ips; or they may be
represented, as often as may be required, on reference-slips. Moreover,
the method of slips is the only one mechanically possible for the
purpose of forming, classifying, and utilising a collection of documents
of any great extent. Statisticians, financiers, and men of letters who
observe, have now discovered this as well as scholars.

The method of slips is not without its drawbacks. Each slip ought to be
furnished with precise references to the source from which its contents
have been derived; consequently, if a document has been analysed upon
fifty different slips, the same references must be repeated fifty times.
Hence a slight increase in the amount of writing to be done. It is
certainly on account of this trivial complication that some obstinately
cling to the inferior note-book system. Again, in virtue of their very
detachability, the slips, or loose leaves, are liable to go astray; and
when a slip is lost how is it to be replaced? To begin with, its
disappearance is not perceived, and, if it were, the only remedy would
be to go right through all the work already done from beginning to end.
But the truth is, experience has suggested a variety of very simple
precautions, which we need not here explain in detail, by which the
drawbacks of the system are reduced to a minimum. It is recommended to
use slips of uniform size and tough material, and to arrange them at the
earliest opportunity in covers or drawers or otherwise. Every one is
free to form his own habits in these matters. But it is well to realise
beforehand that these habits, according as they are more or less
rational and practical, have a direct influence on the results of
scientific work. Renan speaks of "these points of private librarianship
which make up the half of scientific work."[95] This is not too strong.
One scholar will owe a good part of his well-deserved reputation to his
method of collecting, while another will be, so to speak, paralysed by
his clumsiness in that particular.[96]

After having collected the documents, whether copied _in extenso_ or
abridged, on slips or loose leaves, we classify them. On what scheme? In
what order? Clearly different cases must be treated differently, and it
would not be reasonable to lay down precise formulae to govern them all.
However, we may give a few general considerations.

II. We distinguish between the historian who classifies verified
documents for the purposes of historical