Vaticano, Biblioteca Apostolica Vaticana, Reg.lat. 1987

A ninth-century parchment manuscript containing the entire encyclopedia De nuptiis Philologiae et Mercurii, with many contemporary and a few later glosses. It is in fact composed of two parts (book I-II and book III-IX) that originally did not belong together. The joining of these two parts, however, must have taken place at a very early stage, probably already in the ninth century. Both parts are located in Northern or Northeastern France, the first part may have been written in Reims.

Bibliography:

The manuscript consists of 20 quires, 17 of 8 leaves, 2 of 3 and 1 of 10. The first two quires are in fact bifolia, to which single leaves are added: 1² (+0^*) [1-3] and 2(+0^*) [4-6] (see Teeuwen, Harmony and the Music of the Spheres, p. 105, Figure 4). The actual text of Martianus Capella's De nuptiis begins on f 1v, which contains five lines of the opening poem; the poem is continued on bifolium 2^3. Quires 3 and 4 are perfectly regular: 3-4⁸ [7-22]. With these four quires, books I and II are complete.

Quire 5 lacks the first leaf: 5⁸ (-1) [23-29]. This quire opens with book III, On Grammar, and is written in a different script. The first leaf was perhaps cut off in order to make the joining of the two parts from different origins less conspicuous. Quires 6, 7 and 8 are again regular: 6-8⁸ [30-53]. Quire 9 is a quaternio with two singletons: 9⁸ (3|6) [54-61]. Quire 10 is, again, a regular quaternio: 10⁸ [62-69]. A smaller leaf, on which longer glosses were written, was added after the sixth leaf. The quires 11-19 are also regular: 11-19⁸ [70-141]. Quire 20 has five leaves: 20⁴ (+4^*) [142-146] (see Teeuwen, Harmony and the Music of the Spheres, p. 105, Figure 5). The text of the ninth and last book of De nuptiis actually ends at the bottom of f 145v, and 146 contains annotations of a later date. Perhaps, then, the quire was planned as a binio, in which 142^145 were singletons; f 146 is probably a later addition.

Formula: 1² (+0^*) [1-3]

2² (+0^*) [4-6]

3-4⁸ [7-22]

5⁸ (-1) [23-29]

6-8⁸ [30-53]

9⁸ (3|6) [54-61]

10⁸ [62-69]

11-19⁸ [70-141]

20⁴ (+4^*) [142-146]

Quires 3 and 4 both carry an original quire-signature in the lower margin of their last versos: F (14v) and G (22v). The quires 9-19 carry the signatures E-P, also in the lower margin of their last versos: E (61v), F (69v), G (77v), H (85v), I (93v), K (101v), L (109v), M (117v), N (125v) O (133v) and P (141v). The implications of these quire-signatures will be treated below.

The leaves measure ca 295 x 225 mm. They are numbered in ink by a modern hand, in the upper right of each recto. There are no traces of an original foliation. Although sometimes the edges of the actual skin and most prickings alongside the edges of the leaves are visible, there is evidence that at some point small strips were cut from the upper and lower margins, for some of the marginal glosses are no longer complete (e.g. f 44v-45r, f 135, f 145).

The parchment is of medium quality, the thickness not very constant and the hair- and flesh-sides are sometimes clearly (e.g. f 16v-17r), sometimes hardly (e.g. quires 7 and 17) recognizable. The quires of 8 leaves are all arranged according to the pattern HFHF (3-20). The first two quires have the patterns /HF|H (1) and /FH|F (2). Ruling is always executed on the hair-side. There are many edges, tears and holes in the leaves of this manuscript, some restored with new parchment, some not. Edges are seen, for example, at ff 16, 49 (repaired), 59, 76; holes in ff 17, 57, 60, 61, 63, 67, 70, 87, 104, 114, 116, 131, 134, 141, 142 and 143. Pieces have been cut off the margins of ff 83, 117, 135 and 145; ff 17 and 119 have been torn. From the back wormholes penetrate into the manuscript as far as f 107. The last quire is more worn than the others: dark stains obscure the text.

