¹ My references to Köhler, unless otherwise stated, are to this work. The title quoted for it, when necessary, is ‘Die Schule von Tours.’ References to the plates are prefixed with ‘K.’

^1a I will refer to this work as ‘Survey,’ and prefix references to the plates with ‘R.’

^1b Göttingische Gelehrte Anzeigen, 193 (1931), 321–359. I will refer to this conjoint review as ‘G. G. A.’ The editors kindly allowed me to see a proof of it, but the present article was in print before I could avail myself of the criticisms of Köhler and Dom de Bruyne. I can do them only partial justice here.

^1c Mémoire sur l'Établissement du Texte de la Vulgate (Collectanea Biblica Latina, VI), Rome and Paris, 1922, pp. 267–297.

^1d On this see Harvard Theological Review, XVII, 1924, 197–264. A convenient review of recent utterances on critical method is given by F. Peeters, ‘Les différents systèmes de classement des manuscrits,’ in Revue de l'Université de Bruxelles, 1931, pp. 468–485.

² G. G. A., pp. 332–336, 347–351.

³ Several of my reviewers have doubted the existence of my Irish Period in the history of the script, and their criticism may be just. The matter needs further discussion, however, and I will revert to it in the second volume planned for the series, to be called The Earliest Books of Tours.

⁴ I stated also that this hand is likewise found in my No. 32, Paris, Ste. Geneviève 1260, but I am now uncertain of the identity.

⁵ Though not completely certain, I am ready to change my estimate of the script of this book and to associate it with those manuscripts which are both IVa and IVb in style. Hand B (R, Plate XXXI, 3, 4) is very near to the Regular Style. Köhler refers to the book as a “Kuriosum” (p. 43), applying that epithet, I assume, merely to the rude picture of a baptism found on fol. 260 (K, 3e; R, XXXI. 2). I am not sure that this is not a later addition.

⁶ See my article on Dodaldus, written as a supplement for that of Dom Wilmart in Speculum, VI (1931), 587–599.

^6a Of my list of 232 manuscripts, Köhler, in his review (G. G. A., pp. 323 ff.), discusses or mentions 56, most of which he thinks ought to be weeded out. Of these, as he makes clear, I had indicated 18 as doubtful. In the case of three (Nos. 49, 50, 57) it is not the attribution to Tours but the (tentative) attribution to Marmoutier that he calls in question. Seven on the list are books regarded as dubious by me, but as certainly products of Tours by him. There are then but 28 in the list of 232 that he would exclude. 1 am gratified by this result.

⁷ See Köhler, p. 83.

⁸ Köhler speaks (p. 34) as though this oldfashioned idea had been exploded by palaeographical investigation. But eminent palaeographers like Steffens, Chroust, Prou, and Fischer (see The Vatican Livy, p. 23) have not abandoned an Alcuinian date for the Bamberg Bible.

⁹ There has of course been much discussion of the centre or centres with which these works should be associated. I am subscribing to what seem to me the unassailable arguments of the late lamented Rudolf Beer in Monumenta Palaeographica Vindobonensia, I, 1910, 29–68.

¹⁰ First in B. N. lat. 11514, see p. 115. For the Dagulf Psalter, done before 795, see e. g. Mon. Pal. Vind. I, pl. 22, and for the Godescalc Gospels, done between 781 and 783, of which I possess photographs, fol. 4^r.

¹¹ Not 796, as Köhler has it, p. 87.

¹² P. 87.

¹³ Cf. Survey, pp. 41–43, and what is said of the ornamentation of the Golden Gospels preserved in Tours 22, with particular attention to the vine-decoration, pp. 44, 102 f., plate XXXIV.

¹⁴ See Delisle, ‘Les Bibles de Théodulf,’ Bibliothèque de l'École de Chartes, XL, 1879, 5–47. I have had the privilege of examining both of these books with some care, and would express my thanks both to my friends Messieurs Omont and Lauer of the Bibliothèque Nationale and to the Abbé Joseph Vacher, canon of the Cathedral at Le Puy, who put the beautiful volume at my disposal in the summer of 1929 and wrote most informing answers to various questions which I later asked him.

¹⁵ See S. Berger, Histoire de la Vulgate, 1893, p. 148. Of course there is a possibility, to which I referred in my review of Dom Quentin (Harvard Theological Review, XVII, 1924, 222), that the books of Le Puy and Paris were copies made of the original of Theodulf and therefore that they might have been written after his death (821). As I ponder the matter, this possibility seems far from being a probability.

¹⁶ P. 250.

¹⁷ See Survey, Index, p. 235, s. v.

¹⁸ Minor differences doubtless would appear if every page were measured. Those that I selected happened to be exactly the same. We can say at any rate that the same norm was followed in both books. The Paris book has simply lost more of its margins. This is a good example of the necessity of recording the script-space as well as the size of the volume.

