² ‘Judgement's sign: the earth shall drip with sweat;/ Everlastingly the King shall come from heaven, who/ Shall be present to judge bodies and the world …’ (Paris 1154, fol. 122^r). Translation from Dronke, P., Hermes and the Sibyls: Continuations and Creations, Inaugural Lecture (Cambridge, 1990), p. 11Google Scholar. For the complete Latin text, see ‘Sermo contra Judaeos, Paganos et Arianos’, ed. Migne, J.-P., in Patrologiae cursus completus: Series latina (hereafter PL), 221 vols. (Paris, 1844–1864), Xlii, col. 1126Google Scholar.

³ Issues of orality have engaged the attention of a whole generation of historians, anthropologists, philologists and musicologists. For an overview of the developments in musicology, with extensive bibliography, see Levy, K., ‘On Gregorian Orality’, Journal of the American Musicological Society, 43 (1990), pp. 185–92CrossRefGoogle Scholar. For a different perspective in recent literature, see Treitler, L., ‘Oral, Written and Literate Process in the Transmission of Medieval Music’, Speculum, 56 (1981), pp. 471–91CrossRefGoogle Scholar, and ‘Reading and Singing: On the Genesis of Occidental Music-Writing’, Early Music History, 4 (1984), pp. 135–208CrossRefGoogle Scholar.

⁴ Treitler, ‘Reading and Singing’, p. 207.

⁵ Ibid., p. 139.

⁶ Ibid., p. 202. For the complete argument concerning the relation of punctuation signs to notation, see pp. 186–203.

⁷ The dating of this manuscript has divided scholarly opinion. The most significant ascriptions are those of Bischoff, B. (late ninth century), ‘Gottschalks Lied für den Reichenauer Freund’, in Medium Aevum Vivum – Festschrift für Walther Bulst, ed. Jauss, H. and Schaller, D. (Heidelberg, 1960), p. 62Google Scholar, repr. in Bischoff, , Mittelalterliche Studien, ii (Stuttgart, 1966–1967), p. 27Google Scholar, and Chailley, J. (tenth century), L'école musicale de Saint-Martial de Limoges jusqu'à la fin du XI^e; siècle (Paris, 1960), pp. 75–8Google Scholar. The question of origin and provenance is also not straightforward; the general consensus is that it originated from a monastery dedicated to St Martin within Aquitaine and was adapted to use at St Martial of Limoges in the eleventh century. For this see Lauer, P., Catalogue général des manuscrits latins de la Bibliothèque Nationale de Paris, i (Paris, 1939), pp. 421–2Google Scholar, with refinements by Chailley, , L'école, pp. 75–6Google Scholar.

⁸ Chailley, , L'école, p. 73Google Scholar.

⁹ The main study is that of Spanke, H., ‘Rhythmen und Sequenz-Studien’, Studi Medievali, n.s. 4 (1931), pp. 286–320Google Scholar. The main editions are Monumenta Germaniae Historica. Poetae latini medii aevi (hereafter MGH Poetae), iv.ii, ed. Strecker, K. (Berlin, 1914)Google Scholar, and Analecta Hymnica medii aevi (hereafter AH), l, ed. Blume, C. and Dreves, G. (Leipzig, 1907)Google Scholar.

¹⁰ Chailley, , L'école, pp. 73–8 and 123–59Google Scholar. See also Gaborit-Chopin, D., La décoration des manuscrits à Saint-Martial de Limoges et en Limousin du IX^e au XII^e siècle (Paris and Geneva, 1969), p. 45 and 188;Google ScholarCrocker, R., ‘The Repertory of Proses at Saint-Martial de Limoges in the 10th century’, in Journal of the American Musicological Society, 11 (1958), pp. 161–2CrossRefGoogle Scholar; and Stäblein, B., Schriftbild der einstimmigen Musik, Musikgeschichte in Bildern, iii (Leipzig, 1975), pp. 146–8Google Scholar.

¹¹ Chailley, , L'école, p. 158Google Scholar.

¹² Traube, L., MGH Poetae, iii (Berlin, 1896), p. 721Google Scholar. Chailley, , L'école, pp. 74–5Google Scholar.

¹³ For details, see Table 2 below.

¹⁴ The rubric versus de nativitate domini is found in Traube's groups 5 and 8 and in Chailley's 6 and 9. See the ‘Exemplars’ and ‘Liturgical’ groups in Table 2 below.

