128 MB. W. N. SHAW OX HYGBOMETBIC METHODS. 



applying each method. Nothing is said in the letter as to the manner of exposure of 

 the wet bulb, except, parenthetically, that the thermometers were protected from the 

 full force of the wind. It would appear, therefore, that the table should correspond 

 to REGNAULT'S value of A = '00074. 



NOBLE points out that the factors for temperatures below the freezing point do not 

 coincide with those deduced from the Greenwich observations. Without assigning 

 any cause for the difference, he mentions two circumstances which may bear upon the 

 question : 



1. If the air be a little above and has been below 32 F. there will frequently be 

 found a small button of ice at the foot of the wet-bulb thermometer which is not 

 easily perceived, and which will keep it at 32 F. when the temperature of evaporation 

 is really above that point. 



2. The water may be cooled below 32 F. without freezing. 



The hygrometric methods have also been discussed by KAMTZ, in a paper published 

 in 'Kamtz, Repertorium/ vol. 2, pp. 341 to 361. I have (December 20, 1887) only 

 lately seen the original paper ; the following notes are taken from an abstract given 

 in the ' Fortschritte der Physik' for 1861. 



Comparing REGNAULT'S hygrometer in the original form with one consisting of 

 glass test-tubes silvered inside, and with DANIELL'S, and testing the results by the 

 chemical method, KAMTZ prefers the test-tube form of REGNAULT, and considers that 

 the difference of radiation and the difference in the surface may account for different 

 temperatures at which dew is deposited. 



Concerning the psychrometer, we have a discussion of a large number of observa- 

 tions by this method, including the Greenwich observations and some of REGNAULT'S, 

 as well as his own. 



Taking the general theoretical psychrometric formula 



= 



(where e is the required pressure of vapour, e^ the pressure at the temperature of the 

 wet-bulb, T the psychrometric difference, h the barometric height, and a and ft are 

 constants), instead of adopting REGNAULT'S approximate formula 



KAMTZ expands the general form in powers of T : thus 



c = e 1 +otTP 1 + ^T+yT 2 +8 j 



where d = 745 h. 



The various series of observations were then separately employed to determine the 

 constants of this formula, with the following results : 



