Reviews. 36 1
sen's Eddalaeren without warning the reader that Finn's con- ception of the world-tree has been discredited. It should also have been noted that Sir John Mandeville's book is not a record of travel, but a hash of medieval geography and wonders. But these are minor matters.
Corpus Inscriptionum Atticarum : Appendix continens DEFixiONUM TABULAS. Ed. R. Wrench. Bcrolini, 1897,
This volume will be of great service to students of ancient magic. The Attic defixiones number about 200, and are chiefly un- published documents in the possession of the editor. In his preface he classifies and cites in full all the published defixiones from other parts of the Greek world known to him, so that the book is in fact a •' Corpus Defixionum Grsecarum."
These imprecations are always written on leaden tablets, which are then rolled up and made fast with a brass nail. This was one of the ordinary methods of letter-writing (see the business letter outwardly similar to the other tablets quoted here, Prczf. p. ii.), and was doubtless originally selected for the curses owing to its permanency. Subsequently magic virtues were attributed to the material, the writer stating his desire that his victim may be of no price or cold as the lead ; and the magic virtue attributed to the nail which closed the tablet seems to have generated the almost universal phrase Karabw rbv heiva, " defigo M vel. N," " I fix down M or N." The tablets were " posted " to the infernal gods by being placed in tombs or fixed into tomb-stones.
What is the origifial reason of the common practice of writing the curses backwards from right to left ? We find a symbolical virtue attributed to this in some of the texts, the curser wishing that as these letters are backwards, so things may go backwards with the victim; but the primary reason was doubtless simpler, and had reference to the powers to whom the curses were addressed. In a good many cases the inscriptions are unintelligible, i.e. the letters of the formula are, so to speak, shuffled, and in one or two Attic examples (Nos. 55, 76) after each curse the name of the victim is repeated in this puzzle-fashion. These are interesting as showing that the writers at least did not intend the cryptograph