6 TRANSACTIONS OF THE [OCT. 5, 



for degrees of saturation of the air ceases to be regular when the 

 relative humidity becomes as low as 7.8 per cent. This is in 

 agreement with the statements just made. 



In the tests which were carried on, it was also invariably found 

 that if the degree of saturation of the air was altered, some time 

 had to elapse before the hairs became adjusted to the new con- 

 ditions surroundiug them, and therefore ample time was alwaj^s 

 given for the hairs to become adjusted when the hygrometric 

 state of the air was changed. Thus, when the hair hygrometers 

 were placed in an entirely new atmosphere differing 15 per cent, 

 or more in relative humidit}-, 5 to 25 minutes elapsed before the 

 hairs responded perfectly to the change and gave approximately 

 correct readings. It was further observed that the length of 

 this time depended on whether the change was to a higher or to 

 a lower percentage of humidity, and also in what part of the 

 scale the change occurred. 



For example : A change from 



15 to 90 per cent, required about 10 minutes 

 30 to 90 " " " 10 " 

 15 to 30 " " '' 15 " 

 90 to 15 " " " 20 " 

 90 to 30 " " 10 to 15 

 30 to 15 " " about 20 " 



which indicated that it takes longer for the hairs to dry out than 

 for them to take up the moisture, and that the change is slower 

 at the lower parts of the scale than elsewhere. 



A knowledge of the relative humidity of the air is important, 

 not only in various branches of science and arts, but also in the 

 treatment of the sick, particularly in cases of certain pulmonary 

 disorders. 



It is very probable that a perfectly accurate direct reading 

 h3'grometer is difficult to obtain, but this fact effects but little 

 the value of the hair hygrometer, for great exactness is seldom 

 required, a knowledge of the relative humidity of the air to 

 within two or three per cent, being, in most cases, all that is 

 necessary. 



The j>recaution of allowing considerable time to elapse for the 

 hairs of hair hygrometers to become adjusted to a changed at- 

 mosphere, before taking a reading, is only really necessary where 

 a possibility exists that a decided change in therelative humidity 

 has suddenly occurred. 



The table which is given below has been constructed from re- 

 sults obtained bj^ comparing the readings of relative humid it}'^ 

 shown by the hair hygrometer with those calculated from ob- 



