C8 Royal Society : — 



and to eliminate the effect of the cohesion of the glass on the water, 

 as explained in our previous paper on the density of steam. 



Twelve cuhic inches of mercury were measured into the globe, and 

 a file-mark made on the stem, below which, at a distance of 14*45 

 inches, another file-mark was made, affording a fixed point for ascer- 

 taining the correspondence of the upper file-mark with the readings 

 on the cathetometer. 



Let a be the reading on the fixed rod of the level of the column, 

 h the reading of the lower file-mark on the globe-stem ; then b—a 

 = the height of the column of mercury on the globe-stem. 



To correct for temperature, 1\ inches of mercury, enclosed by the 

 oil-bath and its stuffing-box, were corrected for the temperature of 

 the oil, and the remainder of the column for the temperature of the 

 atmosphere at the time. By deducting the column so corrected from 

 the reading of the barometer at the time, the total pressure in the 

 globe is obtained. The readings of the thermometer are corrected 

 for the portion out of the oil-bath. The pressure of mercurial vapour 

 is calculated from data supplied with great courtesy by M. Regnault, 

 and embodying the results of unpublished experiments. The pres- 

 sure of this vapour is assumed to be the same as that in a vacuum, 

 as the vapour in the globe remains still for a sufficient time (it is 

 believed) for saturation to take place. In this view we have been 

 strengthened by M. Regnault's opinion. By deducting the pressure 

 of mercury vapour from the total pressure in the globe, the pressure 

 of the steam is obtained. 



On referring to the experiments contained in the paper, it will be 

 seen that the law of expansion of gaseous bodies is expressed by the 

 formula 



E+t t PV/ ' ' I\Vi- PV ' 



where E is a constant. Taking Regnault's constant 459 as the rate 

 of expansion of air for constant volumes, a remarkable coincidence 

 will be observed in the experiments contained in the paper when 

 reduced to the same standard of value. The values of E thus 

 deduced have been placed in the last column of the calculated experi- 

 ments. They show a decreasing rate of expansion from the satura- 

 tion-point upwards, until at no great increase of temperature the rate 

 of expansion coincides with that of a perfect gas. 



Taking from the Tables the two results, which in each instance 

 represent the case of expansion at the greatest distance from the 

 saturation-point, we have the following values of E : — 



E = (1) 474-48 (3) 466*85 (5) 460*28 



(1) 474*48 



(3) 466*85 



450*11 



451-94 



(2) 455-57 



(4) 464*83 



443-S6 



460*49 



Mean value of E deduced from these numbers =458*74. 



Hence the conclusion which we suggested in our previous paper 