The lay-out of the pages is always in one column, but otherwise there is little regularity. The number of lines per page varies from 23 to 31 and six different types of ruling with a hard point are used.

The first folium (quire 1, f 1) has 25 lines. Its lay-out does not provide an inner margin: the lines stretch from the gutter of the bifolium to the outer margin, marked off with a double vertical through line. The upper- and lower margins are marked off with double horizontal through lines: 0-2/0/2-2/C, measurements ca 170>43 x 28<202>64 mm. Ten lines measure ca 85 mm, the text is written above topline. Prickings for these lines must have been cut off. The next five leaves (quire 1 and 2, ff 2-6) have 29 lines, above topline (ca 87 mm per 10), a single vertical through line marking off the inner margin and a double one marking off the outer margin. The first three and last two writing-lines are projecting into the margins: 1-2/0/3-2C/J, ca 7<165>8>32 x 28<230>35 mm. Prickings for the lines are found along the outer edge of the leaf, and those for the vertical through lines on the first horizontal. Ruling is executed on the hair-side, according to the pattern << (ff 2-3) <<|< (ff 4-6).

Quire 3 has 4 vertical through lines (pricked on the first and last writing line) and two horizontal ones, both at the top of the leaf: 2-2/0/2-0/J (ca 11<7<162>7>25 x 24<245>26 mm). Its 30 text-lines (above topline, 85 mm per 10) are pricked alongside the outer edge of the leaf and ruled on the hair-sides per two bifolia, in what has been called by E.K. Rand “Old Style”: >>>>|<<<<.

The 4^th quire — that is, the first quire of the part which supposedly has a different origin — again has 4 vertical through lines but no horizontal ones: 2-2:E/0/0/J (ca 7<6<171>7>34 x 25<233>36 mm). It has 29 lines (above topline, 84 mm per 10) and is pricked and ruled according to the same pattern.

Quire 5 is different again: it has the same number of vertical and horizontal through lines as quire 3 (2-2/0/1-0/J, ca 12<8<164>10>34 x 20<241>32 mm), but 29 lines (above topline, 86 mm per ten) instead of 30, and these are pricked along the outer vertical through line, instead of along the edge of the leaf. The last two leaves (ff 28-29) have 28 lines instead of 29. They were ruled per two folia, and not per two bifolia, as the pattern of ridges and furrows reveals: />>>|>>>>.

Quires 6, 7 and 8 have the same lay-out as quire 5, but they are ruled per four bifolia: >>>>|<<<<. Quire 6 has 29 lines, above topline, 7 and 8 28. These last two quires also have different measurements: 9<7<166>10>32 x 17<231>47 mm.

For quire 9, again a different scheme was used: double vertical through lines and single horizontal ones above and below the text-lines, 28 lines (above topline). The formulae for this quire are 2-2/0/1-1/J. The two singletons in this quire do not form a proper pair in the pattern of furrows and ridges, for this yields the irregular pattern: >><<|><<<.

Quire 10 is laid out as 5-8, has 28 lines and the same method of ruling per four bifolia as 6-8: >>>>|<<<<. Not only the ruling was done per four bifolia, but also the pricking, as one can conclude from a slanting line of holes for the rules in all the leaves of quire 10. After f 67 (the 6^th leaf of the quire), a smaller leaf is inserted for a long gloss with its own lay-out: no margins are created and it has only writing lines, 48 in number, above topline, 46 mm per 10.

Quire 11 is laid out with two double vertical projecting lines and one horizontal one at the top (as quire 5-8 and 10): 2-2/0/1-0/J. It has 28 lines, which are ruled per two bifolia: >>>>|<<<<.

In quire 12 the one horizontal through line disappears: 2-2/0/0-0/J, it has 28 lines and is ruled per four bifolia: >>>>|<<<<.

