of the Specific Heat of Steam. 339 



relative des corrections produites par les causes perturbatrices." 

 This consideration is still more important when the range of 

 temperature is limited to 25° — only one fourth of the range 

 from which Eegnault deduced his result. The large calori- 

 meter was used onlv in the third and fourth series. 



First series. 



Second series. 



Third series. 



Fourth, series. 



■280 



•546 



■299 



•386 



•465 



•591 



•314 



•381 



•417 



•567 



•340 



•346 



•377 







•375 



•309 



•463 



Mean "567 





•427 









■411 



•463 



Mean -400 





•349 

 •413 







Mean "3853 



Preparatory trials. 



•405 

 •415 







Mean -3721 





That the method of calculation may be perfectly clear, I 

 give the particulars for the first experiment in the fourth series. 

 Superheating to 124°'81, each unit of vapour gave up 646'28 

 units of heat down to 0°. Steam at 100° gave up 636*70 

 down to 0°. 



646-28-636-70 



124-81—100 



= •386. 



The quantity of vapour in this experiment was 102-62 gnus. ; 

 the quantity in the first of the first series was only 8"957 grms. 

 The fourth series was made with an apparatus which was an 

 improvement on that used in the third series ; and, taking that 

 series only, adding the probable amount of moisture which 

 would remain at 100° temperature (say 1 per cent, on the 

 •385), the result is *389; this, on other grounds, I believe to 

 be nearly correct. If, however, neglecting the other grounds 

 for my opinion, we take the mean of the means of the third 

 and the fourth series, we get 



•3721 +-3853 _..,--_ 



~ — 'd/ol. 



This is what Eegnault might fairly have done. 



If we now calculate what Eegnault's experiments would 

 give as the kinetic-theory result, — we get the specific heat of 

 hvdrogen, at page 121, "mean = 3-4090/' and taking the 

 molecular weight of steam =17 - 96, we find the specific heat 