¹⁹ This is another good instance to show the importance of measuring the scriptspace. One could not tell from the present size of these three books that they represent practically the same model; the column in St. Gall 75 is a bit taller, but the width conforms to that of the other two. I give Köhler's measuring of the St. Gall book rather than my own, since he has studied it more carefully. Measurements are bound to differ somewhat, especially in the case of script-spaces, where different pages may be taken by different measurers. Köhler gives only the outer dimensions. His measurements and mine are always within 10 mm. of each other with these few exceptions. I refer to his numbers with my own in parentheses. No. 38 (116). He is wrong with 345 (it is 380). No. 47 (115). I am wrong with 234 (misprint for 134). In our reports of the number of leaves I am wrong in No. 12 (57): 215, not 207 (215 in my notes). I am right in No. 6 (38): 95, not 93. No. 8 (26): 252, not 253. No. 14 (50): 159, not 199. No. 17 (73), in vol. II: 158, not 154. No. 23 (75): 176, not 175. No. 28 (79): 229, not 227. No. 29 (82): 162, not 161. No. 33 (96): 183, not 182. No. 45 (74): 211, not 212. No. 47 (115): 205+4, not 202. No. 54 (132): 153, not 149. No. 56 (108): 139+1, not 139. No. 57 (120) has 140. For the leaves and dimensions of No. 59, see my No. 135. In No. 10 (29), 4 should be added to his 395 (not 396 as printed in my book). In No. 16 (63) he is wrong with 416 and I with 414. To 416, 5 should be added, 421 in all. The lines on a page in No. 23 (75) number 26 (25), not 23.

²⁰ Köhler, pp. 83–88. He admits that Alcuin was not the inventor of a one-volume Bible. See also Dom de Bruyne's remarks on ‘Bibles en un volume’ in his comments accompanying our reviews (G. G. A., p. 352).

²¹ Survey, pp. 38–45.

^21a ‘Turonische Handschriften aus der Zeit Alcuins,’ in the Festgabe für H. Degering, 1926.

^21b Speculum, VI (1931), 598 f.

²² P. 61.

²³ A special study of the art of these two books would be welcome, with reproductions of all the initials that they contain. Both artists are masters of realism and both have a sense of humor differently displayed. In the Ghent manuscript the beasts are less pugnacious.

²⁴ P. 49. This involves a sudden change of pattern into background and vice versa. If, as Köhler maintains, it is a very early trait in the school, vanishing later, it should be noted that it appears in B. N. lat. 1451.

²⁵ P. 66.

²⁶ Is it not dangerous (p. 65) to call the use of animals in the initials of the Bible of Monza “eine vollkommene Neuerung”?

²⁷ Histoire de la Vulgate, p. 129.

²⁸ I, Lieferung XIV, Tafeln 6 and 7.

²⁹ Pp. 41 f.

³⁰ P. 84.

³¹ P. 86.

³² P. 100.

³³ B. N. lat. 17227.

³⁴ Cod. 268. Köhler admits that the evidence of its single initial does not suffice to put this book in the Alcuinian groups. The evidence of the script he does find sufficient. This I should doubt very much. It looks to me rather like a rude attempt made elsewhere to imitate the Regular Style of Tours.

³⁵ P. 47.

³⁶ P. 75.

³⁷ Pp. 72–90.

³⁸ P. 89, n. Cf. R, pl. III, 1.

³⁹ See Survey, p. 83, pl. III, 2.

⁴⁰ For example there appears in the capitula-table on fol. 116^v an elaborate vine with tendrils and a beast-shaped leaf — quite as elaborate a design as that, for example, in St. Gall 75 (K, 1 a). One of the initials (C) in Ghent 102 (K, 11b.) suggests the ancient fish-form of Merovingian times found in the Tours Eugyppius, B. N. nouv. acq. lat. 1575 (in the letter g, fol. 11^r) and in B. N. lat. 1572, fol. 14^r (R, XI, 1).

⁴¹ P. 96.

⁴² P. 161.

⁴³ P. 127.

⁴⁴ P. 150.

⁴⁵ A manuscript discussed by Köhler but not put on his list, because not illuminated, is Paris, Ste. Genevieve 1260 (No. 32). I had put it in the group IVa and identified the hand with L in the Morgan Gospels, not intending thereby to assert that it was written under Alcuin. Köhler (p. 95) makes it probable — though not certain — that, owing to the inclusion in this lectionary of a portion for the festival of All Saints with a vigil, the book was written about 830, “sicherlich nicht erheblich früher.” I have no objection to this dating, though I see no certain reason why it might not be c. 820, the period to which I would assign Harley 2790, 2793, Chartres 3, and perhaps others. I will add that I am now uncertain whether the scribe of this book is the same as L in the Morgan Gospels. In any event Köhler's observation that its script is to be put palaeographically “in die unmittelbare Nähe” of the Bibles of Zürich and Bern is wide of the mark. Its approximate dating should not be used in any way to determine theirs.

