370 



Mr. J. Joly. 



The correction on w 1 is then — 



-Vx 0-000636 gram, 



where V is the volume of the substance in cubic centimetres. 



(2.) If extreme accuracy be desired, it is necessary further to 

 reduce the weight of condensation to vacuo. Calling w 2 the weight 

 deduced after correction (1), the second correction is on w 2 , and 

 is — 



+ w 2 x 0-000589, 



the factor 0'000589 being the absolute density of saturated steam at 

 100° (Regnault). It is assumed that we may for the purpose consider 

 unit weight of water at 100° as equal to unit volume. 

 Finally the true weight of water w is very nearly — 



w=(w 1 - V X 0-000636) x 1-000589. 



Regarding the factor 0*000636 I may mention that a direct experiment 

 made in the calorimeter afforded 0'00062 as the difference in weight 

 of unit volume of steam at 100*2° and air at 17" 7° and 766 mm. The 

 experiment consisted in hanging a heated sphere of blown glass in 

 the calorimeter — it having been previously counterpoised — and 

 admitting steam before the sphere had cooled to steam temperature. 

 When a sufficient time had elapsed to secure equilibrium of tempera- 

 ture the amount necessary to add to the counterpoise to restore 

 balance was ascertained ; this divided by the volume of the sphere 

 afforded the relative density sought.* In the experiment the volume 

 of the sphere was 38*7 c.c, the weight added - 024 gram. 



On the Value of X. 



If the value of \ be calculated from Regnault's formula for, say, 

 the cases of a barometer at 720 and 780, the values 537'2 and 536'0 

 are obtained; at 760 it is 536*5. In careful experiment this variation 

 must be considered. My practice has been to allow for it by the 

 approximate assumption for ordinary variations of pressure that the 

 correction on the value of X for the standard pressure 760 mm. 



.+07(100-* 8 °), 



that is I take 



X = 536-5 +0-7(100-^°) 



where t 2 ° is the temperature of the steam, read directly by a thermo- 

 meter or deduced — as it conveniently may be and with great accuracy 

 . — from the height of the barometer and reference to Regnault's tables 

 for the corresponding boiling temperature of water. 



# See " On some methods of Measuring the Densities of G-ases," by G. F. Fitz- 

 gerald, F.R.S., ' Roy. Dublin Soc. Proc.,' vol. 4, 1885, p. 481. 



