"~'-^! i-j-fc^Tt.;?- MASTER NEGA TIVE ^O. 91-80229-13 MICROFILMED 1991 COLUMBIA UNIVERSITY LIBRARIES/NEW YORK as part of the "Foundations of Western Civilization Preservation Project" Funded by the NATIONAL ENDOWMENT FOR THE HUMANITIES Reproductions may not be made without permission from Columbia University Library COPYRIGHT STATEMENT The copyright law of the United States ~ Title 17, United States Code - concerns the making of photocopies or other reproductions of copyrighted material... Columbia University Library reserves the right to refuse to accept a copy order if, in its judgement, fulfillment of the order would involve violation of the copyright law. AUTHOR: I •a:. an. *f ;*■"«- w |mi^«'' E.A TITLE: Ml M. .j £m \^ JLj • UN TE: -4 O PHICA A , 'fa ^^^^^ COLUMBIA UNIVERSITY LIBRARIES PRESERVATION DEPARTMENT Master Negative H I3I13LIOGIIAPHIC MICROFORM TARGET Original Material as Filmed - Existing Bibliographic Record 06 3M PI 1910 877.17 Z v.l E. A. (Ellas Avery), 1879-1969. Lowe, Eliao Avory^ 1879 1969. ... Studia palaeop;raphica; a contribution to the history of early Latin minuscule and to the dating of Visigothic mss., with seven facsimiles, by E. A. Loew. Vor^elegt am 5. no- veniber liilO. Munchen, Konitrlich bayerische akademie der wissenschaften, in kommission, G. Franz, 1910. vlli, 91 I). 7 facsini. 22i^». (Sitziinpsberichte der Konlglich bave- rischen akademie der wissenschaften, riiilosophisch-pbUologische und hlstoriscbe klasse, jabrg. 1910, 12. abbandlung) l.Paleograpby. Latin. 2. PAleograpiiy, Italian. 3. Paleograpby, Span- ish. L Title, u. Title: Vlsigotblc manuscripts. "- *' * •^' ^ Anothor copy. 1910. Library of Congress Copy i». (Same originally: Ellas Avery Loewj Volume of paxiphjI^^Lb^^ AS182.M823 Jabrg. 1910, 12. abb. Z114.L91 i34bl, Restrictions on Use: TECHNICAL MICROFORM DATA FILM SIZE: j^^^ IMAGE PLACEMENT: lA Cll REDUCTION RATIO: _ ID IIU DATE FILMED: ^//^^/^ INITIALS^_ii/^_4^Hiia«itffc.||lpa^^ J 1 ' 877 17 Columbia ®nit)trs?ftp LIBRARY t^ *^ I' t i I 1 ' - COITEMTS 1. Lowe, E. A. Studia pelaeographia ; a con- tribution to the history of early Latin miniscule. 1910. 2. Gardthausen, Viktor, tostarna; oder, Servius Tullis. 1882. 5. Dasti, Luigi. Corneto Tarqumm. 1878. 4. Sartori, Karl. Studien aus dem gebiete der griechischen privataltertunier . 1893. 5. Schliemann, Heinrich. Bericht iiber die ;rabungen in Troja im jahre 1890. 1891 ausg ^11 -H .1 \ \\ o,\ Sitzungsberichte der Koniglicli Bayeriscben Akadeiuie der Wissenschaften Philosophisch-philologische und historische Klasse Jahrgang 1910, 12. Abhandlung Studia Palaeographica A contribution to the history of early Latin minuscule and to the dating of Visigothic MSS with seven facsimiles . • • by 1 » » » .» • * « » « • . . .E^ .A. Loev;. . > ■ « ( ( ■ * * t « •w- t • ■ Vorgf'leat an'i 5.'-Nx/venfibar 1910 Munchen 1910 Verlag der K5niglich Bajerischen Akademie der Wissenschaften in Kommisaion des G. Franz'schen Verla{,'s (J. Roth) • - V ( To • • •• • • • • • • • • • • • I; the memory of • • •* • • • • • • » » • • • • • •.• • - • LEOPOLD DELIb 1 < I J • "s. '/f* n ,-i»-' Since the epoch-making contention of Scipio Maffei, the illustrious Veronese archaeologian and palaeographer, we have come more and more to recognize with him how important a role was played by the Cursiva Romana, i. e. the notarial script of the early middle ages, in the formation of nearly all types or schools of early minuscule. It was the rise and rapid spread of the Caroline book-hand which proved fatal to the local manner of writing in most centres. The tra- ditional script with its cursive letters and ligatures completely succumbed — in one place sooner, in another later — to the minuscule whose principle was simplicity and clarity. In giving thus a new direction to book- writing, the Caroline reform interrupted a development already past its first stage, and effaced the signs of relationship which united the different pre-Caroline types. Yet we can still realize the closeness of that relationship, and get, as it were, an epitome of the history of early minuscule, by concentrating attention upon one or two typical traits. And for this purpose there is per- haps nothing more interesting or instructive than a study of the usage of i-longa and ii. In the following studies a modest attempt is made to trace the history of i-longa, by giving an account of its cursive origin, its entrance into calligraphic MSS, its rapid spread and short-lived vogue in all but two schools, and the rules which in those tw^o schools seem to have governed its use. This account can be turned to practical use by the philologist. To the palaeographer its value lies in the light it throws on the different types of minuscule in process of formation, and in the explanation it offers for such curious phenomena as the employment of i-longa in early examples of schools so far removed from each other by space and tradition as the Spanish and the north Italian. The remaining and larger part of these studies deals with the history of ti, and tries to show through what medium the fi-ligature was introduced into calligraphy; how it was used in various centres and then discarded by all but the Bene- ventan; how the last-named script reserved it for the specific VI 12. Abhandlung: K. A. Loew purpose of indicating the assibilated sound of ti; how the Visigothic like the Beneventan graphically distinguished the hard" and soft sound of ti; and how this practice furnishes a terminus a quo for dating Visigothic MSS - a criterion whose application will remove some traditional errors from Spanish palaeography and prove its validity in several mooted cases. Incidentally the question of transcribing this ligature will be raised as well as that of a similar form which has been a problem in diplomatics - a form of ^ as yet unrecorded in our literature. The question of phonetics is outside the province . of this investigation. If the data based upon the MSS which served my palaeographical purposes prove also of some value as raw material and evidence to the student of Romanic lang- uages, it will only serve to confirm my conviction that apparently insignificant and usually neglected graphic points have their bearing upon the broader problems of history and philology. To avoid repetition the data for i-longa and ti will be given together; their history will be treated separately. My warmest thanks are due to Professor W. M.Lindsay. These studies have profited from his interest and advice as well as by the information which he put at my disposal with rare generosity. I am also grateful to Professor C.U.Clark for his kindness in permitting me to make use of his valuable collection of Visigothic photographs prior to their publication. Lastly it is my pleasant duty to acknowledge my in- debtedness to the American School of Classical Studies in Rome under the auspices of which I have had the privilege of continuing my studies as Research Associate of the Carnegie Institution of Washington. To the Director of the school and to the members of the committee in America I herewith ex- press my sincere sense of obligation. It is not to be my privilege to put this monograph into the hands of Leopold De lisle. In remembrance of his kind- ness in making public a portion of the results, I do myself the honor of dedicating these studies to his memory. Rome, July 1910. It Studia palaeographica. VII Table of Contents. P «t I. Preface .... i-longa in inscriptions i-longa in cursive Origin of i-longa in MSS Function of i-longa in MSS Summary of usage in MSS Usage in particular schools a) In Visigothic b) In Beneventan i-longa as a sign of relationship between the i-longa and philology .... a) Practical value of rules for i-longa b) Resemblance of i-longa and I c) Interchange of i and g and vice-versa d) Corruptions in text due to this interchange schools Page V 1 2 4 5 7 ^ t 8 9 10 13 13 14 14 15 II. Assibilation of tK The ii-distinction The ligature ti. Its forms Origin of the ligature Usage in cursive Summary of usage in MSS Transcription of ^ . Transcription of a similarly formed £ 16 19 20 20 23 25 26 III. The MS evidence a) Uncial . . . • b) Semi-uncial . c) Early French minuscule 29 30 30 30 ^ -41 7* rt VIII 12. Abhaiidluncr: K. A Loew d) Early Italian minuscule e) The Beneventan MS^S^ . f) German schools g) Insular MSS IT. ti m Visigothic MSS Nature of evidence Works cited MS evidence Evidence of corrections, additions and docuuienis. Results. Criterion for dating .... Four periods in Visigothic writing The fi-criterion tested V. Plates VI. Index of MSS Page 39 46 47 49 52 54 55 56 70 78 80 81 1 - 80 88 .■« ; Studia palaeographica. I. The main function of i-longa with which the student of Latin epigraphy is acquainted is foreign to the i-Ionga of Latin MSS. The i-longa in words like vIxiT, lIberti, dIvo, pkIncipi etc. of Roman inscriptions serves the specific pur- pose of denoting the long quantity of the letter i.^) In Latin MSS i-longa has no reference whatever to quantity. The use of i-longa in inscriptions is, on the whole, optional and not strictly defined. One engraver may use it, another of the same period may not. And the same engraver may use it to indicate the long vowel in one part of the inscrip- tion and not in another. It may be employed at the be- ginning of a line merely as a decorative element, likewise in the middle of the line as in flamIne^) or out of a sense of reverence as in Imperatori^) In MSS, on the other hand — at least in those of certain schools and certain periods - the use of i-longa is obligatory and subject, as we shall see, to definite rules.*) If there are these differences, there is also one important point of similarity. 1) On the subject of i-longa in inscriptions see: Christiansen, De apicibus et i-longis inscriptionum latinai-um (Kieler Disser. 1889), p. 26 sqq. 2) Christiansen, 1. c, p. 28. The Corpus Inscr. Lat. is full of such examples. 3) Ibid., p. 37. *) See below, p. 8 sq. Excepting the brief report of my observations which was made by Leopold Delisle (Comptes-rendus de I'Academie des inscriptions, 1909, p. 775— 778) and reprinted with corrections in the Bibliotheque de I'ecole dea chartes LXXI (1910), 2.^3-235, there exists no connected account of i-longa in MSS. The usual statement found in the descriptions of plates is that i-longa occurs often at the beginning of the word and occasionally in the middle. Sitzgsb. d. philos.-philol. u. d. hist. Kl. Jahrg. 1910, 12. Abh. J I- 2 12. Abhandlung: E. A. Loew The use of i-longa to denote the semi-vocal sound, which in inscriptions is as old as the use of i-longa itself, is a con- stant feature of those MSS which regularly employ i-longa. Such familiar epigraphic forms as eIus, huIus, COnIunx, lUNiUS etc., have their exact graphic equivalent in Latin doc- uments and MSS. Yet there is this difference: the engraver may make a long or a short i in eIus, Iunius etc., but during many centuries the scribe of southern Italy or Spain is obliored to use the long form — as can be seen from the evidence cited below. Against the one point of similarity, then, there are several points of difference, one of which alone is so grave as to make it quite improbable that the use of i-longa in MSS is a direct inheritance from inscriptions. For, if that were the case, should we not expect to find MSS with i-longa used to indicate the long quantity? Such MSS, how- ever, do not exist. Yet a point of contact between the mediaeval and the ancient practice respecting i-longa doubtless exists. It is to be sought, I believe, in the domain of cursive writing. As a matter of fact, we find i-longa in the Pompeian mural inscriptions in cursive used in the manner in which it is later employed in mediaeval documents and MSS, namely, at the beginning of the w^ord regardless of quantity or the meaning of the word, and medially for the semi-vocal sound. ^) In order to see how the ancient cursive practice was taken over and introduced into calligraphy we must examine the connecting link, i. e. the mediaeval or ''later" cursive. Without going too far into detail the usage in the documents may be briefly sketched as follows. The Ravenna documents on papyrus of the 6^^ and 7*^^ centuries*) — and not a few of them have come down to 1) Cf. Christiansen, 1. c, p. 36 and C. I. L. IV, indices, p. 258. *) In fact, i-longa is found also in earlier documents. In Marini's facsimile (Papiri Diplomatici, Rome 1805), pi. 6, No. 82, a. 489 I find Id, lubeatis. But in the still older example of cursive on papyrus, in Strass- burg (Pap. lat. Argent. 1), i-longa is used apparently without any system: domlne, Inimltabili, benlvolentiae etc. Facs. Arndt-Tangl, Schrifttafeln, Heft 2*, pi. 32 A; Steflfens, Lat. Pal.2, pi. 13. Studia palaeographica. " ugi) — show the frequent occurrence of initial and medial i-longa: In, Interfui, Iterum, Ipsum, huIus etc. etc. Marginalia found in 6^^ century semi-uncial MSS written in a slanting uncial-cursive of the same time also show the i-longa initially. ^) North Italian documents of the Lombard regime are con- spicuous for the regularity with which they use i-longa initially and medially. Even in words like ilia the long i is used. The usual examples are: Id, lustitia, huIus etc.^) The earliest south Italian documents show a similar use of i-longa. In the Beneventan centres the practice lasts well into the 13^^ century, and examples are known even in the 14*^ .*) Although no pre-Caroline documents from the papal chancery have come down to us, those of the 9*^ century and after may be assumed to represent an older tradition. They show the use of i-longa initially and medially, as do the Ben- eventan documents, for many centuries.^) The same holds for the non-papal documents of the city of Rome and vicinity.^) 1) They may be studied to advantage at the Vatican library and the British Museum. Facs. Pal. Society, pi. 2, 28; Arch. Pal. Ital. I, pi. 1—6; Arndt-Tangl, 1. c, Heft 1*, pi. I c, 2. 2) 1 refer to marginalia of the type seen in Delisle, Alb. Pal., pi. 7 (MS Lyon 523). Similar cursive exists in Vatic, lat. 3375, Monte Cassino 150, Rome, Basil icanus D 182 and others. 3) Facs. Bonelli, Cod. Pal Lombardo, passim; Schiaparelli, Bullet, deir 1st. Stor. Ital. 30 (1909), 2 plates. *) Facs. Russi, Paleografia e diplomatica de' documenti delle pro- vincie Napolitane, Naples lb83; Codex Dipl. Cavensis, Voll. I— VII, 1873 -1888; Codice Dipl. Barese, Voll. I, IV and V, Bari 1897-1902; Morea, 11 Chartularium del monastero d. s. Benedetto di Conversano, Monte Cas- sino 1892 ; Piscicelli-Taeggi, Saggio di Scrittura notarile, Monte Cassino 1888 ; Voigt, Beitriige zur Diplomatik der langobardischen Fiirsten von Benevent etc., Gottingen 1902 and Archiv. Pal. Ital. Vol. VII (1909), fasc. 31, pi. 20—26. 5) Facs Pflugk-Harttung, Specimina Selecta Chartarum Pontificum Romanorum. Stuttgart 1886; also Steflfens, Lat. Pal.^ pi. 58 and 62. 6) Facs. Hartmann, Ecclesiae S. Mariae in Via Lata Tabularium, Vienna 1895-1901; Fedele, in Archiv. Pal. Ital. VoL VI (1909), fasc. 30 and Vol. VI (1910), fasc. 34. 1* 4 12. Abhandlung: E. A. Loew In the Merovingian documents, of which a considerable number exist in excellent state of preservation, the i-longa plays a rather inconspicuous role.