But what is the evidence in the case? All that can be brought forward is the nature of the ink, and the appearance of the printed pages. As regards the ink, it differed from that used at a later period in Germany, and was certainly much more fluid; this is shewn by its spreading over the edges, and filling up the loops of the letters. By itself, however, the ink proves nothing as to how the types were impressed; that is to be learnt from the appearance of the pages; and this shews that impressions were taken by a rolling process. A wooden or metallic roller, 7 inches long, with a diameter of not less than 3 inches, and covered with two or three folds of fine woollen cloth, rolled, with a sufficient amount of pressure, over the back of the paper when it was laid upon the types after they had been inked,--or an uncovered roller, if two or three folds of woollen cloth were laid over the paper,--would do all that was needed, and be quite sufficient to account for the practice of printing on one side of the paper only. The statement how such an impression might be taken, is not however a proof that such was in reality the method adopted. That proof lies in the fact, that the capital letters of the columns of text on the left hand margin, and the end letters of the lines on the right, are all more or less blurred, and choked with ink, in a way which only such a cylindrical method of printing would effect; that effect being caused by the first and last contact of the roller with the outer edges of the types. In cylindrical machine printing at the present day, where care has not been taken to guard the edges and ends of the types by ‘bearings,’ similar blurrings may at any time be seen. In the admittedly rude appliances of the earliest printers such appearances would under such a process be at least as plainly shewn. And in the _absolute fac-simile_ which Mr. Humphreys gives, on plate 10 of his work, of the page of the _Speculum_ printed from separable letters with black oleaginous ink, those appearances are most plainly visible. Of course it was possible to print sheets of paper in this way on both sides; but care would have to be taken to guard the first printed side by interposing between it and the roller a waste or setting-off sheet, before the process was repeated. And, as already hinted, by the use of ‘bearings,’ the blurring might be avoided. The difference however between such a process and that invented and put in practice by Gutenberg, is as great as that which exists between the battering rams and catapults of the ancients, and the siege trains of modern artillerists.