ed thousand inscriptions nearly all of
which belong to several categories; on the other hand, each monument has
but one place, and a very definite place, in the geographical
order."[99]

The alphabetical arrangement is very convenient when the chronological
and geographical arrangements are unsuitable. There are documents, such
as the sermons, the hymns, and the secular songs of the middle ages,
which are not precisely dated or localised. They are arranged in the
alphabetical order of their _incipit_--that is, the words with which
they begin.[100]

The systematic order, or arrangement by subjects, is not to be
recommended for the compilation of a _Corpus_ or of _regesta_. It is
always arbitrary, and leads to inevitable repetition and confusion.
Besides, given collections arranged in chronological, geographical, or
alphabetical order, nothing more than the addition of a good table of
contents is needed to make them available for all the purposes which
would be served by a systematic arrangement. One of the chief rules of
the art of _Corpus_ and _regesta_-making, that great art which has been
carried to such perfection in the second half of the nineteenth
century,[101] is to provide these collections, whatever the grouping
adopted, with a variety of tables and indexes of a kind to facilitate
the use of them: _incipit_ tables in chronological _regesta_ which lend
themselves to such treatment, indexes of names and dates in _regesta_
arranged by order of _incipit_, and so on.

_Corpus_ and _regesta_-makers collect and classify for the use of others
documents in which, at any rate in _all_ of which, they have no direct
interest, and are absorbed in this labour. Ordinary workers, on the
other hand, only collect and classify materials useful for their
individual studies. Hence certain differences arise. For example, the
arrangement by subjects, on a predetermined system, which is so little
to be recommended for great collections, often provides those who are
composing monographs on their own account with a scheme of
classification preferable to any other. But it will always be well to
cultivate the mechanical habits of which professional compilers have
learnt the value by experience: to write at the head of every slip its
date, if there is occasion for it, and a heading[102] in any case; to
multiply cross-references and indices; to keep a record, on a separate
set of slips, of all the sources utilised, in order to avoid the