¹⁵ Other poets whose versus are found in the collection are: Columbanus (A solis ortu), Angilbert (Aurora cum primo), Prudentius (Germine nobilis), Fortunatus (Pange lingua) and Flavius (Tellus ac aethra). The poets Gottschalk and Paulinus are further highlighted by the placement of their versus (O deus misere and Ad caeli clara respectively) at the head of the collection.

¹⁶ Beyond indicating the ordering of versus, the rubrics also establish their own patterns. Highlighted in the rubrics are two contemporary poets (Paulinus and Gottschalk) and two political figures (Eric, Duke of Friuli, and Hugh, Abbot of St Quentin). This selection betrays an interest in two distinct areas: Aquitaine and Lombardy. Two highlighted ‘modern’ figures – in the sense that they are neither biblical nor canonical – complete the pattern: Charlemagne (as a figure of political unity) and Boethius (as a figure of poetic and musical unity).

¹⁷ For a brief summary of the history and importance of this text, see Dronke, P., Hermes and the Sibyls, pp. 10–11Google Scholar. For a musicological perspective, see Corbin, S., ‘Le cantus sibyllae: origine et premiers textes’, Revue de Musicologie, 31 (1982), pp. 1–10Google Scholar. H. Anglès provides musical incipits for twenty-three Latin versions dating from the tenth to the sixteenth centuries, including a transcription from Paris 1154, in La música a Catalunya fins al segle XIII (Barcelona, 1935), pp. 288–302Google Scholar.

¹⁸ For inventories of two other Carolingian versus collections, see Tables 3 and 4.

¹⁹ The placement of locally significant saints at the end of groups was a common practice in litanies also; for discussion of this see below. The straightforward association of this prosa for St Martial with a provenance of St Martial of Limoges for the manuscript was rejected by Chailley, since this prosa is found in many sources which did not originate there and often show stronger links to institutions connected with St Martin. See Chailley, , L'école, p. 76Google Scholar. Other instances of highlighting specific Aquitanian personalities are found elsewhere in the manuscript: on the rubrics of the versus collection, see n. 16; on the litany (part I), see n. 63; and in the prayer section (part II) the only figure mentioned in the rubrics, and the person whose prayer is placed at the head of the section, is Gregory of Tours.

²⁰ The struggles in the mid ninth century between Charles the Bald and Pippin II were the direct result of Louis the Pious's supporting Charles for the kingship of Aquitaine. For full details, see McKitterick, R., The Frankish Kingdoms under the Carolingians 751–987 (London and New York, 1983), ch. 7Google Scholar.

²¹ Luke 16: 19–31.

²² ‘He acknowledged five brothers in his generation,/ Provided care for them but was not able to provide for himself;/ Because of the crumbs of bread he denied, he is tortured in hell.’ The full Latin text can be found in MGH Poetae, iv.ii, pp. 537–9.

²³ In total, Louis the Pious had three natural brothers and three half-brothers. His natural brothers died too early to represent any threat to his position, but his treatment of his half-brothers on his accession to the imperial throne in 814 was as ruthless as the way in which he dealt with all possible opposition to his rule. Drogo, Hugh and Theodoric were all tonsured and confined to monasteries, and were not released until 822. For the full genealogy, see the insert ‘Die Nachkommen Karls des Grossen 1.-8. Generation’, in Karl der Grosse, iv: Das Nachleben, ed. Braunfels, W. (Düsseldorf, 1967)Google Scholar.

²⁴ ‘He summoned Olofernus the leader of the soldiery:/ Go out, he said, make war against the people of the West … In this city was a multitude of the Jews;/ Worshipping the God of heaven, the Saviour of all,/ Bravely they repelled Olofernus's military’. Latin text reproduced from MGH Poetae, iv.ii, p. 459.

²⁵ See McKitterick, The Frankish Kingdoms, esp. ch. 7.

²⁶ In Brussels Bibliothèque Royale 8860–8867 only a fragment of the text appears (verses 44–50, fol. 1^r). Verona Biblioteca Capitolare 90 contains fifty verses (fols. 14^r– 19^v), whereas Verona Biblioteca Capitolare 85 presents verses 1–11 and 49 (fol. 63^r).