⁴⁶ Pp. 94 f., 145.

⁴⁷ P. 101.

⁴⁸ P. 168.

⁴⁹ Fol. 5^r. Of this I have a photograph.

⁵⁰ Pp. 112–120.

⁵¹ R, XX, 2.

⁵² Survey, p. 94.

⁵³ P. 115.

⁵⁴ K, 29b.

⁵⁵ K, 14a.

⁵⁶ In this design a central vein of a band is bounded by fine contour-lines. See pp. 127, 130 f.

⁵⁷ ‘Rahmentypus,’ see p. 122.

⁵⁸ P. 131.

⁵⁹ One may find plenty of examples of the Intermittent type there. Cf., for example, the Dagulf Psalter (Mon. Pal. Vind. I, Taf. 22, 23). Something that, at any rate in facsimile, looks like an Aderband may be found in the Ada-handschrift (Taf. 9).

⁶⁰ For the contents of the thirteen tables, see Wordsworth and White, I, 7–10. Even in sumptuous manuscripts some of them are combined, though some of the longer ones may be divided into parts. Both of these features appear in the Gospels of Lothair (B. N. 266), which has twelve tables in all.

⁶¹ I should state that not all facts about the canon tables are accessible to me from Köhler's plates or my own photographs and notes.

⁶² P. 144.

⁶³ K, 19b.

⁶⁴ K, 14c.

⁶⁵ K, 29f.

⁶⁶ Bern (K, 19c). Cf. Stutt. (K, 23d). Basel B. II (Bas. B); see K, 54h.

⁶⁷ Bas. A (K, 14c).

⁶⁸ Zür. (K, 17b). Cf. Stutt. (K, 23c), Rorigo Bible (B. N. 3; K, 34a). Leningrad Gospels (Len.; K, 59e).

⁶⁹ K, 19a, b.

⁷⁰ K, 23c.

⁷¹ Gospels of St. Gozlin (Nancy; K, 39b, c, e), Gospels of Prüm (Prüm; K, 97a), Gospels of Lothair (B. N. 266; K, 102a, b.)

⁷² Köhler (p. 138) sees something antique behind the new treasury of forms exhibited by the Bern Bible and the others of this group. I quite agree. The question is at what time this antique source, whatever the form in which it came to Tours, was accessible to the artists and scribes of Tours. The book contained pictures as well as ornamental canon tables. Those in Stutt. and Add. 11848, which are the first of the books of Tours, says Köhler (p. 146), to contain pictures, are obviously from the same source. We await Köhler's second volume for a discussion of their relation to each other.

⁷³ Survey, p. 146.

⁷⁴ Pp. 121, 123.

⁷⁵ P. 110.

⁷⁶ P. 122, Taf. 14a. I have already pointed out the similarity between this initial and the D in Montpellier 412; see above, p. 346.

⁷⁷ Pp. 123 f.

⁷⁸ Köhler, in his only reference to this manuscript (p. 315, n. 1), refers to the order of the books at the end (Apocalypse, Epistles of Paul), in which it agrees with St. Gall 75, as that of a “frühturonische Vorlage.” The copy also reproduces the “Schrift- und Initialschmuck” of this original. If that is so, Köhler's description of what could be done under Alcuin must be enlarged. Note, for instance, the principle of ‘Gliederung’ and the intermittent interlaces in the initial P on fol. 27^v (R, LXXXVI).

⁷⁹ Besides Köhler, Dom Wilmart came to this opinion, as he states in a letter of March 1926.

⁸⁰ Cf. especially Angers 3–4, 5–6, 18, specified by Dom Wilmart, and to them may be added 19, 21, 22, 23. The School of St. Aubin deserves special study, particularly in relation to that of Tours.

⁸¹ See above, p. 328, on No. 13.

⁸² Ibid.

⁸³ There are at least a dozen in each of the Bibles above mentioned.

⁸⁴ P. 161. He states that his acquaintance with this manuscript is limited to the photographs furnished by the Pierpont Morgan Library, and he refers (p. 379) to my article in the Miscellanea Ehrle, IV, 1924, 89–104.

⁸⁵ If Part A was done under Alcuin, then gold initials were used at that time. It is just as sure that they are part of the original execution there as it is that they replace red initials in Part B.

⁸⁶ P. 120, n. 1.

⁸⁷ Survey, p. 131.

⁸⁸ P. 161.

⁸⁹ Another descendant of this original is Troyes 29, which because of its meagre illumination is not treated by Köhler. It shows connection with B. N. lat. 68 in the quire-contents. Its script seems to belong in Period V (No. 93). It is a good example of a later book that, after elaborate ornamentation had come in, could still be ornamented in a very simple style.

