132 M|{. W. N. SHAW OX HTGROMBTBIC METHODS. 



formulae and their application. He first explains that the assumptions made by 

 AUGUST and others in order to obtain a theoretical formula cannot practically be 

 justified, and he then deduces the formula originally given by BELLI* and supplies it 

 with the more accurate values of the constants obtained since BELLI'S time, and 

 reduces it by omitting negligible quantities to the form : 



. fl _ (0-237 B - l-79)(< - Q 

 377-2 - 0-490 t' 



Following REGNAULT'S suggestion that the theoretical formulae should be regarded 

 only as general forms, he has the three following : 



REGNAULT'S form, /=/' A (< f)B . . .... (I. 



BELLI'S form, /=/ - ( ~- + '-*><' ~~ f) (2.) 



AUGUST'S form, f = f - * ( '~_f- . .T (3.) 



+ n(t-f)'V (4) 



n _ 

 to these he adds as a suggestion : 



where A, m, n, p are constants to be determined by experiment. 



He then applies these four formulae to the computation of the pressure of aqueous 

 vapour from psychrometric observations, determining the constants by the method of 

 least squares, using the results given by REGNAULT'S hygrometer as standards. He 

 then tabulates the errors ; if we denote by A/*j A/ 2 . . . the errors from REGNAULT of 

 the formulae (1) (2) ... we get the following values for the mean absolute error, i.e., 

 the mean of the errors without regard to sign : 



A/! = 0'33 mm. 

 A/ 2 = 0'32 

 A/3 = 0-29 



He concludes that his own suggestion most nearly represents the actual case, but 

 at the same time he remarks that the psychrometric method is, so far as he can see, 

 not capable of giving the pressure of aqueous vapour to within O'l mm. 



Certain practical points with reference to psychrometer observations are discussed 

 in a paper by CANTONI.* By a series of observations upon the temperature of the air 

 as recorded by thermometers exposed freely to the sun, or protected by various screens 

 or metallic tubes, he shows that the best results are obtained without screens, provided 

 that a current of air is made to pass over the bulbs, and a similar advantage results 



* See PKRNTEB'S paper, quoted below. 



