e reprinted and re-edited every few years. But our
editions should be _good_ ones. 'A good edition should be a complete
edition, ungarbled and unabridged.'[28] Perchance you may prefer to have
them, if it be possible, in the original editions? If so, you will be
wise in your generation, but your purse will need to be a long one
indeed.

Remember that the first edition is not necessarily the best. It
may be, but in the great majority of cases it is not. In addition to the
inevitable clerical mistakes and printer's errors which are almost always
corrected in the second and subsequent editions, the author or editor
frequently interpolates matter which the publication _de ipso_ has
brought to his notice by reviews or correspondence. This is notably the
case in large and important works. 'Scott's Last Expedition,' published
in two large octavo volumes in 1914, rapidly passed through five editions
the same year, corrections being incorporated in each successive edition
(thereby distinguishing them from mere 'impressions'); so that the fifth
edition remains the best, being the most correct. On the other hand, in
the second edition an author sometimes omits passages or makes drastic
emendations from prudential reasons. Then it is that the first edition is
to be sought for in preference to all others, for this alone contains the
author's true opinions on certain subjects. Such instances the
book-lover gradually learns in his journey through the world of books.

But I repeat that, apart from this question of first or later issue, our
editions should be good ones. Good editions are not merely luxuries. The
better the type and paper, the greater our ease in reading, and--most
important of all--the consequent safeguarding of our eyesight.

It is not only type and paper, however, that constitute a good edition.
In addition to these requisites it must contain the recognised text
complete, it must be in a seemly and convenient shape, neither
extravagant nor blatant, and it must not contain a long list of errata.
Of the many qualities that go to make up a good edition, after paper and
print, these are perhaps the most important. But there is another
immediate consideration: _shall it have notes?_ And this raises such a
momentous point that I almost hesitate to approach it. The answer must be
qualified. Provided always that the edition has been superintended (I use
the word advisedly) by a _recognised_ scholar, and that the notes are
f