The quires 13 to 19 have this same lay-out with two double vertical through lines and no horizontal ones: 2-2/0/0-0/J. All these quires with the exception of quire 19 have 28 lines, above topline; quire 19 has 29 lines. In the quires 13 to 17 the rules are not, however, pricked along the fourth vertical through line (as in the quires 6-12 and 18), but the third. In quire 19 the prickings must have been placed along the edge of the leaf and cut off, for they no longer remain. Quire 13 was ruled per two folia (the bifolia being already folded): >>>>|>>>>, just as quire 5. The quires 16 and 17 are ruled per two bifolia (as 3-5 and 11): >>>>|<<<<; 14, 15, 18 and 19 per four bifolia (as 6-8, 10 and 12): >>>>|<<<<.

In quire 20, finally, two different ruling-patterns are used: ff 142-144 have 31 lines (height per 10: 78 mm), two double vertical through lines and no horizontal ones: 2-2/0/0-0/J. The measurements, however, are different: 10<6<179>8>13 x 30<224>41 mm.

Ff 145-146 seem to have been cut to size for this manuscript, but originally laid out for a larger codex: on f 145 29 lines are used for text, although more lines can be seen in the lower margin. The text on f 145 (the whole leaf is used for the last section of De nuptiis: the explicit stands at the very bottom of f 145v) does not fit the ruling-pattern; f 146 shows that there are in fact prickings for 23 lines (height per 10: 95 mm). This same leaf reveals two double vertical (pricked in the upper and lower margins) and two double horizontal through lines.

From this long description of different types of lay-out used in the Vatican manuscript Reg.lat. 1987 (for a more systematic rendering see Figure 6 in Teeuwen, Harmony and the Music of the Spheres, p. 109) the following conclusions can be drawn. First of all, it is clear that the general hair-flesh arrangement is HFHF (Rand's Rule I), and that ruling is always executed on the hair-side. Where the prickings for these ruling-patterns are not found along the edges of the leaves, a template must have been used.

Secondly, two main types of ruling are used: per two and per four bifolia. In the quires 3, 4, 11, 16, 17 and 20 pairs of bifolia are ruled,[1] in 6-8, 10, 14, 15, 18 and 19 whole quires.

Thirdly, there are two main types of lay-out: with double vertical through lines on both sides of the text and either with or without one horizontal through line at the top of the leaf. The quires 4 and 12-20 correspond with each other in lay-out: 2-2/0-0. The quires 5-8, 10 and 11 follow the other pattern: 2-2/1-0.

Fourthly, the main numbers of lines are either 28 (7-18) or 29 (1, 2, 4-6, 19).

These observations lead to the conclusion that there is no clear system behind all these variations. The variations in the four aspects mentioned do not occur in coherence with each other, but at random. This and the fact that there are two different series of quire-signatures suggests that the codex in its present form is a composition, as was noted already by Préaux and Leonardi.[2]

The actual break occurs between quires 1-4 on the one hand (containing the text of book I-II), and quires 5-20 on the other (book III-IX). The quire-signatures of 3 and 4, F (14v) and G (22v) respectively, make it clear that these were originally the 6^th and 7^th quire in a no longer existing codex. The first two quires, thus, are a ninth-century addition (ff 2-3) and an eleventh-century addition (f 1). Préaux suggested that they consist of pieces of what originally must have been quire E, but since ff 5^6 form a bifolium, this is not possible. Préaux furthermore suggested that quires 1 to 4 were originally ff 33-40 (E), ff 41-48 (F) and ff 49-56 (G), and that the text of De nuptiis originally opened on f 35v. With this, however, he implies that in this original manuscript a text covering the leaves of four quires (A-D) must have preceded De nuptiis.

Quires 5-8 are the quires A-D of a second manuscript, but these quires are not numbered. The quires 9-19, however, carry the signatures E-P. These then must have constituted a manuscript which contained only the books on the arts (III-IX). Although it cannot be proved by means of quire-signatures, it seems to me that quire 20 (a binio with the last sections of book IX) might originally have been part of the same source as the quires 1-4; this, at least, is suggested by its script, which appears to be slightly older and resembles the first four quires more closely.[3]

To conclude: the first four quires (and perhaps the last quire) of this manuscript seem to stem from an older manuscript than the quires containing the books on the artes. If this manuscript is indeed a joining of two already existing, but incomplete or otherwise unsuitable manuscripts, it must have been a joining of 1) an older manuscript containing at least the first two books of the encyclopedia, preceded by another text that filled four quires, and 2) a newer one, containing only the books on the seven liberal arts from De nuptiis. Perhaps an old, severely damaged manuscript (quires 1-5 and perhaps 20?) were taken as the starting point, and quires 5-20 (or 19) were newly created to complete it. We can be certain, in any case, that the joining of these two parts took place in the ninth century, for the glossing of the whole was done in one ninth-century hand.

