252 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 51 



For the value of this quantity we obtain 



c v = 0.1685 + 0.000175 y' 



by considerations quite analogous to those that Hann has adopted 

 in his determination of the value of c p . 



At the temperatures with which we have to do y' seldom exceeds 

 the value 10 and then only slightly, and since r varies about the 

 value 600 therefore the quantity 



cot a„ = 



1000 c, 



can be considered as constant and without incurring important 

 error can be taken as 



cot a v = 3.5, 



a value that is generally a little too small, but is nearer the truth than 

 3.6, which is considerably too large. 



For temperatures below o°C, we must use, instead of a v , a value 

 a* v determined by the equation 



cot a* v = — — _ = 4.0 

 1000c„ 



where / is the latent heat of melting ice. 



If we introduce the angle a v into the construction of fig. 30, 

 or its equivalent value into the formula if we prefer numerical 

 computation, then we can utilize the same method as above 

 described. 



In this method T 2 is the end of the abscissa belonging to the 

 temperature t 2 and y l — y 2 = F, / is the quantity of precipitated 

 water. 



We see from this that only a part of the excess of vapor over the 

 amount required for saturation is precipitated while the remainder 

 remains vapor in spite of the discontinuance of supersaturation, 

 since the air now needs a greater quantity of water for its satura- 

 tion in consequence of the rise in temperature. 



But a rise in pressure accompanies this rise in temperature because 

 of the unchanged volume, at least during the first instant, which 

 is given by the equation 



ft _ 273 - t 2 

 ft 273 - t t 



Zeit. d. Oest. Gesell. fur Met., 1874, p. 374. 



