ubmitted to the Author, until these have been made: a
careful Reader in the Printing Office will also sometimes draw the
Author's attention to some Word or Sentence, which appears to be
susceptible of improvement, and which might otherwise have passed
unnoticed; this is, however, not always done, unless requested.

In Correcting a Proof, for the reasons already given, as few alterations
as possible should be made; when these are, however, unavoidable, it
would be advisable to observe this Rule, namely--always if possible, to
insert in a Line or Page, as much as is taken out, or vice versa; this
is in a great majority of instances very practicable; and the advantage
of it is, that it will avoid what is technically called _Overrunning_.
This will, perhaps, be best explained by referring to the Corrected
Proof (p. 40) in the 3rd line of which, it will be seen that the word
_for_ is marked out, and the word _of_ inserted in its stead; which, it
will be perceived by the opposite Revised Page, has occasioned no
alteration beyond the line; but at line 17 there is an insertion marked
without an omission; which would have rendered it necessary to carry as
many lines as were inserted to the next Page, if the Page had been
previously filled up in the usual way. This is called _Overrunning_, and
often requires that each subsequent Page should be altered to the end of
a Chapter, or if the work is continuous, to the end of all that has at
that time been set in Type.

There is also another point to be observed; which is, that where Revises
are considered necessary, as few as possible should be required, each
Revise requiring the repetition of the process already described in
striking off a Proof, and which will not only occasion additional
Expense, but will also frequently cause considerable delay in the
progress of the Work. Generally speaking, if the Corrections are clearly
marked, and not very numerous, the final Revision may be safely
entrusted to the care of a skilful Printer. If any error should escape
the notice of the Author, or Corrector, and be Printed off, it may be
corrected by Re-printing the leaf in which it occurs, which is called a
Cancel. This is, however, seldom necessary, when the error is clearly
typographical.

It is frequently a convenience to the Author to have two proofs of each
sheet, one to be returned corrected, the other to be retained for
reference.

It is not, perhaps, generally known, that Works Prin