photographic rendering of Contrast. 635 



Now 00' is a length on the log. exposure axis of the 

 characteristic curve and therefore represents the difference 

 between two values of logl£ (intensity x time), i.e. the 



logarithm of a mere ratio or abstract quantity, while log^ 

 must be the logarithm of an intensity. 



Similarly O'O" represents the logarithm of another ratio, 

 while log P is again the logarithm of an intensity. 



I submit therefore that the authors are wrong in saying 

 tliat " to choose the point 0" is equivalent to deciding 

 upon particular values for the printing and viewing lights 

 and the reduction factor K," as they state in the succeeding 

 paragraph. For similar reasons the lengths marked on their 

 diagram (fig. 4} cannot possibly represent the quantities 

 logEj, logP, log I, log V, and logK, which they are 

 said to do. 



The remainder of the same paragraph reads : " The more 

 nearly K is to unity the more nearly will the light from the 

 positive be not only the same in gradation as that from 

 the subject but also the same in absolute intensity." This 

 appears to be open to the interpretation that K has some- 

 thing to do with the gradation of the picture, but I do not 

 suppose the authors mean to imply that such is the case* 

 They apparently consider that correct picture-making must 

 be governed absolutely by the rule (p. 188) : — " It is necessary 

 to have the same ratio between light emitted from two portions 

 of our final picture as that which fell in the camera from 

 the two corresponding portions of the subject" ; or (p. 197): 

 "The only condition which seems to us useful is that the 

 gradation in the issuing light (I) shall be the same as that 

 from the original source (Ej). This is the condition which 

 we have taken as fundamental." 



Now this is undoubtedly a more stringent requirement 

 than is desirable, as Hurter and Driffield or at least the latter 

 (see " Control of the Development Factor/' V. C. Driffield, 

 Photographic Journal, Jan. 1903), and all subsequent workers 

 have long recognized. The authors do not tell us how they 

 would compress the gradation in a print of a subject whose 

 range of contrasts was appreciably greater than the printing- 

 medium will yield, so we are left to speculate as to whether 

 or no they concur with ihe generally adopted rule, which 

 may be expressed as follows : — The relationship between the 

 intensities of the original and those of the positive copy 

 should comply with I C = KI " where K and n are both 

 constants, of which K may have any value while n must 

 be greater than unity when the contrasts of the copy are 