⁹⁰ See Nos. 35, 49, 50, 57, 70 in Survey.

⁹¹ See on Nos. 29 and 50.

⁹² The index of Libri Confraternitatum Sancti Galli (Mon. German. Hist., 1884, p. 410) contains over forty occurrences of this name in its different forms. There are two called Amalricus in the list of the monks of Tours.

⁹³ The Basel Bible, we may note, does not in its fragmentary condition offer a chance for comparison with the Bibles here discussed, with the single exception of B. N. lat. 11514. What common ground is there covered shows no coincidences in quirecontents. This lack of data is no disproof of the assumption, if we care to make it, that this book too was written at Marmoutier.

⁹⁴ Köhler (p. 87) admits the possibility that alongside the ornamentation that he calls Alcuinian another style, “reichere, prächtigere,” may have been in vogue, but declares that there are no proofs of its existence. I hope that the facts which I have already presented and those to which I am coming will at least make clear that the question deserves further investigation.

⁹⁵ P. 173.

⁹⁶ P. 168.

⁹⁷ P. 168.

⁹⁸ On fol. 2^r (not reproduced by either K or R ), beato is the first word of the third line of the heading, which is in faded red majuscules of the mixed sort appropriate for the Embellished Merovingian Style. The B is larger than the other majuscules in the line, as is the initial P of praefatio in the first. The N introducing the first line of the text is about twice as large as the P or the B. Like them it is a single letter in red, with the finials and forkings characteristic of the Embellished Merovingian Style.

⁹⁹ P. 183.

¹⁰⁰ Pp. 181, 186.

¹⁰¹ P. 182.

¹⁰² K, 37b.

¹⁰³ See above, p. 337.

¹⁰⁴ P. 185.

¹⁰⁵ E. g. K, 38c, e. Compare also the T (K, 37c; R, XCIX) with the simpler sort of the same design in Bern 165 (K, 31g; R, LXXVI).

¹⁰⁶ Histoire de la Vulgate, p. 249: “mais s'ils [les Évangiles de saint Gauzelin] n'ont pas été écrits de sa main, ils sont certainement l'oeuvre d'un de ses meilleurs disciples.”

¹⁰⁷ P. 206.

¹⁰⁸ Pp. 202, 207 f.

¹⁰⁹ P. 208.

¹¹⁰ Pp. 195, 197.

¹¹¹ P. 195. See below, p. 390, n. 285.

¹¹² Pp. 194, 198, n. 1.

¹¹³ K, 45e, 46e, and 44c, where the later illuminator made a ligature of L and I, not noticing that the I had already been included in the rest of the word (l)iber.

¹¹⁴ See above, p. 354.

¹¹⁵ Survey, p. 136.

¹¹⁶ P. 211.

¹¹⁷ K, 58; R, LIX.

¹¹⁸ Mon. Pal., Series I, xviii, Taf. 2–5.

¹¹⁹ Survey, p. 47.

¹²⁰ Ibid. p. 49.

¹²¹ Köhler (p. 209) speaks of the doubts raised “von Paläographen” against a dating of Bamb. in Alcuin's time. This is too sweeping a statement. Some palaeographers would not date it so early,- and some would. See above, p. 336, n. 8.

¹²² The letter is not addressed to this Ambrosius, as Köhler implies (p. 209).

¹²³ Survey, p. 118.

¹²⁴ P. 210.

¹²⁵ Pp. 209 f.

¹²⁶ Pp. 350 f.

¹²⁷ Bamb. 356 × 116 in a column; Zür. 362 × 112.

¹²⁸ Bamb. 50 and 51 (45, 46, 47, 48, 49); Zür. 50 (49).

¹²⁹ Zür. uses fol. 4^v for the poem, but Bamb. inserts a special leaf for it.

¹³⁰ In the subsequent parts of these Bibles, so far as I have examined, one may note the planning of the same matter, but with more of an attempt on the part of Zür. to fit an important ending in the text to the close of a quaternion. Thus we find Kings ending quire XVII on fol. 134^v, since foll. 129–130 were added in the middle of the gathering, making it a quinion, in order to end the book with the gathering. In Bamb. the text runs on into quire XVIII and ends on fol. 136^v. In the same way in Zür. Daniel ends quire XXV on fol. 198^v, while in Bamb. it runs on into quire XXVI. After the disposition of the matter has thus been altered in the two manuscripts, we should probably not expect to find further coincidence.

¹³¹ Such, for example, as the Bern Bible. This book has not quite the same scriptspace (382 × 130 in a column) but the number of lines is virtually the same, 51 (52). The planning of the matter is obviously not the same as in Zür. and Bamb. Ruth does not end quire XII but comes to a close in quire X, fol. 75^r, Kings beginning on fol. 75^v.

¹³² See below, pp. 389 f.

