266 Prof. Apjohn on the Specific Heats of the Gases, 



/" = /' - '^^^^P 



whereas, assigning to a its ordinary value '267, my expres- 

 sion, when brought to this shape, would be 



Now, the only way of accounting for this difference is by 

 supposing Dr. August to consider the value of a, the specific 

 heat of air, to be not '267 but 'S^9, an assumption which it is 

 scarcely necessary to say is altogether inconsistent with ex- 

 periment. Professor Graham observes of this formula, " It 

 was employed by Humboldt and G. Rose in their recent ex- 

 pedition to Siberia, and (as I was assured by the latter) with 

 excellent effect." This testimony would have startled me 

 much had it been stated that contemporaneous observations 

 were uniformly made with a condensation hygrometer. But 

 as this is not asserted, I feel the less reluctance in declaring 

 that the formula in question, however satisfactory it may seem 

 to the eminent philosophers just named, is undoubtedly er- 

 roneous. In an observation of theirs quoted by Professor 

 Graham, t being 74<°-9, t' 56°-9, and p 30-17, by the method 

 of August, f" is inferred to be •22678, whereas by mine it 

 would be "26596. Adopting the former value the dew-point 

 f would be 35°'6, while, if the latter be correct, it is 40°*1. 

 Now from the many and severe tests to which I have put my 

 own method, I do not hesitate to assert that the first-mentioned 

 determination of the point of deposition makes it at least four 

 degrees Fahrenheit too low. 



But to return from this digression to the subject of specific 



heats. From equation (A) or/" =/' ^ so ^^ 



readily deduce the following expression for the specific heat 

 of a gas, viz. 



{f'-f)"e 30 

 48 5 c/ p 



which, when the gas is perfectly dry, or, in other words, 

 wherey" = becomes 



f'e SO 



4t8s a p 



This is the specific heat of the gas under a given weight. 

 Hence as the specific heats of equal volumes of different gases 

 are equal to the specific heats of equal weights multiplied by the 



