200 
NATURE 
[JUNE 29, 1899 
light, but almost perfect absence of any light whatever ; 
in other words, instead of getting yellow we get black. 
Let us consider first how a picture in colour might be 
produced by diffraction. Place a diffraction grating 
(which is merely a glass plate with fine lines ruled on its 
surface) before a lens, and allow the light of a lamp to 
fall upon it. There will be formed on a sheet of paper 
placed in the focal plane of the lens, an image of the 
lamp flame, and spectra, or rainbow-coloured bands on 
each side of it. Now make a small hole in the sheet of 
Lamp Flame 
Vv 
Sa dupes 
paper in the red part of one of these spectra. This hole 
is receiving red light from the whole surface of the 
grating, consequently if we get behind the paper and 
look through the hole we shall see the grating illuminated 
in pure red light over its whole extent. This is indicated 
in Fig. 1, where we have the red end of the spectrum 
falling on the hole, the paths of the red rays from the 
grating to the eye being indicated by dotted lines. Now 
the position of the spectra with reference to the central 
image of the flame depends on the number of lines to the 
inch with which the grating is ruled. The finer the 
ruling the further removed from the central image are 
the coloured bands. Suppose now we remove the grating 
in Fig. 1, and substitute for it one with closer ruling. 
The spectrum will be a little lower down in the 
diagram, and instead of the red falling on the hole, there 
will be green ; consequently, if we now look through 
the hole, we shall see this grating illuminated in green 
light. A still finer ruling will give us a grating which 
will appear blue. Now suppose that the two first gratings 
be put in front of the lens together, overlapping as 
shown in Fig. 2. This combination will form two over- 
lapping spectra, the red of the one falling in the same 
place as the green of the other, namely on the eye-hole. 
The upper strip, where we have the close ruling, sends 
Fic. 2. 
green light to the eye and appears green ; the under strip, 
with the coarser ruling, sends red light to the eye and 
appears red, while the middle portion, where we have 
both rulings, sends both red and green light to the eye, 
and in consequence appears yellow, since the simul- 
taneous action of red and green light on any portion of 
the retina causes the sensation of yellow. In other 
words, we have in superposed diffraction gratings a 
NO. 1548, VOL. 60] 
structure capable of sending several colours at once to 
the eye. 
If we add the third grating, we shall see the portion 
where all three overlap illuminated in white, produced 
by the mixture of red, green and blue light. 
Three gratings with 2000 lines, 2400 lines, and 2750 
lines to the inch, will send red, green and blue light in 
the same direction, or, in other words, to the same spot 
on the screen behind the lens. 
Suppose, now, we have a glass plate with a design of a 
tulip, with its blossom ruled with 2000 lines to the inch, 
its leaves ruled with 2400, and the pot in which it is 
growing ruled with 2750 lines, and place this plate before 
the lens. On looking through the hole we shall see a 
red tulip with green leaves growing in a blue pot. Thus 
we see how it is possible to produce a coloured picture 
by means of diffraction lines, which are in themselves 
colourless. Those portions of the plate where there are 
no lines send no light to the eye, and appear black. 
We have now to consider how this principle can be ap- 
plied tophotography. That photographs whichshow colour 
on this principle can be made, depends on the fact that 
a diffraction grating can be copied by contact printing in 
sun-light, on glass coated with a thin film of bichromated 
Fic. 3. 
gelatine. The general method which I have found best 
is as follows. Three gratings ruled on glass with the 
requisite spacing were first prepared.? 
To produce a picture in colour three negatives were 
taken through red, green, and blue colour filters in the 
usual manner. From these three ordinary lantern-slide 
positives were made. A sheet of thin plate-glass was 
coated with chrom gelatine, dried, and cut up into pieces 
of suitable size ; one of these was placed with the sensi- 
tive film in contact with the ruled surface of the 2000-line 
grating, and the whole covered with the positive repre- 
senting the action of the red light in the picture. An 
exposure of thirty seconds to sunlight impressed the 
lines of the grating on the film in those places which lay 
under the transparent parts of the positive. The second 
grating and the positive representing the green were now 
substituted for the others, and a second exposure was 
made. The yellows in the picture being transparent in 
both positives, both sets of lines were printed superposed 
in these parts of the picture, while the green parts re- 
ceived the impression of 2400 lines to the inch only. 
The same was done for the blue, and the plate then 
washed for a few seconds in warm water. On drying it 
appeared as a coloured photograph when placed in front 
of the lens and viewed through the hole in the screen. 
1 These gratings were ruled for us on the dividing engine at Cornell 
University, through the courtesy of Prof. E. L. Nichols. 
