It is equally clear from the Haarlem tradition that the art of casting metal type was the second stage in the invention, a development or outcome of the primary idea of "movable letters," for Junius says that Coster "afterwards changed the beechen characters into leaden, and the latter again into tin ones." Theod. Bibliander, in 1548, was the first to speak of movable wooden type and to describe them. First they cut their letters, he reports, on wood blocks the size of an entire page; but because the labor and cost of that way was so great, they devised movable wooden type, perforated and joined one to another by a thread. Bibliander does not say that he had ever seen such type himself, but Dan Speckle or Specklin (d. 1589) who ascribed the invention to Mentelin, asserts that he saw some of these wooden type at Straussburg; and Angelo Roccho asserted in 1591 that he had seen at Venice type perforated and joined one to another by a thread, but he does not state whether they were of wood or of metal. There is a theory also that between block-printing and printing with movable cast-metal type there was an intermediate stage of printing with "sculpto-fusi" type; that is, a type of which the shank had to be cast in a quadrilateral mold and the characters or letters engraved afterwards by hand. This theory was suggested by some one who could not believe in wooden type and yet wished to account for the marked irregularities of the type used to print the earliest books. Granting that all the earlier works of typography preserved to us are impressions of cast-metal type, there are still differences of opinion, especially among practical printers and type-founders, as to the probable methods employed to cast them. It is considered unlikely, although not impossible, that the invention of printing passed all at once from xylography to the perfect typography of the punch, matrix, and mold. The types that Coster made and used were supposed to have been manufactured in one of three or four probable ways. Bernard believed that the first movable cast-metal type were molded in sand, since that method of casting was known to the silversmiths and trinket-makers of the fifteenth century. In substantiation of his theory he exhibits a specimen of a word cast as a unit for him by this process, roughly similar to a modern linotype slug. A second suggested mode was that of casting in clay molds, by a method very similar to that used in the sand process, and resulting in like peculiarities and variations in the type. Ottley, in his "Invention of Printing," was the chief exponent of this theory. He believed that type were made by pouring molten lead into molds of clay or plaster, after the ordinary manner used from time immemorial in casting statues and other articles of metal.