194 TRANSACTIONS OF THE [mAR. 1, 



more easily form a judgment concerning tlie equality of the two 

 luminosities. 



Mace de Lepinay and Nicati in their experiments* used a 

 Rumford photometer in which the surfaces to be compared were 

 ver^'^ small. The angle subtended by them at the eye was about 

 fort3'-five minutes. They found that the accuracy with which 

 the luminosities of differently colored surfaces could be com- 

 pared was thus increased. The results, however, were now 

 found to vary with the absolute size of the retinal image. 



Professor Mayerf uses the effect of simultaneous contrast to 

 destroy the difference in color between the two surfaces to be 

 compared. This method, however, is only applicable when the 

 difference in color is not very great. 



The methods of Virerordt| and Draper§ avoid the necessity 

 of comparing the luminosities of two differently colored surfaces. 

 The validity of the methods, however, depend upon the assump- 

 tion that the quantity of light which must be added to a lumi- 

 nous, surface in order to be just perceptible is, under all condi- 

 tions, the same fraction of the total luminosity. The assumption 

 is no longer considered valid. 



I will now consider the photometric methods which do not 

 depend upon what I have called the principle of equal or uni- 

 form luminosity. The first of such methods used depended upon 

 the extinction of the source of light by the introduction of 

 translucent media. This method seems to have been employed 

 as early as 1700 by Francois Marie. He introduced pieces of 

 glass between the light and the observer until the light was ex- 

 tinguished. The value of the luminosity was then judged by the 

 number of pieces of glass used. The same method but with 

 various refinements comes up again and again in the history of 

 photometery. To all such photometers there is this serious 

 objection, that the limit of sensitiveness of the eye to light is so 

 variable that the error thus introduced may be three or more 

 times the quantity to be measured. 



A photometric method depending upon the distinctness of 

 vision was first suggested by Celsius (1701-1744). It has been 

 very carefully worked up by Maci de Lepinay and Nicati. || As 

 a measure of visual acuteness (" accuete visuelle ") they take 

 the angle subtended at the eye by a black character, as a printed 



♦Annales de Chimie et de Physique, 5th series, Vol. 30, page 151. 1883. 

 + American Journal of Science. Vol.46. 1893. 

 iPogg. Ann. 1869. 



I Phil. Mag. 5th series. Vol. VIII. 



II Annales de Chimie et de Physique. 5th series, Vol. 21, page 30, and Journal de 

 Physique 1882-1883. 



