He adopted a single fixed point, the freezing-point of water, a constant more difficult to determine accurately than the melting-point of ice; he ascertained that alcohol diluted with one-fifth water expanded from 1000 to 1080 volumes between the freezing- and boiling-points of water, and so he took zero for the lower and 80 for the higher temperature, dividing the intervening space into 80 parts.
Réaumur showed much ingenuity in his experimental work, but in his theoretical conclusions he made serious errors; he believed erroneously that he obtained the temperature of boiling water by immersing the unsealed alcohol tube in boiling water, and he ignored the influence of air-pressure, although Fahrenheit's experiments were generally known. For these and other reasons his thermometers were not satisfactory; with bulbs three to four inches in diameter the instruments were too large for standardizing smaller ones, and attempts to transfer Réaumur's scale to mercury thermometers led to confusion. Lambert, writing in 1779, gives for Réaumur three scales: first, that intended by him, second, the so-called Réaumur scale applied to mercury thermometers, and third, the scale actually obtained by