²⁷ Witness to this are the current ‘debates’ surrounding the status of oral texts and the role of the scribes in the recording process. The most recent exchange has been between Jeffrey, P., Re-envisioning Past Musical Cultures (Chicago and London, 1992)Google Scholar, passim, esp. ch. 2, and Treitler, L. in his review of this book, Journal of the American Musicological Society, 47 (1994), pp. 137–71CrossRefGoogle Scholar.

²⁸ Exceptional in this regard is McKitterick's suggestion that the linguistic ‘vulgarisms’ found in Hugh's lament and Aurora cum primo may be read as indications of the lay orientation of the manuscript. See McKitterick, R., The Carolingians and the Written Word (Cambridge, 1989), pp. 230–1CrossRefGoogle Scholar. For further evidence suggesting a lay orientation for this manuscript, see Barrett, S., ‘The Writing of Paris 1154’, unpublished M.Phil, diss. (Faculty of Music, Cambridge University, 1996), ch. 2Google Scholar.

²⁹ For a parallel situation in relation to Charlemagne's lament A solis solis ortu usque, see Table 3 below; this appears in the fourth group, which is otherwise wholly centred on Christ. A similar process occurs in Verona 90, where the same versus appears in a series focused on the events of Easter. For the inventory of this latter manuscript see Meersseman, G., ‘Il codice XC della capitolare di Verona’, Archivio Veneto (5th. ser.), 104 (1975), pp. 14–23Google Scholar.

³⁰ Of the other collections, the Leiden collection contains no concordances and Verona 88 contains only four.

³¹ Bischoff, , ‘Gottschalks Lied’, Mittelalterliche Studien, ii, p. 26Google Scholar.

³² The collation is as follows: I⁵ (1^r–5^v), lacks 1, 2 and 8; II² (6^r–7^v); III⁸ (8^r–15^v); IV⁶ (16^r–21^v); V–XI⁸ (22^r–76^v). Signatures are found from 7^v, beginning with III (not II); also 15^v IIII; 45^v, VIII; 60^v, IX (probably in error); and 76^v XII.

³³ The evidence for a Laon provenance by the early tenth century is presented in Contreni, J., The Cathedral School of Laon from 850 to 930: Its Manuscripts and Masters, Münchener Beiträge zur Mediävistik und Renaissance-Forschung 29 (Munich, 1978), pp. 160–1Google Scholar. In Wilmart's opinion, which is recorded on the opening flyleaf, the manuscript originated from Tours. This ascription is doubtful since the script displays none of the characteristic features of Tours – for which see Rand, E., A Survey of the Manuscripts of Tours, i (Cambridge, Mass., 1929), ch. 2. As regards date, the ninthcentury dating of the manuscript by Contreni (p. 160, n. 15) remains the most recent and best-informed opinionGoogle Scholar.

³⁴ The full gathering structure is: I–V⁸ (1^r–40^v); VI⁴ (41^r–44^v), lacks 3–6. The signatures are: fol. 15^v IIII (not II); 24^v, V; 32^v, VI; 40^v, VII; and 44^v, VIII.

³⁵ Of added importance in relation to these two versus is the dedication of a general versus, Iste confessor, to a specific saint, St Germanus. This could only be the result of the local significance of the saint. The implications of this for the provenance of the manuscript remain an open question.

³⁶ Although there is little immediate difference in appearance, certain features suggest a change of hand in Brussels 8860–8867: from I to III, only an uncial a is used, and there are few ligatures; from IV to XI, the oc form of a is also employed, the left-hand shaft of the n extends beneath the line, and there is an expanded range of ligatures (including rr, ae and r-et).

³⁷ Both Verona 88 and Verona 90 are composite manuscripts containing much material besides versus. Within the main versus collection of Verona 90, six different scribes were identified by G. Meersseman in ‘Il codice’, p. 24. For a more recent palaeographical account of the work done by different scribes in this manuscript, see Borders, J., ‘The Cathedral of Verona as a Musical Center in the Middle Ages: Its History, Manuscripts, and Liturgical Practice’, Ph.D. Diss. (University of Chicago, 1983), pp. 467–84Google Scholar. Verona 88 has received less palaeographic attention, but its versus collection also appears to have been compiled by several scribes. It has been described as showing evidence of a group of scribes ‘d'un nombre incertain, qui doivent être localisés vraisemblablement aussi à Saint-Denis’. See Meersseman, G., Les capitules du diurnal de Saint-Denis (Cod. Verona. Cap. LXXXVIII, Saec. IX), Spicilegium Friburgense 30 (Fribourg, 1986), p. 13Google Scholar.