¹³³ P. 211.

¹³⁴ P. 212.

¹³⁵ P. 215.

¹³⁶ P. 222.

¹³⁷ P. 224.

¹³⁸ P. 226.

¹³⁹ K, 58g.

¹⁴⁰ See above, p. 354. I note incidentally that B not N is illuminated in Bern 4, fol. 83^r.

¹⁴¹ P. 217. In V, fol. 39^v; in H, fol. 175^v; in A, fol. 192^v. For touches of this design see K, 58k and 1. The spaces between the interlaces are clear enough in the photographs. Here is a delicate point where art no less than palaeography needs life-size reproductions.

¹⁴² P. 231.

¹⁴³ P. 232, and see below, p. 389.

¹⁴⁴ P. 232: “darüber lassen Palaögraphie und Ausstattung keinen Zweifel.”

¹⁴⁵ P. 233.

¹⁴⁶ P. 357.

¹⁴⁷ K, 57e.

¹⁴⁸ R, LXXXV. 2.

¹⁴⁹ In the present case we do not need the actual size of the letter to be sure of the facts, since we are dealing with proportions. Occasions might well arise, however, when the exact dimensions would be necessary to know — another argument for the need of life-size reproductions in art as well as in palaeography.

¹⁵⁰ I regret that I have no photograph of this letter in Zür. It would be interesting to try to fit it into the space in the Bamberg book.

¹⁵¹ K, 33a.

¹⁵² R, LIX. I.

¹⁵³ It would be interesting to trace the history of the two forms in Carolingian illumination. The scanty evidence at my disposal does not warrant the conclusion that the form with the smaller upper loop is consistently earlier than the other.

¹⁵⁴ P. 210.

¹⁵⁵ Pp. 227 f. Still, gold and silver are lavishly used.

¹⁵⁶ Fol.7^r. Not in K or R.

¹⁵⁷ See above pp. 354–356.

¹⁵⁸ P. 229.

¹⁵⁹ P. 229.

¹⁶⁰ Or 837. I will not dispute Köhler's assertion (p. 230) that despite the apparent evidence of the list, the date might have been after 837 (though not very probably).

¹⁶¹ Pp. 230 f.

¹⁶² Survey, p. 149.

¹⁶³ P. 247.

¹⁶⁴ P. 248.

¹⁶⁵ Survey, p. 150.

¹⁶⁶ P. 251.

¹⁶⁷ P. 240.

¹⁶⁸ Pp. 238–240.

¹⁶⁹ P. 251.

¹⁷⁰ P. 252.

¹⁷¹ P. 251.

¹⁷² Survey, pp. 155 f.

¹⁷³ Neues Archiv, XXVII, 1901, 264–285.

¹⁷⁴ Foll. 238 and 239 are the only pages on which I noticed t' in the text — once on each page.

¹⁷⁵ See above, p. 367.

¹⁷⁶ K, 74.

¹⁷⁷ Fol. 324^r; R, CXXXI. 2; not in K

¹⁷⁸ P. 255.

¹⁷⁹ Survey, p. 132.

¹⁸⁰ P. 235.

¹⁸¹ E. Dümmler, Geschichte des ostgothischen Reiches, I, 1887, 61.

¹⁸² P. 236.

¹⁸³ Survey, p. 132.

¹⁸⁴ I hope to present a comparative study of these two books at a later time.

¹⁸⁵ P. 289.

¹⁸⁶ See above, p. 335.

¹⁸⁷ For instance, he likens the style of B. N. 261 to a translation of an ode of Horace into Vulgar Latin (p. 292). How often is a facile wit tempted into epigram! Like Ovid, ingenio periit. At least a word of praise would be appropriate for the rose-window into which Q has been transformed, in different ways, in B. N. 261 (K, 119d) and Add. 11849 (K, 122a). I confess also to a sense of pleasure at turning the leaves of B. N. 263 and 267. Beatty 8 is surely, at least in its script, one of the great books of Tours.

¹⁸⁸ P. 293.

¹⁸⁹ Survey, pp. 66–68. But this is a false confidence, as is apparent in Köhler's review (G. G. A., pp. 326, 332).

¹⁹⁰ Survey, p. 144.

¹⁹¹ ‘Le Lectionnaire de Saint-Père,’ Speculum, I, 1926, 269–278; cf. Ibid. VI, 1931, 577, note 2.

¹⁹² Survey, p. 133.

¹⁹³ Speculum, I, 277.

¹⁹⁴ Ibid., plate opposite p. 270.

¹⁹⁵ P. 243.

¹⁹⁶ P. 244, “Die ungewöhnliche Eintragung des Namens ist, wie aus der abweichenden Tintenfarbe ohnehin wahrscheinlich wird, nicht als Künstlersignatur, sondern als Besitzvermerk anzusehen. Damit entfallen auch die Gründe die Dom Wilmart veranlasst haben, die Entstehung der Handschrift in die Frühzeit des 9. Jahrhunderts zu rücken. Das ist aus kunstgeschichtlichen Gründen nicht möglich.”

