January 6, 1922] 



SCIENCE 



21 



brightness distribution is symmetrical about a 

 vertical semicircle passing through the sun. 

 Such measurements were made on days that 

 were (1) perfectly clear, (2) overcast with 

 thin clouds or dense haze, (3) completely 

 overcast with clouds or dense fog, so that 

 neither sun nor blue sky could be seen, (4) 

 overcast with clouds from which rain or snow 

 was falling, and (5) partly overcast, in an 

 irregular manner. 



On clear days it was found that the sky 

 brightness at Washington has somewhat the 

 following distribution: The brightest part of 

 the sky is, of course, that close about the sun. 

 The darkest part is that in the solar vertical 

 about 90° distant from the sun. In general, 

 the sky increases in brightness toward the 

 horizon, although there is a "dark valley" ex- 

 tending from the dark point in the solar ver- 

 tical to a point about midway between the 

 sun and the horizon. This distribution agrees 

 closely Avith that observed by Dorno at Davos, 

 Switzerland, except that the Swiss sky is 

 brighter than that at Washington. This differ- 

 ence in brightness is probably the result of 

 secondary reflection of light from the Alpine 

 snows. In comparison with observations made 

 at Chicago University and on the roof of the 

 Federal Building in "Loop" district of Chi- 

 cago, it was found that the distribution there 

 is much the same, except that the horizon 

 opposite the sun is darker at Chicago than at 

 Washington. This is attributed to smoke, 

 from which the Washington atmosphere is par- 

 ticularly free. 



The brightest type of skj' measured at 

 Washington is that completely overcast with 

 thin clouds or dense haze. With clouds from 

 which rain is falling, the distribution is about 

 the same as with thin clouds, but its intensity 

 is only half as great. 



Measurements of the illumination on hori- 

 zontal and vertical surfaces were made at 

 Washington and at the two Chicago stations 

 mentioned above. It was found with respect 

 to the variations with change of solar altitude 

 that the illumination on horizontal surfaces 

 increased mai'kedly with increase of solar alti- 

 tude ; but in the case of illumination on vertical 

 surfaces the difference between a surface facing 



the sun and one oppositely directed grows less 

 with increase of solar altitude. Moreover, 



The daylight illumination on a vertical surface 

 facing opposite the sun, and with an unobstructed 

 exposure to the sky, in the Loop district of Chi- 

 cago under summer conditions as regards smoke, 

 averages only about two thirds as intense as 

 illumination on a similarly exposed surface at 

 Washington under similar sky conditions with 

 respect to clouds, except when the sun is more 

 than 40° above the horizon and the sky is clear. 



The equation. 



Tan — h/io^l/{l + tanas), 

 is given for computing the shading effect of 

 buildings on the opposite side of the street. 

 d is tlie angular lieight of a building as seen 

 from the center of a window across the street, 

 the width of the street being w. The horizontal 

 angle between a normal to the window and a 

 line joining a point p on the building opposite 

 is .X, and h is the height of the obstructing 

 building above the point p. The author gives 

 a table showing the relation between x and 6 

 for various values of li/w. Attention is 

 directed to the fact that the horizon is the most 

 effective illuminating agent for vertical sur- 

 faces, hence buildings and other objects on the 

 horizon are the most serious obstacles in the 

 question of illuminating rooms through ver- 

 tically placed windows, especially with a clear 

 sky. 



Two interesting examples of the relation 

 between electric light load and sky brightness 

 are given. At Washington, on July 15 and 

 29, 1921, there occurred thunderstorms about 

 2 :30 p.m. and noon, respectivelj^. On the 

 former occasion, the daylight intensity fell 

 rather quickly to about one foot-candle and the 

 sudden increase in electric light load caused by 

 the nearly simultaneous turning on of thou- 

 sands of electric lights was sufficient to put 

 the power plant out of commission. The sta- - 

 tistieian for the company states that 



During the day in the business section a sudden 

 increase in current consumption occurs when the 

 day light illumination intensity falls below l,.jOO 

 foot-candles. Tlie lower the intensity, the higher 

 the current consumption, but fluctuations in 

 intensity above 1,800 foot-candles have only a 

 negligible effect. 



