176 THREE-DIMENSIONAL PHOTOGRAPHYtion is to reproduce that scene so that all relative distances, in all
directions, will remain visually the same as in the original.
This in turn involves the existence of an imaginary space lattice,
which is a huge skeleton enclosure founded by imaginary lines,
much as the lines of latitude and longitude are shown upon maps.
It makes no difference how large or complex this lattice is, because
it only exists in your mind. Within this skeleton we now consider
the step of placing the reproduction of our scene. It is evident that
no matter how many slides are shown, it is desirable to have the
images of all objects which happen to be just ten feet away, always
fall into precisely the same position within our imaginary frame-
work-or to put it more clearly, they should appear to be ten feet
away, not eight nor 12 but ten.
If you have the conditions of projection fixed by having the
projector at a distance which will just fill the screen; and you are
using some definite slide separation it follows that every slide pro-
jected shall have its infinity homologous separation exactly the
same as all of the others. So we come to the conclusion that the
same type and size of mount shall be used for all slides regardless
of the distance of the object shown in the slide. (We shall discuss
closeup separation as a part of an arbitrary modification of para-
projection).
It has already been explained that for realistic reproduction it
is necessary that the distances in depth and the distances in width
of the scene shall retain their original ratios. Of course the distance
in depth is the parallax effect and as parallax is fixed at the time
of the camera exposure, it is impossible to alter the distances in
depth displayed by the slide. Many of you will at once exclaim
about that obvious error; but it is not an error, it is absolutely
true. If you photograph a six-foot log end-on, that log is going to
have a length of six feet in the picture. After reading the explana-
tion you will agree, because our problem is the very elusive one
of relative sizes.
The width (or height) of the field is extremely variable, far too
much so for convenience, because every time you go nearer to or
farther from the screen, the screen becomes larger or smaller. If
you set up two sheets of cardboard of exactly the same size, and
place one of them twice as far away as the other, both being so
placed that they can be seen simultaneously, it is obvious that the