Laiv for True Photographic Rendering of Contrast. 195 



of the characteristic curve lying outside the latitude 

 of the paper and thus depart from direct proportional 

 rendering.''' 



This conclusion on their part illustrates a prevalent fallacy 

 that the range of contrast on the exposure scales of the 

 negative and of the positive must be alike. Consider, how- 

 ever, a paper for which the latitude is 0'75 in log units and 

 for^which <y 2 I s 1*5. This is shown in tig. 8, where 



tan 2 = y 2 =l'5. 



The curve of the negative which will be correctly printed 

 by this paper under these conditions has 



71=^=^=0*67, 



72 



1-5 



and reproduces a range ab on the logEi scale which is given 

 by 0-75 tan 2 = O-75 x 1*5 = 1-12. This corresponds to a 

 ratio of 1 : 13 in exposure units. Thus though the exposure 

 latitude of this paper is only 1 : 5"6, it will correctly repro- 

 duce a latitude in the camera exposure of 1 : 13 and not 

 only the 1 : 5*6 as these authors take for granted. 



Case II. — It is not compulsory, however, to work in the 

 straight por-ions of the curve as Hurter and Driffield thought 

 to be necessary. Examination of the curve B'B in fig. 9 

 will show that this possible extension is of limited avail- 

 ability. No sensitive material known will yield such a 

 curve as is shown in this figure throughout its whole 

 length ; but this is necessary to give a true photographic 

 reproduction. 



In fig. 9 we see that the curve of the paper pqr will fit 



