1895.] The Latent Heat of Evaporation of Water. 223 



of course, can only be cleared up by a direct determination of the 

 capacity for heat of water over the range to 100. 



In Paper L, I give further evidence in support of the equality of 

 the two units. 



Considering the extreme attention which has been given during 

 the last few years to the determination of the electrical units, it is 

 strange that so little has been done regarding what is perhaps as 

 important a unit, viz., that of heat. I venture to appeal to the Royal 

 Society to take steps to place our knowledge in this respect 011 a 

 firmer basis, and I would go so far as to express my belief that the 

 method of measuring small differences of temperature described in 

 Paper L, and also in the communication printed in the ' Philosophical 

 Magazine ' of this month, points out a way to the solution of some of 

 the difficulties. 



Section III. 

 The density of water- vapour at different pressures can be obtained 



from the thermodynamic equation L = - (s s) 2. In Paper L, I 



J diL 



have given the density at different pressures thus obtained by the 

 substitution of my values of L and J. 



We can also obtain what Winkelmann terms the "theoretical 

 density " by assuming that water-vapour behaves as a perfect gas, 

 having the same molecular weight. I have shown that if we take the 

 most recent determinations of the atomic weight of oxygen,* the 

 " theoretical density " of water- vapour at low pressures is almost 

 identical with the density as deduced from the thermodynamic 

 equation. At higher pressures (above 140 mm.) the density appears 

 to remain nearly constant, and is about T02 times as great as the 

 " theoretical density." 



These conclusions are confirmed by a study of the " volume energy " 

 of water-vapour at different temperatures. 



Conclusion. 



The results obtained by Dieterici at C., by Regnault at tempera- 

 tures 63 to 100 C., and by myself at intermediate temperatures, are 

 represented with great accuracy by the formula 



L = 59673- 0-6010 0. 

 * Scott, 'Phil. Trans.,' A, 1893, p. 507. 