^) It is manifestly not at home there. It may be observed initially here and there. Often enough it is found in the body of a word at the end of a syllable, or at the end of a word, e. g. nostrl. This use, it should be noted, is also found in some semi-uncial MSS and some French 8*^ century minuscule MSS which recall semi- uncial, e. g. Epinal 68. But the Italian practice found its way across the Alps. Initial i-longa may be seen quite frequently in many diplomas 2) and other French and German^) documents of the Caroline age and later, but its use is inconstant. The Spanish notaries, as far as I can judge from the rather inadequate facsimiles of Merino and Munoz y Rivera,*) make constant use of i-longa initially and medially for j — precisely in the manner of the 8*^^ century north Italian notaries. The practice lasts as long as the Visigothic script remains in vogue. With this rapid survey before us we are more in a po- sition to discuss the question of the origin of i-longa. If we consider on the one hand the utter absence of i-lonsra in the oldest Latin MSS in uncial and semi-uncial from the 4'*" to the 7'^ century, and its gradual and tentative entrance only into uncial and semi-uncial MSS of the recent type i. e. of the 8^*" and 9*^ centuries; and on the other hand its very frequent and continued use in cursive documents dating from the 6^'' to the 9^^ century (in many cases even much later than the 9*^ century), it seems reasonable to explain the presence of i-longa in most of the pre-Caroline MSS in min- uscule as the result of direct imitation of the cursive. Nor ^) Faca. Lauer-Samaran, Les Diplomes originaux des Merovingiens. ^) Fac8. V. Sybel & Sickel, Kaiserurkunden in Abbildungen (Berlin 1880—1891) especially Lieferung I and III; also Schiaparelli, Archiv. Pal. Ital. Vol. IX (1910), fasc. 33, pi. 1 — 12. ^) For German documents see facs. in Chroust's Monumenta Palaeo- graphica. *) Merino, Escuela Paleografica, 1780 and Munoz y Rivera, Paleo- grafia Visigoda, Madrid 1881. ^ •« »• ; Studia palaeographica. & would i-longa in this respect present an exceptional pheno- menon. An examination of the extant examples of early minuscule of the 7*^ and 8^'' centuries shows that often enough the calligraphic scribe of those centuries did not hesitate to appropriate from the domain of the notary many another feature beside the i-longa. The fact is familiar to the palaeo- grapher. He thinks at once of the open a, the broken c, the peculiar t, as well as of the more striking ligatures of fi, riy ti, te, ta, til etc. Moreover a comparison of the calligraphic pro- ducts in minuscule of the 7^^ and 8**^ centuries with the notarial documents of the same period will convince any observer that the calligrapher borrowed freely from the notary. It is hardly necessary to demonstrate that the reverse was not the case. For the careful methods of the calligrapher were not suited to the rapid, economical and practical methods of the notary ; whereas the calligrapher, in his efforts to form a minuscule script, that is a more economical script, took over cursive liga- tures and cursive forms of single letters because they were more easily traceable and thus more economical. Finally, con- siderable light is thrown upon the origin of i-longa by the fact that it flourishes in MSS which employ cursive elements, and that it is avoided in MSS in which cursive elements are few or wanting altogether. In other words, the company in which we find i-longa is a fair indication of its origin.^) In view of the above considerations there can hardly be any serious doubt that i-longa came into MSS from the cursive. The primary purpose which i-longa served in cursive writing can only be conjectured. The fact that it is most frequently found at the beginning of a word suggests that it owes its origin to the desire of facilitating the reading; the appearance of the long form of i indicating at once the be- 1) See below, p. 12. In Paris 653, a north Italian MS of about 800 A. D., this point is clearly illustrated. On fob 6^ two hands can be seen. The first used the a'-ligature and the i-longa regularly. The other hand used neither. Cf. plate 2. This facsimile 1 owe to the kindness of Prof. W. M. Lindsay. / ,'A 6 12. Abhandlung: E. A. Loew ginning of a word. Whereas the book-hand with its scrip- tura continua neglected such aids, partly no doubt for reasons of symmetry, in cursive, on the other hand, where symmetry played no role, where words were often abbreviated by any capricious suspension, and a short letter like i could be easily overlooked, the use of a long form of the letter i initially must have been of signal assistance to the notary who had to read or copy the document.^) Perhaps this need of giving more body to the small letter i was first felt in words in which letters with short strokes followed initial i, as In, Im- peratoris^) etc. By analogy its use may have spread to any word, so that in the 8^^ century north Italian documents lUe and ibi are written with i-longa as well as in, imperatoris etc. But we find i-longa in documents not alone at the be- ginning of the word, but also in the body. The reasons sug- gested above for using i-longa initially are in so far applic- able to its use in the body of the word as the long form of the letter here also facilitated reading. A consideration, how- ever, of the examples of medial i-longa shows that with this form of the letter went a specific pronunciation. The writing of hiiltis, cuius, fualor, leluniis, makes it clear that the long form of i has reference to its semi-vocal sound. Whatever may have been the reasons for the employment of i-longa in cursive, the important fact remains that in many pre-Caroline documents the long form is constantly used in these two ways: initially, and also medially for the semi- vocal sound. 1) In this connection it is interesting to cite Zangemeister's opinion respecting the purpose of i-longa in the Pompeian mural incriptions in cursive: "Patet maxime in eis (sc. inscriptionibus parietariis Pompeianis), quae cursivis litteris exaratae sunt, inscriptionibus i saepe productam esse non alia de causa nisi ut eius litterae forma niagis plane et per- spicua redderetur". C. I. L. IV, indices, p. 258. *) At any rate, it is a striking fact that i-longa clings longest to such words as in, ita etc. even in scripts which had given up its regular employment. ,J r Studia palaeographica. • It is precisely this use of i-longa that we encounter in MSS. From data given below ^) the course of i-longa in MSS may be sketched as follows. Unknown to the oldest types of uncial and semi-uncial, it gradually enters into their more recent types and is used there tentatively and irregularly. 2) The earliest minuscule MSS of Italy, France and Spain, those MSS which are occasionally styled "half-cursive" or "minuscule- cursive" make constant use of i-longa. The regular use of it which is observable in 8*^ century north Italian cursive docu- ments has its exact parallel in contemporaneous north Italian MSS. In France the i-longa is a feature of those pre-Caroline minuscule types which still cling to the cursive elements, e. g. the Luxeuil type and the c< type. During the 8*^ century it already begins to lose ground in France, so that many a Corbie MS of the icb type either lacks it entirely or uses it sparingly. In time it is practically eliminated from French calligraphy by the Caroline reform. To the compact, orderly and neat Caro- line script such a trait as i-longa manifestly appeared uncalli- graphic and was therefore avoided. Its employment in Italy lasts as long as Caroline influence does not interfere. When the scriptoria of northern and central Italy adopted the Caro- line script, i-longa was given up along with the other cursive features which formed part and parcel of the native hand. In southern Italy, however, as well as in Spain, the foreign forces never possessed sufficient energy to modify the local scripts. The old cursive practice of using i-longa, therefore, continued as long as the native script remained in use. The manner in which i-longa was used in MSS has in a general way already been indicated. But two schools de- mand our particular attention, for in Visigothic and Beneventan calligraphy the regular employment of i-longa lasted for over four centuries and died out only when the scripts went out of fashion. In the case, therefore, of these two schools it is ad- 1) See the evidence cited in the Ust of MSS p. 29 sqq. 2) The presence of i-longa in an uncial MS is an unfaihng sign that it is of the recent type. 4 4 8 12. Abhandlun^: E. A. Loew visable to illustrate somewhat more fully the rules which governed the use of i-longa. In Visigothic. I. At the beginning of a word i has the long form. e. g. lam, Ibi, Iccirca, Id, lecit, Ignem, Ihs, lUe, Impar, In, lovita, Ipse, Ira, Iste, Itinera, lus etc. Exception. When initial i is followed by a tall letter the use of i-longa is not obligatory, e. g. ibi, ihs, ille (written with a short i). 11. Semi-vocal i requires the long form.^) e. g. malas, alebat, prolciatur, alt, galus, elus,^) leluniis etc. The Spanish scribe adhered to these rules with unusual strictness. If he wrote in or huiiis with a short i it happened through inattention or slavish copying from an original which did not use i-longa. In any case he was breaking a rule of 1) It is interesting to note that Isidore does not speak of i-longa as a means of denoting semi-vocal i. He would perhaps have mentioned it, if scribes and notaries of his time had made such use of i-longa. His statement, however, is merely an excerpt from an earlier writer: *'t litteram inter duas vocales constitutam bis scribi quidam existimabant ut Troiia, Maiia, sed hoc ratio non permittit. Nunquam enim tres vo- cales in una syllaba scribuntur. Sed i littera inter duas vocales con- stituta pro duplice habetur". Etymol. 1, 27, 11. 2) Much light upon Visigothic palaeography was thrown by Delisle's description of the Silos MSS in Melanges de paleographie et de hihlioyraphie. From what he says of i-longa (p. 56) it appears that he failed to realise the rules governing its use: "On trouve I capital tres allonge non seule- ment au commencement du mot, raais encore a la fin, surtout quand la desinence est figuree sous nne forme ahbreviatire: el^ pour ejus'' (italics are mine). The fact that the form is abbreviated is a matter of indif- ference, eius would have the i-longa even if written out. When the ws was abbreviated the i-longa naturally remained. But i-longa at the end of a word is absolutely foreign to Spanish calligraphy. Mufioz, Paleografia Visigoda, has nothing on the regularity of i-longa in Visigothic MSS. IV V Studia palaeographica. 9 the script. I have noted such irregularities in very few MSS.^) The utter neglect of the rule in these cases was a proof that the scribe was laboring under foreign influences. Here mention should be made of a type of i-longa pe- culiar to Spanish MSS. It is a long i with a forked top resembling on the whole a tall y. It is frequently found in the word ait. Examples are cited below in the list of Spanish MSS. In Beneventan. The two main rules for initial and medial i-longa which prevailed in Visigothic scriptoria hold for Beneventan.^) There is, however, this difi'erence between the Beneventan and the Spanish scribe: the former was more averse to using i-longa before a shafted letter. He regarded it as un calligraphic and therefore eschewed it. It is only in very few Beneventan MSS — and these are all of the early period, i. e. of the 8*^^ and 9^'' centuries — that we find initial i invariably long. The rule is to write short i when the following letter has an upper or lower shaft, e. g. ibi, ihs, illi, ipse, ire (the r has a shaft), iste etc. Another exception to the main rule of initial i-longa occurs when the preposition precedes the noun which begins wath i, e. g. ad imagincm, In itaJiam. In such cases the scribe was accustomed to run the noun and the preposition together, and as he wrote them together he regarded the phrase as a unit and therefore wrote short i. This circumstance, it may be noted in passing, seems to confirm what has been said of the purpose of i-longa, namely, to call attention to the begin- ning of a word. On the other hand, the use of i-longa in 1) e.g. Paris 10876 and 10877. See below list of Spanish MSS. 2) How little the rule for medial i-longa was recognized by Ro- stagno (Praefatio, p. IX, to the Leyden reproduction of the Tacitus MS, Floren. Laur. 68, 2) is seen from his words: "i grandi, quae vocatur, usus est non nunquam librarius ineuntibus vocabulis, cum praesertim subeat u littera: semper, ut quidem, post u in vocabulo cuius; item in iniuria, obiectare, maior, coniugium, coniunctio cet.". 10 12. Abhandlung: E. A. Loew Studia palaeographica. 11 delude, exinde is doubtless due to the inveterate habit of writing in with the long form of i, the excuse being furnished by the composite character of the two words. It is possible to cite not a few instances in which Bene- ventan scribes break the rules. But this is mainly the case during the formative and uncertain period of the script, i. e. durinor the 8^^ and 9**" centuries. The careful scribe conscien- tiously observed them, and the best possible proof that they ivere rules of the script is furnished by the autograph of Leo Ostiensis (Monacensis 4623). In making the additions and corrections in his chronicle of Monte Cassino Leo was hard pressed for space. The long form of i is certainly not the most economical. Yet in all the pages of small and crowded writing the above rules are carefully observed. We have seen that in at least two scripts i-longa was a constant feature for several centuries. In this respect the Visi- ffothic and Beneventan are different from other hands. We have also seen that the use of i-longa in both these schools was governed practically by the same rules (rules which al- ready obtained in the 7^^ and 8^^ century documents) and that of the two the Visigothic showed stricter adherence to the rules. The question which naturally arises — and it is one of no little interest to palaeography — is this: did the Visi- gothic serve as a model to the Beneventan?^) If it were not for the fact that nearer and more likely models existed, the answer to the above question would have to be an unqualified affirmative, considering the importance and vogue of Spanish literature in the 8*^ century just when the Beneventan script was springing into life. But the south Italian minuscule could easily borrow the use of i-longa from its own notarial products; and if it went farther for its models, north or central Italian documents as well as MSS of the 7*^ and 8*^ centuries could have supplied them. This being ^) The reverse is out of the question, since the Beneventan as a script was just beginning its existence when the Visigothic had already reached maturity. *x .