¹⁹⁷ See Survey, p. 127; plate LXXVII. 1.

¹⁹⁸ Ibid., p. 106; Köhler, p. 39.

¹⁹⁹ Philologische Wochenschrift, 1926, 879. He writes ‘Adelricus me fecit’ on one of the bands of the pediment of the aedicula containing the marks of the dramatis personae (fol. 3^r).

²⁰⁰ Such as the Gospels of Lothair and of Laon (see Köhler, p. 287). The art of the book of Chartres is obviously simpler than that in these beautiful productions.

²⁰¹ P. 314, ‘Kontrolmittel.’

²⁰² P. 335.

²⁰³ P. 313. I wonder if Dom de Bruyne, despite his remarks in Revue Bénédictine, XXXV, 1923, Bull. lit. chrét., [72], would quite subscribe to the statement that he has “accepted (angenommen)” Dom Quentin's new “Auffassung vom Verhältnis der Handschriften zueinander und von der Geschichte des turonischen Textes.”

²⁰⁴ P. 315.

²⁰⁵ P. 316.

²⁰⁶ P. 315 with n. 1.

²⁰⁷ P. 316; Berger, Histoire de la Vulgate, pp. 301–306.

²⁰⁸ P. 315, n. 1.

²⁰⁹ Above, p. 350.

²¹⁰ Apocalypse is the last book, beginning on fol. 202^r of vol. 2 and going nearly to the end on fol. 206^v. Fol. 207 was added later. A hand of the eleventh century, which has supplied lacunae elsewhere in the book, writes the short amount of the text of Apocalypse lacking. He also writes out the Epistle to the Laodiceans. Fol. 207^v is blank, except for scribblings. The argument for the Epistle to the Romans begins on fol. 175^v and the text on 176^v. The order Apocalypse, Paul is cited by Berger (p. 332) as found in “S. Gall. 75. Souvigny. Clermont 1. Angers 2.3. B. N. 25.” I wonder if Köhler was misled by this statement, which follows the former designation of the MSS. of Angers. The old Angers 1 is now 1–2; the old 2 is now 3–4 etc. Berger is referring to the present 3–4, not to the present 2. See the Catalogue des Départements, XXXI, 190. The other manuscripts here associated with that of St. Gall seem neither of Tours nor important.

²¹¹ See below, p. 389.

²¹² K, 8a; R, XLIII. 1.

²¹³ .

²¹⁴ Beginning on fol. 17.

²¹⁵ Fol. 18 begins with Canon Secundus.

²¹⁵ a See G. G. A., p. 354.

²¹⁶ Numbered pages 17–18.

²¹⁷ Pp. 357 f.

²¹⁸ On the Bibles of Monza and St. Gall, see Köhler, p. 317. The Basel Bible is incomplete at the beginning.

²¹⁹ See Berger, p. 307.

²²⁰ I am ashamed to say that I took only occasional notes on this point. It deserves a fresh investigation.

²²¹ Pp. 323–335.

²²² I agree with Berger's general attitude toward this question. See Histoire de la Vulgate, p. 228.

²²³ The latter arrangement seems the normal one.

²²⁴ Köhler adopts the notation of S. Beissel, Geschichte der Evangelienbücher in der ersten Hälfte des Mittelalters (Ergänzungshefte zu den Stimmen aus Maria-Laach, Heft 92 and 93), Freiburg, 1906, pp. 334 f.

²²⁵ Beissel, p. 332, IX, lists Zür. with Monza as having 77 chapters. But Monza apparently has no capitula for Matthew. Beissel is obviously wrong.

²²⁶ Pp. 351 f.

²²⁷ P. 329. The number 28 corresponds to that in the Codex Cavensis and in manuscripts associated with the revision of Theodulf. See Wordsworth and White, Novum Testamentum Latine, I, Oxford, 1889–1898, 38.

²²⁸ Loc. cit.

²²⁹ I, 18 f.

²³⁰ So in Mor. (fol. 70^r), Harley 2790 (fol. 89^r, and Grandval (fol. 348: see Wordsworth and White, p. 175); Grandval innovates by putting the capitula for all four gospels together on foll. 347–348^v. The title reported by Köhler (from Beissel, p. 335), “De Baptismo Johannis in Jordane,” is apparently incorrect.

²³¹ Wordsworth and White, p. 175.

²³² Ibid.; Beissel, pp. 332 f.

²³³ Wordsworth and White, p. 187.

²³⁴ St. Gall 75, Bas. A., B. N. 17227.

²³⁵ Bern has 74 numbers with 78 in the text, and Zür. has 72 in the text.