³⁸ In Verona 90, for example, versus 7–17 are all hymns to saints. Verona 88, with only seventeen versus, is too small for the identification of any coherent design.

³⁹ Chailley, , L'ecole, p. 73Google Scholar.

⁴⁰ PL LXXXIII, cols. 827–49.

⁴¹ The extract from the Synonyma closes with the words in voluntate sunt … This appears to be not a matter of loss or later removal, but the result of a deliberate editing technique on behalf of the compiler of the manuscript. The same technique is also used in the joining of parts I and II of this manuscript (see Table 7 and note). For a full discussion of this technique and its structural implications, see S. Barrett, ‘The Writing’, pp. 56–60.

⁴² For a full discussion of the stilus Isidorianus and all other stylistic and structural aspects of the Synonyma, see, principally, Fontaine, J., ‘Isidor de Seville auteur “ascétique”: les énigmes des Synonyma’, in Studi Medievali (3rd ser.), 6.2 (1965), pp. 163–95Google Scholar.

⁴³ Ibid., pp. 171–3.

⁴⁴ These changes in voice are marked by the rubrics RATIO and HOMO in part III of Paris 1154.

⁴⁵ ‘O Lord, the glorious creator of all things,/ I cry to you, sighing with tearful and bitter cries/ Grant pardon to the penitent, O Christ.’ Latin text reproduced from AH xix, ed. G. Dreves (Leipzig, 1895), p. 42.

⁴⁶ For the most recent discussion of whether the De Consolatione Philosophiae is intrinsically Christian in outlook, see Chadwick, H., Boethius: The Consolation of Music, Logic, Theology and Philosophy (Oxford, 1981), pp. 248–53Google Scholar. Chadwick's conclusion is that the ‘Consolation is a work by a Platonist who is also a Christian, but it is not a Christian work’ (ibid., p. 249).

⁴⁷ ‘The man who wants to be powerful/ Must tame his high spirits’. Both the Latin text and its translation are found in Rand, E., Stewart, H. and Tester, S. (transl.), Boethius: Tractates, The Consolation of Philosophy, Loeb Classical Library 74 (Cambridge, Mass., new edn 1973), pp. 252–3Google Scholar.

⁴⁸ Augustine's Iudicii signum as found in De civitate dei (XVIII: 23) represents a translation from the Greek original. His attempt to render the Greek acrostic in Latin reads: IESUS CREISTOS TEVD NIOS SOTER.

⁴⁹ For a brief summary of the distinctive quality of Carolingian penance and the role of penitentials within it, see Pierce, R., ‘The Frankish Penitentials’, Studies in Church History, 11 (1975), pp. 31–9CrossRefGoogle Scholar. For a comprehensive study of this field and further bibliography, see Kottje, R., Die Bussbücher Halitgars von Cambrai und des Hrabanus Maurus (Berlin and New York, 1980)CrossRefGoogle Scholar.

⁵⁰ The foundations for research into florilegia were laid by H. Rochais in his account of the manuscript base of ‘ascetic’ florilegia from the Sententia of Isidore to the end of the eleventh century. See Rochais, H., ‘Contribution à l'histoire des florilèges ascétiques du haut moyen-âge latin: le Liber Scintillarum’, Revue Bénédictine, 63 (1953), pp. 246–91CrossRefGoogle Scholar. On the distinctive quality and role of florilegia in Carolingian society, see McKitterick, R., Frankish Church and Carolingian Reforms 789–895 (London, 1977), ch. 5Google Scholar.

⁵¹ See McKitterick, Frankish Church, ch. 5. For a further consideration of the issue of lay use and the appearance of florilegia in manuscript compilations, see idem, The Carolingians, pp. 266–70.

⁵² This broader definition of florilegia, defined from a Carolingian rather than a patristic perspective, is that emphasised by McKitterick, in Frankish Church, pp. 155–83Google Scholar.

⁵³ For a brief account of the origins and contents of libelli precum, see Driscoll, M., ‘Penance in Transition: Popular Piety and Practice’, in Essays in Medieval Liturgy, ed. Larson-Miller, L. (forthcoming)Google Scholar. For the opportunity to read this essay prior to publication, grateful thanks are extended to Michael Driscoll, Gunilla Björkvall and Ritva Jacobson.

