﻿Measurement of Light. 1 167 



that from l\ 2 to that from B 2 , the illumination from 1m and 

 \\ 2 will not be always equal to that from B I and B 2 . " 



Tliis statement is most surprising, especially if taken in 

 connexion with a previous statement that, illumination is 

 "characteristic, not of the surface illuminated, btrt of the 

 circumstances in which it is placed/' These two statements 

 are nut easily reconciled, for it is easily possible to imagine 

 two pairs of photometric surfaces, one pair of high reflexion 

 ratio and one of low reflexion ratio, so that in the former case 

 the addition law of illumination may he found to hold, while 

 in the latter case with the same illuminations it does not. Surely 

 if the statement first quoted be true, the second is refuted. 



It seems, on the whole, more consistent to regard illumin- 

 ation as independent both of the nature of the surface 

 illuminated and of such essentially ocular phenomena as the 

 Purkinje effect. The latter can, and must, be reckoned with 

 when considering brightness, and the statement first quoted 

 above is quite intelligible if read as referring to the addition 

 of brightness, and if the words "brightness due to" be substi- 

 tuted for '"illumination from."' 



It seems, then, that we have to choose between a system of 

 definitions depending on (a) brightness, and (/>) luminous 

 intensity or luminous flux, as the fundamental magnitude. 



It seems to the writer at least a doubtful contention that 

 it is more logical to start with the magnitude actually 

 perceived and measured and to work back to the magnitude 

 in which the unit is maintained. In either case the physical 

 laws connecting the various magnitudes in the chain have to 

 be known and, in fact, they must be described, at any rate by 

 implication, in the definitions of the dependent magnitudes. 



The gain, if any, in logical security seems to be more than 

 counterbalanced by a very marked loss of " concreteness " 

 — never a pronounced characteristic of formal definitions. 

 The mind naturally finds it most easy to form a picture of 

 the magnitude in which the unit is maintained, that being a 

 phenomenon having the closest association with a concrete 

 object. In the case of photometry it would seem that the 

 luminous intensity, or candle-power, of a source in a given 

 direction is far more readily understood as a basis of 

 definitions than is the brightness of a surface viewed in a 

 given direction. For the natural physical order is (a) 

 emission of luminous flux by a source owing to its luminous 

 intensity, (6) incidence of this flux at a surface, (/•) brightness 

 of this surface due to the illumination and the power of the 

 surface to reflect light. This is, then, the order in which the 

 mind expects the magnitudes to be defined, and it appears to 

 the writer the preferable order for that reason. 