²³⁶ So B. N. 250 (fol. 26^r), 260 (fol. 109^r, Bas. B. (fol. 165^r). Beissel (p. 335) has Zachariae sacerdoti apparuit (Gabriel) Angelus, and Köhler (p. 324) omits Gabriel altogether.

²³⁷ Wordsworth and White, p. 275.

²³⁸ Ibid. pp. 274 f.; Beissel, pp. 332 f.

²³⁹ 72 in Monza (text) and Prüm; 74 in Bern; 78 in Bern (text).

²⁴⁰ Wordsworth and White, p. 305.

²⁴¹ P. 324. He did not have time to take full notes on this manuscript (p. 327, n. 1) — nor did I.

²⁴² Köhler notes a sprinkling of numbers in the margins of St. Gall in an unsystematic fashion that does not seem to him worth recording in his list (p. 325). Now and then a number is “übersprungen.” He concludes that St. Gall has two different systems mixed and that Monza replaces (“ersetzt”) this mixture by a unity. I should call both Monza and St. Gall very imperfect representations of the original system of Alcuin.

²⁴³ 36 in Prüm.

²⁴⁴ 35 in Q, 36 in P: see Wordsworth and White, p. 506.

²⁴⁵ Wordsworth and White, p. 493.

²⁴⁶ Köhler's hypothesis of an Insular source for some features of the art of Tours and for the capitula (pp. 326 f., 331, 334) receives a setback, he thinks, from the evidence of the capitula in Matthew (p. 332). A comparison of the titles in the Irish manuscripts with those in the Grandval Bible (Wordsworth and White, pp. 19–39) tends rather, I believe, to establish a relationship — not perhaps of quite the kind postulated by Köhler—between Alcuin's Bible and certain English and Irish sources.

²⁴⁷ This investigation is urgent and promising. Even now certain results are patent — such as the use of an entirely different system in later books where other influences are apparent in the art. See Köhler, p. 330, on B. N. 261 and Add. 11849. What I have written here was in print before I could take advantage of Dom de Bruyne's comments in G. G. A., pp. 354 ff. We may hope for a final treatment of the subject from him. Meanwhile I fail to find in the facts presented by him evidence to overthrow my main conclusion as stated here.

²⁴⁸ P. 336.

²⁴⁹ Köhler (p. 319) finds that Grandval and Rorigo belong with the earlier Bibles, just because they omit the letter of Eusebius, but that this evidence can no longer be retained since “nach dem feineren und deswegen dem anderen übergeordneten Kriterium” (i. e. the use of capitula) they belong in the later group. I differ in estimating the value of the former criterion as Köhler applies it, but, supposing it established any fact whatsoever, that fact cannot be thrown away just because a “finer” criterion discovers apparently contradictory facts. Fact must be adjusted to fact. The omission of a prefatory piece is per se as important a fact as the omission of a sentence of the text.

²⁵⁰ P. 338, n. 1.

²⁵¹ Die Trierer Adahandschrift, bearbeitet und herausgegeben von K. Menzel, P. Corssen, etc., Leipzig, 1889, pp. 29–61. I will cite this book as Adahandschrift.

²⁵² The use of Corssen's selections is the more regrettable since in B. N. 274, veritably a pivotal manuscript, on the borderland between the old text and the new, the passage from St. Luke is missing. See also Dom de Bruyne's comments in G. G. A., p. 357.

²⁵³ P. 241.

²⁵⁴ See Harvard Theological Review, XVII, 1924, 197–264.

²⁵⁵ P. 347. One Irish (though not exclusively Irish) spelling that the scribes of Tours never ventured to abandon is ‘abhominationem’ in Matt. 24,15. Only B. N. 47 and the second hand in the Gospels of Nancy venture to correct.

²⁵⁶ P. 340.

²⁵⁷ Köhler suggests that Tours 23, probably a gift to St. Martin's, may have been one of the sources of such influence (p. 343). B. N. 261, he finds, diverges entirely from the text of Tours, while Add. 11849 is mixed (p. 330). A typical reading is No. 117, Matt. 24, 13, ‘permanserit’] Tur. Ir: ‘perseveraverit’ B. N. 261, 267, Add. 11849. Corssen cites other manuscripts apparently of a Franco-Saxon character (Adahandschrift, p. 51).

²⁵⁸ P. 336.

²⁵⁹ P. 339.

²⁶⁰ Nos. 34, 45, 91, 141.

²⁶¹ Nos. 1, 5, 19, 47, 76, 83, 159.

²⁶² I was influenced by the fact that the Gospels of St. Gozlin, one of the earliest books to contain clear signs of N. E., contains besides the usual prefaces those of Theodulf. See Köhler, p. 179.

²⁶³ Nos. 1, 5, 19, 45, 47, 76, 83, 141.

²⁶⁴ Nos. 34, 159.

