Edition 4 is made up in the same way as edition 2. The copy in the library at Lille contains some leaves with text printed upon both sides, seemingly by an error of the printer. The very fact of their existence shows that it was possible to print the text on both sides of the leaf. There must therefore have been some reason other than the ignorance or incapacity of the printer for printing these books on one side only, or, as it is called, anopisthographically. There can be very little doubt that Mr. Sotheby is correct in his conjecture, that ‘the then usual process of taking off the wood engravings by friction, rendered it impossible to effect two impressions back to back, as the friction for the second would materially injure the first. On this account, and on no other, we presume, was the text printed only on one side.’ In the Lille copy above mentioned, two leaves, 25 and 26 (the centre sheet of the third quire), contain printed on their other side the text, not the illustrations, of leaves 47 and 62 (the first sheet of the fifth quire.) From this we learn three things of great importance--1. That the text and the cut were not printed at the same time, and that the text was printed first. 2. That the printer could print the text, for which he used movable type, on both sides of the paper. 3. That the book was printed, not page by page, but two pages at a time. Mr. Ottley was strongly of opinion, after careful examination, that the book was certainly printed two pages at a time. He says, ‘The proofs of this are, I think, conclusive. The upper lines of the text in those two pages always range exactly with each other.... Here and there, in turning over the book, we observe a page printed awry or diagonally on the paper; in such case, if the other page of the same sheet be examined, the same defect will be noticed. Upon opening the two Dutch copies of the edition, which I shall hereafter show to be the fourth at Harlem, in the middle sheet of the same gathering we find, upon comparing them, the exact same breadth and regularity of the inner margin in both, and the lines of the two pages range with each other exactly the same in both copies, which could not be the case had each page been printed separately.’