S. P. Langley — Energy and Vision. 365 



object and limits of our experiments and the limiting positions 

 of the screen, so nearly coincident with a geometrical cone, 

 that, as the slider is carried away from the slit, the light may 

 be treated as diminishing proportionally to the inverse square 

 of the distance from the slit to the screen. The nearest posi- 

 tion of the screen brings it within 20 cm of the slit, the farthest 

 is over 300, so that we have the power of diminishing the light 



(300\ 2 

 - — or over 225 times. This however is by no means 



a sufficient range for the comparison of the light in the 

 yellow green with that in the extreme red ; and because the 

 graduated rod was not long enough to thus give the desired 

 range, a photometer wheel was introduced in some of the 

 measures between the siderostat mirror and the remote slit (sj. 

 This photometer wheel is capable of reducing the light from 

 *50 to *05 or further, and is more fully described in Memoirs 

 National Academy of Sciences, vol. iii, Memoir on the Temper- 

 ature of the Moon. We have, then, without altering the slit, 



•225 

 a range of adjustment through over - — or over 4500 times. 



The slit s t where the light first enters has doubly moving jaws, 

 controlled by a micrometer screw. Its standard opening in 



these experiments for light comprised between ^=0 / " , 40 (violet) 



and ^=0^*65 (red) was 0'l mm , but it has been opened for sup- 

 plementary experiments to 5 mm , so that we have by opening or 

 closing it a range of light from 50 to 1. It was, however, con- 

 stantly kept at the standard opening of # l mm until the main 

 series of experiments was completed, so as not to vary the light 

 by attempting to reset it by the screw. Admitting, however, 

 that for any given prism, transmitting any given ray, the light 

 is sensibly proportional to the width of the slit (which may 

 vary from 50 to 1), to the disposition of that coming through 

 the photometer wheel, which may vary from 20 to 1, and to 

 the inverse square of the distance of the slider from slit # 2 

 (225 to 1), we have a possible range of 50x20x225 = 225,000 

 to 1. This, however, it will be understood, has only been em- 

 ployed in our supplementary measures. 



In the following table all observations, whether made with 

 or without the photometer wheel, or with a wide slit, as in the 

 case of the supplementary observations in the most feebly 

 luminous portions at the extremities in the spectrum, have 

 been reduced to these standard conditions : 



Photometer wheel absent; 



Slit fo) 0-l mm wide; 



Slit (s a ) l ram wide; 



Slider with logarithm table at 1 meter from slit s a . 