²⁶⁵ Nos. 91, 94, 170, 181 (note that the Codex Vallicellanus, in accordance with Berger's statement [p. 203] that its text, though originating in Tours, is crossed with some influence from the north of France, contains the reading of N. E.), 195.

²⁶⁶ P. 384.

²⁶⁷ P. 161.

²⁶⁸ See above p. 351.

²⁶⁹ No. 191, bonam (et) confessam (confertam). The reading of A. T. was apparently bonam confersam.

²⁷⁰ No. 122.

²⁷¹ I put the reading of MT first in each case. No. 48, illis (eis); 60, dicant (dicunt); 104, ex eis (ex illis); 105, persequemini (persequimini); 129, omnia (omnia haec, haec omnia)— see below, p. 391, n. 291; 135, MT omits the questionable sentence — see below, p. 391, n. 291; 147, unguerent (ungerent); 151, obstipuerunt (obstupuerunt); 157, crediderunt (crediderant); 186, aufert (auferet); 188, aufert (auferet); 189, nihil inde sperantes (nihil desperantes); 196, poteris (potes, potest).

²⁷² No. 200, audit (audiuit).

²⁷³ No. 148.

²⁷³ a Dom de Bruyne, in G. G. A., p. 358, concludes too hastily, I believe, that MT is not “un vrai ms alcuinien.” Perhaps his opinion may prove correct when all the facts are in. It would be profitless to discuss here the agreements and the differences in my comments on the text of Alcuin's revision and those of Dom de Bruyne. Both are obviously of a tentative character —mémoires pour servir.

²⁷⁴ P. 347.

²⁷⁵ P. 347: “das älteste, uns bekannte Exemplar von Alkuins Bibel und muss die Grundlage für alle auf seine Textemendation gerichteten Erörterungen bilden.”

²⁷⁶ P. 129.

²⁷⁷ See above, p. 361.

²⁷⁸ Köhler can only point out (p. 344, n. 3) that this book is the latest of the group. But if both it and Bamb. were done in the latter half of the nine-year abbacy of Adalhardus, no great length of time intervened between their appearance.

²⁷⁹ P. 343.

²⁸⁰ P. 345, n. 2.

²⁸¹ Ibid. n. 3. Note also that Bas. B. and B. N. 260 have the order Plures, Novum, Sciendum; see above, p. 375.

²⁸² By an unfortunate misprint he has Bern for Bamberg here (p. 345).

²⁸³ Ibid. n. 1.

²⁸⁴ In No. 122 they agree with Monza. On p. 317 Köhler admits that Bamb. and Zür. may go back to the same original — this on the strength of the introductory matter. See also Berger, p. 229. Also see above, pp. 357–359.

²⁸⁵ Grandval, I may note here, is consistently A. T., as is appropriate in accordance with my view that the script is earlier than some of the illumination. See above, p. 355. It contains no N. E. reading that is not found also in one or more of the earlier A. T. books.

²⁸⁶ For other clues leading towards Ireland, see above, pp. 379, 380, 381, 383, 385.

²⁸⁷ See Dom Quentin, Mémoire sur l'Établissement du Texte de la Vulgate, 1922, pp. 290–293, and my review of this work. Köhler apparently would agree with this estimate of the nature of Alcuin's edition and of the manner in which its variants were taken; see p. 345.

²⁸⁸ Read, appropriately, by Mor and MT.

²⁸⁹ Mor and MT have debitor est.

²⁹⁰ For example, Nos. 1, 21, 22, 30, 40, 55, 56, 58, 73, 83, 86, 90, 133, 134, 141, 171, 172.

²⁹¹ I will refer the reader also, for a few among many examples, to Nos. 22, 120, 129, 134, 135. The reading of Mor is Alcuinian in all of these, and so is that of MT in all but Nos. 22 and 135, where the previous text is retained.

²⁹² P. 22.

²⁹³ P 342.

²⁹⁴ Nos. 28, 45, 46, 73(?). The A. T. reading is given by m. 2 in Nos. 46, 73.

²⁹⁵ For instance, Nos. 1, 5, 7, etc.

²⁹⁶ Berger (p. 215) speaks of the extremely close connection between B. N. 3 and B.N. 1.

²⁹⁷ See above, p. 342.

²⁹⁸ P. 60. Köhler means that the scribes of Alcuin in this book were feeling their way and experimenting with certain features of script only later elaborated. I mean that the scribes of St. Gall may be imperfectly reproducing the full-fledged Embellished Merovingian Style.

²⁹⁹ See above, p. 342.

³⁰⁰ See above, p. 347.

³⁰¹ On some of the matters specified above, I hope for the coöperation of B. M. Peebles of the Harvard Graduate School of Arts and Sciences and of Professor L. W. Jones, author of The Script of Cologne, soon to be published by the Mediaeval Academy of America.