550 AVRANCHES MANUSCRIPT OF VAGARIUS October between 50 and 51, 59 and 60, 87 and 88, 88 and 89, 89 and 90, 153 and 154, 169 and 170. Another obvious fact is that, since the codex was in its present binding, three considerable series of folios have been wrenched out, namely at the beginning, between fos. 57 and 58, and at the end. Moreover, as the number of the book is written in coloured Roman numerals at the top of each recto, disorder in the binding is at once apparent. The folios are normally in sheaves of eight. Fo. 1 a begins with the rubric De Summa Trinitate (Cod. 1,1; V, W, 1, 1), followed by Imperator cunctos populos in alternate blue and red capitals. Fo. 1 is the first of a complete sheaf of eight ; the missing Prologus must have occupied 4-5 folios (3 J in P, probably 4 in V). If as in W, P there was an index (in B at the end ; V, R defective) before the prologue, we may conjecture an avulsion of 8 folios, a complete sheaf. Book 1 (fos. 1 a-12 a) and book 2 (fos. 12 a-34 b init.) are complete. Book 3 (fos. 34 b-57 b) is incomplete, owing to a gap created by an avulsion between fos. 57 and 58, fos. 56 and 57 being torn in the process. The gap is : fo. 57 b expl. ' . . . non est ambiguum . . . ' (Cod. 3, 36, 18 ; V, W, 3, 49), fo. 58 a inc. ' . . . qui in earn rem inpensi sunt . . . ' (Dig. 17, 2, 52, 15 ; V, W, 4, 38). Thus book 3 lacks 6 titles and a fraction at the end, and book 4 lacks 37 titles and a fraction at the beginning. The lost matter occupies V fos. 62 a, 1. 17-93 b, 1. 30, say 31 i folios. This means 35-40 missing folios in A, and as fos. 56 and 57 are the first of a sheaf of which the rest is gone and fo. 58 is the second of a sheaf complete but for one folio, the probability is that the actual number of missing folios is 39, made up of the 6 + 1 odd folios and 4 complete sheaves of 8. Book 4 (fos. 58 a-83 b, col. 1) is in the remainder complete, as is book 5 (fos. 84 a-94 a). So is book 6, occupying fos. 94 a-114 b, with omission of fos. 110-13 which are correctly shown by the coloured numeral IX, recto top, to belong to book 9. Book 7 (inc. fo. 115 a) is incomplete, even after the correct order of the folios is restored. The correct order is : 115, 116, | 119, 120, 117, 118, 123, 124, 121, 122, | 131-8 | 125 (book 7 expl. fo. 125 b). The gap is between fo. 116 b expl. ' ... in suo ultimo elogio . . . ' (Cod. 7, 15, 3 ; V, W, 7, 11) and fo. 119 a inc. ' . . . tempus certum uel incertum . . . ' (Cod. 8, 54 (55), 3 ; V, W, 7, 19), covering the matter of V fo. 146 b, 1. 21, to fo. 153 b, 11. 31, 32, i. e. over 7 folios. There is no sign that the present binding of A ever held the missing folios, which probably made a complete sheaf of 8, as fos. 115-16 are the 7th and 8th of a complete sheaf (fos. 105-16, fos. 110-13 interpolated), and as fo. 119 seq. in the order given above could be rebound as another complete sheaf. Book 8 occupies fos. 125 b, 126-30, 139-68 a, but there is a gap between fo. 127 b (expl. ' . . . post annum uero de eo quod . . . ' Dig. 43, 16, 1 pr. ; V, W, 8, 12 init.) and fo. 128 a (inc. . . . ' possessionem estimandam quanti ipsa res est ... ' Dig. 43, 17, 3, 11 ; V,' W, 8, 14 fin.), corresponding to V fo. 171 a, 1. 6 to fo. 172 b, 1. 16, equivalent say to 2 folios of A. These 2 folios would complete a sheaf with A fos. 125-30 as its central double sheet ; but there is no external evidence of a tear from the present binding. Book 9 runs without a break from fo'. 168 a to fo. 172 b (expl. ' ... in hac lege culpa lata . . . ' Dig. 48, 8, 7 ; W, 9, 13 ; V def.), the end of the work
