onable size. Preference must be given to those which
are essential, the number of those which are cited merely on methodical
grounds being made as small as possible. It is hardly necessary to add
that the negative evidence furnished by quotations is generally of
little value; one can seldom, that is, be certain that the lexicographer
has actually found the earliest or the latest use, or that the word or
sense has not been current during some intermediate period from which he
has no quotations.

Lastly, a much more important place in the scheme of the ideal
dictionary is now assigned to the _etymology_ of words. This may be
attributed, in part, to the recent rapid development of etymology as a
science, and to the greater abundance of trustworthy data; but it is
chiefly due to the fact that from the historical point of view the
connexion between that section of the biography of a word which lies
within the language--subsequent, that is, to the time when the language
may, for lexicographical purposes, be assumed to have begun, or to the
time when the word was adopted or invented--and its antecedent history
has become more vital and interesting. Etymology, in other words, is
essentially the history of the _form_ of a word up to the time when it
became a part of the language, and is, in a measure, an extension of the
history of the development of the word in the language. Moreover, it is
the only means by which the exact relations of allied words can be
ascertained, and the separation of words of the same form but of diverse
origin (homonyms) can be effected, and is thus, for the dictionary, the
foundation of all _family history_ and correct _genealogy_. In fact, the
attention that has been paid to these two points in the best recent
lexicography is one of its distinguishing and most important
characteristics. Related to the etymology of words are the changes in
their form which may have occurred while they have been in use as parts
of the language--modifications of their pronunciation, corruptions by
popular etymology or false associations, and the like. The facts with
regard to these things which the wide research necessitated by the
historical method furnishes abundantly to the modern lexicographer are
often among the most novel and interesting of his acquisitions.

It should be added that even approximate conformity to the theoretical
requirements of modern lexicography as above outlined is possible only
under condition