Psi/c/iroiuefric Furniuld. 155 



no difference, and only (> diffur l)y inoic than 0.1^; the n)ea.n dif- 

 ference is 0.040, while the observahlc limit is O.Oo'. In the other 

 case of linen and silk, 36 show a zero difi'ercnce, and only 9 show 

 above 0. 1° difference, while the averaj^e is 0.015'^. ("onibininjx these 

 with the observations already piiblished, we find the average differences 

 in the two cases to be resj)ectively 0.03O and O.Ol^, while out of the 

 large ninnber of differences only 10 in one case and 11 in the other 

 are greater than a tentli f)f a degree. 



We must conclude from these figures that the value of the co- 

 efficient -q is not perceptibly dependent on the nature of the covering 

 used for the wet-l)ulb thermometer, this result holding true through- 

 out a considerable range of temperature above zero, and under prac- 

 tically all conditions of humidity. 



The full details of the observations are contained in the next table. 

 Under the heading /' are given the mean values, where necessary, of 

 the wet-bulb readings ; the pressure is given co^rrected for tempera- 

 ture. 



Table II. 



