The collation of a book is the enumeration of the number of leaves according to the way in which they are arranged in quires, and this collation should be given whether the quires are signed or not. If there are signatures, there can be no difficulty in counting the number of leaves which go to each quire; but when there are no signatures, as is the case with most books before 1475, the collation is a more difficult matter. The first thing to be looked at, if the book has no MS. signatures, is the sewing, which shows us the centre of the quire,[42] and we can then count from sewing to sewing. This gives us only the halves of two quires; we must then have recourse to the watermarks. In a folio, if one leaf has a watermark, the corresponding leaf which forms the other half of the sheet has none. Again, in a quarto, corresponding leaves have either no watermark, or each half a one. Judging from the sewing and the watermarks, there is rarely any difficulty in making out the collation, the first and last quires being the most difficult to determine with accuracy; the others present no difficulty. It is thus always best to settle the arrangement of the interior quires first, and work from them to the outer ones, which are more likely to be mutilated. [42] It was the custom of many binders in the earlier part of the present century, when they had to rebind an old book, to separate all the leaves and then fix them together in convenient sections, entirely ignoring the original “make up.” A very large number of books in the British Museum were thus misbound, and even the celebrated Codex Alexandrinus was treated in this way. This method of collation by the watermarks is very often useful for detecting made up copies. For instance, in the copy of the thirty-six line Bible in the British Museum, the first and last leaf of the first quire have each a watermark, showing absolutely that one of the two leaves (in this case the first) has been inserted from another copy. In many old books which have been rebound, the outside pages of the quire are very much smoother and more polished than the rest, and may thus be distinguished by touch. This, though a pretty certain test, may mislead, if the book has been misbound, and should only be used in conjunction with the other methods. A little practical work will soon enable the beginner to find for himself various small points, all of which, though hardly worthy of a lengthy description, are useful in giving information, but are only useful when they have been acquired by experience.