[sHAW] IMPROVED METHODS IN HYGROMETRY 89 
chloride and gelatine, and cupric chloride and gelatine were found 
also to be suitable but not so sensitive as the phosphorus pentoxide. 
The cobalt and the cupric chlorides change colour as the vapour pres- 
sure varies and roughly indicate the humidity in that additional way). 
A piece of clean filter paper 7 X 16 cms. was moistened with this 
solution and suspended from the arm of a balance in a bell jar. There 
were 83-9 sq. cms. of surface per gram of solution when the vapour 
pressure was equivalent to 2-10 cms. of mercury. The surrounding 
air, or any given samples of gas could be passed through the jar 
in a slow and regular manner by means of a faucet suction pump, 
or any similar apparatus. The figures in the table below illustrate 
the constant behaviour of this arrangement with a balance measuring 
to within 0-01g. The first three sets were taken on consecutive 
days in September 1915, and the last two in March, 1916. The 
temperature was in each case 23-0°C, which determined the constancy 
of vapour pressure of the saturated supplies of air prepared on the dif- 
ferent occasions. The dry sample was bubbled through pure con- 
centrated sulphuric acid and the saturated was bubbled through 
a long tube of distilled water. The other samples were of laboratory 
air which happened to have the same vapour pressure on these chosen 
days—this was tested simultaneously by chemical methods. 
TABLE ILLUSTRATING THE CONSTANCY OF WEIGHT OF THE SOLUTION ON THE 
“MOISTENED PAPERS’’ UNDER SAME HUMIDITY CONDITIONS. 


Air Dry. | Air Saturated at Relative Humidity 
at 23-0°C. 57% at 23-0°C. 


(V.p. negligible.) | (V.p. 2-10 cms.) (V.p. 1-20 cms.) 
_ | a 
1. 1-26 grams. 2:65 grams. 2-00 grams. 
Sept, 1915.2 1328, .“ Dass) L900 LE 
Ser USA DC TE DOUTE 
March, 1916/4. 1-26 “ 268) BAW, & 
De, lo 2 Gn 5 Bo sil be POP 

In this particular case it will be noted that the change in weight 
is nearly proportional to the vapour pressure. 2-06 grams for a vapour 
pressure of 1-20 cms. would have corresponded to a proportional 
change, since 
ZA0 2567 = 1726 
120 0 2:06 — 1-26 approximately. 
