DORSET: LUMINESCENCE MEASUREMENTS 5 



conditions may be practically attained by the use of a small 

 pocket flashlight suitably screened and provided with a ruby 

 bulb. 



In order that the luminosity of the preparation shall be due 

 solely to the radioactive excitant, it is necessary that the prepa- 

 ration shall have been excluded from all light except its own 

 for at least several hours before the measurement is made. 

 During the measurements the milk-glass -window must in every 

 case be screened from all light coming from the preparation; 

 otherwise a false setting will be obtained. The error due to 

 this cause may easily amount to over 10 per cent in the lumi- 

 nosity. 



Another error that must be carefully guarded against at these 

 luminosities is that due to the "Purkinje effect." By this is 

 meant the fact that two adjoining luminous surfaces of differ- 

 ent colors that appear to be of equal brightness when viewed 

 from a certain distance will not appear to be of equal brightness 

 when viewed from a different distance. This means that when 

 we are dealing with different colors our judgment of equality of 

 brightness is a function of the intensity of the illumination. 

 This effect becomes very pronounced at low intensities. A 

 further difficulty is introduced by the fact that under such 

 conditions the judgment of equality of brightness depends upon 

 the peculiarities of the eye of the observer; different observers 

 judge differently. On account of both of these reasons it is 

 important that the illuminated milk-glass be a very close color 

 match to the material being measured. This is true even when 

 the luminosity is so low that it produces practically no color 

 sensation, properly so called. Under such conditions the color 

 match must be determined by spectroscopic examination; indeed 

 this is the best procedure in all cases. 



By the addition of a phosphoroscope and a constant source 

 of illumination this photometer can be readily adapted to the 

 study of the intensity of the luminescence excited by light, 

 especially when the spectrum of the luminescence is a single 

 broad band. If the luminiscence consists of a number of nar- 

 row bands scattered throughout the visible spectrum an exact 