⁵⁴ Wilmart, A., Precum libelli quattuor aevi karolini (Rome, 1940)Google Scholar.

⁵⁵ Salmon, P., Analecta liturgica: extraits des manuscrits liturgiques de la Bibliothèque Vaticane, Studi e Testi 273 (Vatican, 1974), pp. 120–94Google Scholar.

⁵⁶ The place of versus within the florilegia tradition has not been previously noted. From the concordance patterns of the individual versus found in Paris 1154 it becomes clear that this context, rather than that of hymnic collections or unexamined notions of ‘independent’ versus collections, is the strongest in terms of manuscript evidence. For a partial account of the concordance patterns and the nature of the manuscripts in which they are contained, see S. Barrett, ‘The Writing’, pp. 104–8. Further to this, within the Bern collection the hymnic influence of the libelli precum is evident in the ‘liturgical’ group, whereas in the Brussels collection the ascetic element is strongly represented by the ‘moralistic’ group.

⁵⁷ This occurs in the first two books of Alcuin's De vitiis et virtutibus, PL, CL, cols. 613–38. For a discussion of the history of the eight vices and the place of this work within it, see McKitterick, Frankish Church, ch. 5.

⁵⁸ One does not have to travel far to find precursors for this: Paris Bibliothèque Nationale lat. 1153, a ninth-century florilegium from St Denis, combines a libellus precum, a litany of saints and Isidore's Synonyma. For a description of this manuscript, see Lauer, P., Catalogue général, i, pp. 420–1Google Scholar.

⁵⁹ For a broad discussion of the function of litanies, including penitential use, see Lapidge, M. (ed.), Anglo-Saxon Litanies of the Saints, Henry Bradshaw Society 106 (London, 1991), pp. 1–13Google Scholar. In relation to three Anglo-Saxon manuscripts of the tenth and eleventh century, Lapidge notes that ‘the litany of the saints was adapted for confessional purposes, in which the recitation of the litany formed a central feature of the act of contrition’ (p. 46).

⁶⁰ For the use of this rubric in connection with litanies in the context of florilegia, see Alcuin's Officia per ferias, PL, ci, col. 546, and Wilmart, , Precum libelli, pp. 134–5Google Scholar. For liturgical use, see, for example, Antiphonale missarum sextuplex, ed. Hesbert, R.-J. (Rome, 1935), at no. 201Google Scholar.

⁶¹ Of the twenty-six litanies edited and discussed by M. Coens, the longest are those contained in florilegia: such as no. I, Cologne Cathedral 106 (275 entries), no. 6, Bavarian State Library 27305 (c. 350 entries); and no. 8, the libellus precum of Fleury, Orléans 184 (250 entries). See Coens, M., ‘Anciennes litanies des saints’, Recueil d'études Bollandiennes, Subsidia Hagiographica 37 (Brussels, 1963), pp. 129–322Google Scholar.

⁶² These principles are noted by Coens in relation to the first of the litanies he discusses, Cologne Cathedral 106. Ibid., pp. 139–49.

⁶³ This sequence, unnoted by Chailley, is as follows: (fol. 2^r) Cromaci I., (fol. 4^r) Fotine II., (fol. 5^v) [erased]. III.,(fol. 7^v) Aget. IV., (fol. 8^r) Senoni. V., (fol. 10^r) Claudiane. VI., (fol. 11^v) Leobine .VII., (fol. 13^r) Brigida. VIII., and (fol. 15^r) Daria. VIIII. Since the saints in this sequence do not commonly appear, the reasons for highlighting them appear to have been local. It has been tentatively suggested elsewhere that these saints display some connections with Tours, although this is far from proven. See S. Barrett, ‘The Writing’, pp. 18–19.

⁶⁴ The attachment of rubric Oratio ad Sanctam Mariam to Domine Ihesu Christe rex virginum is in fact an error. This rubric applies to the prayer Sancta Maria misericordissima, often found in this sequence. See, for example, the Libellus Sacrarum Precum, PL, ci, cols. 1399–1400.

