86 Transactions of the Royal Microscopical Society. 
What that colour should be is incapable of determination gene- 
rally, as the capacity for appreciating light varies with different 
individuals. 
If, for instance, we take 43 in the spectrum as being suffi- 
ciently luminous for vision, we find the maximum, as far as seeing 
is concerned, to be 118,000 to the inch; but as the non-luminous 
chemical rays remain in the field after the departure of the visible 
spectrum, a photographic image of lines much closer together than 
those named might be produced. _ 
How little is gained in “ resolving” power by such an excessive 
aperture is at once seen when it 1s considered how slowly the 
sines of the larger angles increase, a reduction of the angle from 
180° to 128° causing a reduction of 10 per cent. only in the 
resolving power, with an immense increase in the general utility 
of the glass; or, if reduced to 1063°, we still have a resolving 
power equal, on the same hypothesis, to 94,400 lines to the inch. 
The next experiments are made with crossed gratings, and give 
equally important results. 
These gratings are prepared by ruling two sets of lines through 
silver films as before, one set being ruled on the under side of a 
thin glass cover and the other on the slide; the two pieces of glass 
with the limes in contact at an angle of 60° are cemented together 
with Canada balsam, and of course give rhomboid markings over 
the entire structure (Fig. 13). 
4th Hauperiment.—The object of the first experiment with 
crossed gratings is to show that with a certain arrangement of the 
incident light both sets of real lines disappear, and are replaced 
by one set of perfectly distinct spurious lines parallel to a diagonal 
of the rhombic figure (Fig. 14). This is effected by using a single 
slit stop, with the slit in the direction of one of the diagonals, when 
the spurious lines will appear parallel to the other diagonal, and 
therefore at right angles to the slit. 
If a crossed slit, as in Fig. 15, is used, two sets of spurious 
lines willappear at right angles to each other (Fig. 16), although 
the real lines, from which they originate, are at an angle of 60°. 
The reason of this will be at once understood, from the previous 
experiments, to arise from the admission of two sets of spectra, 
the directions of which are parallel to the diagonals. 
An experiment identical in principle is shown on a rectangular 
grating, Fig. 17, in which with a slit admitting one spectrum, such 
as is seen in Fig. 19, both sets of lines (vertical and horizontal) 
disappear, and are replaced by one set (Fig. 20) intersecting the 
squares formed by the real lines, and therefore closer in the pro- 
portion 1 : 4/2. 
5th Eaperiment.—The object of this experiment, which is per- 
haps the most important of all, is to show that with only one row 
