324 Photographic Values of Moonlight and Starlight. [Mar. 19, 



positions on the first night differ so little from it that it is unneces- 

 sary to repeat the chart. I have to thank the Astronomer Royal for 

 these charts, he kindly having had them prepared for me from 

 Phillips' atlas. I have, however, reduced them from the elliptical 

 form to that of a circular disc. 



Fig. 3. 



The light from Jupiter has been estimated, and it would not be 

 far wrong to assume that it is equivalent to a candle placed at 

 800 feet from the screen. It may, therefore, be neglected in taking 

 into account the light from the stars, and much more so may Mars 

 and Neptune. It must be recollected that the exposure was made to 

 the plate in a horizontal position. If the stars were uniformly dis- 

 tributed in the hemisphere the measured light would be but half of 

 the actual light, since it would strike the plate at an angle, except 

 at the zenith. Besides this, however, we have to take into account 

 the atmospheric absorption, and taking the most favourable coefficient 

 for the plates used we shall find that only about 25 per cent, of uni- 

 formly distributed starlight would be effective. Total starlight would 

 thus be 4 x O'OOIMS or 0*006 standard candle at 1 foot distant from 

 the screen. Taking moonlight as 0'266 of a standard candle, we find 



