and the Psychrometer. . 447 



each of which has usually had great uncertainties entering into 

 it. It is believed that the errors entering into t and t' have 

 been reduced for each individual observation to less than '1°. It 

 is surprising to see how swing after swing may be made with 

 the sling psychrometer resulting in absolutely no change in the 

 readings. The same cannot be said, however, regarding the 

 determination of the dew-point; in this errors will creep in on 

 different days of '3°, and in consequence the value of A may be 

 changed slightly. When there is a small difference between t and 

 t' there is great difficulty in getting uniform results, but when 

 {t—t') is small a larger difference from the mean in A produces 

 a much smaller effect upon the deduced moisture content than 

 when {I— V) is large. With {t—i') equal to or greater than 15° 

 there is no difficulty in getting A within two of the fifth place 

 again and again and on different days. The vapor tensions 

 employed in computing the table have been taken as recently 

 computed by Broche from Regnault's observations. 



One of the more interesting questions which arises in connec- 

 tion with the formula is the effect of bight above sea upon the 

 relation between the dp : t, and P. Glaisher, after bringing 

 together many thousands of comparisons between a Daniel's 

 hygrometer and the psychrometer, at various hights in India and 

 elsewhere, decided that bight had no appreciable effect upon 

 the deductions to be made with the psychrometer. Although 

 these observations were made without any uniform ventilation 

 for the psychrometer and cannot be regarded as giving absolute 

 results at either sea level or on mountain tops, yet it would 

 seem as though they were of sufficient accuracy to enable rela- 

 tive comparisons so far as the question of hight comes in, i. e., 

 if the same instruments observed in precisely the same way at 

 two hights give the same results, we may infer that there is no 

 difference in the relations of the quantities measured, although 

 large errors may occur at both hights. The question of a pos- 

 sibly better ventilation at the higher station need not be con- 

 sidered, as this would have tended to lower the wet-bulb read- 

 ing and would have been in exactly the direction of theory at 

 high stations. The exposures and amount of moisture at the 

 two hights would have a slight influence, but in a long series 

 of observations there would be an elimination of these effects. 

 Dr. Pernter has published a series of observations with a ven- 

 tilated psychrometer,' 55 ' and Regnault's condensing hygrometer 

 at a barometric pressure of about GOO™ 11 . These observations 

 gave a mean value of A as -001042 ; while the value of A at 

 sea level adopted by Dr. Pernter is -000843. If we multiply 



* This psychrometer was in a ventilated screen of Prof. Wild's form, but 

 Prof. Wild thinks that the screen used was his earlier form with a clock-work 

 ventilator, which did not give a very perfect ventilation. 



