738 Lord Rayleigh on the General 



If we abandon the restriction that the photographic process 

 is to be the same in the two operations, simple conclusions of 

 generality can hardly be looked for. But the problem is 

 easily formulated. AVe may write 



f=Mef), t = t"=f 2 (e't'), . . . . (9) 



where e, e' are the exposures^ not generally equal, and/i, f 2 

 represent two functions, whose forms may vary further with 

 details of development and intensification. But for some 

 printing processes f\ might be treated as a fixed function. 

 It would seem that this is the end at which discussion 

 should begin. When the printing process is laid down and 

 the character of the results yielded thereby are determined, 

 it becomes possible to say what is required in the negative; 

 but it is not possible before. 



In many photographs it would appear that gradation tends 

 to be lost at the ends of the scale, that is in the hioh lights 

 and deep shadows, and (as a necessary consequence, if the 

 full range is preserved) to be exaggerated in the half-tones. 

 For some purposes, where precise reproduction is not desired, 

 this feature may be of advantage. Consider, for example, 

 the experimental problem, discussed by Huggins, of photo- 

 graphing the solar corona without an eclipse. The corona is 

 always present, but is overpowered by atmospheric glare. 

 The problem is to render evident a very small relative dif- 

 ference of luminous intensity. If the difference is exag- 

 gerated in a suitably exposed and developed photograph, so 

 much the better. A repetition of successive copyings might 

 render conspicuous a difference originally invisible. At each 

 operation we may suppose a factor a to be introduced, a being 

 greater than unity. After n copyings dt/t becomes a n dt/t. 

 Unless the gain each time were very decided, this would be 

 a slow process, and it would be liable to fail in practice owing 

 to multiplication of slight irregular photographic markings. 

 But a method proposed by Mach * and the present writer t 

 should be of service here. By the aid of reflexion light at 

 each stage is transmitted twice through the picture. By this 

 means alone a is raised to equality with 2, and upon it any 

 purely photographic exaggeration of gradation is superposed. 

 Three successive copyings on this plan should ensure at least 

 a ten-fold exaltation of contrast. 



* Eder's Jahrbitch f. Photoqraphie. 



f Phil. Mag. xliv. p. 282 (1897) ; Scientific Papers, vol. iv. p. 333. 



