trically. The mouth has alone to be watched. When the
mouth is adjusted to the lower fiducial line, the scale is exact. It
is a great help having to attend to no more than one varying element.
The only inconvenience is that the image does not lie in the best
position on the plate when the point between the eyes occupies its
centre. This is easily remedied by using a larger back with a
suitable inner frame. I have a more elaborate contrivance in my
apparatus to produce the same result, which I need not stop to
explain.

For success and speed in making composites, the apparatus should be
solidly made, chiefly of metal, and all the adjustments ought to
work smoothly and accurately. Good composites cannot be made without
very careful adjustment in scale and position. An off-hand way of
working produces nothing but failures.

I will first exhibit a very simple but instructive composite effect.
I drew on a square card a circle of about 2-1/2 inches in diameter,
and two cross lines through its centre, cutting one another at right
angles. Round each of the four points, 90 deg. apart, where the cross
cuts the circle, I drew small circles of the size of wafers and
gummed upon each a disc of different tint. Finally I made a single
black dot half-way between two of the arms of the cross. I then made
a composite of the four positions of the card, as it was placed
successively with each of its sides downwards. The result is a
photograph having a sharply-defined cross surrounded by four discs
of precisely uniform tint, and between each pair of arms of the
cross there is a very faint dot. This photograph shows many things.
The fact of its being a composite is shown by the four faint dots.
The equality of the successive periods of exposure is shown by the
equal tint of the four dots. The accuracy of adjustment is shown by
the sharpness of the cross being as great in the composite as in the
original card. We see the smallness of the effect produced by any
trait, such as the dot, when it appears in the same place in only
one of the components: if this effect be so small in a series of
only four components, it would certainly be imperceptible in a much
larger series. Thirdly, the uniformity of resulting tint in the
composite wafer is quite irrespective of the order of exposure. Let
us call the four component wafers A, B, C, D, respectively, and the
four composite wafers 1, 2, 3, 4; then we see, by the diagram, that
the order of exposure