00 



S6 



NATURE 



[October 26, 191 1 



teachers of both geology and geography; and their 

 value would be much greater to British students if the 

 titles of the plates were repeated in English, for which 

 abundant space on the explanatory sheets. 

 1 s are short and necessarily technical, and 

 their translation, unaided by context, would be often 

 too difficult for the students to whom these photo- 

 graphs should be of most value. J. W. G. 



THREE-COLOUR KIN EMATOGRAPHY. 



IN spite of the large amount of work which has 

 been done upon three-colour photography, it is 

 doubtful whether the results obtained by the newer 

 methods are in any way superior to those obtainable 

 by the original three-colour projection process used 

 by Maxwell in his classic lecture. 

 ' In three-colour photography, three negatives are 

 taken, one through a red, a second through a 

 green, and a third through a blue-violet filter. Having 

 secured the negatives, we may then proceed to pro- 

 duce a three-colour photograph by one of two methods : 

 either we may prepare black positives from our nega- 

 tives and project the three positives by means of the 

 three primary colours, or we may prepare from each 

 negative a print in some substance such as bichromated 

 gelatine, which can be dyed in the colour comple- 

 mentary to the taking filter. The superposition of 

 these three-coloured positives will then give a coloured 

 picture. It is found, however, in practice, that the 

 first method, which is known as the additive process, 

 gives results appreciably superior to the second or 

 subtractive process. The cause of this lies in the fact 

 that the dyes used for preparing the prints cannot 

 completely absorb the colours which they are required 

 to absorb, and, at the same time, completely transmit 

 that portion of the spectrum which they should trans- 

 mit. While the yellow dye is fairly satisfactory in 

 this respect, and the red dye efficiently transmits red 

 light, the red dye always fails to transmit sufficient 

 blue light, and the light blue dye, used for the prints 

 from the red negative, never transmits either sufficient 

 blue light or sufficient green light, the deficiency in the 

 transmission of green being usually very marked, and 

 resulting in the reproduction of all dark green objects 

 as browns, or at best, blacks. 



For the first method or additive process it is, of 

 course, necessary to use three separate projection 

 systems, and owing to the great loss of light through 

 the filters these are advisedly supplied from three 

 separate light sources. Unfortunately the unsteadi- 

 ness of arcs makes it very difficult, if not impossible, 

 to use three separate arcs, and consequently the best 

 svstem is probably that arranged to use three large 

 Nernst burners, the triple lantern being- arranged so 

 that the three projection systems stand vertically one 

 above the other, and the three positives, printed on 

 one plate, can be inserted together and focussed 

 simultaneously, the adjustments for register being 

 made once for all upon the lantern, so that subse- 

 quent slides register automatically. If an arc lamp 

 is used, as in Mr. Ives's apparatus, only one lamp 

 can be used, and the light from this must be divided 

 into three by means oi a system of reflectors, a pro- 

 , dure which unfortunately" is frequently very waste- 

 ful of light and leads only to unsatisfactory results. 



Attempts to apply triple projection methods to kine- 

 matographv have frequently been made, but have not yet 

 come before the public, except in the " Kinemacolor " 

 method, which uses only two colours, abandoning 

 the third, and so eliminating much complication. A 

 two-colour method can, of course, at best be only a 

 compromise, but it would seem at present to be the 

 atisfactory for kinematograph work. 

 NO. 2191. VOL. 87] 



In kinematograph work it is not necessary to pro- 

 ject the colours to be combined simultaneously, as in. 

 the ordinary projection method. The colour> can be 

 successivel} projected, the observer combining them 

 by persistence of vision. The three nej 

 taken in turn on one film, a rotating shutter being 

 arranged behind the lens of the camera carrying the 

 three filters; but if the image is not to flicker stronglj 

 on the screen, it will be necessary for all three filters 

 to be exposed within the ordinary period of exposure 

 of a single kinematograph picture, which involves an 

 exposure for each filter of about one-fiftieth of a 

 second, a rate of work which is not only a gn at 

 strain on the film and apparatus, but represents ver\ 

 brief exposures when it is remembered that the nega- 

 tives must be taken through strongly absorbing red 

 and green filters. 



Attempts have been made to project the three 

 images simultaneously by means of three separate 

 optical systems, as in ordinary three-colour work, but 

 the difficulties of register have been found verj 

 serious, although it is not at all clear why this should 

 be so if both positive and negative sets of pictures are 

 handled simultaneously on films of three times the 

 normal width. 



In the " Kinemacolor " method, the spectrum is 

 divided into two, one taking-filter being orange-red 

 and the other blue-green. In order to get a satisfac- 

 tory rendering of greens, the latter taking-filter must 

 be rather greener than a true blue-green filter would 

 be, and in the resulting picture blue is little distin- 

 guishable from blue-green, while a yellow is rendered 

 rather as an orange than a true yellow. The results 

 are, however, surprisingly good considering the 

 theoretical difficulties of the method and the extreme 

 badness of ordinary two-colour work, and they will 

 doubtless be even better when the promoters of the 

 process have learned that above all things they must 

 avoid in their pictures the two colours which they are 

 using for projecting. These two colours are a pure red 

 and a verv bright blue-green, or minus-red colour, and 

 in some of the pictures at present there is a tendency 

 to use these two colours in the composition of the 

 original scene, so that one gets little girls dressed 

 in red blouses and blue-green skirts, which colours 

 the process renders with only too distressing fidelity. 

 Considering the excellent way in which the process 

 reproduces subdued colours, such as browns, 

 really seems no need for this. A much less serious 

 defect in the process, and one which is, of course, 

 unavoidable., is that when anything moves rapidly it 

 is projected as a series of coloured images, so 11: 

 horse's leg on the screen may appear as a series 

 of alternately red and blue-green legs. 



There is no doubt, however, that kinemacolour i* 

 a success, and is a striking testimony to the 

 practical results which may sometimes be obt. 

 from a theoretically inaccurate system. 



In a reci nl patent it has been suggested that th< 

 successive and synchronous methods of projection 

 might be combined, two projecting systems being i 

 with two sets of pictures, one set representing only 

 one-colour element, such, for instance, as red, whili 

 the other is composed alternately of pictures of thi 

 other two-colour elements, blue and green. A 

 stationary red filter is used in front of one projecting 

 lens, while the second film is projected through a 

 rotating filter, part green and part violet. This would 

 appear to be somewhat more complicated than tlv 

 two-colour system, while the colour rendering should 

 undoubtedly be superior; but the method does not 

 appear to have, at present, reached the stage of prac- 

 tical use. 



C. E. Kenneth Mees. 



