1921 AVRANCHES MANUSCRIPT OF VACARIUS 547 I do not wish to anticipate the account of these manuscripts, which I hope to publish in an edition of Vacarius's work for the Selden Society. But it seems right to offer the following descrip- tion of A, first because it has never been described, and my examination of it has yielded some results ; also because it may give me the opportunity of publishing a specimen of the post- Vacarian glosses of A, which, while they have only a remote interest for an editor of Vacarius, may be of some importance for the general history of the gloss. In order to make clear what follows, a few details about Vacarius's work are necessary. It consists of a text in nine books based on the first nine books of the Code of Justinian ; the books are subdivided into titles composed of extracts from the corresponding titles of the Digest and Code. The text is supplemented by a gloss, which forms an integral part of the original work. This gloss consists mainly of further extracts from the same, and sometimes from other, titles, and also from the novels, the last generally in the summary forms (Authenticae Communes). But this Vacarian gloss also contains unsigned original glosses. They are not very numerous, but they are the most important and characteristic part of the work, which for the rest is a ohrestomathy. Naturally all the manuscripts contain an accumulation of later glosses, and the main problem of an editor of Vacarius is precisely to distinguish the primitive Vacarian gloss. One might expect A to be of less assistance in this matter than it actually is. Layers of subsequent gloss which would be distinguishable in their original manuscript might easily become indistinguishable in a copy. But as a matter of fact the organization of A is directed precisely to keeping the original Vacarian mass distinct. I have collated book 1 of A as to the text and the main gloss carefully, and the rubrics of the other books generally, with occasional titles more in detail. There is a surprising agreement between V, W, and A. One of the problems raised by V is whether its numerous interlinear glosses are Vacarian. A's interlinear glosses are different, but some of V's interlinear glosses occur in A's main gloss, a fact which shows at least their considerable antiquity. W does not help here, as Wenck did not publish its interlinear glosses. There are details in which A agrees with V against W, but on one more important point A follows W against V. I think it clear, for reasons which I will explain shortly, that A is not descended from either V or W, but that a close connexion exists between it and P. The connexion between P and the undoubtedly Norman A recalls StolzePs conjecture * that JP is the Le Bee codex reported 1 Zeitschrift fur Rechlsgeech. vi. 243. Nn2
