case, for example, with the charters issued from the chancery of a prince, care will have been taken to place at the head of each slip the date (expressed in modern reckoning) of the document entered upon it. Nothing is then easier than to group in chronological order all the slips, that is, all the documents, which have been collected. The rule is to use chronological classification whenever possible. There is only one difficulty, and that is of a practical order. Even in the most favourable circumstances some of the documents will have accidentally lost their dates; these dates the compiler is bound to restore, or at least to attempt to restore; long and patient research is necessary for the purpose.
If the documents are not dated, a choice must be made between the alphabetical, the geographical, and the systematic order. The history of the Corpus of Latin inscriptions bears witness to the difficulty of this choice. "The arrangement according to date was impossible, seeing that most of the inscriptions are not dated. From the time of Smetius it was usual to divide them into classes, that is, a distinction was made, resting solely on the contents of the inscription, and having no regard to their place of origin, between religious, sepulchral, military, and poetical inscriptions, those which have a public character, and those which only concern private persons, and so on. Boeckh, although he had preferred the geographical arrangement for his Corpus Inscriptionum Græcarum, was of opinion that the arrangement by subjects, which had been hitherto employed, was the only possible one for a Latin Corpus...." [Even those who, in France, proposed the geographical arrangement] "wished to make an exception of texts relating to the general