554 



NATURE 



[October 6, 1904 



useful process, that is, a method in which the coloured 

 image that falls upon the prepared surface produces on 

 it its own colours at once, is that suggested about 

 twenty-five years ago by Charles Cros. The receiving 

 surface is coated with a red, a yellow, and a blue dye 

 of suitable tint, each of which fades quickly when ex- 

 posed to white light. As it is the light that is absorbed 

 that causes the fading and not the reflected light, on 

 exposure to light of any given colour the only dye or 

 dyes that remain unbleached are those that reflect the 

 same colour as the incident light : a red light, for 

 example, causes the yellow and blue dves to 

 fade, but not the red, and this colour therefore 

 remains in those parts upon which it impinges. The 

 primary difficulty is to prevent further change in the 

 resulting picture, for the very essence of the process 

 consists in the fugitive nature of the dyes employed. 

 This principle of work has recently engaged the atten- 

 tion of several investigators, but no satisfactory method 

 has yet been arrived at. 



All practical methods of colour photography, that is, 

 methods that are practised for the sake of the results 

 that they furnish, are indirect. The light from the 

 coloured object does not produce colour at all. By 

 dividing the light into three suitable parts or colours 

 each may be photographed separatelv so as to give a 

 record of the distribution of its own colour, and by the 

 use of these records a compound print may be made 

 with three suitable colours. Three colours are used 

 for the same reasons that led to the theory that the 

 normal eye can distinguish onlv three colours — the 

 three colour sensations — all the vast variety of tints 

 being due to the excitation of one or more of these 

 simultaneously and in due proportions. Although it 

 has been proved that three colours are sufficient, four, 

 and even five, have occasionally been used to overcome 

 the difficulties of the simpler method. But none of 

 these methods must be confused with the procedure in 

 chromolithography, in which a dozen or more colours 

 may be used, their choice being chiefly, if not entirely, 

 empirical. 



If any two or all three of the colour sensations are 

 excited to the same extent, the same colour effect will 

 be produced whatever the character of the light that 

 causes the excitation. Therefore, the fact that two 

 colours are not distinguishable from each other bv the 

 naked eye is no proof that they are really the same; 

 spectroscopic analvsis of the two coloured lights may 

 reveal a great difference between them. In reproduc- 

 ing colour by photography, therefore, it is not neces- 

 sary (and not often possible) to reproduce the original 

 colours ; it is sufficient to produce a colour that the eye 

 cannot distinguish from the original, that is, one that 

 affects to the same extent each of the three sensations. 

 It was by the application of this principle that Ives 

 worked out the essential conditions for his chromo- 

 scope. It has been stated that Ives's actual apparatus 

 does not illustrate this principle as completely as has 

 been claimed, but whether this is so or not does not 

 affect the principle itself nor the usefulness of it. 



The conditions obtaining in Ives's chromoscope, and 

 in the method of making coloured transoarencies (or 

 lantern slides) which Sanger-Shepherd has made a 

 commercial success, are the two simplest illustrations 

 of three-colour photography. In the first case the three 

 lights are added, for each is transmitted to the eye 

 independently of the others, while in the second case 

 the colours are superposed and the light that passes 

 is only that which is absorbed by neither of the three. 

 In the first case the result is the sum of the transmitted 

 lights, while in the second it is the sum of the absorp- 

 tions that has to be considered. Practically speak- 

 ing, though not quite actually, the colours used in the 

 second case have to be complementary to those in the 



NO. 1823, VOL. 70] 



first case. The first thing to be done is to settle on 

 the three fundamental colours. Ives and others have 

 sought to follow the three-colour sensation spectrum 

 curves founded on the Young-Helmholtz theory of 

 colour, as drawn by Clerk-Maxwell and later by Sir 

 William Abney. Colonel A. F. von Hiibl, in his "Three- 

 Colour Photography," just published in English 

 (translated by Mr. H. O. Klein), arranges a diagram in 

 which the normal spectrum forms a circle with purple 

 between the violet and red, white being at the centre 

 of the circle, and the remainder of the space being 

 filled with whitish tints of the periphery colours. He 

 says that an infinite number of systems of three 

 theoretically correct fundamental colours may be 

 selected by taking any three that are 120° distant from 

 each other. But as blue is the darkest colour it must 

 be one of the three fundamentals, and, of course, red 

 and green follow. Hiibl asserts that Ives's curves are 

 entirely different from those which the theory of Helm- 

 holtz requires, and are based on Maxwell's colour- 

 mixing experiments, in which three spectrum colours 

 were assumed as fundamentals. Hiibl, by means of a 

 similar diagram, but with black (total absorption) at 

 the centre, finds the three fundamental colours for 

 superposition. " Three narrow-banded colours, situ- 

 ated in the colour circle at 120° from each other, 

 form a suitable colour system for trichromatic printing, 

 and a great number of such theoretically possible sys- 

 tems can be ascertained." But here again we are 

 limited to one system. .As " vellow cannot be pro- 

 duced by pigment mixtures . . . this colour must form 

 one of the fundamental colours." "The yellow must 

 be absolutely correct and must not be of a reddish 

 tint," or any neutral pure yellow would be missing in 

 the print. Both methods, therefore, in both cases give 

 no choice for a perfect system, and the practical results 

 as obtained bv those who have worked them out on 

 these rather different plans are very similar. 



But it is not simply a matter of dividing the light 

 into three suitable parts. The object of the division is 

 to get a photographic record of each, and as no photo- 

 graphic plate yet made has a sensitiveness to various 

 colours proportional to their brilliancy to the eye (or 

 rather will give densities by exposure and development 

 proportional to these brilliancies), the colour screens or 

 filters that divide the light into the selected funda- 

 mental parts have to compensate for the deficiencies of 

 the plates used. Now. this compensation can only be 

 done by reducing the light that produces an excessive 

 effect ; it therefore always leads to the necessity for a 

 lengthened exposure. Practically, a lengthened ex- 

 posure means a more costly procedure, if onlv because, 

 for the same capital outlay in apparatus and accom- 

 modation, less work can be done in a given time. 

 But, even disregarding such considerations as these, 

 ordinary plates are so little sensitive to red that it 

 would be hardly possible to get a photograph on them 

 of the red image, because the very protracted exposure 

 would give the opportunity for all sorts of interfering 

 circumstances to produce their characteristic errors. 

 The sensitising of the plates to be used for the green 

 and red elements becomes, therefore, an esserftial part 

 of the procedure. About two years ago. Dr. Miethe, of 

 Berlin, showed some results of three-colour work that 

 attracted considerable attention, the superiority of 

 which was partly due to the use of ethyl red, a cyanine 

 derivative, instead of the cyanine invariably used until 

 then. Plates treated with this sensitiser give with the 

 prismatic spectrum an almost even density from nearly 

 C to the violet, the deficient sensitiveness in the green 

 that most sensitisers (including cyanine itself) give 

 being hardly appreciable. Last year a still better sensi- 

 tiser was introduced, namely, " orthochrome T," and a 

 few months ago this was found to be surpassed by 



