"Rotulus Beratinus", a Greek Liturgy Roll Daly, Lloyd W Greek, Roman and Byzantine Studies; Spring 1972; 13, 1; ProQuest pg. 109 Rotulus Beratinus, A Greek Liturgy Roll Lloyd W. Daly THREE PIECES of a parchment roll containing the text of the lit­ urgy of the Greek Orthodox Church known as the Liturgy of St Basil the Great have recently been brought to my attention. They are at the time of this writing the property of Mrs Theodore Ku­ chova of Atlantic City, New Jersey, and were brought with her about fifty years ago at the time of her immigration from her original home at Berat in southern Albania, about forty-five kilometers northeast of Valona. No further information as to their more precise provenience is available, but one piece (2V, see PLATE 6) bears in its margin a clearly modern note in purplish ink and in the modern Albanian orthog­ raphy which was codified in 1908. The meaning of the note appears to be that the manuscript of the liturgy is one that was used by a priest named Vangjeli.1 Berat, once the capital of a Turkish sanjak and the see of an Ortho­ dox metropolitan, is not unknown for its manuscripts. It has given its name to one of the few New Testament MSS written in letters of silver on purple parchment, the Codex Beratinus, von Soden's d7, and was, at least at one time, the home of another MS of the Gospels which has been called the Codex Aureus Anthimi, written in letters of gold on purple parchment. It would appear that these treasures were taken, around 1900, to the skete of St Andrew on Mount Athos, too late to be recorded in Lambros' catalogue of the Athonite MSS.2 In 1885 Pierre Battifol visited Berat with the specific objective of studying the Codex Beratinus. He reported that many of the MSS in the archbishop's palace were in a wretched state of neglect and disrepair. 1 For the reading, interpretation and information about this note I am deeply indebted to Professor Kostas Kazazis of the University of Chicago, who was kind enough to offer his help on the basis of a transcription from the original. 2 Cf L. Petit, EchO 4 (1900/01) 64 and, for further bibliography on the MSS of Berat, M. Richard, Repertoire des bibliotheques et des catalogues des manuscnts grecs (Paris 1958) 51-52. 109 110 ROTULUS BERATINUS, A GREEK LITURGY ROLL Among other items he describes a roll containing the liturgy of St John Chrysostom in letters of silver on purple parchment, which he judged one might date as early as the twelfth century.3 Although the dimensions of this roll are apparently about the same as those of the pieces here to be described, and although Battifol mentions that the piece he describes was mutilated at the bottom, there is no question of the identity of the two, since their content as well as their style is dif­ ferent. There is also no reason to suppose that our pieces were part of the much-neglected archiepiscopal collection. There were several other churches in the city, in which any surviving MSS may be sup­ posed to have been equally neglected. Rolls of this kind are not uncommon. In fact the collections of the great libraries of Europe contain numerous manuscript examples which preserve one or another of the forms of the Greek Orthodox liturgy in copies dating from as early as the XI to as late as the XIX century.4 The format of these parchment rolls (€l>..7]7'ap£(t,) is unique. The skins are cut to maximum length, glued together to form a strip, some­ times as long as twelve meters (Paris.suppl.gr. 578),5 and attached at one end to a knobbed spindle (Vat.gr. 22816 and B.M.Add. 27563). The lines of writing are at a right angle to the long axis of the strip, con­ tinuous from the top to the bottom end of the strip, at which point the parchment is reversed and the lines of writing continued from bottom to top of the verso. Just as with the codex, one set of ruling serves for both sides, but, since the letters are, as usual in Greek MSS, pendent from the ruling, the lines of writing on the verso fall in the spaces between the lines on the recto, with the result that the eye­ confusing show-through of ink, which is an inevitable result of the codex form, is minimized. I do not, of course, suggest that this techni- 3 P. Bauifol, "Les manuscrits grecs de Berat d'Albanie," ArchMiss, 38 SER. 13 (1887) 437-66. For the description of the liturgy roll see p.444. 'E.g. Paris, Bibliotheque Nationale; London, British Museum; the Vatican Library, and, of course, the libraries of the monasteries of Mt Athos and the National Library and the Byzantine Museum in Athens. Examples in codex form are much more numerous. 