﻿582 Dr. K Campbell and Mr. B. P. Dndding on 



first pair does not form a photometric pair with a member of 

 the second. Since any surface (except an absolutely black 

 one) can be used as a P.S. with a suitable pair, this law 

 naturally leads us to regard illumination as something 

 characteristic, not of the surface illuminated, but of the 

 circumstances in which it is placed. We find, further, that 

 among the most important of these circumstances are the 

 positions of the surfaces relative to the sources and the nature 

 of these sources. It is these laws which give us the first 

 clues to a theory of illumination ; but such a theory, though 

 it is a very useful guide in seeking a satisfactory system of 

 measurement, is best excluded entirely from any descrip- 

 tion of it. 



5. We must now define addition. The illumination on a 

 surface X from the sources A and B is equal to the sum of 

 the illumination of X from A and the illumination of X from 

 B, if A and B^ when tbey are illuminating the surface, are 

 always in the same physical state and in the same position 

 relative to the surface whether they are acting together or 

 singly. 



With this definition, the first law of addition is true ; 

 cutting off the illumination from a source always decreases 

 brightness. But the second law is not true in all circum- 

 stances ; it is not true, for example, when the Purkinje 

 effect is apparent. For, if Pj and R 2 are red sources, B x and 

 B 2 blue sources, and if the illumination from Pj is equal to 

 that from B x and that from P 2 to that from B 2 , the illumina- 

 tion from P 2 and R 2 will not be always equal to that from Bi 

 and B 2 . On the other hand, if all the sources are of the same 

 colour, the second law is true ; and it is true apparently if 

 the sources, though of different colour, give sufficiently great 

 illumination and fulfil some other minor conditions. The 

 Purkinje and allied effects are simply failures of the neces- 

 sary laws of photometric measurement, and any complete 

 statement concerning such measurement must include the 

 proviso that the conditions are such that these effects do not 

 enter and that the laws of measurement are true. It must 

 be insisted that the only logical way to describe these effects 

 is in terms of the failures of the laws of measurement which 

 they involve ; to describe them in terms of the measurements 

 which they make impossible, though it may be convenient 

 and conduce to brevity for general purposes, is utterly 

 ludicrous if precision is important. 



6. We have now defined completely the magnitude illu- 

 mination, and can proceed to measure it and to state signi- 

 ficantly and prove experimentally the following important 