< the case, and as no actual proof exists that the Beneventan took over the practice of i-longa or any other calligraphic feature from Spanish calligraphy, it is more reasonable to ex- plain the matter somewhat thus: as the Beneventan has many cursive elements which are not found in Visigothic, the pre- sence of i-longa must be regarded in the same light as the presence of the other cursive elements, namely as a remnant of the traditional Italian minuscule in which cursive features, adapted to calligraphic purposes, played a large role. If it is true that the Beneventan does not depend upon the Visigothic for its use of i-longa, the same can be said with even greater emphasis of the north Italian schools. For if we assume for a moment for the sake of argument the direct dependence of north Italian upon Spanish MSS with regard to this point, we are at a loss to explain the same use of i-longa in contemporaneous north Italian documents. And no one would try to maintain that Italian notaries copied from the Spanish. The opposite is not only more probable, but doubt- less was the case. The Spanish notary built upon Roman tradition; his model was the Italian notary. The knowledge of the i-longa w^hich the Spanish notary had he owes to his Italian cousin. The knowledge of it possessed by the Spanish scribe is doubtless knowledge gained from the notary. And the same conditions which made the Spanish scribe turn to cursive for new material also made the north Italian scribe borrow from cursive. And that he really did so can best be illus- trated by two concrete examples. It is impossible not to realize the points of similarity between the Ambrosian Josephus on papyrus of the 7^^ century and the Ravenna documents of about the same period. It would almost seem that the calli- grapher in this case also filled the post of notary. The fact that interests us now is that the Ambrosian MS, whose style is little removed from a cursive document, uses the i-longa regularly at the beginning of a word and medially when semi- vocal, i. e. precisely in the manner of later Spanish scribes and notaries. The Ravenna notary certainly did not learn from the S 12 12. Abhandlung: E. A. Loew V Studia palaeographica. 13 Spanish; nor was the scribe of the Amhrosianus under any obligation to a Spanish scribe; for even the existence of a Visigothic minuscule at that date can only be assumed, not demonstrated. But a more cogent example is that furnished by the 8^^ century north Italian MS Vercelli 183 (see plate 1). Several other MSS — for instance, those from Bobbio,^) — might also be pressed into service to illustrate my point. But I single out Vercelli 183 because its north Italian origin as well as its dependence upon notarial writing is practically demon- strable. First of all the general impression of the script bears distinct resemblance to the writing in north Italian documents of the Lombard regime, the main difference being that the MS is orderly and calligraphic, and manifestly the work of an expert scribe. But the scribe attempted to use a certain form of -s* (cf. plate I, line 11) which is almost unique in MSS.^) This form of the letter, however, is not rare in north Italian docu- ments of the S^^ century. Here we have, as it were, caught the scribe in the act of appropriating a cursive element. Now this scribe makes constant and regular use of i-longa initially, and medially when semi-vocal. The contemporaneous north Italian notary does precisely the same. Far from explaining this fact as due to the influence of Spanish models — and it is important to note that both the abbreviations and the ortho- graphy show no trace whatever of Visigothic influence — the above considerations force us to admit that the writer of Ver- celli 183 merely took over i-longa as he did the singular form of ^, from the cursive writing practiced in his region. The use of i-longa, therefore, in all the schools is due merely and entirely to the influence, mediate or immediate, of cursive upon calligraphic writing. With this in mind, we can easily understand how the Caroline reform which banished cursive elements from the book hand, was inimical to the use of i-longa; also, how its use happened to remain a feature of •^, \ \ > *) See below the list of Italian MSS. ') See below, p. 26 sqq. ] Beneventan writing, which is par excellence the script which calligraphicized cursive elements ; and lastly how two such distant schools as the north Italian and the Spanish used the i-longa in precisely the same way. Maffei's view of the common origin of the different types of minuscule is instructively borne out by the results of this little investigation of the use of i-longa. i-longa and philology. Heretofore our considerations have been purely palaeo- graphical; but the question has also its practical side. Some of our important authors have come down to us through the medium of Beneventan or Visigothic transmission. When such a text depends mainly upon a single MS, and that MS is in a bad state of preservation — I need only mention the Annales and Historiae of Tacitus, Varro's de Lingua Latina and the fragments of Hyginus in Beneventan writing — its editor will not fail to profit from the rules formulated above (cf. p. 8 sq.). For some of the errors which creep into the text are manifestly due to ignorance of these rules. No less a philologian than Halm, in his edition of the fragments of Hy- ginus (Monacensis 6437) misread i-longa for an I. His un- familiarity with another rule in Beneventan, that of the lig- ature n, was the cause of two errors in one word. Halm gives malorum where the scribe wrote maiori^) with i-longa as is required by the rules of his school. In a passage in the Historiae of Tacitus (IV, 48, 10) editors have wavered between the readings ius and uis.^) Its last editor, Andresen, gives: legatorum ius adoleuit. The Ben- eventan MS upon which the text is based (Floren. Laur. 68, 2) is hardly legible on that page as the ink has grown very pale. It was in fact illegible in the time of the humanists, as appears 1) Cf. Kellogg, in Amer. Journal of Philology XX (1899) 411. 2) Cf. Andresen, In Taciti Historias studia critica et palaeographica II (1900) p. 13. s / 14 12. Abhandluncr: E. A. Loew ^\ Studia palaeographica. 15 from the interlineal transcription of the text.^) But the two words are impossible to confuse in Beneventan, for ius must be written with i-longa and ^ds must begin with a short letter. The MS, even in its present state, shows plainly that the first letter was short, in which case the correct reading is uis and not ius — correct at least palaeographically.^) The resemblance of i-longa to the letter I could not but become a stumbling-block to ancient copyists in whose schools i-longa was not a rule. After the 9*'' century a continental scribe copying from a Beneventan or Visigothic original could easily mistake aiebat for aJehat, maias for ynalas, obiedat for obledat etc. Consequently editors must be mindful of this source of error, particularly if there is reason to believe that the archetype was Visigothic, Beneventan or in early pre- Caroline minuscule.^) The fact that i-longa did service for semi-vocal i in Spanish and Beneventan calligraphy may in a measure account for the relatively frequent confusion of i and g in the MSS of those two schools. Owing to similarity of pronunciation this inter- chancre is by no means uncommon in other schools.*) The ^) The partial disappearance of the ink is noticeable in a great number of Beneventan, especially Cassinese MSS of the lltl» century. It was evidently due to the manner of treating the parchment then practiced, for the ink has grown pale on one side of the leaf, the other, the hair-side, having retained the ink much better. 2) Cf. the Leyden reproduction of the MS in the De Vries series: Codices Graeci et Latini photographice depicti, torn. VII, 2, fol. 94^, col. 2, line 21. 3) Cf Tafel, Die "Dberlieferungsgeschichte von Ovids Carmina Ama- toria (Miinchener Diss. 1909) pp. 27 and 36. *) On the confusion of i and g owing to the similarity of sound see the following works whose title in full is given on p. 16 n. 2: Corssen, tJber Aussprache etc. P, 12G8qq.; Schuchardt. Yocalismus I, 65, see p. 70: "Im gotischen Alphabet ist G — J; zu des Ulfilas Zeit mufi also <7 vor e und i allgemein wie j gelautet haben"; Bonnet, Le Latin de Gregoire etc., p. 173 sq.; Haag, Die Latinitat Fredegars, p. 867; Carnoy, Le Latin d'Espagne etc., p. 154 — 5. /- ancient grammarians had already treated of semi-vocal i.^) And the interchange between semi-vocal i and g is evidenced by inscriptions, e. g. GEN for IAN (VARIAS) or GEIUNA for IEIUNA.2) But in MSS we find not only g for semi-vocal i, but also i-longa i. e. semi-vocal i for g.^) The latter type of error seems to me less likely in a script in which the semi- vocal i has not a distinct graphic form. It is the presence of the graphic distinction between semi-vocal and vocal i which often occasions the use of i-longa for g on the part of the Visi- gothic and Beneventan scribes. I cite the following examples from Beneventan MSS: Monte Cassino 332, saec. x, p. 13 dilesta for digesta, p. 38 quadralesime ; Floren. Laur. S. Marco 604, saec. xi, conlule for coniuge; Monte Cassino 289, saec. xi, agebat for aiebat, progecit for proiecit; Oxford Bodl. Canon. Class. 41, lulera for iugera; Monte Cassino 303, saec. xi, in. lesserunt for gesserunt; Floren. Laur. 68, 2 (Tacitus), saec. xi, lestus for gestus etc. The confusing of semi-vocal i and g is not as familiar to editors as one might expect. An instructive case in point has been kindly brought to my attention and has since been pub- lished by the Reverend Dom De Bruyne.*) He points out that in the important MS Jc of the gospels (Turin G VII, 15) the passage Mark XV, 11 is thus given: 'sacerdotes autem et scribae persuaserunt populo ut magis agcrent barabbam di- mitte nobis'. Puzzled by the word agerent some editors, as ? ^) Cf.Keil, Gram. lat. I, 13; VI, 333; Isidor. Etymol. I, 27, 11. 2) C. 1. L. V, 1717; XII, 2193, 934, 3189, 649 etc. See also Pirson, La langue des inscriptions latines de la Gaule, p. 75: 'Ti-longa ayant fini par tenir lieu du jod dans les documents de la decadence". ^) The use of g for j in Visigothic Verona 89 was noted by the editors of the Nouveau Traite (III, 449 nota). *) Cf. Revue Benedictine XXVIl (1910) 498. > 16 12. Abhandlung: E. A. Loew Burkitt and H. v. Soden, rejected it altogether and substi- tuted dicerent; another editor, W. Sanday, explained agerent as used in a "special sensed 'But the original reading was manifestly aierent.'^) 11. Assibilation of ti. The ti-distinction. As there were two distinct sounds of ti, methods were in time adopted by both scribes and notaries of graphically mark- ing the difference of pronunciation. «) In some schools the dis- tinction between soft and hard tl came to be represented by two different forms. Where that did not happen, cl often did service for assibilated ti. The practice of the various centres in this respect is on the whole sufficiently consistent to allow us at times to derive ideas of the provenance of a MS by a 1) Another instance cited by De Bruyne is that of agis for ais which also proved a source of worry to two editors. Cf. 1. c, p. 498 There are other biblical passages where the confusion occurs m parts of the verb aio. Cf. Wordsworth and White, Novum Testaraentum I, 7o7. Bonnet (1. c, p. 173) mentions similar corruption in the texts of Gregory's Historia Francornm. 2) On the phonetic value of assibilated ti and its interchange with ci see- Corssen, tJber Aussprache, Vokalismus und Betonung der lateinischen Sprache ^ (18n8) 22 sqq. The second edition, 1868-70. 1 did not have at hand; Schuchardt, Der Vokalismus des Vulgarlateins 1(1866) 155 sqq, III (1868) 317- Joret, Du c dans les langues romanes (Pans 1874) p. 66 sqq.; Seelmann, Die Aussprache des Lateins (Heilbronn 1885) p. 320; Bonnet, Le latin du Gregoire de Tours (Paris 1890) p. 170 sqq. and p. 751 1 assi- bilation de ci et ti est un fait accompli'' scil. in the time of Gregory of Tours. See also: Haag, Die Latinitat Fredegars, in Romanische Foischungen X (1899) 864 sq.; Pirson, La langue des inscriptions latines de la Gaule (Brussels 1901) p. 71 sqq.; Carnoy, Le latin d'Espagne d'apres les inscriptions (Brussels 1906) p. 141 sqq.; see also Meyer-Liibke m Grobers Grundrifi der romanischen Philologie I (Strafiburg 1904-6) 4/5. Studia palaeographica. 17 study of its ti usage. This point has heretofore received less attention than it merits.^) As I shall often have occasion to speak of assibilated and unassibilated ti^ it is advisable to make the points clear at the outset. The difference in the pronunciation between assibilated and unassibilated ti may already be observed in Roman inscrip- tions of the 2°^ century.^) The question received due attention from the grammarians. We have longer or shorter treatment of it by Consentius^), Pompeius*), Servius in his commentary of Donatus^), Papirius^) and Isidore"^). Other anonymous gram- marians of the later middle ages also touched upon the sub- ject.^) I select for quotation the passage from Papirius who wrote about 400 A. D.: 1) In giving the arguments against the Italian origin of the famous Missale Gallicanum from Bobbio (now Paris 13 246) Traube never men- tioned the fact that such spelling as PoHcio, tercia etc. was un-Italian and particularly typical of French MSS of that time. Cf. L. Traube, Palaographische Bemerkungen, in Facsimiles of the Creeds, edited by A. E. Burn, p. 45 sq. 2) Cf. Ferd. Schultz, Orthographicarum Quaestionum Decas, Brauns- berger Programm, Paderborn 1855; and E. Hiibner, Neue Jahrbiicher LVII, 339 sq. 3) Keil, Grammatici Latini Y, 395. *) Keil, 1. c. V, 104; V, 286. I quote this excerpt: "fit hoc vitium (iotacismus), quotiens post ti vel di sequitur vocalis ubi s littera est, ibi non possumus sibilum in ipsa i littera fa cere quoniam ipsa syllaba a litteris accepit sibilum etc.". ^) Keil, 1. c. IV, 445 "iotacismi sunt, quotiens post ti vel di syl- labam sequitur vocalis etc.". See also Keil, 1. c. V, 327. «) Keil, 1. c. VII, 216. For this citation I am indebted to Dr. P. Lehmann. '^) Etymologiae I, cap. 27, 28 = Migne, Patrolog. Lat. 82, col. 104, "y et z litteris sola Graeca nomina scribuntur. Nam justitia z litterae sonum exprimat, tamen, quia Latinum est, per t scribendum eat. Sic militia, malitia, neqiiitia et caetera similia". 8) Cf. Thurot, Notices et Extraits des MSS etc.. Vol. XXII, part 2 (1869) p. 78, who gives the following excerpt from the 10*^ cent. MS Paris 7505. "Nunquam enim T ante duas vocales, I post ipsam, priore non Sitzgsb. d. philos.-pliilol. u. d. hist. Kl. Jahrg. 1910, 12. Abb. 2 r ^^ 12. Abhandlung: E. A. Loew ''Justitia cum scribitur, tertia syllaba sic sonat, quasi constet ex tribus litteris /, z et /, cum habeat duos, t et i. Sed notandum quia in his syllabis iste sonus littorae z inmixtus inveniri tantum potest, quae constant ex t et i et eas sequitur vocalis quaelibet, ut tatius et otia justitia et talia. Excipiuntur quaedam nomina propria, quae pere- grina sunt. Sed ab his syllabis excluditur sonus z litterae, quas sequitur littera L ut otii iustitii, item non sonat z\ cum syllabam ti antecedit littera s, ut istius castius.'"^) The statement of Papirius describes exactly the method of distinguishing the two sounds of ti which was followed by mediaeval scribes and notaries as far as that method can be derived from graphic distinctions. There is only this difference: in the case of ti followed by i no exception was made. The rule was simply this: ti before any vowel has the assibilated sound. When preceded by the letter 5, ti has the' unassibilated sound.*) tamen s precedente venire potest ut species, glades . . . ocium spacium . . . tercius nisi sint primitiva a quibus T retineat, ut scientia a sciente sapientia a sapiente etc. On same page '7 ergo « precedente sonum non immutat, ut molestia, modestia, itstio, quaestio etc.". Cf. also p. 144-5. ^) See preceding page, note 6. 