6 This exceeds somewhat the maximum length of known Greek literary rolls. Cf F. W. Hall, Companion to Classical Texts (Oxford 1913) 7, who mentions the British Museum roll containing the text of Hyperides, which is about 28 feet in length. • For an illustration cf Franchi de' Cavallieri and Lietzmann, Specimina codicum graecorum Vaticanorum (Berlin 1929) pU5, and for a full description A. Turyn, Codices graeci Vaticani saeculis Xill-XIV scripti (Vatican City 1964) 23-25. LLOYD W. DALY 111 cal feature has anything to do with the choice of this format for liturgy rolls.7 What originally dictated the use of this format is not, so far as I am aware, a matter of record. There are few examples earlier than the XII century. No doubt the wear to which a liturgical work would be subject can account for the disappearance of earlier copies. Still we have no assurance that the format has a history more ancient than the preserved specimens. There is a natural inclination to suppose that this use of the roll form would be a survival from the ancient papyrus book-roll, and religious conservatism might be invoked to explain the survival. But the papyrus book-roll was a very different thing, its lines of writing parallel to the long axis of the roll, in columns at a right angle to the long axis, and the verso left blank. Thus, while it is chrono­ logically antecedent to the liturgy roll, its characteristic format will hardly serve to explain that of its later congener.s There is, indeed, a parallel to this format in the Latin Exultet Rolls of south Italy, which have their text disposed in the same manner as the Greek liturgy rolls but contain illustrations alternating with sec­ tions of text, oriented in the opposite direction to the text, so as to be visible to the congregation as the roll was slowly unwound by the officiating priest and allowed to drop over the edge of a lectern.9 The examples of these rolls at Bari and Pisa, e.g., are no older than the XI century and can hardly be regarded as antecedents of the Greek rolls. The most that can be said would seem to be that they are similar. There is, of course, the possibility, since the Exultet Rolls are of south Italian origin, that they have been influenced by Byzantine liturgy rolls. If the use of this particular format goes back, on either side, to much earlier Christian usage, the probability that it is of eastern origin increases. If again this roll form should prove to be of early Christian origin one might look to Hebrew practice for its inspiration. The Torah does, of course, persist in roll form, but it follows the Graeco-Roman prac- 7 For a description of the roll form and a brief listing of examples cf V. Gardthausen, Griechische Palaeographie2 I (Leipzig 1913) 152-54. 8 Cf E. M. Thompson, Introduction to Greek and Latin Palaeography (Oxford 1912) 50-51, who speaks of the survival of the roll form, but with a changed direction of writing, in the Middle Ages. It may be questioned whether this is properly a survival or a different form with or without ancient models. t For the format of these rolls and scenes showing their use cf Myrtilla Avery, The Exultet Rolls of South Italy (Princeton 1936). 112 ROTULUS BERA TIN US, A GREEK LITURGY ROLL tice, with its lines of writing parallel to the long axis. This is also the case with the Dead Sea Scrolls. One other parallel occurs to me, and that is the so-called Ravenna Roll. This parchment piece is represented by a remnant 19 cm. wide and 3.60 m long. It is now in the Archivio del Principe Pio at Lugano. It was written originally on only one side and contains a series of Latin prayers associated with the Advent, or perhaps more probably with Christmas, disposed as in the Greek liturgy rolls. It has been associated with Ravenna because of the content of the letters which have been copied on its verso in a hand later than the original of the recto. The large, bold Latin uncials of the recto have been dated by E. A. Lowe to the early VIII century.lO Although it was originally written on only one side, it is in its format the closest parallel to the Greek rolls. If Lowe's dating can be pressed for that high a degree of accuracy the roll would have been written under the Exarchate, and similarity to Greek rolls would not be surprising.H In this context, however, it would be most probable that the Latin roll was made in imitation of Greek rolls. If this be the case we have testimony to the existence of the roll format for the Greek liturgy as early as the VII century. Thus I can only raise the question of the origin of this