368 SECTIONAL TRANSACTIONS.—A. 
to be nearly a straight-line one. From this comparison it appears that 1 milli- 
gramme of dust per cukic metre has the same effect as about 10,000 particles 
per cubic centimetre ; thus 101° smoke particles weigh 1 mg. approximately. The 
size of suspended dust particles is fairly uniform, but tends to increase during 
smoke fogs, probably due to their rapid formation giving insufficient time for 
grading by settlement. 
29. Mr. J. Parrerson.—Upper-Air Observations in Canada. 
Upper-air observations were commenced in Canada in 1911, but were partially 
interrupted by the war. It has not yet been possible to get balloons for 
carrying instruments equal to those of pre-war days; there are, however, good 
prospects of overcoming this handicap in the near future. During the past 
year an automatic apparatus for calibrating the meteorographs has been installed 
and the Dines meteorograph simplified. The results of the sounding balloon 
ascents during the past five years and the observations with pilot balloons in 
the Arctic will be discussed, together with the prospects of permanently 
extending the field of observations in the upper air to this region. 
30. Prof. H. H. Kimparu.—The Determination of Daylight Intensity 
from Automatic Records of Total Solar and Sky Radiation. 
Colour temperatures of sunlight and skylight, and the spectrum energy 
curves of radiation from the sun and from the sky, have been utilised to 
determine approximately the spectrum energy curve of the total radiation 
received on a horizontal surface, and its variation with atmospheric trans- 
missibility and the solar zenith distance. 
A comparison of these latter curves with the curve of ‘ visibility of radia- 
tion ’’ permits a prediction to be made of the variations to be expected in the 
ratio between the intensities of the vertical components of daylight and of 
the total solar and sky radiation. . 
This ratio has also been determined experimentally by comparing photo- 
metric measurements of daylight illumination on a horizontal surface with 
continuous records of the total solar and sky radiation made by a U.S. Weather 
Burean thermoelectric pyrheliometer horizontally exposed. 
The above investigations have been confined to skies that were either cloud- 
less or else completely covered with clouds. 
31. Prof. W. J. Humpnreys.—Rainmaking. 
Several of the more persistently urged schemes for producing rain are con- 
sidered in respect to the underlying principles involved, and measured quanti- 
tatively to determine the question of their practical use. 
These schemes include, especially : the production of loud noises; the use 
of chemicals; mechanical or forced convection; fog-collecting screens; dusting 
the sky; spraying liquid air on to clouds, and sprinkling clouds with electrified 
sand. 
None of these rainmaking methods is practicable in the commercial sense 
of the term; but each, when treated quantitatively, is full of meteorological 
interest. 
32. Prof. C. F. Marvin.—Let us Simplify the Calendar and Publish 
Statistical Data in Standardised Summaries. 
Great masses of statistical data covering the fields of meteorology, yields and 
prices of crops, business and economic conditions, panics, &c., are now being 
annually accumulated. 
To be fully useful to students for analysis in any of these fields these data 
need to be appropriately assembled, summarised in suitable units of time and 
sectional area or representative groups, and promptly, regularly, and syste- 
matically published. This is not now being done sufficiently, either with 
reference to much available data of past years or comprehensively for the 
future. 
Our complex and awkward calendar, with months of unequal lengths made 
up of four weeks plus 0, 1, 2, or 3 days, apportioned according to the envy 
