proceedings: philosophical society 255 



quartz plate between them, the path of the hght being parallel to the 

 optic axis of the quartz, and the thickness of the quartz as well as the 

 angle between the principal planes of the nicols being properly chosen. 

 If three nicols are placed in series in the beam, one quartz plate being 

 placed between the first and second nicols and another quartz plate 

 between the second and third nicols, the approximation to a desired 

 spectral energy distribution may be made still closer. 



The rotatory dispersion of quartz has been previously used by others 

 in "chromoscopes," etc. The novelty of the present communication 

 consists soleh^ in showing how the method maj^ be used in producing 

 "artificial daylight," and in presenting precise specifications for pro- 

 ducing results. 



This method, of course, is not adapted to illuminating large surfaces 

 and so is not a commercial competitor with the blue-glass method 

 or other "artificial daylight." It is, however, very well adapted to use 

 with instruments (photometers, microscopes, etc.) where the quartz- 

 nicol system may be inserted between the eyepiece of the instrument 

 and the observer's eye. 



The chief advantages of this method over the blue-glass method are : 



1 . A much more accurate reproduction of the desired spectral energy 

 distribution. The distributions obtained by the use of blue glass (e.g., 

 Luckiesh's "Trutint" or Corning "Dayhght") are always distorted 

 from the desired distribution by a sharp maximum at X = 570 mm as 

 well as by a rise in the red for X greater than 660 jjlijl. 



2. Certain reproducibility and definiteness of specifications. 



3. Adjustabilit}^ By varying the angle between the principal 

 planes of the nicols, the distribution may be slowly changed by known 

 amounts. 



Discussion: The paper was discussed by Messrs. Sosman, Critten- 

 den, Burgess, and White. 



The third paper, on A simplified form of Robinson's anemometer, 

 was presented by Mr. B. C. Kadel. This paper was illustrated by 

 lantern slides. 



To begin with, the observer is assumed to have a watch or clock 

 available with which to measure a suitable time interval. The next 

 step is to arrange the anemometer to make electric contact at short 

 intervals, the signal being made audible to the observer by means of a 

 door bell, buzzer, or telephone receiver. The most convenient interval 

 is found as follows: 



Distance distance D d . , . , 



-y^. = — -. or 7^ = - m which 



Time time T t 



D = the linear unit chosen as a measure. 

 d = the travel of the wind between signals. 

 T = the time unit chosen as a measure. 

 t = the time the signals are to be counted. 



