1920 REVIEWS OF BOOKS 263 since the very first page of this volume will provide them. In two succes- sive lines of the Terre date of Shropshire are two place-names, Laverleg' and Cheworthin. In the first case the chancellor's roll gives the correct reading Claverleg' ; but in the second case the reading of that manuscript is also corrupt and gives ' Chelewrthin ' ; so that for a correct reading the editor has been compelled to go back to the roll of 25 Henry III, which reads correctly * Chesworthin '. Under Yorkshire, again in the Terre date (p. 22), the pipe roll reads ' Et in quietancia terre Osberti de Brigham quam primus Henricus rex dedit Waltero filio Walteri filio Algrim de Frimareis ' ; and here the chancellor's roll is again right, and omits ' Osberti ' and later reads ' Waltero filio Osberti ' instead of ' Walteri filio Walteri '. And so we get a clear proof that the chancellor's roll is an independent text, not a copy of the pipe roll, as the Dialogus suggests. It is unnecessary to multiply instances ; these are enough to show that, at any rate in the reign of Henry III, the pipe and chancellor's rolls are not free from corruption, and that the task of constructing a correct text will test the critical skill and knowledge of any editor, especially when he is dealing with the numerous personal and place-names contained in them. They are, in fact, no more to be trusted than any other texts derived by copying from other manuscripts, but share the common weaknesses of their kind. Dr. Cannon has made no attempt to construct a critical or correct text or to indicate to the reader which of the variant readings cited by him is to be preferred. All that he has aimed at is to give an accurate representation in print of the pipe roll with a complete collation of the readings of the chancellor's roll and very occasional quotations from other rolls. He was well aware that there were errors in his text. He does not seem to have understood that these errors were the necessary result of his determination to concentrate his efforts on the transcription of his manuscripts rather than on the criticism of them, and that he was in fact exposing himself without defence to the snares that medieval scribes delight in laying for the unwary. On the first page of his edition the reader will find the name ' Alicie de Harcuret ' with a number of variant readings. It need hardly be said that ' Harcuret ' is not a likely form of Harcourt ; nor is it in the original roll. The misreading, which need trouble no one of experience, arises from the fact that the pipe roll scribe makes an e which is indistinguishable from a c, and the editor finding
- Harec't ' has convinced himself that the e was a c and the c an e, and that
the mark of contraction was over the first letter. On page 89 the reader will find the name ' Leonius de Maliniers ' ; it would be unfair to blame any one for misreading this name, which has been misread by many readers, until it has finally ended in the absurd form Manvers, absurdly conferred upon an English peer. But it may be worth while to point out that the true form of the name is Malnuers, or Maunuers, and to add the guess that its origin must be looked for in the department of the Orne and in the charters of St. ifivroul. This is, of course, the result of leaning too heavily on the staff of palaeography, and forgetting that no amount of care in reading is an efiicient substitute for a knowledge of the possible forms of words. In itself the blunder is wholly unimportant, but it is the type of a blunder that runs through the book and now and then