⁶⁵ It has been argued that litanies, libelli precum and versus collections have their roots in Insular traditions. Lapidge explains that the form of the litany was first established in Anglo-Saxon England in the early part of the eighth century (Litanies, pp. 13–25). Driscoll proposes that libelli precum have their roots in Celtic and Anglo-Saxon booklets of confessional and penitential prayers (‘Penance in Transition’). In addition, Bourgain has suggested that the earliest versus collection is a manuscript written at Lorsch at the end of the eighth century, whose contents are based on a lost exemplar made in seventh-century England: see Bourgain, P., ‘Les recueils carolingiens de poésie rythmique’, in De Tertullien aux Mozarabes, ii, ed. Holtz, L. and Fredouille, J.-C. (Paris, 1992), pp. 117–18Google Scholar. In each case, however, surviving precursors are limited in number whereas Carolingian witnesses are numerous. As such, although Insular precursors may be assumed, it can be said that these traditions burgeoned in written form only in the Carolingian era.

⁶⁶ This opinion was first expressed by Handschin, J., ‘Über Estampie und Sequenz II’, Zeitschrift für Musikwissenschaft, 13/1 (1930), p. 122 n. 1Google Scholar.

⁶⁷ The relative heighting of scribe A's notation implies a late-tenth-century dating. Against this stands Chailley's objection that a non-diastematic notation was employed for reasons of restricted space: see Chailley, , L'école, p. 77Google Scholar. This can be rejected through an identification of scribes D and A as one and the same person. The notations of these scribes are exactly the same in formation and ink colour, the only difference being the compacted presentation of D. Further, the notation and the text of D are identical in ink colour, relative size and thickness of stroke. Since Deum time was the first textual addition made on fol. 105^v and was not laid out with sufficient space for diastematic heighting, scribe A/D may be assumed to have been writing before the introduction of exact heighting. As for the other scribes, since their work was (with one exception) undertaken after that of scribe A, the difficulty of dating remains. Nevertheless, the fact that not one of the remaining nine scribes attempted even a careful partial diastematy suggests that much of it was added before the eleventh century.

⁶⁸ The hand of scribe A is neat and compact, with thin shafts and stable forms. Scribe B, on the other hand, displays thick strokes and forms which are slightly varied on each appearance.

⁶⁹ Of the remaining scripts, E is a significantly later addition (probably of the eleventh century) and was added at the same time as the textual addition over which it appears: the ink colour and the thickness of stroke are exactly the same. This leaves scribes C and F. The latter displays features of French notation as mentioned above in relation to scribe K and is most likely by the same hand; the construction of its forms, the delicacy of stroke and the light brown quality of the ink are all identical. Even if the musical notation is not by the same hand, it remains most probable that it was added in the same layer as K and hence after the work of scribe A. The hand of scribe C is unsteady, with a limited range of forms and uneven construction. Although this makes it difficult to date, the evidence of the rest of the manuscript suggests that it was added much later, since the additions by less well formed hands in the prayer section and the Synonyma are all of the eleventh century.

⁷⁰ The later additions, all with musical notation of the St Gall type, are found on fols. 15^v, 22^r, 66^r and 74^v–76^r. In addition to the versus Fulgentibus palmis, Avarus maximam and Mendaces ostendit noted by Van, der Gheyn, Catalogue des manuscrits de la Bibliothèque Royale de Belgique, ii (Brussels, 1902), pp. 289–92, also recorded are the versus Hierez runetal Salve linguas (fol. 15^v), Beati ergo corpus, a text referring to St Otmar the first Abbot of St Gall (fol. 76^r), and the Agnus dei and Gloria in excelsis (fol. 76^v)Google Scholar.

⁷¹ See ‘A Canterbury Classbook of the Mid-Eleventh Century (the ‘Cambridge Songs’ Manuscript)’, Riggs, A. and Wieland, G., Anglo-Saxon England, 4 (1975), pp. 113–30Google Scholar. Also Page, C., ‘The Boethian Metrum Bella bis quinis: A New Song from Saxon Canterbury’, in Boethius – His Life, Thought and Influence, ed. Gibson, M. (Oxford, 1981), pp. 306–11Google Scholar.

⁷² See Contreni, , The Cathedral School, pp. 160–1Google Scholar.

⁷³ For further examples of this, see S. Rankin, ‘Doodles in the Margins: Musical Notation at Fleury’, paper given at Symposium ‘Die Erschliessung der Quellen des mittelalterlichen liturgischen Gesanges’ (1996), to be published in series Wolfenbütteler Mittelalter-Studien.