50 Mr. P. J. Hartog and Dr. J. A. Harker on 



the amount of heat required to raise one gramme of water 

 through i° of the hard-glass mercury thermometer in the 

 neighbourhood of 1 5° C. 



It will be observed that our results differ from those of 

 Regnault by more than 2 per cent. We have failed so far 

 to discover any explanation of this difference, either by 

 ascertaining an appreciable error in our own work or in that 

 of the great French physicist. Professor Osborne Reynolds 

 has suggested that radiation from the inner surface of the 

 central glass tube must cause a certain amount of conden- 

 sation on its surface, and that the water so condensed 

 would run down into the worm, and thereby cause an error 

 of calculation possibly sufficient to account for the difference 

 between our results and those of Regnault. The heat 

 radiated into a vacuum by a square centimetre of glass at 

 ioo° has been determined by Gratz (Wied. Ann., Vol. XL, 

 p. 913), and the absorption by glass of heat radiated from a 

 Leslie cube heated to ioo° by Melloni (quoted by Wiillner, 

 Lehrb. d. Experimental Physik, Vol. III., p. 197), so that we 

 can form an estimate of the loss due to this source, and 

 calculation shows that the loss falls in all probability well 

 within the error of experiment. This conclusion is borne 

 out by the fact that the amount of water condensed by 

 radiation must be proportional to the duration of the experi- 

 ment ; while a glance at the table shews that, other things 

 being approximately the same (see e.g. Expts. III. and IV.), 

 the results of experiments, calculated without taking account 

 of this correction, are independent of duration. We are, 

 however, obliged to Professor Reynolds for his criticism, 

 and shall meet it by making use of metal vessels in the 

 experiments which are in progress. 



It may here be pointed out that the determinations of 

 Regnault for the latent heat of steam at o° have not been 

 confirmed by subsequent observers [see Winkelmann 

 {Wied. Ann., Vol. IX., p. 208, 1880) and Dieterici (ibid, Vol. 



