1891.] The Numerical Registration of Colour. 227 



paper, and called a focometer. It has been constructed to the author's 

 designs by Messrs. Nalder Brothers, to whom sundry of the mechani- 

 cal details are due. 



The paper also describes the results obtained with the focometer 

 upon various lenses, some of them being microscope objectives, others 

 camera lenses. The author finds in several of these lenses that the 

 principal planes are crossed : the distance between the symmetric 

 points being less than four times the focal length. In some other 

 lenses which are achromatic in respect of bringing all rays to a com- 

 mon principal focus, the positions of the principal planes are different 

 for rays of different colours. In one lens, a microscope objective by 

 Reichert, the principal planes are not only crossed but are actually 

 at a greater distance apart than the two principal foci. The paper is 

 accompanied by a sheet of full-size drawings showing the construc- 

 tion of the instrument and its details. 



[V. " The Numerical Registration of Colour. Preliminary Note." 

 By Captain W. DE W. ABNEY, C.B., R.E., D.C.L., F.R.S. 

 Received February 6, 1891. 



The Committee of the Royal Society on Colour Vision having put 

 ito my hands the determination of the colour of certain signal glasses, 



memorandum was drawn up on the method of the numerical regis- 



ition of colours and submitted to them. They considered that it 

 lould be submitted to the Royal Society, and having slightly 

 modified, it, it is presented as a preliminary note of a part of a paper 

 which will be subsequently submitted by General Festing and myself 

 as Part III of " Colour Photometry." 



It must be premised that a colour is determined when its hue, its 

 purity, and its luminosity are known, the last constant being its 

 comparison with the white light before its passage through a trans- 

 parent coloured body, or with white light reflected from a white 

 surface if it be an opaque coloured body such as a pigment. 



There has hitherto been a certain amount of difficulty on the part 

 of normal-eyed persons in stating the exact hue of compound colours 

 in terms of any standard ; in fact, I believe, except by the method 

 given in the Second Part of " Colour Photometry " (' Phil. Trans.,' A, 

 1888), there has been no exact means indicated of reproducing a 

 colour from measurements made. The method which will be described 

 can take the place of the previous plan for certain purposes, more 

 particularly when it is the impression on the eye which has to be 

 considered. Any colour can be reproduced from the registration 

 numbers with the greatest exactness. 



To pei-sons who are totally colour-blind to one sensation, viz., the 



