0 CCC, 300
III, 3 XL, 40 CCCC, 400
IV, 4 L, 50 D, 500
V, 5 LX, 60 DC, 600
VI, 6 LXX, 70 DCC, 700
VII, 7 LXXX, 80 DCCC, 800
VIII, 8 XC, 90 CM, 900
IX, 9 C (centum), 100 M, (mille), 1,000
X, 10

Now, when the early printers came to apply dates of publication to the
books they issued, (and here is where their methods of notation become
most important to librarians) they used precisely these methods. For
example, to express the year 1695, they printed it thus: MDCVC, that
is--1000+500+100+100-5. But the printers of the 15th century and later,
often used complications of letters, dictated by caprice rather than by
any fixed principles, so that it is sometimes difficult to interpret
certain dates in the colophons or title-pages of books, without
collateral aid of some kind, usually supplied to the librarian by
bibliographies. One of the simpler methods of departure from the regular
notation as above explained, was to substitute for the letter D (500) two
letters, thus--I[inverted C], an I and a C inverted, supposed to resemble
the letter D in outline. Another fancy was to replace the M, standing for
1,000, by the symbols CI[inverted C]--which present a faint approach to
the outline of the letter M, for which they stand. Thus, to express the
year 1610, we have this combination--CI[inverted C] I[inverted C] CX,
which would be indecipherable to a modern reader, uninstructed in the
numerical signs anciently used, and their values. In like manner, 1548 is
expressed thus: MDXLIIX, meaning 1000+500+40+10-2. And for 1626, we have
CI[inverted C] I[inverted C] C XXVI.

As every considerable library has early printed books, a librarian must
know these peculiarities of notation, in order to catalogue them
properly, without mistake as to their dates. In some books, where a
capricious combination of Roman numerals leaves him without a precedent
to guide him to the true date, reference must be had to the
bibliographies of the older literature, (as Hain, Panzer, etc.), which
will commonly solve the doubt.

As to the mechanics of catalogue-making, widely different