SECTION III, 1916 [85] TRANS. R.S.C. 
Improved Methods in Hygrometry. 
By A. NorMAN SHAW, B.A. (Cantab.), D.Sc. (McGill). 
Presented by Dr. H. T. Barnes, F.R.S., F.R.S.C. 
(Read May Meeting, 1916.) 
The object of this paper is to describe experiments with several. 
hygrometers of simple construction which apparently give results 
of accuracy greater than that given by the instruments in common 
use. The lack of refinement in ordinary humidity determinations 
is generally recognized, so the account of an attempt at its improve- 
ment needs no extended introduction. A brief summary of the 
weaknesses and limitations of the customary methods is given here. 
$ 1. SUMMARY OF THE LIMITATIONS OF THE HYGROMETRIC METHODS 
IN GENERAL USE. 
The following conclusions have been drawn from a series of tests 
performed with various hygrometers during the past year, and are 
in general agreement with existing records of the subject. 
(a) “Chemical” and ‘‘ Vapour Pressure’ Methods. These require a 
considerable amount of time for the taking of an observation, and are 
usually elaborate. If the humidity remains constant it can be deter- 
mined by some of these methods to within 0-1% of its true value. 
This requires, however, expert attention and “research” conditions. 
Under ordinary circumstances it is difficult to eliminate errors of less 
than two or three per cent. These methods are almost useless for the 
examination of a rapidly varying humidity, giving merely the average 
value during the time of observation. 
Such methods as those of Brunner, Schwackhofer, Svenson 
and Edelmann may be classified under this heading and, as far as 
the criticism in regard to time, complexity and accuracy is concerned, 
those methods which involve the measurement of the amount of 
additional vapor necessary to produce saturation, may also be included. 
(b) Dew-Point Methods. Under the most favourable circum- 
stances an accuracy of greater than 1% cannot be assured. If the 
relative humidity is less than 20%, or if the temperature is below 
40°F. errors as large as 10% may easily occur. 
(c) Wet-and-Dry-Bulb Methods. Under the best conditions a 
good constancy of repetition may be obtained but an absolute accuracy 
of within 2% cannot be assured. If care of the wet sack and attention 
