later.
When there are three documents in a family their mutual relationships are sometimes harder to specify. Let A, B, and C be the documents. Suppose A is the common source: perhaps B and C copied it independently; perhaps C only knew A through the medium of B, or B knew it only through C. If B and C have abridged the common source in different ways, they are evidently independent. When B depends on C, or vice versâ, we have the simplest case, treated in the preceding paragraph. But suppose the author of C combined A and B, while B had already used A: the genealogy begins to get complicated. It is more complicated still when there are four, five, or more documents in a family, for the number of possible combinations increases with great rapidity. However, if too many intermediate links have not been lost, criticism succeeds in disentangling the relationships by persistent and ingenious applications of the method of repeated comparisons. Modern scholars (Krusch, for example, who has made a speciality of Merovingian hagiography) have recently constructed, by the use of this method, precise genealogies of the utmost solidity. The results of the critical investigation of authorship, as applied to the filiation of documents, are of two kinds. Firstly, lost documents are reconstructed. Suppose two chroniclers, B and C, have used, each in his own way, a common source X, which has now disappeared. We may form an idea of X by piecing together the fragments of it which occur imbedded in B and C, just as we form an idea of a lost manuscript by comparing the partial copies of it which have been preserved. On the other hand, criticism destroys the authority