Casting in sand was an art not unknown to the silversmiths and {10} trinket-makers of the fifteenth century, and several writers have suggested that some of the early printers applied this process to typefounding. M. Bernard[20] considers that the types of the _Speculum_ were sand-cast, and accounts for the varieties observable in the shapes of various letters, by explaining that several models would probably be made of each letter, and that the types when cast would, as is usual after sand-casting, require some touching up or finishing by hand. He shows a specimen of a word cast by himself by this process, which, as far as it goes, is a satisfactory proof of the possibility of casting letters in this way.[21] There are, indeed, many points in this theory which satisfactorily account for peculiarities in the appearance of books printed by the earliest rude Dutch School. Not only are the irregularities of the letters in body and line intelligible, but the specks between the lines, so frequently observable, would be accounted for by the roughness on the “shoulders” of the sand-cast bodies.[22] An important difficulty to be overcome in type cast by this or any other primitive method would be the absence of uniformity in what letter founders term “height to paper.” Some types would stand higher than others, and the low ones, unless raised, would not only miss the ink, but would not appear at all in the impression. The comparative rarity of faults of this kind in the _Speculum_, leads one to suppose that if a process of sand-casting had been adopted, the difficulty of uneven heights had been surmounted either by locking up the forme face downwards, or by perforating the types either at the time of or after casting, and by means of a thread or wire holding them in their places. The uneven length of the lines favours such a supposition, and to the same cause Mr. Ottley[23] attributes the numerous misprints of the _Speculum_, to correct which in the type would have involved the unthreading of every line in which an error occurred. And as a still more striking proof that the lines were put into the forme one by one, in a piece, he shows a curious printer’s blunder at the end of one page, where the whole of the last reference-line is put in upside down, thus:― [Illustration] {11} A “turn” of this magnitude could hardly have occurred if the letters had been set in the forme type by type.