format with­ out being able to provide any answer. As more attention is paid to the evolution of book forms and as new material and evidence becomes available an answer may be forthcoming. My purpose here is to describe the roll fragments mentioned at the outset (see PLATES 5 and 6). The maximum original width of the pieces is 25 cm. The roll was at some time cut into pieces. Piece la is 70.5 cm. long, piece Ib is 75 cm., and piece 2 is 33 cm. Pieces la and lb were originally continuous. Something is lost between 1 band 2, and 2 was not the last piece in the strip. The right-hand side of piece la has been badly eroded, perhaps by mice, and thus the text is mutilated down to its 29th line on the recto and to a corresponding point on the verso. The overlap for gluing pieces together is 1 cm. wide. 10 Originally published by A. Ceriani and G. Porro, II rotolo opistografo del Principe An­ tonio Pio di Savoia (Milan 1883) with facsimiles. It is discussed with bibliography by H. Le­ clerq, s.v. ROULEAU DE RAVENNE in DACL. Lowe's description is in Codices Latini Antiquiores III, no.371. See also K. Gamber, Codices Liturgici Latini Ar.tiquiores (Freiburg 1963) no.270, with additional bibliography. 11 Even though, as Gamber observes, the Roman rite had been introduced at Ravenna at least a hundred years before the time of the roll. ROTt:LLS BERATI~US, PART OF 1 bv (Photo red [iced abollt one-half) DALY PLATE 5 .. ~ PLATE 6 DALY " N ~ ~ "5 0 >- F-< ~ ~ ~ '" A. ::: c CI'J" ...... ;:l ~ ~ ~ '::i F-< -~ § ~ ~ ~ , o::l ;::: CI'J C ;:l ....:I 'Q ;:l S F-< 0 r::G LLOYD W. DALY 113 The ruling was done with a dry point pressed so heavily as almost to pierce the parchment in places. The prickings are visible where the margin has not been disturbed. The vertical rulings for the single column of text define a column with lines 12.5 cm. wide. Parallel to these rulings on either side is a second vertical ruling which defines a part of the margin 2.5 cm. wide, within which large initial capitals are written. The lines of writing are thus 12.5 cm. long, and the horizontal rulings for them are at intervals of 1.5 cm. The letters themselves, in­ sofar as they are confined between two lines, are about 0.5 cm. high, while the majuscule headings for individual prayers occupy a full line space 1.5 cm. high. The writing itself is, as previously mentioned, pendent from the lines. The majuscule headings are written in typical ornamental Byz­ antine uncials, with frequent abbreviations and monogrammatic combinations involving small medial and superscript letters. These headings, the large initials in the rubrics and the whole of the abbre­ viated €KcpcfJ are written in red and washed with gilt. The script of the text is a striking and harmonious product of a hand that has full command of both uncial and minuscule forms, and com­ bines them in a proportion in which the uncial element is very heavy indeed. In fact every letter of the alphabet appears in both uncial and minuscule forms. This reintroduction of uncial forms in the already firmly established minuscule script is notable in the second half of the X century and becomes most marked in the XI.12 Our case is an ex­ treme one and seems to me to find close parallels in such dated MSS as Vat.Ottob.gr. 422 (A.D. 1004), Vat.gr. 1675 (A.D. 1018), Modena Bibl. Es­ tense Ms.gr. 230 (A.D. 1050), Vat.Pal.gr. 259 (A.D. 1054), B.M. Add. 19352 (the Theodore Psalter, A.D. 1066), Meteora Metamorphosis Cod. 548 (A.D. 1089).13 The script of the text is, however, further differentiated. The text of the long prayers, those introduced by a majuscule heading, are writ­ ten in the mixed script just referred to. That of the ceremonial direc­ tions (rubrics) and of the shorter prayers associated with them, on the other hand, is written entirely and exclusively in the type of uncial 12 cf E. Follieri, "La reintroduzione di lettere sernionciali nei piu antichi manoscritti greci in minuscoIa," in Bulletino dell' Archivio Paleografico Italiano 1 (1962) 15-36. 18 Cf Franchi de' Cavallieri and Lietzmann, op.cit. (supra n.6) pIs. 18,20,23; Palaeographi­ cal Society Facsimiles of Mss. and Inscriptions I (London 1873-83) p1.53; and New Palaeographi­ cal Society Facsimiles, 2nd SBR. I (London 1913-30) p1.5; Kirsopp and Silva Lake, Dated Greek Minuscule Manuscripts to the Year 1200 X (Boston 1939) p1.755. 