The dissimilarity of the small types has been made greater by faults of type-founding and of presswork. In all copies of the _Speculum_ the careful observer will see the impressions of types with imperfect faces. There are many half-formed letters, with little peculiarities of appearance which can be satisfactorily explained only by the conjecture that the types in leaving the mould, carried with them the impress of defects in the matrices. We can see that the types were unequal in height, and that the over-high types have been flattened out under impression. This flattening-out of the soft metal has produced a strange appearance of compactness, making letters that were really separate seem connected. The ink, which [p300] was sometimes thin and gray and sometimes thick and strong black, was applied by an imperfect method which has filled the counters of some letters until they are almost illegible, while it has not fairly covered the faces of other letters. The singular irregularities of a collection of types, apparently new on one page and worn-out on another, which have provoked the astonishment of many critics, are chargeable, not to the condition of the types, but to faulty methods of inking and impression. Few persons have a proper notion of the changes that can be given to the appearance of the best modern types by substituting wet for dry paper, hard for light impression, and thin for thick ink.[155] How the types of these and of other early books were founded cannot be learned from the vague descriptions of the early chroniclers of typography. We have to conjecture the process from the workmanship of the books. The discrepancies in the bodies and the imperfections of the faces indicate that the process was rude and unscientific, and that the mould was not of metal. It is possible that the maker of these types followed the example of other founders in metals, and made types in moulds of sand.[156] There are some peculiarities in his types which almost confirm this conjecture. The difficulty encountered in fitting matrices to these moulds, or in adjusting the mould of the face of the letter in proper position on the body, a difficulty that calls for no explanation, may be the reason why the types are so often out of line, crookedly set on body and of irregular height to paper. The feebleness of the sand mould, its liability to damage, and the necessity for its frequent renewal are, possibly, the reasons why we find in the [p301] impressions of the unknown printer types of so many bodies, and with such singular defects.[157] The rounded edges, spotted stems and deficient lines of many of the letters seem the faults of types unskillfully founded in moulds of sand, from metal insufficiently hot, poured in without the force that is needed to make it penetrate all the finer lines of the matrix.[158]