AA ROYAL SOCIETY OF CANADA 
a broken line in Fig. 18; it ought to consist of equal black and white 
squares. The original tone, 0:25 L, which is one-quarter the brightness 
of white paper, is represented by equal black and white squares, which 
are one-half the brightness of white paper. In the deep shadows, the 
dots disappear altogether and detail is entirely lost. 
In actual practice the result is slightly different. A shorter exposure 
is given in order to produce fine black dots in the high lights, and all the 
black dots are slightly enlarged. The light tones are improved, but there 
is a corresponding loss in the shadows where more white dots disappear. 
The proportion between the opaque and transparent lines of the screen 
may be varied. As the width of the opaque lines decreases, the light 
tones of the print are improved ; the dark ones become worse. Fig. 19 
illustrates an extreme case, the proportion between the transparent and 
opaque lines being 3415. The light tones are very much improved, and 
the original tone, 0°5 ZL, is correctly represented by equal black and white 
squares, but all the original tones less bright than 0:33 J are translated 
into solid black. From the foregoing we may conclude that an original 
in which the dark tones predominate must be copied with a screen having 
wide opaque lines, while a subject in which light tones predominate is 
better translated by a screen having thin opaque lines, I am inclined to 
believe that the opaque lines should not be much more than one-half and 
not much less than one-third the width of the transparent lines. 
A larger diaphragm than the correct size, or, what is the same thing, 
adjusting the screen further away from the plate, produces approximately 
the same effect as a screen with thinner opaque lines; the opposite effect 
is produced by a smaller diaphragm or by placing the screen closer to the 
plate. An examination of a number of good prints shows that this mode 
of adapting their screen to the character of the original is resorted to by 
some of the best operators, but it must be used sparingly, because it is 
apt to produce reticulation. A reference to Fig. 15 shows what.is taking 
place. With the larger diaphragm, the four triangular portions which 
are lighting the point O under the opaque lines, are increased in area, 
precisely as they would be if the opaque lines were thinner, while with 
the smaller diaphragm their area is reduced as it would be by wider 
opaque lines. 
We already know that with the chess-board screen we must look to 
changes in the size of the diaphragm to produce 
the effect of unequal ruling of the cross-lined 
screen. An increase in size, as in Fig. 20, produces 
illumination at O under the centre of the centre 
of the opaque square by the four uncovered cor- 
ners of the diaphragm. Making the area of these 
; four corners one-eighth of the total area of the 
Fie. 20. aperture gives the dots of Fig. 18 ; the dots of Fig. 
19 are obtained by a still larger diaphragm, the four uncovered corners 