114 ROTULUS BERA TIN US, A GREEK LITURGY ROLL lettering found in the mixed script, and the individual letters are kept separate from one another, with very few ligatures or connecting strokes.14 Perhaps the most notable feature of the prayer script is the line­ initial (and occasionally medial) alpha with a leftward pointing hook at the bottom, a hook which, from the point of view of letter-forma­ tion, may have been originally part of a ligature with a preceding let­ ter, but by the time of this manuscript had become independent and usually line-initial, although it occurs once with a preceding minus­ cule gamma which touches the loop of the alpha without making any connection with the hook. Gardthausen noted this letter-form and observed that, while it occurred in X-century examples, it was com­ mon in the XI century.IS Tenth-century examples are, in fact, not un­ common,16 and its use extends into the XII. While bearing in mind the caveat of E. M. Thompsonl7 to the effect that the type of writing used in sacred books is extremely conservative and that it is wise to exercise some hesitation in attempting to fix the exact date of such a manuscript, it seems to me that a date somewhere near the middle or in the latter half of the XI century would not be unreasonable for the script of this roll. The closely similar script of the Theodore Psalter of 1066 is, after all, also that of a sacred book. The wide-sweeping circumflex used may suggest a somewhat later date, and it cannot be denied that archaizing conservatism is a factor to be reckoned with. There is, however, no reliable criterion for assessing this factor, and one is reduced to weighing such subjective considera­ tions as the ease and naturalness with which a scribe appears to use his script. Obviously an accomplished scribe can master any script of which he has adequate examples, and our scribe would appear to be competent enough. What we have, then, is either a later scribe's imi­ tation of an XI-century hand or a piece of XI-century work, and I in­ cline to believe the latter is true. If this date be acceptable, our roll will then be among the earliest of such texts.IS a Minuscule forms of zeta, nu, xi, rho occur once each and the digraph stigma appears once in these passages. 16 Gardthausen, op.cit. (supra n.7) 218. 16 Cf. L. Th. Lefort and J. Cochez, Palaeographisch Album (Louvain 1932) pIs. 18 (913/4), 33 (954), 35 (961), 62 (992), 70 (994). 17 op.cit. (supra n.8) 222. 18 Vat.Barb.gr. 336 (formerly II1.55, formerly 77) e.g., which is dated to the vm or IX cen­ tury, is a codex. LLOYD W. DALY 115 Among other features of this script one may note the occurrence of exaggeratedly tall ligatures, as between aK, €'T}, 'T}K, vt/J, etc. and the use of a stroke over vowels at the end of lines to indicate the addition of nu. Both these features are to be found also in Vat.attob.gr. 422, whose similarity to our MS was pointed out above, and which was written on Mt Athos. This is not good enough evidence on which to base any conclusion, but it would not be surprising if our roll were of Athonite origin. As for the text of the Liturgy of St Basil preserved in the fragments of this roll, it does not vary greatly, so far as the prayers are con­ cerned, from that of Trempe1as in his TPE'iC A€£Tovpyta£.19 The por­ tions of the text preserved are the following (Tr= Trempelas, with reference to his page and line; Be= Beratinus, with reference to pieces):2o Be recto 1a= Tr 167,4 Kat to 168,7 (see 71,6) capKLKa'ic Be recto 1 b = Tr 71,6 bn(}V/LtaLc to 76,6 eX/Lap[ TW~OV Be recto 2=Tr 180,16 TeVV 8ov~wv to 180,28 ~af1wv Be verso 2= Tr 185,3 lKEclaLc to 185,14 aVTa Be verso 1b= Tr 190,3 a7To~oylav to 190,32 (see 129,19) 86g'T}c Be verso 1a= Tr 129,19 rilc f3acLAEtac to 129,23 Aacr and 195,5 XaPLCTOV/LEV to 192,13 E[lc 7TEPL7TOt'T}CLV The variants in the prayers are the following: Tr 68,1-2 Kat • •• alwvwv: om. Be Tr 70,6-7 Kat ••. alwvwv : om. Be Tr 73,1 a/LETp'T}TOV : a/LETpOV Be Tr 73,2 avaAAo£wTwC: -TOC Be (with B) Tr 73,3 iXP'T}/LdmCaC : -T'T}cac Be Tr 75,1 EEpacpEt/L: cEpacpt/L Be Tr 180,17 7TpoKaTaYYEAAwv ~/L'iv: ~fl'iv om. Be (with H) Tr 185,4 7TPOKEKO£/L'T}/LEVWV : KEKO£/L7J/LEVWV Be (with I F) Tr 190,6 &: J)JI Be (with B) Tr 128,2-3 vvv ••• alwvwv : om. Be Tr 190,26-7 aKaTaKptTwc : -TOC Be Tr 190,27 cov: om. Be (with B) , EV- 19 Texte und Forschungen zur byzantinisch-neugriechischen Philo!ogie, Nr.15 (1955) 161-93. 20 Because of the practice of not repeating the text of identical prayers one must, in using Trempelas' text, refer back to that of the Liturgy of St John Chrysostom for the complete text. The references in parentheses are to page and line in that liturgy. 116 ROTULUS BERATINUS, A GREEK LITURGY ROLL Tr 129,13 Kat aE' ... alwvwv: om. Be Tr 129,23 a'lp.aToc: a'lp.aT6c cov Be (with A) Aa