86 THE ROYAL SOCIETY OF CANADA 
to its ventilation are neglected, errors of greater than 5% may develop 
under apparently normal conditions. If either the relative humidity 
or the temperature is low, or if the relative humidity is approaching 
100% the errors may be larger than 10%. The former conditions 
for such errors are very common during the winter in cold climates. 
Inside steam heated buildings, for example, the relative humidity 
will, when the temperature is below —20°F. outside, be sometimes 
as low as 5% and yet be indicated as high as 20% on an instrument 
which without. alteration will record satisfactorily under average 
summer conditions: ; 
(d) Hygroscopic Expansion Methods. Hygrometers of this kind 
especially those employing hair, gut, or horn need very frequent 
recalibrations and in a few months may become almost useless. 
After subjection to extreme conditions there are often large ‘after- 
effects” which vary in a complicated manner. If, however, these 
instruments are compared frequently with others over a large range 
of humidities they are the most convenient because their sensitiveness 
is great, the time needed for a determination of humidity is short, 
and the procedure merely that of reading a scale or chart. They 
have, perhaps, been over maligned by many observers, but it must 
be admitted that the necessity for frequent recalibration renders 
them unsuitable for accurate work outside the laboratory. 
§ 2. A SATISFACTORY ABSORPTION HYGROMETER. 
Among the instruments examined was a slightly modified copy 
of a hygrometer recently developed by Dr. E. K. Rideal and Mr. 
A. Hannah.! This instrument was found to be so convenient that it 
was adopted as one of the absolute methods in these experiments for 
checking determinations made by other means. The results obtained 
with it show an accuracy apparently greater than that claimed by 
its designers, and as the method is a new one it seems desirable to 
draw further attention to it. 
Figure 1 is a photograph of a simple form of the instrument con- 
structed in this laboratory. There is also a sheet of glass which is 
placed in front of the bulbs and gauge in order that there may be no 
heating effect when the observer is bending towards the scales. (A 
more compact form which can be immersed and operated in a constant 
temperature jacket is better suited for experimental tests in hygrom- 
etry over large ranges of temperature and humidity, but is 
unnecessary for the ordinary determination of the humidity in 
any place sheltered from the sun and from strong winds) A 

1 Rideal and Hannah. The Analyst, vol. XL ,No. 467, p. 41, Feb. 1915. 