2) In his Praefatio (p. IX) to the Leyden reproduction of the Medi- cean Tacitus (Flor. Laur. 68, 2) Prof. Rostagno tried to formulate the rule governing the use of the two kinds of ti, but he was not successful be- cause he failed to realize that it was a case of graphically representing a phonetic distinction as appears from his words : "subeunte enim vocali, ti litterae uno ductu (i. e. our ti ligature which in Beneventan is reserved for the assibilated sound) per compendium scriptae exstant, exceptis qui- dem, ut par est, comparativis adjectivorum in — estus - ustus desinen- tium, ut iustjor f. IK A. XII, 40, 7 etc. Cf. questjorc f. 9v, XII, 26, 1, et ita passim". The reason why the Beneventan scribe used the ordinary ti in the above examples is explained in the citations from Papirius. The scribe also wrote istius and hostium with the ordinary ti for the same reason that he thus wrote iustior and quaestiore, i. e. for phonetic reasons, since ti followed by a vowel is unassibilated when an s precedes. The statement in Muiioz y Rivero's Paleografia Visigoda, p. 105, is inexact and suggests that he also missed the essential point in the matter. Studia palaeographica. 19 As will be seen from MS evidence adduced below many centuries had to pass before the phonetic distinction between the two sounds of ti was graphically reproduced.^) The Ligature ti. Its Forms. In rapid writing the letter t particularly lends itself to combination with the following letter. The cross-beam of t, by being drawn down, readily forms part or even the whole of the next letter. The ligatures te, tu, tr and ta amply illustrate this tendency, but whereas they furnish examples of partial coincidence, we have in the ligature % complete coincidence, since the continuation of the cross-beam constitutes the letter i. Cursive t standing by itself would look thus: cC. By drawing down the horizontal stroke without removing the pen we ^q\ 8 Thus arose a form which plays an interesting part in Latin palaeography. There are several w^ays of forming the ligature §. It may be made in t^vo strokes, or without removing the pen. The latter way is more usual in cursive, the former in MSS. An analysis of the ligature shows that the upper arc or semi- circle corresponds to the cross-beam of the t, and that the point where the curves meet corresponds to the point where the vertical and horizontal strokes of the t meet. In some cases the scribe or notary begins with this point of juncture. First the lower half-curve is made, then the pen is placed at the initial point and the upper loop with its tail or continu- ation is formed. In either cases the pen starts at the top and forms larst the two half-loops, like broken c, then the pen is placed at the same point and the vertical line representing the cross-beam of t and the letter i is traced. If made without removing the pen, the ligature began at the point where the two curves join, but after forming the lower curve the pen ') The spelling ci for ti is much older than the conscious attempt to represent the two sounds of ti by two distinct forms. But ci for soft ti, instructive as it is phonetically, is after all misspelling. 20 12. Abhandlung: E. A. Loew was not lifted up, but returned to the starting-point in a straight line, then continued as in the case above, thus producing a form resembling oj- Another form of the ligature ti which deserves mention occurs in the earliest cursive extant, especially in the Ravenna documents and later in Insular MSS. It differs from the forms already described in lacking the upper half- curve. It resembles somewhat the letter q with the vertical stroke extending above the loop, thus: CI. Origin. The ligature of t and i is so obviously of cursive origin that no demonstration of the fact is necessary.^) It is sufficient to remember that the ligature is found in documents as early as the 5*^ century when no MS used it, and that the first MSS which show the ligature are practically written in cursive. As in the case of i-longa, here too a brief survey of the manner in which the notaries of the different centres used the ligature may be found instructive, for the light thrown upon the relation between cursive and calligraphic writing. Usage in Cursive. A form of the /i-ligature is already found in the well-known letter on papyrus (Pap. lat. Argent. 1) of Strassburg.2) It is used regardless of the sound: scholas- ^icos, sugges^ione. It is used indifferently in a document of 489 reproduced by Marini (Papiri Diplomatici, pi. 6, no. 82), The celebrated documents of Ravenna of the 6^^ and 7*^ cen- turies make very frequent use of the ligature regardless of the ti-distinction: designaif^s, mancipa^ioni, testis, pre^o etc.^) In the peculiar uncial-cursive of the 6^'' century which is found in many semi-uncial MSS as marginalia, the ligature is found: uigin^i in Paris 12097;*) ui^iatis, u^ilitas in Lyon 523.^) ^) Not all ligatures are necessarely cursive. Combinations of o and s, u and .«?, n and t are peculiarities of uncial writing, just as the combination of i and t at the end of a line is typical of Spanish min- uscule, but hardly of its cursive. 2) For facs. see p. 2, note 2. ') Cf. p. 3, note 1. ♦) Facs. Delisle, Le Cabinet des MSS, pi. Ill, 3. 5) Fac8. Delisle, Alb. Pal., pl. 7. Studia Palaeographica. 21 The ligature § is a constant feature in the documents of the Lombard regime. It is used indifferently: ^bi, uindi^onis, por^onem, ex^nia^onem, Ius^§a etc.^) I found § used indifferently in several 8*^^ century central Italian documents preserved in the Archives of Lucca.^) In the Merovingian documents, how^ever, % is rarely used.^) I noted it in a document of 688: quolibe§psa' = quolibet ipsa.*) The spelling ci for assibilated ti is the rule rather than the exception in these documents. In some diplomas of Charlemagne § still occurs e. g. comi^bus, institu^s (a. 775); auctorita^s (a. 775); pala§o(a. 775); pra^s, tradi§onis(a. 782).^) It is only rarely to be seen in later diplomas. I noted trini- ta^s in one of the year 902. The ligature % is found in St. Gall documents of 752, 757, 772 and 797, used indifferently: agents, pra^s, dona^ionem etc.^) The reign of Charlemagne may be said to mark a turning- point in the history and function of the ligature ^. The in- fluence of the Caroline reform in writing drives out the lig- ature. This is more noticeable in France than in Italy. The notaries of Italy however begin about the year 800 to reserve the ligature for the assibilated ti — a practice which lasts for centuries. Thus in Tuscan documents ^ is still found in the 11*^ century;''^) in southern Italy some notaries use it in the 13**^ and even in the 14*^ century, always for assibilated ti.^) The ^) Bonelli, op. cit. passim see p. 3, note 3. 2) Examples are the documents *L 75, a. 713— 4, *N 100, a. 773, *B 65, a. 773, * G 46, a. 807. ^) Facs. Lauer-Samaran, op. cit., p. 4, note 1. 4) Facs. Arndt-Tangl, Heft P, pl. 10. ^) Facs. V. Sybel and Sickel, Kaiserurkunden in Abbildungen. The five diplomas cited are reproduced respectively in Lief. I, 2; Lief. Ill, 3; Lief. I, 3; Lief. I, 4 and Lief. I, 13. 6) Facs. Arndt-Tangl, Heft III*, pl. 71 and Steffens, Lat. Pal.* pl. 38. ■^) Facs. Collezione Fiorentina, pl. 36 of a document of 1013. One of the earliest instances of the ligature for soft ti is in a Pisan doc- ument of 780, facs. Collez. Fior., pl. 29. ^) Cf. works cited p. 3, note 4. 22 12. A})handluno:: E. A. Loew o same js true of the peculiar script of the papal chancery. We find the ligature in the oldest extant documents as well as in papal bulls of the 11*^ century — always for the soft sound of ti.^) As soon as the characteristic script is supplanted by the papal minuscule the ligature disappears and somewhat later the ^-distinction.^) The same is true of the cursive written by the notaries of the city of Rome and vicinity.^) In a doc- ument of 1083 the ^i-ligature still has its traditional use;*) in documents of tte early 12**' century we begin to miss both the liorature and its distinctive function.*) It is important to note however that during the 11*^ century we find in documents of northern Italy and Ravenna a ligature of ci which is strikingly like the ligature of ti. That the ligature represents ci and not ti is established beyond a doubt by the cir- cumstance that when the same word is used in the same doc- ument by a hand writing ordinary minuscule or when it is repeated by means of tachygraphic signs, ci is used and not ti.^) 1) See facs. in Pflugk-Harttung, op. cit., p. 3, note 5. A papal bull of 1098 still has the ligature. Cf. ibid., pi. 47. 2) For I noted that the ti-distinction is carefully observed in two documents of 1127 and 1138 written in ordinary or papal minuscule. Facs. Steffens, Lat. Pal.^ pi. 80 and 81*. 5) Facs. Hartmann, op. cit., p. 3, note 6 and Fedele in Arch. Pal. Ital., Vol. VI (1909) fasc. 30 and fasc. 34 (1910). *) Hartmann, op. cit., pi. 26. ^) Hartmann, op. cit., pi. 27, a. 1107 and pi. 28, a. 1110- 6) Professor L. Schiaparelli who has kindly called my attention to this fact, furnished me with these examples: a document of Pa via of Dec. 1029, now in the Archives of Nonantola, has de^ma (I do not at-' tempt to give the exact forms of the ligature) ti^nenso, fa^as, sancti quiri^, and the tachygraphic signs give queric/. In a document of Pia- cenza of Dec. 31, 1007 we have Domini^ which must be expanded by ci. Cf. Schiaparelli , Tachigrafia, Sillabica (Rome 1910) p. 38. Other doc- uments have pec/a, tert^a in tachygraphic signs, and in the text pe^a, ter^a. Signor Pozzi who is working upon the later Ravenna documents has given me numerous instances of the ligature for ci and not ti in Ravenna documents. To him and Professor Schiaparelli 1 here express my warm thanks. Studia palaeographica. 23 The Beneventan notary practices the ti-distinction even as early as the end of the 8**' century,^) though the indifferent use of the ligature occurs during the 9*^ century. Later the notary shows the same care in distinguishing the two sounds of ti as the scribe. The practice lasts as long as the peculiar script remains in use.^) Spanish notaries, as far as I can judge from an examin- ation of facsimiles, observe the fi-distinction. It should be noted that at first (during the 8^** and 9^^ centuries) ^ serves for assibilated ti^ and later, that is during the 10*^ and 11*^ cent- uries, qc) performs that function precisely as in Visigothic MSS. The more recent Visigothic documents show a marked ten- dency toward employing ci for soft ti.^) So much then to give an idea of the wide use of ^ in documents and of its specific function in many of them since the time of Charlemagne. Usage in MSS. We are now ready to examine its use and function in MSS. This examination will help to bring out the closeness of relationship which existed between cursive and calligraphic writing. From the evidence given below the history of this ligature and of the ^i-distinction in Latin MSS may be sum- marized as follows. In the oldest MSS in uncial and semi-uncial we find neither ^ nor the ^^-distinction. In the earliest French min- uscule MSS of the 7"' and 8*^ centuries % is used indifferently. It is still found in some MSS of the Corbie lcd type, but the great majority of them do not employ it. In a number of MSS of the early Caroline epoch, MSS which still use the open a 1) Cf. Cod. Diplom. Cavensis I, pi. 1. 2) For other facsimiles see works cited p. 3, note 4. 3) Cf. Merino and Munoz cited p. 4, note 4. See also below, part IV, where Spanish usage is discussed. 24 12. Abhandlung: E. A. Loew Studia palaeographica. 25 and the n-ligature, the form § is still to be found, but always used indifferently. With the spread of the Caroline minuscule its use gradually dies out. It is scarcely found in MSS written after the beginning of the 9*'' century. Its presence in a French MS is a fair hint of its date. As for the ^i-distinction in French MSS, the practice ap- parently never took root. It is only in a few MSS of the 8*^^ century, and only in portions of these, that the attempt to observe the distinction is noticeable.^) Curiously enough, ^ stood for the hard sound and ordinary ti for the soft sound of ti. Of no small importance, on the other hand, is the fact — which doubtless stands in some causal relation with the absence of the ti-distinction — that ci often stood for soft ti. The ligature § is manifestly at home in Italy. We find it already in the earliest examples of Italian minuscule where (as in contemporaneous documents) it is used indifferently for both the soft and the hard sound. At about the end of the 8*^ century both in north and south Italy attempts are made to observe the ti-distinction, reserving § for the assibilated sound. The ligature % disappears from the north Italian scrip- toria during the first decades of the 9*^ century, owing to the influence of the Caroline reform. In south Italy, on the other hand, where the Caroline reform did not penetrate, § remained. Its one function was to represent assibilated ti. In Spanish calligraphy § is in reality but a makeshift, occuring chiefly at the end of a line because space was wanting for the normal ti. To make the distinction between the two sounds of ti other means were used (see below, Part III). As in Beneventan, here too ci is rare. It becomes frequent as soon as the Visigothic gives way to the ordinary minuscule in which the two sounds of ti are not differentiated. The absence of such spelling as nacio, leccio in Beneventan and Visigothic MSS is directly and causally related to the 'i !» : \\ presence of distinct forms for differentiating the assibilated and unassibilated ti.^) Of this there can be no reasonable doubt. Insular MSS do not make the ^^-distinction. The form of the ligature used in them is probably of semi-uncial origin, and is found in MSS posterior even to the 9^*^ century. The transcription of the ligature. In view of what has been said of the ligature the question of how it should be transcribed may seem gratuitous. Yet this is not the case. For scholars are not at one on the subject. There are those who transcribe the ligature by means of d.^) That this is incorrect is proven not alone by the origin of the ligature which is simply a combination of t and i but by the fact that for generations scribes and notaries used the ligature in words like satis, tihi^ peccati as well as in words like natio or uenditio etc. There are, to be sure, cases where notaries used a ligature like this for ci,^) but in MSS this is hardly possible. That in Beneventan the ligature may never be transliterated by ci is proven by the fact that words like provincia, specie, Decii, socio, atrocius etc. are Avritten with ci and practically never with the ligature. We see then that the Beneventan scribe made a careful distinction between ci ») Cf. MSS: Paris 121G8; Laoii 423; Laoii 137; Paris 8921. ^) This observation was already made by Mommsen in his de- scription of the Beneventan MS. Vatic, lat. 3342. See the preface to his edition of Solinus, p. CIV, where he quotes Traube, Roma nobilis, p. 13, note 7. See also Bluhme in Pertz' Archiv V, 259. 2) Cf. Federici's description of Rom. Casanat. 641 ^ in Archiv. Paleot^r. Ital. Ill, fasc. 22, also op. cit., Vol. III. Notizie dei facsimili, p. XIII, published in 1910. I find the ligature transcribed by ci in the word Translatio occuring on fol. 31 of the Beneventan MS in the library of H. Y. Thompson. See A descriptive catalogue of fift}'- MSS in the collection of Henry Tales Thompson (1898) p. 87 sqq. 3) See p. 22, note 6. 26 12. Abhandlung: E. A. Loew Stiidia palaeographica. 27 and soft ti. And the fact that he (as well as the Visigothic scribe) possessed a special way of writing assibilated ti doubt- less accounts for his rarely writing ci for ti, so that such spelling as nacio, leccio, pocius, which fill the pages of early French MSS, are practically a rarity in Beneventan or Visigothic.