The text is written in Carolingian minuscule. As was already suggested in the evaluation of the different ruling-patterns, many different hands worked on this codex, which is moreover composed of parts from different manuscripts. There are many certain and even more possible changes of scribes, writing letters which vary in height between 1.5 and 3 mm. A few of the possible breaks are listed here:

• f 1r-v: A later addition to this manuscript, written in two eleventh-century hands, which can be identified as belonging to the school of abbot Wichard of Ghent.
• f 4r; l.5: From f 4r (quire 2) onwards the letter seems to be written in a more vertical direction; the scribe uses a different kind of capital A.
• f 22v-23r: This scribe uses a more open g, cc-a next to two-storeyed a. The cauda of the e-caudata is almost closed.
• f 41v;l.23-f 42r;l.10: A thinner pen is used, which is possibly the one of the glossator.
• f 43r: The letters become sloppier, lopsided and irregular.
• f 50v: The scribe uses tidier, more regular script.
• f 102v; l.2 and 21: The lines 3-20 are perhaps written on rasura; the letters grow larger and sloppier at first, then smaller and more regular.
• f 120r; l.7: At the beginning of book eight a small, rather flat script takes over, which uses ligatures that could perhaps be dated slightly later, but which uses cc-a's as well.
• f 132v: The script is written with a thinner pen, on slightly undulating lines.
• f 138v: The script becomes increasingly irregular and sloppy.
• f 142r: From here (quire 20) a small, tidy and regular letter is written, almost vertical and with slightly clubbed ascenders.

Some of these hands use ligatures that could point to a dating in the earlier 9^th century, such as an nt-ligature (e.g. f 70v; l.12), ligatures with an r that stretches beneath the ruling: rt (e.g. f 96v; l.21) and cre (e.g. f 4r; l.9). The number of corrected word divisions is striking; they are mainly executed by the hand that wrote the text. This could be a second characteristic which points in the direction of an early dating of at least part of the manuscript, for it suggests a close relation with the (sixth-century) original in scriptio continua. Another notable feature occurs in the Greek words: these are most often marked with a line over the Greek capital letters (e.g. f 2r; l.17, 19).[4]

The glosses are almost entirely written in one and the same hand,[5] or in two hands that resemble each other closely. In the glosses one can observe variations in ink-colour, from brown to black, but it is not possible to point to clear changes of hands. The glossator sometimes shapes his glosses in the form of a V. There are a few additions from later (11^th- and 12^th-century) hands, e.g. on f 6v, f 28r-v, ff 71v-72r, ff 86v-90v. The letters of the glossator(s) measure only 1 mm, his g's are 3-shaped with open tails, his r protrudes beneath the (imaginary) line and he has a characteristic preference for giving serifs to letters such as i, n, m. The books I, II, VIII and IX are most densely glossed. Sometimes a hand that closely resembles the one that wrote the text bridges small lacunae in the margins. F 1r-v, finally, is written in two eleventh-century hands, which belong to the school of abbot Wichard of Ghent. These same hands can be recognized on the last folia.

On the first folium of the manuscript the poem Versus de astris et signis caeli and the refrain of the poem Scande caeli templa are written, in an 11^th-century, “Wichardian” hand. To these songs musical notation is added, which was first studied by Bannister in his Monumenti Vaticani de Paleografia Musicale Latina,[6] and more recently by Corbin.[7] The neumes are in Lorrainian notation and since the text of the song was obviously written in Ghent, the musical notation may very well share this place of origin. There are no specific characteristics that suggest otherwise.

