308 Life and Letters of Francis Galton 



wasted too much time already on what may be of no use. So I simply send the enclosed. Abney 

 used a rotating cylinder with a black and white drawing wrapped round it, in order to get the 

 photographic equivalent to each combination of black and white*. 



Culirzder 

 with f»per rouftd it 



What good news about Pearl. I return his letter. I will shortly send the centile paper and 

 diagram. Some delay has necessarily occurred about it. I am not idle but get through things 

 now so slowly. 



Best Christmas wishes to you all, and may you enjoy cake with the F. G. cut. My Nature 

 of last week has miscarried so I have not yet seen my own paragraph, though others, like your- 

 self suspecting me, have written to me about it. The post is just going out so I conclude now, 



Ever affectionately, Francis Galton. 



To obtain the mean tint of a rectangvl.ar picture. 



Mount the picture on a [rotating] cylinder with axis vertical [? horizontal], in front of a 

 camera. The dark slide of the camera to have a narrow vertical slit. Take an exposure — then 

 cap. Move screen the width of the slit and take a second exposure ; again cap. Repeat the process 

 until the sum of the widths of the slits is equal to the length of the picture. Print off. The 

 print will be streaky and of same width as the original was long. Mount the print crossways 

 on the same cylinder as before and proceed as before. The result will be a plate of a uniform 

 tint, the mean tint of the original. F. G. Dec. 25, 1906. 



Galton's plan to get a mean tint is suggestive although it is not quite 

 clear how he proposed to carry it out in practice, especially in dealing with 

 the relative mean tints of engravings, say, of different sizes, or of piebald skins. 

 Would a whole series of cylinders be needful to fit subjects of different heights, 

 or must the subjects first be reduced to a standard size by photography ? How 

 in practice would such reduction affect the relative tints of the two engravings? 

 Again I do not follow the necessity for the slit, or how it is to be moved. A 

 photograph of the engraving on the rotating horizontal cylinder would give 

 vertical streaks on the plate. A print from this, which must be taken under 

 stringently standardised conditions, could then be put crosswise on a cylinder 

 of proper size and again photographed to obtain a uniformly tinted negative 

 and thence a print. The " greyness " of this print would have — with ab- 

 solutely standardised conditions — some relation to the average tint of the 

 engraving, but I cannot see that they would be the same. Supposing the 

 lighting always (artificially) the same and the exposures identical, it would 

 be possible to compare the " greyness " of the prints thus defined with those 

 obtained from known amounts of black and white on the cylinder, and thus 

 form a scale. As I have said, Galton is here suggestive, and the problem is 

 of some practical importance, but it needs much experimental work before 

 it can be considered solved. 



(13) Worh and Correspondence of 1 907. The year 1906, owing to reasons 

 in part indicated, had been a year of stress and change for both Galton and 



* A similar arrangement was adopted in the Biometric Laboratory for tint comparison judg- 

 ments in 1894 (see Phil. Trans. Vol. 186 A (1895), p. 392). It is still used in the Anthropo- 

 metric Laboratory attached to the Biometric Laboratory. 