^) The transcription of the ligature ^ in documents was some years ago the subject of lively dispute.^) Without entering the discussion I may state that I hold with Lupi against Paoli that the ligature § should be rendered by ti regardless of what its probable pronunciation may have been. When such extraordinary forms are encountered as ac^ione, with the superfluous i, or a^Qo in which the ligature has plainly the value of z and not of soft fi, the editor ought to call attention to that fact. ^) The instance just mentioned of a§§o for az^o brings up an interesting question. Is it not possible that in such a case we have perhaps a reminiscence of a form of ^ which vanished in time, but the use of which in documents *) There is a form of t in Visigothic which strongly resembles c, one must therefore be skeptical of transcriptions with ci for soft ti, if the MS is Visigothic. 2) Cf. C. Paoli, Miscellanea di paleografia e diploniatica. TI, ZI, Z in Archivio Storico Italiano, Serie IV, Vol. 16 (1885) p. 284 sqq.; C. Lupi, Come si debba trascrivere il nesso TI, in Archiv. Stor. Ital., Ser. IV, Vol. 20 (1887) p. 279 sqq.; ibid. Paoli's reply. Paoli transcribes the ligature reg- ularly with zi when it is assibilated. Cf. Collez. Fiorent., plates 21 and 29. Other Italian diplomatists transcribe the ligature by ii. Cf. Fedele, Archivio della R. Societa Romana di Storia patria XXI (1898) p. 464 and Schiaparelli, Bulletin© dell' Istituto storico Italiano. No. 30 (1909) p. 53. 3) The question deserves further investigation. I learn through the courtesy of Dr. F. Schneider that this strange phenomenon is to be noted in a Tuscan document of 1043. Cf. Quellen und Forschungen XI (1908) p. 33. Curiously enough, I have found two instances of superfluous i after the ti ligature on a single page (uitiium, quotiiens) in the Ben- eventan MS Paris 7530 (Monte Cassino), saec. VIII ex. This page, fol. 222, is being reproduced in Part 1 of the Scriptiira Bencventana. 1 1 , of the 8*^ century is fully attested? This form of ^, by reason of its resemblance to the usual form of the ligature ti has presented considerable difficulty to editors who usually tran- scribe it by ti. The two forms are made precisely alike only that the ^ has an affix, as in capital Q, which consists of a wavy line made from left to right, thus: a . Examples of its use are to be seen in Bonelli, Codice Paleografico Lombardo. As this feature is scarcely known I give here some instances, and point out where Bonelli reads erroneously. doc. a. 748 Bonelli, pi. 6, line 5 pezola; line 8 pezola (Bonelli petiola), „ „ 9, line 9 peza, „ » 12, line 1 zenoni (Bonelli tzenoni); line 2 pezola (Bonelli petzola), „ „ 16, line 15 florenzione (Bonelli Flor- doc. a. 765 doc. a. 769 doc. a. 774 „ entione). Schiaparelli (in Bullet, dell. Istit. Stor. Ital. 1910, No. 30) noted this curious letter in two documents, and even called attention to the difference between it and ordinary §, but he did not feel justified in transcribing it differently. doc. a. 742, pi. 1, line 3 peza (Sch. petia), doc. a. 758, pi. 2, line 15 pezola (Sch. petiola). A fortunate find has furnished me the evidence which establishes to a certainty that this form is to be regarded as the letter ^ and not as the ligature ti with a meaningless appendage. In the important MS Vercelli 183, saec. viii (it has ni == nostri, no = nostro, nm = nostrum etc.) this form of ^ occurs many times. ^) It differs from the ligature, which also occurs continually in the MS only in the matter of the affix. Ex- amples are: f. 99^ ^elo: f. 104^ e^ecMel, acha^ etc. ; f. 91^ ^osinio. 1) Cf. Plate 1, line 11. 28 12. Abhandlunt? : E. A. Loew The regular use of this form of the letter ^ in a perfectly calli- graphic book furnishes one of the clearest illustrations of the dependence of early minuscule upon cursive. The scribe of Yercelli 183 Avas evidently bold in employing this letter. For it appears that the form never got naturalized in calligraphy. On careful enquiry I find that Vercelli 183 is practically unique in its use of this ^. Through the kindness of Professor Lindsay I learn that in a fairly similar form it also occurs in the north Italian 8*** century MS Milan Ambros. C. 98 inf. This form of the letter is not mentioned in our texts on palaeo- graphy. -ft t Studia palaeographica. 29 The Evidence. a) ti in Latin MSS. b) i-longa in Latin MSS. \ •>^ 1. To illustrate the usage of ti and i-longa I give only one or two typical examples which I noted on examining the MS. In some cases I have had to depend on photographs. To dis- tinguish such evidence from that based upon a study of the whole MS, I prefix an asterisk (*) to MSS actually examined. 2. The form of % used in the examples is the most common. No attempt could be made to reproduce the different varieties found in the MSS. 3. By § used indifferently I mean that the ligature is not reserved exclusively either for assibilated or for unassibilated ti. 4. The date ascribed to a MS is an approximate one. To avoid ambiguity it may be stated that saec. vm in. = P*- third of 8^^ century ; saec. Ylil ex. = last third of the century; saec. vni post med. = 2"^ half of the century; saec. vni/ix = ca. 800. 5. The MSS are arranged as far as possible according to countries, in groups which present common graphic features. It is hoped that this attempt at classifying MSS in early Latin minuscule will prove helpful. Inexpensive facsimiles of these MSS will be made accessible to the student in an extensive collection now in press. ) V) ii 30 12. Abhandlung: E. A. Loew Uncial MSS. a) In the oldest type the ligature % is not found. But in the more recent type it slips in occasionally at the end of a line for lack of space, e. g. *Lucca 490 saec. viii/ix in the uncial part: paren§bus. b) The i-longa is lacking in the oldest type of uncial. However, in MSS of the vii^'" and viu*'^ centuries it is not in- frequently used, thus showing the influence of notarial upon calligraphic writing, e. g. Paris 1732: In, lelunio; *Vatic. lat. 317 : lelunii passim, i-longa initially, passim by one scribe; *Vercelli 188 initially passim; Paris 13246: In, lelunauit, hulus etc. * Vatic, lat. 5007 (Naples): In, hulus etc. Semi-uncial 3fSS. a) In the oldest kind § does not occur. In the recent type it is occasionally found at the end of a line, e. g. *Novara 84 saec. vnr. b) i-longa is not used in the oldest kind. In the more recent type it occurs, e. g. Cambrai 470 initially often ; *Rome Sessor. 55 (2099): In. loseph, malore; Ambros. S 45 sup. often initially; Lyon 523, initially passim; *Vatic. Regin. lat. 1024 (Spanish) often initially; Autun 27 (Spanish) often initially: In, ludaei, Ipse, Imago, also medially: elus. In St. Gall 722 it occurs initially, but also finally after t: repletl. In Autun 24 it is also used in other parts beside the beginning: Itlnerls etc., in this respect recalling Merovingian cursive. Early French Minuscule. Paris 8913 saec. vii. The script is very cursive. a) g is rarely used: con^geret, coUegis^s. The ordinary forms of t and i are used for both the soft and hard sounds. But ci occurs for assibilated ti\ hospicio, sullercia. b) Initially often: In, Introeat, luxta; but ilia, ibi with short i. I Studia palaeographica. 31 *Paris 17655 saec. vii ex. The writing hardly ditfers from that of Merovingian diplomas. a) ^ used indijfferently: mon^um, al^tudinem. I noted ci for assibilated ti in the uncial portion: commemo- racione (f. 2). b) Initially and medially: In, cuius, elus; occasionally short: iniurias. *Paris 9427 Luxeuil type, saec. vii/viii. Lectionarium Gallicanum. a) ^ used before a consonant: sa^s, sta^m. Assibilated ti is often represented by ci: pacientiam, adnunciavi, sici- antem, leccio etc. b) Initially and medially: In, Ita, Ille, oblecit etc. *Verona XL (38).^) Same type. saec. vii/viii. a) § occurs for assibilated and unassibilated ti^ but the ordinary ti is more usual: senten^iam and sententiam; seme^psam and semet ipsam, tokens and fa^gat. b) Initially and medially: In, lob, Ipse, Iste, alt, elus, lustum, Indicium etc. but illius with short i. St. Paul in Carinthia MS XXV^. Same type. saec. vii/viii. a) 5 used indifferently: sapien^a, noc^bus; scien^a, repen§na. b) Occasionally long initially: In, but ipse, ilium, eius with short i. M Verona XL is in precisely the same script as Paris 9427. By means of internal evidence the French origin of the Paris MS is estab- lished beyond a doubt. Graphic features point to France also as the home of the script, since it resembles French cursive much more than Italian. Then too, the style of ornamentation and the orthography — the use of ci for assibilated ti — strongly favor France. These con- siderations seem so grave that I feel justified in differing with Traube according to whom the Veronese MS was written in Verona. See Vor- lesungen und Abhandlungen II, 28. There seems to be a slight incon- sistency in this passage for the same MS is spoken of as a "Kursivschrift eigener Art" and then again as an example of "Scriptura Luxoviensis". 32 12. Abhandlunf? : E. A. Loew Studia palaeographica. 33 *Ivrea 1. Same type. saec. vii/viii. a) § used for assibilated and unassibilated ti'. inimi- ci^as and occults, silen§o and u^lis. The ordinary ti is also used for soft ti: etiam. The ligature § occurs for a: quanto^us, ami ^§ as. b) Initially and medially: In, Iterum, Illius, Idolatriam, Ipse, Illos; alt, hulus, conlugum etc., yet cuius with short i. *London Add. MS 11878. Same type. saec. viii in. a) % used indifferently: tempta^onis, u^, sen^t. b) Initially: In; medially not always: elus but cukis. *London Add. MS 29972.^) Same type. saec. viii in. a) 9 used indifferently: quo ^ ens, men^mur, ^bi. The ordinary form of ti is also used for assibilated ti\ etiam. b) Initially the rule; medially occasionally: In, cuius etc. but also cuius. Fulda Bonifatianus 2. A similar type of writing but some- what more recent than that of the preceding MSS. a) § used indifferently: ra^o and ni^tur, despera^onis and praesen^s. Frequently ci is used for soft ti: uicia. A corrector changed it to uitia. b) Often long in the word m, but not always. Wolfenbiittel Weissenb. 99. Similar type. saec. vin in. a) § used indifferently: ressurec^onem, u^que; laeti^am, lus^. ^ occurs for ci e. g. suspi^onem. b) Initially: In, Ihm, lam, lusti even lUe, yet ipsius w^ith short i. *Munich 29033 (fragment). Similar type. saec. viii. (Formerly served as fly-leaves of Munich 14102). a) § used indifferently: tempta^o, mit^t, confes^im, bap^sta; ci occurs for assibilated ti: spacium. Also ^ used for d: deli^osa. b) Often long initially: Iter, Ingressus, lam, lussit; but ille, ipse, iustus with short i. ^) Similar writing maybe seen in Vatic. Regin. lat. 317, e.g. the additions on ff. 31^', 180, 160^' etc. \ *Admoiit (Abbey) Fragm. Prophet.^) Similar type, saec. VIII. a) § used indifferently: adflic^onis, sabba^ , por^s, uic^mam etc.; ci occurs for soft ti: poenitenciam, con- tricione, oblacionem (corrected to oblationem). b^ Initially otten; occasionally also medially: In, Ipsa, luxta, malestate ; but ibi, illut, ipse, maiestas with short i. Wtirzburg Mp. Theol. Fol. 64 ^ Similar type. saec. viii. a) % used indifferently: gen^um, tribula§one, gen^bus, ul^mum; ci occurs for soft ti: cognicio, tribulacione, per- secucionem, adnunciate etc. b) Initially occasionally long, more often short: In, but also in, iudicium, huius with short i. *Vienna 847 ff". 1^, 5^ G^. saec. viii. a) § occurs for the hard sound: peccan§; ci is often used for assibilated ti: accio, legacio. b) Initially and medially: In, lusticiam etc. *Paris 12168. « type. ca. a. 750. The angularity of the two parts of a is characteristic of this group. a) One scribe regularly used § for unassibilated sound: res^§t, procrea^s and ordinary ti for assibilated: otium, potius.^) But ci often occurs for soft ti. Another scribe (after f. 68) uses § indifferently. It is evident that the first scribe was trying to make a strict distinction between assibilated and unassibilated ti. Curiously enough, the form he chose for hard ti became in other schools the regular form for soft ti. b) Commonly in the word m, otherwise often short: ita, iudas. ^) The fragments show two contemporaneous hands. The usage cited is true only of one scribe, the other does not employ the ^^-ligature nor the same form of a. His writing makes a more recent impression and most likely represents the more modern style. The same scribe, 1 believe, wrote the biblical fragments now in Munich (MS 29158). 2j My attention to this regularity on the part of the first scribe was called by Prof. W. M. Lindsay. Sitzgsb. d. pbilos.-pUilol. u. d. hist. Kl. Jahrg. 1910, 12. Abh. 8 34 12. Abhimdlunj?: E. A. Loew Studia palaeographica. 35 *London Add. MS. 31031. (Same type.)*) ca. a. 750. a) § often for uiiassibilated ti: ads^^s, §bi, peccas^; ci very often for assibilated ti: inius^cie, explanacio etc. b) Usually short. This cursive element is slowly being eliminated from the book hand. Laon 423. (Same type.) ca. a. 750. a) The first scribe (jff. 1 — 17) has ^ for unassibilated ti and ordinary ti for assibilated: supers^tiose, inues^ga- tione etc. The other scribes use § indiflPerently. Here it may be fair to suppose that the first scribe was con- sciously making a distinction between the two sounds of ti.^) Laon 137. (Same type.) ca. a. 750. a) % is used indifferently, although it seems that here and there an eftbrt was made to have it represent only the hard sound, e. g. pes§lentia, res^tutione. *St. Gall 214. (The /-type.) saec. vm. The characteristic letter is ?, which has a distinct bend in the middle, somewhat like broken c. The script is related to the Corbie ab type. See p. 36. a) § not used. Ordinary ti is used for assibilated and unassibilated ti, but ci often occurs for the soft sound: cicius, perdicione. b) Initially often, but in, impleri, ignorat; occasionally also medially: cuius, elus. *London Harley 5041. (Same type.) saec. viii. a) % not used. Ordinary ti for assibilated and unassibil- ated sound. b) Used occasionally: lam, malor. Often short, even in the word in. Chateau de Trousseures. Same type. saec. viii. Nov. Testam. See catalogue of sale, pi. 2 (Paris, Leclerc, 1909). ^) To judge from a small facsimile, the Cambridge MS Corpus Christi College K 8 belongs in this class of MSS. 2) Knowledge of this and the next MS 1 owe to the kindness of Prof. W. M. Lindsay. a) § occurs for hard sound: §bi; ci is used for assibil- ated ii: narracio, depraecacio. b) Initial i has a somewhat longer form: In. *Paris 14086. Similar script, saec. viii. a) ^ occasionally for assibilated sound: praesump^onis; but ci is very frequent for soft ti: senciant, paenitenciam etc. b) Initially. *Berne 611. Similar script, saec. viii. a) § is used indifferently: legen^um, praeposi^onum, pon^fex; ci very often occurs for soft ti: noticiam, moni- cione, quociens. Ordinary ti is also used for the soft sound. b) Initially as a rule; medially occassionally : In, hulus, cuius; but also eius with short i. Here and there the i-longa extends below the line: ejus, jejunij. *Bamberg B V 13. Similar script. saec. viii/ix. a) No §. No distinction between the two sounds. b) No i-longa. *Paris 12598. saec. viii ex. a) § used for unassibilated ti, ci often occuring for assibilated: ^bi, pe^cionibus, adfleccione. b) Found here and there initially and even medially: elus, leluniis; but as a rule i-longa is not used. *Vienna 1616. saec. viii ex. a) § used for unassibilated ti: u^, bap^zatus, casti- ta§s; ci often occurs for assibilated ti: tristicia, poncio, gencium, damnacionis etc. b) Initially, but ilia with short i; medially as a rule: malestas, hulus, lelunii, lelunare etc. Epinal 68. saec. viii (a. 744). A type of pre-Caroline minus- cule out of which the Caroline developed. The cursive elements are few; the general impression is that of a modified semi-uncial. ■ 36 12. Abhandlung: E. A. Loew a) g seldom occurs: imperii ssimis; no distinction is made, but ci is often used for soft ti: laeticia, uiciis, uiciata etc. b) Initially and medially by one hand: In, lam, elus- demetc; short i initially and medially, by another. The cursive portion has i-longa. The use of i-longa in the body of the word, at the end of a syllable e. g. lacrl- marum recalls certain semi-uncial MSS and Merovingian cursive. There are a number of MSS of the type of Epinal 68. ♦Oxford Bodl. Douce f. 1 (fragments), saec. viii post med. This script is the immediate precursor of the Kb type which is manifestly only a further development of it. Very typical is the letter a which in combination is often suprascript and has the first curve turned leftward at the top. Otherwise the a is shaped like two adjacent c's. The b has already the form found in the Corbie MSS of the ab type. a) % used indifferently: poten^am, securita^s. Ordin- ary ti is often used for the assibilated sound. % occurs for a, e. g. fa^at. b) not used: in, huius, maiestatem — all with short L *Vatic. Regin. lat. 316. Same script, saec. viii post med. The MS is in uncial, but several lines occur in this type of minuscule on if. 2^ and 46. a) g used: substangalem, tempta^one; ci occurs for soft ti: tercia. b) A slightly longer form of i occurs initially: In. Brussels 9850-52. Corbie-script,^) icb type. saec. viii ex. Most of the MSS of this type are of the early ninth century, a few are of the end of the 8*^. Studia palaeographica. 