graduated thermometer, the temperature of boiling water is found to be invariable. Not only does the thermometer immersed in water keep for any number of hours of boiling the height it had reached when the first bubbles came up, but it ascends to the same point every time it is placed in boiling water. If Amontons had added the proviso that the pressure of the atmosphere should be the same in all the experiments, which we know now is indispensable, he would have been rigorously exact.
When we take a bulb of air connected with a manometer, mark carefully the pressure which it sustains when it is plunged into boiling water, and then the pressure at which, under other circumstances, it reaches the same volume, the ratio of that pressure to the former may be regarded as expressing the ratio between the temperature to which the air was raised under the latter condition to the fixed temperature of boiling water. This ratio will be the same, whatever thermometer, constructed in the same way, we may use. In this way we have a sure means of obtaining instruments that can be compared with one another.
Amontons proposed for a thermometer, as Drebbel did, a mass of air maintained at a constant volume under a variable pressure. The rule by which he attached a certain degree of temperature to each degree of heat and cold, or a larger number for more intense heat and a smaller for cold, is the same rule to which Desormes and Clément on the one hand, and Laplace on the other, returned a century afterward; and is the rule proposed in the works of Sadi Carnot, Clausius, and Lord Kelvin as the measure of the absolute temperature.
The profound reasons which cause us to prefer the definition of temperature proposed by Amontons to every other could not be divined at the beginning of the eighteenth century. The large size and inconvenient shape of Amontons's instrument, and the necessity of taking account of the variations of atmospheric pressure in interpreting its indications, prevented its general adoption; and the Florence thermometer was still preferred. Spirit thermometers, that could be compared with one another, were in demand. Réaumur furnished them.
Réaumur observed, in 1730, that a thermometer placed in freezing water went down to a certain degree, and remained fixed there as long as the water was not wholly solidified. The temperature of water in process of congelation was therefore always the same, and fixed. As physics has advanced, some corrections have been made in this law, and causes have been discovered that make the point of congelation of water vary; and physicists have been induced, in view of it, to take as their fixed temperature, instead of the freez-
