°
C., and if we take off all the pressure, the water ought to boil and
freeze at the same time. This actually happens in the Carré ice-making
machine. The question now arises, "Why does the water freeze in the
Carré machine?" All substances require certain amounts of heat to enable
them to take and to maintain the liquid state if they are ordinarily
solid, and the gaseous state if ordinarily liquid or solid, and the
greater the change of state the greater the heat needed. Moreover, this
heat does not make them warm, it is simply absorbed or swallowed up, and
becomes latent, and is merely necessary to maintain the new condition
assumed. In the case of the Carré machine, liquid water is, by removal
of the atmospheric pressure, coerced, as it were, to take the gaseous
form. But to do so it needs to absorb the requisite amount of heat to
aid it in taking that form, and this heat it must take up from all
surrounding warm objects. It absorbs quickly all it can get out of
itself as liquid water, out of the glass vessel containing it, and from
the surrounding air. But the process of gasification with ebullition
goes on so quickly that the temperature of the water thus robbed of heat
quickly falls to 0° C., and the remaining water freezes. Thus, then, by
pumping out the air from a vessel, _i.e._ working in a vacuum, we can
boil a liquid in such exhausted vessel far below its ordinary boiling
temperature in the open air. This fact is of the utmost industrial
importance. But touching this question of latent heat, you may ask me
for my proof that there is latent heat, and a large amount of it, in a
substance that feels perfectly cold. I have told you that a gasified
liquid, or a liquefied solid, or most of all a gasified solid, contains
such heat, and if reconverted into liquid and solid forms respectively,
that heat is evolved, or becomes sensible heat, and then it can be
decidedly felt and indicated by the thermometer. Take the case of a