37 1) The name originated with Traube. The script is very conventional and shows a high point of development. a) ^ used indifferently: pa^en^ssima. b) Initially often, but not medially. *Paris 3836. (Same type.) saec. viii ex. a) ^ used indifferently: sen^endum, proba^s; ci often occurs for assibilated ti: racione, penetenciam etc. b) Not used regularly. *Paris 8921. (Same type.) saec. viii ex. a) ^ is not used. However it is evident that the dis- tinction between the two sounds is striven after. When the ti is assibilated the i is extended below the line (as later in Visigothic MSS); when it is unassibilated the usual form of the i is retained. This distinction is ob- servable in many parts of the MS. I cite these examples: f. 31^ an^iocensis but cottinensis; f. 32"^ e^iani but ex- titerit; f. 45 deuo^ionis, persecu^^onis but multis. (Yet I noted nescieu^^bus); f. 138^ Lauren^ms but surentinus; pro- iectifius but hostiensis; f. 140^ etiam but sanctitas. ci is not infrequently used for assibilated ti, b) Often initially and medially: huIus, cuius etc. Turin D V 3. Same type. saec. viii ex. a) § occurs for unassibilated ti: omnipoten^s, pro- sequen^s; ci is used for assibilated ti: milicia, pocius, racioni, graciarum etc. b) Initially: In, lohannis; not medially: huius, cuius. *Paris 11627. (Same type.) saec. viii/ix. a) No %. No distinction. b) Often used, but not regularly. *Paris 11 681. (Same type.) saec. viii/ix. a) No §. No distinction. b) Only occasionally. *Paris 12134. (Same type.) saec. viii/ix. a) No. §. No distinction. b) Often initially. 38 12. Abhantlluni,' : E. A. Loew Studia palaeographica. 39 *Paris 12135. (Same type.) saec. viii/ix. a) No §. No distinction. b) Occasionally. *Paris 12155. (Same type.) saec. viii/ix. a) No §. No distinction. b) Used irregularly. *Paris 12217. (Same type.) saec. viii/ix. a) No §. No distinction, c'l occurs for soft ti. b) Hardly used. *Paris 13048. (Same type.) saec. vm/ix. a) No 5. No distinction. b) Often initially, but irregularly. *Paris 13 440. (Same type.) saec. ix in. a) No 9. No distinction. b) Rarely used. *Paris 11529 — 30. (Same type.) saec. ix in. a) No §. No distinction. b) Often used, but not regularly. *Paris 17 451. (Same type.) saec. ix in. a) No §. No distinction. *Paris Nouv. Acq. 1628 ff. 15—16. (Same type.) saec. ix in. a) No §. No distinction. *Bamberg B III 4 fly-leaf. (Same type.) saec. ix in. a) No ^. ci occurs for soft ti. ♦London Harley 3063. (Same type.) saec. ix in. a) No %. No distinction. b) Used initially; not medially. There are doubtless many other French MSS of the pre-Caroline or early Caroline epoch — it would hardly be necessary to enumerate them even if I were able to do so — which employ % indifferently. Gradually, however, this cursive element altogether disappears from the book-script. The i-longa, especially in the word in or otherwise at the beginning of a word stays longer than §. But it too was practically rejected, although it crops up here and there at all times. Early Italian Minuscule, *Milan Ambros. Josephus on papyrus. (North Italy.) saec. VII. a) § used indifferently: repe^^one. No distinction is made between soft and hard ti, b) Regularly initially: In, Ipse, Itaque; even lUud, Die, Ibi; medially regularly for the semi-vocal sound: pelor, hulus, cuius, alt, Inlurias etc. *Milan Ambros. C 105 inf. (Bobbio.) saec. vii/viii. a) 9 used indifferently: prae^o, merits, reper^. No distinction. b) Initially and medially: In, Ipsa, malorem etc. *Naples IV A 8. (Bobbio.) saec. vii/viii. a) % used indifferently: muni^onem, sta^m. Inno- cent us, iacen^bus. No distinction. b) Initially and medially: In, lacentibus, prolecerunt. *Vienna 17. (Bobbio.) saec. vii/viii. See preceding MS of which it formed a part. *Milan Ambros. D 268 inf. (Bobbio.) saec. viii in. a) § used indifferently: e^am, uirtu^s, men^s, con- ten §oni. No distinction. b) Initially and medially: Ihs, Illud, cuius, malestatem, alt. Where the scribe had made it short initially, the corrector made it long. *Milan Ambros. C 98 inf. (Bobbio.) saec. viii. a) ^ used indifferently: digna^one, sapien^bus. No distinction. b) Initially the rule, even Illo, Ipso, Ihs, Ibi etc. Medi- ally not always: Inluria, hulus, malestate, malor; but also huius, eius. 40 12. Abhandlung: E. A. Loew *Vatic. lat. 5763. (Bobbio.) saec. viri. a) g used indifferently: no gg a, con gnent. No distinction. b) Initially: Ignem, Inter, ludea; medially not always: cuius but ejus and eius. Wolfenbtittel Weissenb. 64. (Bobbio.) saec. viii. This MS belonged with the preceding. a) 9 used indifferently: to^us, alterna^o, gra^a, noc^s, 111 g mum, gberis. No distinction. b) Initially the rule: Id, Ipse, Igne; also used medially: cuius. Turin A II 2. (Bobbio.) saec. viii. a) 9 used indifferently: prae^i, ui^o, ci^a, ^bi, gen^s. b) Initially: In, lustis, Ipse, but ille; medially: hulus, leluniis, delnceps, but ejus. Turin G V 26.^) fol. 5\ (Bobbio.) saec. viii. a) ^ used indifferently: essen^a, extan^bus. b) Long in in (no other words occur). ♦Milan Ambros. L 99 sup. (Bobbio.) saec. viii. a) g used indifferently: stul^^a, dis^ncgonem. No distinction. b) Initially and medially: In, Ipsa, even Illos; hulus, sublectis, alunt etc. *Milan Ambros. B 31 sup. (Bobbio?) saec. ix in. a) g is used for assibilated fi. but ordinary H is also thus used: ra^onis, but fluctio, tertia, sapientia. No strict distinction. b) Initially and medially: In, Inluria, cuius. *Verona I fol. 403\ 404 ^ (Verona.) saec. vii. An interesting example of north Italian cursive. Very characteristic is the letter n which somewhat resembles our capital M. 1) A good example of Bobbio cursive may be seen in Milan Ambros. S 45 sup. (Bobbio) p. 44. to which Professor Lindsay has kindly called my attention. Studia palaeographica. 41 a) % occurs: temperan^a. No ti-distinction: nequitia.^) b) Initially, medially (regardless of sound) and even finally: Iniquitus, Ita, lUi; sublecti, erit, nequitia; mel, del, fieri, subiectl.^) *Verona III. (Verona.) saec. vrii in. A curious minuscule derived from half-uncial and the cursive noted in Yerona I fol. 403'', 404^. It has the same form of n. a) 9 not used. b) Initially in the word in. *Verona XXXIII. (Verona.) saec. viii in. An excellent example of half-uncial passing into minuscule, a) ^ not used. b) Not used. * Verona XLII. (Verona ')) saec. viii in. Half-uncial passing into minuscule. a) § rarely used, e. g. at end of lines: lus^fi cationis. b) Initially and medially: In, Ille, elus. * Verona II fol. P. (Verona.) saec. viii. Cursive. Characteristic letters are: Z, j), r, g and the ligature w^. a) 5 used indifferently: na^ones, gen^bus, polluis^s. No distinction. b) Initially: In. *Verona IV fol. 6, 6^. (Verona.) saec. vm. Similar cursive. a) § used indifferently: men^s, ^bi, uinc§, pronun- §ans etc. No distinction. b) Initially often: In, lusto, ludaei, lussit, but illas, ignis. ^) The word otmm is spelled ozium, the z havinor the same form as in the word zelus. Assibilated ti must accordingly have had the pronunciation of z. 2) A similar use of i-longa is to be noted in Milan Ambros. 210 sup. p. 46^ written in a very old type of cursive. The peculiar form of n found in the Veronese MS is also to be seen on this page. The //-ligature is used indifferently: uenera^one, salu^s, men^s. Examples of i-longa are: lam, sublacere, hulus, oratlone, deuotlone, coelestl. 3) The MS has the Veronese ss which resembles ns. 42 12. Abhandlunj?: E. A. I.oew * Verona XXXVII fol. 169\ (Verona.) saec. viii. Similar cursive. a) 9 used indifferently: ter^o, dedica^onem, la^tu- dinem, can§co. No distinction. The ligature occurs for ci: prouin^ae. b) Initially: In, loacliini, ludae. *Verona XXXVIII fol. 118. (Verona), saec. viii in. Transition script. This well-known page furnishes one of the earliest examples of Veronese minuscule with the typical g, r, p and l. a) ^ not used. b) Initially and medially: In, Ignes, Illi, Ita, elus, prolecta. *Verona LXII. (Verona.) saec. viii. Calligraphic minuscule which is manifestly derived from the above mentioned Veronese cursive. It has the characteristic /, p, r, g, the ligatures nt, ae, ss (resembling ns) and the superior a. a) 9 used indifferently: nup^is, leon^o, merits, legi- §mam, con^nen^ae. No distinction. b) Not used: in, coniugium etc. with short i. *Verona LV. (Verona.) saec. viii. a) % used indifferently: mundi^a, ui^a, §morem, per- §naciae. No distinction. b) Initially often, but not regularly: In, Ita, Indicium, but also iustus, iustitiae, ignis, iram, ilia etc. *VeroDa LXI fol. 1. (Verona.) saec. viii. a) ^ not used. b) Initially and medially: In, elus, conlunctio, alt. *Verona CLXIII. (Verona.) saec. vin. a) § occurs occasionally. It is used indifferently : gra^a, rogan^, po^us, adduces. b) Initially and medially: In, luuat, lacit, cuius, lelunas. *Verona XV marginalia. (Verona.) saec. viii. a) § used indifferently: §bi, facien^bus. b) Initially and medially: In, lacobi, malori. Studia ptilcieot^^riiijhicci. 43 *Carlsruhe Reich. LVII. (Verona.^)) saec. viii. a) g used indifferently, more often for soft ti: e^am, egyp^is, ciuita§. b) Used irregularly: In, Inter, but ingressu, imperium, cuius with short i. *Paris 653. (Verona?) saec. viii/ix. See plate 2. a) % used by one hand (fol. 1 — G"") for assibilated ii: graham, ignoran^a, but partis. Distinction made. The new hand on fol. 6^ knows neither § nor the ^i-distinc- tion: e^?am, uoca^i b) Used by the first scribe (who knows §): In, Ipse, Ihra, Ita etc. The second scribe does not use it. *Vercelli CLXXXIII. (Vercelli?) saec. viii. See plate 1.^) a) g used indifferently: ui^a, u§, mul^. No distinction. b) Initially always: In, Ipso, Illi, Ibi etc.; medially regul- arly for the semi-vocal sound: elus, hulus, cuius; also when in occurs in the body of a composite word, e. g. delude. See discussion on p. 12. *Vercelli CCII. (Vercelli?) saec. ix in. a) 9 used indifferently: ra^one, mulQ. No distinction. b) Usually in the word m, otherwise not employed: In but ius, ita, cuius etc. 1) The MS has the curious .s-.>? resembhng )is — a feature to be noted in several Veronese MSS. 2) Knowledge of this palaeographically most interesting MS I owe to the kindness of Father Ehrle, Prefect of the Vatican library. Through the great courtesy of Mgr. M. Vatasso T have the privilege of repro- ducing the MS. Several full-page facsimiles of this MS as well as of others from the chapter library of Vercelli will be given by Mgr. Vatasso in a forthcoming work. We have no positive evidence that this and the following two MSS were actually written in VerceUi. Since they are manifestly of north Italian origin, the probability is that they were. I mention in passing that the marginalia of VerceUi CLVIII are in a hand which is not Italian. I take it for Visigothic. The rules for i-longa are, as may be expected, carefully observed.' 44 12, Abhandlunff: E. A. Loew *Vercelli CXLVIII. (Vercelli?) saec. ix. a) § regularly reserved for assibilated ti] and ordinary ti for unassibilated. Distinction made. b) Initially often: In, lam, but ilium, ihm; medially not used. *Novara 84. (North Italy). saec. viii/ix. a) 9 used indifferently: peniten^a, na^uita^s. No distinction. b) Usual with m, otherwise rarely used: In, but iam, ita, huius. Milan Trivulziana 688. (Novara.) saec. viii/ix. a) % used indifferently: li^gia; ordinary ti for soft sound: cautioni; ci for soft ti: admonicionem. b) Initially frequent thoucjh not always: In, ludiciis, luret, but index. *Paris Baluze 270. (North Italy.) saec. viii/ix. a) § used indifferently: ra^o, mul^s. b) Rarely used: In but also in with short i. Breslau Rhedig. R 169 f. 92^. (Aquileia?) saec. viii ex. a) § used before consonants: ^berii. ci is used for assibilated ti: tercie, nupcie. b) Initially the rule: Ilium, circumlbat, Ihs etc. Modena I N 11. saec. viii/ix. a) % used indifferently. No distinction. b) Initially: In, ludaica. *Lucca 490. saec. viii/ix. a) % used indifferently: lus^^am, rait^tur, §meas etc. No distinction. b) Not used. *Roine Sessor. 55 (2099) ff. 89 to end. saec. viii ex. a) § used indifferently: enun^are, is§s, dis^inc^one. No distinction. b) Not used as a rule: in, indicaret, coniungas. Studia palaeographica. 45 *Rome Sessor. 94 (1524) part I = pp. 1 — 32. saec. viii/ix. a) 5 used indifferently, but preferably for soft ti: ui^a, faculta^bus. No strict distinction: pretiosus, fortia. b) Initially and medially: lam (corrector changed to iam), leluniis, cuius etc. *Rome Sessor. 66 (2098). saec. ix. a) § where used has soft sound, but no strict distinction is observed between assibilated and unassibilated ti: in- nocent am, but definitione. b) Initially the rule; medially rarely. *Rome Sessor. 40 (1258). saec. ix. a) § used for assibilated ti. Distinction observed: scien^a, adtingeret. b) Initially and medially: In, Ire, but illius; hujus, elus etc. *Rome Sessor. 41 (1479). saec. ix. a) § for assibilated ti. Distinction observed. b) Initially, the rule; but ipse, illi; medially not al- ways: huIus and huius, maior. *Rome Sessor. 96 (1565). saec. ix. a) § for assibilated ^i. Distinction made: prophe§ am, ^ibi. b) Not regularly used: In, but also in, huius, adiunxit etc. *Rome Sessor. 63 (2102). saec. ix. a) ^ for assibilated ti. Distinction usually observed: po§us, tanti. b) As a rule not used. In the more recent MSS of this school — for the above named Sessoriani are supposedly all from Nonantola — § and the ^i-distinction and i-longa are all given up^). The same is true of the MSS of Vercelli, Novara, Bobbio, Verona, Lucca and other Italian centres. These elements dis- appear as soon as the Caroline minuscule prevails. ') Is it possible that we have a revival of the practice in the MS *Bologna Univ. 1604 (Nonantola) saec. Xl/XII, or is it a case of copying? I noted ra/tonis (with i drawn down) but utique (with short i). 