Both original punctuation and punctuation added by later scribes are found in the manuscript. Apart from the usual medial and low points (with or without commas, generally added by the glossator), the scribes use question marks. An addition that is probably later is the small upside down triangle consisting of three points or two points and a comma, used for longer rests (e.g. at the end of a book) or in between verse-lines (e.g. f 120v). The scribes write majuscules for subtitles and explicit/incipits, mainly using rustic capitals and occasionally uncial E, M, D, and F (protruding beneath the line).[8] Coloured majuscules are used for the first explicit/incipit, sometimes for the initial letters of verses or prose-passages (e.g. f 2r, 6r). The glossator writes key-words and subtitles in rustic capital.

Some of the majuscules, as already mentioned, are decorated with colour. The majuscules on ff 2r and 7r are written in red ink; the bows of some majuscules are coloured with red (ff 115r-v, 120v, 124v, 125v) or green (f 6r) ink. On f 1r a figure of the planets is drawn in red and green ink, and in the title of the poem below, these same colours are also used.

Unfortunately, there are few sources which reveal the history of Vatican, Reg.lat. 1987. As was argued before, it consists of two parts (books I-II & III-VIII), but these were joined already in the ninth century — before the main layer of glosses was written. In the eleventh century the manuscript was present in the library of the monastery of St. Peter in Ghent, as is clear from the presence of an 11^th-century hand from the writing school of abbot Wichard (988/994-1058). On f 146v one can decipher a later confirmation of ownership:[9]

Liber Sancti Petri gandensis ecclesie. Servanti benedictio tollenti maledictio. Qui folium ex eo tulerit vel curtaverit anathema sit

(Book of the church of Saint Peter of Ghent. A blessing to the keeper, a curse to the thief. May he who took or cut a folium from it be damned.)

Préaux speculatively suggested that the book was given to the monastery of St. Peter by one of its sister-abbeys, Corbie or Reims.[10] In the sixteenth century the codex was obtained by the humanist scholar Franciscus Nansius (1525-1595), who left his mark of ownership on f 1r.[11] Later, the manuscript became part of the collection of the Swedish queen Christina, who — as is well known — donated a large part of her book-collection to the Vatican. Thus it entered the Vatican collection in 1690, receiving its present signature Reginense Latinus 1987. A stamp on f 146v confirms this.

Apart from the ninth-century traces of usage, there are some 11^th- or early 12^th-century hands recognizable in the manuscript, in which occasional additions or marginal remarks were written.

The ninth-century manuscript today has modern boards, covered with reddish brown leather with a golden ornamental edge. In their present form the boards measure 306 x 224 mm and the spine shows five raised bands (spaced 50-50-45-47-48-66 mm).[12] On the spine, the signature is printed in gold letters: Reg 1987. The codex has red-and-yellow silk sewn headbands. The spine is — especially at the head — quite gravely affected by worm-holes.[13] It has two paper fly-leafs, both empty.

The manuscript contains:

• f 1r-v: Additions in an eleventh-century hand: a figure of the planetary orbits, made by use of a pair of compasses; the song Versus de astris et signis caeli (in several manuscripts attributed to “Priscianus Grammaticus”)[14] and the refrain Scande caeli templa (Martianus Capella, book II.117 ff.), both with musical notation.
• f 1v: The first 5 verses of De nuptiis, book I in the same eleventh-century hand.
• f 2r-145v: Martianus Capella, De nuptiis Philologiae et Mercurii, in a ninth-century hand:

  books I.6-II: I 1v-12v; II 12v-22v;

  books III-IX: III 23r-41v; IV 41v-57v; V 58r-79r; VI 79r-102r; VII 102r-120r; VIII 120r-132v; IX 132v-145v.

• f 146r-v: Additions in the same, eleventh-century hand: on the recto a Greek alphabet with numerical values, figure of the proportions of the intervals bisdiapason, diapason, diapente, diatessaron and tonus; on verso a partly erased possession-mark of St. Peter in Ghent and a stamp of the Biblioteca Apostolica Vaticana.