^{page 235 note *} Transactions of the Royal Society of Edinburgh, vol. xv., p. 411.

^{page 235 note †} Annales de Chimie, 3^me Serie, vol. xi., p. 332.

^{page 238 note *} The reason for this change in the index error, according to the two hypotheses, will be seen in the concluding paragraph of this paper.

^{page 238 note †} I still retain the doubt expressed in my former paper, as to whether the boiling point can be taken correctly to represent the temperature of steam whose elasticity is that of the atmospheric pressure at the time. This doubt is confirmed by the difference of M. Regnault's and Magnus's Tables of Elasticity, as also by experiments of a different kind. I take this opportunity of adding, that I have obtained true ebullition of water in an exhausted receiver at the low temperature of 46°; the syphon-gauge then stood at 0·25 inch, being ·06 below the elasticity of vapour, at that temperature, as given by M. Regnault.

^{page 238 note ‡} The superabundance of heating power, and the mass of liquid in ebullition, I consider very important to the good result. No other portable apparatus that I am aware of, gives so ready means of adapting the force of the flame to the circumstances of the case. With a common spirit-lamp in fixed position below a boiler, it is next to impossible to regulate the rate of boiling, especially in an exposed situation. Mine is also the only instrument, so far as I know, which can be used in a gale of wind.

^{page 241 note *} A few points marked by the letter r, calculated from Regnault's formula, are inserted in the figure for the sake of comparison.

^{page 241 note †} The entire series of Dr Hooker's observations is best represented by 548 feet for 1°, when we include the (somewhat doubtful) highest observations. This agrees almost exactly with my earlier determination.

^{page 242 note *} In the calculation of De Saussure's observation on Mont Blanc, in my former paper (Ed. Trans., vol. xv., p. 414), a slight mistake occurs. The depression of the boiling point should be 26°·81, instead of 26°·71, giving 545·9 feet for 1°, agreeing strikingly with the other results.

^{page 242 note †} It happens by a double fortunate coincidence, that the coefficient of T², is equivalent to unity in both cases, when the proper reductions are made.

^{page 243 note *} If we aim at representing M. Regnault's Table only, between the temperatures 212° and 192°, the coefficient should be only 535. The difference in the kind of glass used in constructing thermometers would alone account for the variation. The reason of the different coefficient in the formulæ (A) and (B), arises from the different inclination of the tangent line PT, and the intersecting line p q, in Plate III., fig. 3, as explained in the succeeding paragraph of the text.