46 12. Abhandlung: E. A. Loew Hot nan SchooL No very ancient minuscule MSS are known. Those that are posterior to the 9^*^ century lack the ligature § and ob- serve no fi-distinction. The i-longa is not seldom used initially. The Beneventan or South Italian School ^) a) I. In oldest minuscule MSS (saec. viii) § is used indiffer- ently, e. g. Monte Cassino 753: ui^is and mit^tur. Bam- berg H J IV 15: no§§iam. II. In Paris 7530 saec. viii ex. § is regularly reserved for assibilated ti, and the distinction is strictly observed. Although in some MSS of the 9^'' century insecurity is still to be noted (e. g. Vatic. 3320, where a later corrector often changed tio to ^o, and Naples VI B 12) the majority of the MSS show perfect knowledge of the two uses of ti. From the 9^^* to the 14^^ century the form ^ is regularly used for assibilated, and the normal form for unassibilated ti. This is one of the main rules of the Beneventan script. A scribe rarely wrote ord- inary ti for ^. I have noted but few cases, e. g. Rome Valli- cell. D. 5, saec. XI in.: unguentiam; Vatic, lat. 595: petiit, changed by corrector to pe^it, and some cases in Floren. Laur. 68, 2. ^) Occassionally too, we find ci for ti. This occurs so seldom that it is without doubt the result of slavish copying from an original in which vi stood for assibilated ti — and such spelling was certainly not unusual in the schools north of the Beneventan zone. Examples are: Monte Cassino 5: precio corrected to pre^o; Monte Cassino 295: uicia corrected to ui§a; Vatic, lat. 3973: ueneciis, and Vatic. Borgian. 339: cicius. On the other hand, there is nothing surprising if *) The following summary is based upon an examination of over three hundred Beneventan MSS. ^) Cf. Andresen, In Taciti Historias studia critica et palaeographica I (1899), p. 8. Studia palaeographica. 47 we find the ligature § for ci. I noted perni^e in Monte Cassino 187, saec. ix. An 11*^^ century corrector wrote ci for the ligature. b) For the usage of i-longa in Beneventan MSS see p. 9-10. Visigothic Minuscule. a) The frequent occurrence of § is noticeable only in the oldest MSS, e. g. Verona LXXXIX (where it is used indiffer- ently) and Autun 27 + Paris Nouv. Acq. 1628 — 9 (where there is a tendency to reserve the ligature for the assibilated sound). In MSS of the 9^^ or 10^^ century § is found here and there at the end of a line to save space. It does not form part of the calligraphic hand. The distinction between assibilated and unassibilated ti was in time graphically represented. As this question is of importance in dating Visigothic MSS, it has been treated separately and at greater length below. See part IV. b) For the usage of i-longa in Visigothic MSS see above p. 8 — 9. The MS evidence is given in part IV. German Schools. Early Minuscule MSS from German centres have as a rule neither ^ nor the ^i-distinction, nor the i-longa — owing most likely to Caroline influence. Nevertheless in several MSS of the transition period^ is found, along with other cursive features such as ri and te. Its presence, therefore, may safely be taken as a hint of the date of the MS. I noted § sparingly used in the following MSS. *Munich 4547.^) (Kysila-group.) saec. viii/ix. a) used for hard sound: ^meret (f. 11), uerita^s (f. 12), inmaculag (f. 12), ^bi (f. 22) etc. b) i-longa is not used. ^) Dr. Wilhelm of the University of Munich places the Kysila-group of MSS in the region of Utrecht. This judgment is based upon litur- gical and philological evidence furnished by the MSS themselves. 48 12. Abhandlung; E. A. Loew *Munich 4549. (Same group.) saec. viii/ix. a) % used indifferently: ui^is, impagen^ae, men^, cura^s, la§tat etc. b) Initially here and there; not medially. *Munich 4542. (Same group.) saec. viii/ix. a) § occurs for the assibilated sound, but chiefly the ordinary U: sapien^am (f. 139^) corrup^onem (f. 132') but next line: corruptione, with ordinary tl. b) Initially in the word in; not medially. *Munich 14421. saec. viu/ix. a) 9 like the ligature te is found chiefly at the end of the line, and is used indifferently: stul§ (f. 9^'), dix- eri^s (f. 12^), uerita^s (f. 15^), laeti^a (f. 24), captiui- ta§s (f. 43 in middle of line) etc. b) Not used. *Munich 4564. saec. ix. Hand A is calligraphic, B more cursive. a) Not used by hand A. Hand B used § indifferently: cotgdiae, ora^one (f. 220), benedici^s, facials (f. 220^), turba^onem (f. 221^). b) Not used. *Munich 6277. saec. ix. a) § used indifferently: opera^o, per^mescat, in^mo (f. 50), iusti^e, ni^tur, desperationem with ordinary ti (f. 50') etc. b) Not used. *Munich 6402. saec. ix. a) Where found % usually has the assibilated sound: por^o (f.45), gra^a (f. 51^), e^am, genera^o (f. 52) etc. But talen^ (f. 53^). Ordinary ti is chiefly used for either sound, yet ci occurs for ti: praecio, praeciosi (f. 61). b) Here and there it crops up, but manifestly due to the exemplar: malor and maior (f. 53"^). *Munich 4719^. saec. ix. a) % used indifferently: contesta^o, perseuera^, opta^o, obstina^s. b) Not used. Studia Palaeographica. 49 In MSS of the St. Gall, Reichenau and Chur districts no ^^'-distinction is observed. In many of them, however, ci takes the place of assibilated ti — a practice already noted in numerous French MSS, which probably served as models for the Swiss. ^) The ligature ^ occurs only here and there, used indifferently. As a rule i-longa is not employed; occasionally it is found at the beginning of a word, and less frequently in the middle. The following early examples have been examined^): § St. Gall 70, §238^), 44, 914, 185, §731, §3483), § 722; Berne 376 3); § Zurich Cantonsbibl. CXL^), § Cantonsbibl. (Rheinau) 30; § Einsiedeln 27, § 347 3), 199^), §281 3) and 157. 3) Insular Schools.^) a) The form of the ^i-ligature found in Insular MSS, as has been mentioned above, differs from § in that the upper loop or curve is missing (see p. 20). The form could easily have arisen from semi-uncial t combining with i. The ab- sence of the form ^ in pure Insular products may be regarded as one of the many proofs of the peculiar origin — in which cursive played no part — of the Insular writing. The ^i-lig- ature, where found, is used indifferently. No distinction be- tween the assibilated and unassibilated sounds is made. b) It is fair to say that i-longa — which as has been shown is of cursive origin — is foreign to Insular MSS. It ^) Historical and graphic considerations suggest Burgundian in- fluence. Further investigation may disclose relations between Luxeuil and Chur or some other Swiss centre. I suspect that the MSS Berne 611 and St. Gall 214 are Swiss products formed under the influence of Luxeuil. 2) MSS preceded by § have ci for soft ti. 3) In this MS ^ used indiff'erently is occasionally found, especially at the end of a line. *) Cf. facs. in Lindsay, Early Irish Minuscule Script, Oxford 1910. Sitzgsb. d. philos.-pbilol. u. d. hist. K1. Jahrg. 1910, 12. Abh. 4 50 12. Abhandlung: E. A. Loew is often found initially, but not with any apparent regularity. Medially it is used but rarely.^) I o-ive a few examples. For the ^i-ligature I use italics. *Bodl. Douce 140 f. 100\ a) can^icum, b) not used. The Book of Dimma. a) fueri/is, b) used initially, often. The Book of Mulling, a) uultis, b) used initially, often. *Vatic. Pal. lat. 68. a) adnun^iauit, deniergen/is, b) used in in. * Vatic, lat. 491. a) grafias, pieta^is, b) not used. *London Cotton Tib. C II. a) potesta/?', b) used with in. *Paris 10 837. a) fimoreni, agapi/i. *Vatic. Pal. lat. 235. a) Hhi, fon^ibus, b) not used. *Vienna 16. a) repeti^ione, ^ibi, b) In long. Insular in- fluenced by Italian cursive. Turin F IV 1 fasc. 6. a) indigna^ionein, tihl, mortis, b) In long. We have seen, then, that the ^i-ligature originated in Italian cursive of the early middle ages. We have found it in all those types of pre-Caroline minuscule which obviously base upon cursive, and the usage in the MSS corresponded to that of the documents. We missed it, on the other hand, in most of the MSS from about the beginning of the 9^^ century. This circumstance can be attributed to but one cause — the Carohne script-reform. The hypothesis is confirmed by the consideration that many MSS of about the year 800, written in north Italy, France and Germany show traces of the aban- doned practice. They are the MSS of the transition period. Still more cogent evidence is furnished by the fact that in the Beneventan centres where the Caroline influence did not reach, the ^i-ligature continued in use along with several other ^) I have found i-lonj^a medially in *Palat. 202 delnde; *Bodl. Laud, lat. 108 lelunandura. I believe that in all such cases foreign influence is responsible for the i-longa. Studia palaeographica. 51 cursive features which elsewhere were abolished. Doubtless for similar reasons ^ is found in many Visigothic MSS, though relegated, to be sure, to a place of insignificance. The history of §, then, is a kind of epitome of the development of Latin minuscule in its first important stage. We have seen, also, that the spelling ci for soft ti is a characteristic of early French, not of early Italian and that the graphic distinction of assibilated and unassibilated ti was regularly practiced in but two schools, the Beneventan and the Visigothic; although the usefulness of distinguishing in script the two sounds of ti was elsewhere recognized — as several instances clearly show — before the practice became a law of the Spanish and south Italian minuscule. From all this the palaeographer may draw a practical hint or two for dating and placing MSS. For example, the regular use of % in a French MS is a fair sign that the MS was written some time before the middle of the 8*^ century.^) Its sporadic appearance, on the other hand, suggests that the MS belongs in the period of transition, i. e. about the year 800. The frequent use of ci for soft ti in a pre-Caroline MS points to French origin rather than to Italian or Spanish.^) And certain corruptions in the text due to the ligature ^ permit a surmise as to the probable nature of the archetype.^) M The same is true for Visigothic MSS. -) See p. 20, note 1. An editor collating a Visigothic MS must be on his guard against mistaking for c a certain form of t w^hich occurs in ligatures. Even Maffei misread ci where the MS has ti. Cf. Spagnolo, L'Orazionale Gotico-Mozarabico etc. estratto dalla Rivista Bibliofirafica Italiana (10—25. Aug. 1899} p. 8, line 11. For precmm read pre^mm. ^) 1 refer to cases where the text has g for ti, an error due most likely to copying, from an original which had ^, by a scribe unaccus- tomed to the ligature. An instructive example is cited by Traube, Text- geschichte der Regula S. Benedict!, p. 85. 52 12. Abhandlung: E. A. Loew IV, Ti in Spanish MSS. In Yisigothic calligraphy the manner of writing ti is of signal interest and importance. After a certain time the Spanish scHbe, just as the Beneventan, used two distinct forms for as- sibilated and unassibilated tL From evidence given below it will be seen that it is possible to fix with some degree of precision the period when the custom of making the distinction was introduced into Visigothic book- writing. In other words, a cri- terion for dating can be won. The assibilated and unassibil- ated forms differ but slightly.^) In the case of unassibilated ti the normal forms of t and i are retained. In the case of as- sibilated ti the i is prolonged below the line and often turned in instead of out (cf. plates 5, 6 and 7), the whole difference lying in the form of the i, the letter t suffering no change. The Spanish form for assibilated ti (o)) corresponds, then, to the Beneventan for unassibilated. But the form §, which is 1) This perhaps explains how it happened to escape the attention of palaeographers. Steffens has noted the (^-distinction in his description of Escor. T II 24 (formerly Q II 24). That he too failed to realize that it was as much a scribal rule in Visigothic as in Beneventan is seen from the fact that in his introduction he speaks of the ^-distinction in Beneventan M6S but not in Visigothic. I believe that Delisle's report of my observations on the subject (Comptes-rendus de FAcademie des inscriptions, 1909, pp. 775 -778 and Bibliotheque de Tecole des chartes LXXI (1910) 233-235) is its first formulation in palaeographical liter- ature, for there is no mention of it in Munoz y Rivero, Ewald and Loewe, Wattenbach, or in the earlier writers on Spanish palaeography. It is a curious fact that even Paoli with whom the question of assibil- ated ti was a matter of keen interest made no reference to the distinc- tion in his description of the Visigothic MS Floren. Laur. Ashb. 17- Cf. CoUezione Fiorentina, pi. 33. Studia palaeographica. 53 regularly reserved for assibilated ti in Beneventan calligraphy, was not unknown in Spanish MSS. However, whereas in Ben- eventan it was a constant feature of the book-hand, in Spanish it was in time avoided. For, excepting the oldest known Visigothic MSS (Verona LXXXIX and Autun 27 + Paris Nouv. Acq. 1628 — 9) which employ ^ frequently, we find it chiefly at the end of a line, where economy of space demanded the shorter form, or in additions entered in cursive where % is usually confined — as is the case in Italian cursive — to re- presenting the assibilated sound. It is needless to say that the custom of graphically dis- tinguishing the two kinds of ti in the Visigothic book-hand, which dates, as will be seen, from about the end of the 9"' century, is in no wise a reflection of a change of pronun- ciation then taking place in Spain. The rule given by Isidore, bishop of Seville, for the orthography .of such words as iiistitia^ militia etc. — to the effect that they should not be written with a ^ as they were pronounced but with a t as was Latin usage — shows that three centuries prior to the introduction into calligraphy of the graphic distinction between assibilated and unassibilated ti, the difference in their pronunciation w^as already an accomplished fact.^) And we know from inscriptions that the assibilation of ti must have taken place at quite an early date.^) That the graphic distinction should have fol- lowed centuries after the phonetic change may be natural enough — we encounter the same phenomenon in Italy — but it is important to observe that the distinction was prac- ticed in cursive writing long before it was employed in calli- graphic products, and that the manner of representing the distinction in Spanish cursive (§ for soft ti) was the same as that employed in Italian cursive and in Beneventan book-hand — facts which seem to speak for the Italian origin of the 1) Isidor. Etymol. I, XXVlI, 28. See above, p. 17, note 7, where the passage is quoted. 2) On the assibilation of ti in the Latin-speaking countries see the works cited above, p. 16, note 2. 54 12. Abhandlung: E. A. Loew Studia palaeographica. 