We doubt, therefore, whether, in types liable to these accidental depressions of surface, a small artificial hole thus easily simulated would be of any service as a guide to the compositor. A more probable explanation of the appearance seems to be that the head of a small screw or pin, used to fix the side-piece of the mould, projecting slightly on the surface of the piece it fixed, left its mark on the side of the types as they were cast, and thus caused the circular depression observable in the illustrations.[47] Before leaving this subject it may be remarked that the clear impression of the printed matter, despite the laid-on types, which must in either case have been a thin sort, is strong evidence of the softness of the metal in which the fount was cast. The press appears to have crushed the truant types down into the letters on which it lay, and, unimpeded by the obstacle, to have taken as good an impression of the remainder of the forme as if that obstacle had never existed. * * * * * The quantity of type with which the earliest printers found it necessary to provide themselves, turns, of course, upon the question, did the first printers print only one page at a time, or more? M. Bernard considers that the Gutenberg _Bible_, which is usually collated in sections of five sheets, or twenty pages, containing about 2,688 types in a page, would require 60,000 types to print a single section; and if sufficient type was cast to enable the compositors to set one section while another was being worked, the fount would need to consist of 120,000 letters. Others consider that two pages, requiring, in the case of the Gutenberg _Bible_, only 6,000 types, were printed at one time. But even this estimate has been shown to be opposed to the evidence afforded by a considerable number of the incunabula, respecting which it is evident only one page was printed at a time. On this point we cannot do better than quote the words of Mr. Blades. “The scribe,” he says, “necessarily wrote but one page at a time, and, curiously enough, the early printers here also assimilated their practice. Whether from want of sufficient type to set up the requisite number of pages, or from the limited capability of the presses, there is strong evidence of the early books from Caxton’s press having been printed page by page. . . . . Instances are found of pages on the same side of the sheet being out of parallel, which could not occur if two pages were printed together. . . . A positive proof of the separate printing of the pages may be seen in a copy of the _Recuyell of the Histories of Troye_, in the Bodleian; {27} for the ninth recto of the third quaternion has never been printed at all, while the second verso (the page which must fall on the same side of the sheet) appears properly printed.”[48]