In the meantime several other improvements along this line were undergoing experiment. Firmin Didot, (1764-1836), a printer of Paris, cast type of a hard alloy, and when his book-pages were composed, made an impression of them on a sheet of soft lead, thus forming a mold. Molten metal was then poured into a shallow tray, and just as this was on the point of solidifying, but still plastic, the lead-mold of the book-page was pressed on the soft metal in the tray. This process called Polytypage, was but partly successful; it could be used only for small pages, and the plates were too often defective. A process similar to this is what Lambinet thought the printers of the latter half of the fifteenth century might have used as one of the probable methods to cast their metal types. These and other experiments, however, were leading to the real stereotyping process which developed later. Early in the nineteenth century, Earl Stanhope, of England, re-introduced Ged's stereotyping process with many improvements. One or more pages of type were locked in a chase, the surface of the type being oiled to prevent the subsequent mold from sticking. The mold was made by pouring a semi-fluid composition of plaster-of-paris mixed with a little fine salt to make the plaster settle solidly. While the plaster was still soft, it was carefully pressed down and rolled smooth on top to give a uniform thickness to the mold and to expel any air there might be in the plaster. When the plaster became solid, it formed a perfect matrix of the type pages. The moisture in those early plaster molds was expelled by baking them in an oven for three or four hours. A later method for drying was practiced by suspending the mold directly over the metal-pots or to float them on the surface of the molten metal. By this means the drying could be accomplished in a half-hour or so. In the process of casting, several of these plaster molds were placed side by side face downward in a special casting-pan. The pan was one and three-quarters or two inches deep, and a lid on the pan screwed down on the back of the molds. By means of a crane the casting pan with its molds was then lowered into the pot of molten metal which ran into the pan at the corners and sides. The mold was allowed to remain ten minutes or so in the metal-pot, or until the face of the inverted mold was entirely filled with the metal. [Illustration] A later method of casting from a plaster mold was to place it in a frame with a smooth, flat plate opposite the face of the mold and to enclose the open space at one end and on the two sides. The casting space thus formed was then turned with the open end up and metal was poured in with a ladle, in a manner similar to the method still employed for casting job-work stereotypes. The distance between the flat plate and the mold was adjusted to make a stereotype plate of the required thickness.