But then the axis of the emergent pencil, instead of returning (as we see it in fig.

1) _towards_ the axis of the telescope, would depart as much _from_ that axis.

Thus there would be no point on the axis at which the eye could be so placed as to receive emergent pencils showing any considerable part of the object.

The difference may be compared to that between looking through the small end of a cone-shaped roll of paper and looking through the large end; in the former case the eye sees at once all that is to be seen through the roll (supposed fixed in position), in the latter the eye may be moved about so as to command the same range of view, but _at any instant_ sees over a much smaller range.
To return to the arrangement actually employed, which is illustrated by the common opera-glass.

We see that the full illuminating power of the telescope is not brought into play.