55 custom. This supposition becomes more convincinor when we remember that the Spanish scribe invented a new form for denoting assibilated ti, and that this form is found in Visi- gothic MSS a good century after the Beneventan scribe was making the distinction. That the practice of making the ^i-distinction in Visigothic MSS dates from about the year 900 is established beyond a reasonable doubt by the evidence of over one hundred MSS listed below. A word as to the nature of the evidence. It is furnished by two sources: the MSS themselves, and facsimiles of MSS. As for facsimiles, in the case of some MSS I was dependent upon one only; in other cases, however, photographs of several pages or even of the entire MS were at my disposition. More MSS might easily have been added without modifying results, but I pr'eferred to use only those dated by recognized author- ities, thus avoiding as far as possible basing an argument upon dates for which I alone was responsible. I also hesitated to use facsimiles when it was not clear whence they were taken, as in older books on Spanish palaeography. Notes furnished me by others were used only when supplemented by facsimiles. 1 am aware that the evidence supplied by facsimiles of one or two pages of a MS is not necessarily conclusive, as it may represent (as it sometimes does) the usage of one scribe and not of another. But whereas this evidence taken by itself might seem of questionable worth, its weight as supplementary evidence when used in connection with facts gathered from the MSS themselves will not be gainsaid. The fact that the usage found in the facsimiles is not at all at variance with the usage noted by me in the MSS is a guarantee of their value. However, the brunt of the argument will be borne by the forty- five MSS actually examined by me — MSS which are fairly representative of the different phases of Spanish calligraphy. In the following list the MSS are arranged approxim- 1) I have examined photographs of at least fifty MSS not included in my list. In these MSS the ti-usage agreed with that of the MSS whose evidence is given below. ately in chronological order. In most cases my date is iden- tical with that of others. In the few instances where the difference of opinion is essential the reasons for my date are ^riveii after the list.^) I give first the usage of ti, with ex- amples taken from the MS or from a facsimile. The itali- cized ft represents the ordinary form of t and L For th3 ligature § and the assibilated form of ti I have tried to reproduce the typical form found in the MS. After ti I give the i-longa usage. I also noted the use of the forked i-longa (shaped like a tall y). The form of the shafts of tall letters is given because of its value as a criterion for dating. Lastly, it seemed helpful to give some literature, for the sake of quick orientation. I gave that which I had at hand, without going out of my way to make researches extraneous to the purposes of this study. The references frequently cited appear under the following abbreviated forms: Beer. Handschriftenschiltze Spaniens, Vienna 1894. Beer-Diaz Jimenez. Noticias bibliograficas y catalogo de los codices de la santa Iglesia Catedral de Leon, Leon 1888. Bibl. P. L. H. Hartel-Loewe, Bibliotheca Patrum Latinorum Hispaniensis, Vienna 1887. Cat. Add. A Catalogue of the Additions to the MSS of the British Museum. Delisle-Melanges. Melanges de paleographie et de biblio- graphic, Paris 1880. Eguren. Memoria descriptiva de los codices notables conser- vados en los Archivos ecclesiasticos de Espafia, Madrid 1859. Exempla. Ewald et Loewe, Exempla Scripturae Visigoticae, Heidelberg 1883. Merino. Escuela Paleografica, Madrid 1780. Munoz. Munoz y Rivero, Paleografia Visigoda, Madrid 1881. I) See p. 81 sqq. 56 12. Abhandlung: E. A. Loew N. A. Neues Archiv der Gesellschaft flir iiltere deutsche Ge- schichtskunde VI (1880) p. 219-398 = P. Ewald, Keise nach Spanien im Winter 1878 — 79. *) MSS actually examined are starred. 1. *Verona Capitol. LXXXIX. saec. viii in.^) ut vid. a) No ti-distinction: pa/ienfie, unique, Hh\. Noteworthy is the relatively frequent occurrence of §. It is found passim on every page and is used indifferently: nequi^e (begin, of line), frucgficet (middle of L), men ^bus (middle of 1.), conscien^ia (middle of 1.). These four examples are taken from one page. In contemporary marginalia: ius^^am etc. Later MSS use § only occasionally at the end of lines. b) Rule observed. 2) Cf. Maffei, Opusc. Eccles., p. 80, pi. IV, no. 18 (whence Nouveau Traite III, 449, pi. 60); idem, Istoria Teologica (Trento 1742) pi. IV, part XVII and XXI; a poor facsimile also on p. CXXXI of Thomasii Opera omnia studio et cura Josephi Blanchini, Tom. I (Rome 1741); Spagnolo, L'orazionale gotico-mozarabico etc., 1) On f. 3^ (lower right hand corner) there is a rather obscure entry of a personal character ending with the words: in XX anno liut- prandi regis, i. e. the year 732. As the upper half of the page has the same kind of writing as the body of the MS, the above entry — if indeed we may regard it as chronicling an actual fact which then took place — gives us the terminus post quern non, and the mention of Luitprand would connect the MS with north Italy. It must be con- fessed that the first impression is that the MS belongs in the 9^^ century, — it is carefully and regularly written — but being a liturgical book, special pains may have been taken with it, which would account for the impression. Furthermore the rather frequent occurence of certain ligatures, especially of §, also favors the earlier date. I prefer to leave the question of the date undecided. The matter deserves further in- vestigation. 2) For the rules of i-longa in Visigothic MSS see above, p. 8—9. Studia palaeographica. 57 estratto dalla Rivista Bibliografica Italians (10—25 Aug. 1899); Ferotin, Liber Ordinum, p. XV, note 2. 2. *Paris Nouv. Acq. Lat. 1628 (ff. 17-18). saec. viii ut vid. a) No ti-distinction. In the more cursive portions § is used indifferently: ter^a, eviden^ssime. b) Rule observed. Occasionally even Ilia. Cf. Delisle, Les vols de Libri au seminaire d'Autun (Bibliotheque de I'ecole des chartes LIX (1898) 386-392.1) 3. Escor. R II 18. ante a. 779. a) No fi-distinction in minuscule portion: resurrec^/one, ter^io. In cursive parts the distinction is usually made, § or similar forms representing the soft sound: Ius/«*§am, e^am. Yet exceptions occur: segon^ia. b) i-longa rule observed in cursive and minuscule: In, Ipsa, Ibi, cuius; but ilia. Also i-longa with forked top: acala. Cf. Exempla, pi. IV— VII, whence Arndt-Tangl, Schrifttaf.*, pi. 8b; N. A. VI, 275; Bibl. P. L. H., p. 130; Steffens, Lat. Pal.^, pi. 35. 4. Madrid Tolet. 2. 1. saec. vni ex. ut vid. Now kept in Vitrina 4'', Sala I^. a) No f^-distinction: pafienter, terfia and sep^ima. b) Rule observed: Isti, malor, caIn, elus, even Illi; caIn with forked i-longa. Cf. Exempla, pi. IX: Bibl. P. L. H., p. 261; Mufioz, pi. VIII— IX. The date there given (10*^' cent., p. 119) is impossible. The date a. 708 given by Merino (p. 55) is likewise untenable. On the inscription at the end of the MS, which has been the cause of erroneous dating, see Berger, Hist, de la Vulg., p. 13. ^) These leaves as well as ff. 21—22 of Paris Nouv. Acq. 1629 formed part of Autun 27 which unfortunately I have seen only in facsimiles. Professor Lindsay kindly informs me that the distinction is usually made in ihe minuscule part of the MS, but not as in later Visigothic MSS, the assibilation being represented by § or some similar form. But cases of ^ for the hard sound as well as of ordinary ti for the soft sound also occur. It is very important to note that no distinction is made in the cursive portions. 58 12. Abhaudlung: E. A. Loew 5. Madrid Tolet. 15. 8. saec. viii ex. ut vid. Now kept in Vitrina 4^ Sala 1*. a) No fi-distinction: ter/ia, gra/issima. In the later ad- ditions in cursive the distinction is made as in Escor. U II 18. The use of § in the word den^bus (Exempla, pi. XII) recalls older cursive where no distinction is made and ^ is used indifferently. b) Rule observed, even lUic, Ille, but ilia also occurs. Cf. Exemi)la pi. X— XII, whence Arndt-Tangl, op. cit. pi. 8 c; Bibl. P. L. H., p. 291, "saec. viii/ir'; Beer, Codices Graeci et Latini photographice de- picti, Tom. XIII (Sijthoff, Leyden 1909), Praefatio p. XXI\', whence Ihm, Pal. Lat., pi. VII. (). Leon Eccl. Cathedr. 15. saec. ix. (Clark's photos.) a) No fi-distinction: erudi^ionis. an^iociam. b) Regular, even lUis and Illi. Cf. Beer-Diaz Jimenez, p. 16 sq., who date the upper script in the 10^^ century: "medio vel de- clinante IX. saec.'\ p. XVI of Prooemium to Legis Romanae Wisigothorum fragmenta ex codice palimp- sest© sanctae Legionensis ecclesiae protulit, illustravit ac sumptu publico edidit regia historiae Academia Hispana, Matriti (1896); Theodosiani libri XVI, edd. Mommsen et Meyer I, 1. p. lxx. 7. *London Egerton 1934. saec. ix in. ut vid. a) No ^i-distinction: cifius, diui/iis and anfiquissima. b) Rule observed: Idem, Iberiam, hulus, even lUe. Cf. Cat. Add. (1854-1875) p. 916; Facs. in Cat. of Anc. MSS in Brit. Mus. II, pi. 36. 8. *Moiite Cassino 4. saec. ix. See plate 3. a) No ^i-distinction: sapien^iam, ^ibi. :But in cursive marginal notes entered apparently by a later hand ^ is regularly used for assibilated U: senten/iam. b) Rule observed. Usually lUe, but occasionally ilia, ilium. Cf. Bibliotheca Casinensis I, 97 and facsimile. The date (saec. vii) can hardly be correct. Studia palaeographica. 59 9. *Monte Cassino 19. saec. ix. a) No ^i-distinction: ratio and re^inere. But cursive additions by a later hand have § to mark assibilation. b) Rule observed, even Ilia, also alt. Cf. Bibliotheca Casinensis I, 233 and facsimile. Their date is saec. vii, which is hardly possible. 10. Escor. & I 14. saec. ix ut vid. a) No ^i-distinction : inven^ione and dogmafibus. b) Rule observed : Id, In, Ignem, cuius, delude, even Ibi. Cf. Exempla, pi. XIII; N. A. VI, 250; Bibl. P. L. H., p. 70 and earlier Pertz' Archiv VIII, 815; Rev. Bened. XXVII (1910) p. 2. 11. Madrid Tolet. 14. 24 (now 10018). saec. ix ut vid. a) No /^-distinction: gra/^a, iumen/is. b) Rule observed, even Illis, Illorum. Cf. Exempla, pi. XVIII; N. A. VI, 318; Bibl. P. L. H., p. 290. 12. *Paris Lat. 2994 (part II). saec. ix ut vid. a) No /i-distinction : conpara/ione and pecca/i. b) Rule observed, even Ille, pro(h)Ibeant, coltu. Cf. Delisle, Melanges, p. 54 and Facs. de Pecole des chartes, pi. 281. 13. Paris Lat. 8093. saec. ix ut vid. (Vollmer's photos.) a) No /^-distinction : sep/ies, Ingen/ia and fluc/^bus. b) Rule observed, even Illi. Cf. De Rossi, Inscriptiones Christianaell, 292 (where Delisle in his description dates the MS saec. viii); Vollmer in M. G. H. Auctt. Ant., T. XIV, p. xix & xl. 14. *Paris Lat. 4667 a. 828. a) No ^i-distinction : Induc^ione and u^ilitafis. b) Rule observed: Ipsius and usually Ille but also illis. Cf. Nouveau Traite III, 327 and pi. 52; Delisle, Melanges, p. 54; Steffens, Lat. Pal.^, pi. 49; Prou, Manuel de Paleographie^ (1910), pi. V, no. 2. 60 12. Ahhandlunff : E. A. Loew Studia palaeographica. 61 15. Paris Lat. 12254. saec. ix iit vid. a) No /i-distinction: lec^ionis, utilis. b) Regular. Cf. Delisle, Le cabinet des manuscrits III, 229 (where no mention is made of the MS being Visi- gothic. His description is: t'critiire du VHP siecle). For files, see pi. XVIII, 4. 16. Leon Eccl Cathedr. 22 (CVI). post a. 839. (Voll- mer's photos). a) No ^^-distinction: dignaf^'onis and is^^s. b) Rule observed. Cf. Eguren, p. 7S — 9; Beer-Diaz Jimenez, p. 23 '^a. 839"; N. A. XXVI, 397: M. G. H. Auctt. Ant., T. XIY, p. XXXVIII, "saec. x in." and p. xl. 17. Leon Eccl. Cathedr. Fragm. no. 8. saec. ix ut vid. (Yollmer's photos.) a) No ^^-distinction : gra^iae, peten^i. b) Regular. Cf. Beer-Diaz Jimenez, p. 43: "s. x" and M. G. H. Auctt. Ant. T. XIV, p. xxxviii sq.: "saec. x". The script is of the oldest type. 18. Barcelona Rivipullensis 46 (flv-leaves). saec. ix. a) No fi-distinction: gentium, composifio and uagan^^bus. b) Rule observed. Ibi but ille. The MS presents several features unusual in a Visigothic MS, e. g. abbreviations of 2)rae and tiir and the Caroline symbols for nosfri, per and pro. Cf. Beer, Die Handschriften des Klosters Santa Maria de Ripoll, I 33 and pi. 1. (Sitzungsberichte d. Kais. Akad. d. Wiss. in Wien. Vol. 155 (1907), 3. Abh. 19. *Berne A 92. 3. saec. ix ut vid. a) No fi-distinction: mali^ia and Ira^i, damna^^one, mor^i- ferum. b) Rule observed. Cf. Steffens, Lat. Pal.^, pi. 35. 20. Madrid Univ. 31. saec. ix. (D. De Bruyne's photos, of entire MS.) a) No ^i-distinction: letitisL^ humilia^io and ves^imen^is. At the end of a line the ligature § is used for assibil- ated ti; oran^um, exulta^one. b) Rule observed, even Illius (often) and alt. Cf. Facs. in Merino, pi. VI; Berger, Hist, de la Vulg. , p. 22. The date (saec. x) in V\^attenbach, Anleit. z. lat. Pal.*, p. 22 is hardly possible. 21. *Siguenza Capitol. Decretale 150.^) saec. ix ut vid. a) No ^i-distinction: lustifisL. But at end of line, for economy of space, § is used for soft t: tradi^onum. Cf. preceding MS. b) Rule observed. Ihu, Ipsa and Ilia. Also ludalsmo; als. In the last two examples the i-longa splits at the top and resembles a tall y. Cf. De Bruyne and Tisserant, Une feuille arabo- latine de Tepitre aux Galates, in Revue Biblique, July 1910 (with facsimile). 22. *Paris Nouv. Acq. Lat. 238. saec. ix. a) No ^i-distinction: discref^one and sta^im. b) Rule observed, lUae but also ille: Ihu and ihu. Cf. Delisle, Melanges, p. 60—1 : "du x« siecle". 23. Escor. PIG. saec. ix. a) No ^^-distinction: contempla^ione and dedi^i. b) Rule observed. Cf. Exempla, pi. XXVI: "saec. fere decimo"; Bibl. P. L. H., p. 100: "saec. x — xi". The script is de- cidedly against this recent date. h^ ^) These few leaves were formerly attached to the cover of "De- cretale 150" in the chapter library of Sigiienza, where they were dis- covered by D. De Bruyne. They contain a unique specimen of the Latin and Arabic versions of St. Paul's Epistles, and for the present are pre- served in the Vatican library. 62 12. Abhandlung: E. A. Loew 24. Albi 29. saec. ix. a) No ^/-distinction: to^ius, par/ibus, orien^is. § is used indifferently but more often for soft tL b) Keguiar, even delnde, deinc, prolnde. Cf. Facs. in Catalogue general des nianuscrits des bibliotheques publiques des departements I (1849) 487. 25. "^La Cava I (formerly 14) Danila Bible, saec. ix post med. a) No //-distinction: genersitione and eun//bus. b) Rule observed: Ibi, Ibant, but illuc. Cf. Facs. in Sylvestre, Paleogr. Universelle III, pi. 141 and two plates in Cod. Diplom. Cavens., Tom I, Mano- scritti Membranacei, p. 1, where it is put in the 8*^ century. For its proper date see A. Amelli, De libri Barucli vetustissima latina versione etc. Epis- toLa ad Antonium M. Ceriani (Monte Cassino 1902) ])p. 7 and 14; Berger. Hist, de la Vulg., p. 15. This is by far the finest product of Spanish penmanship and book-decoration known to me. 26. Madrid. Univ. 32. saec. ix ut vid. (D. De Bruyne's photos.) a) No fi-distinction. b) Rule observed. Cf. Facs. in Merino, pi. VI; Berger, Hist, de la Vulg., p. 15 et sqq. 27. Toledo Capitol. 99. 30. saec. ix. a) No /^-distinction: e/iam, at/aigo. b) Rule observed. Cf. Exempla, pi. XV^I. 28. '^Paris Nouv. Acq. Lat. 2168. saec. ix ut vid. a) No /i-distinction: pes/ilenfia. b) Rule observed, even lUis. Cf. Delisle, Melanges, p. 76 '*du x® siecle". 29. "^T nchester John Rylands Library MS Lat. 116. saec. IX ex. ut vid. (Lindsay's photo.) a) No ^/-distinction: lustitla, mentis, cogita^ione. Studia palaeographica. 63 b) Rule observed: Iste, Ipse, Ideo, Ille, but more often ille; also ihs. i-longa with forked top in alt, esalas etc. Cf. Facs. in New Palaeographical Society, pi. 162. 30. *London Add. MS 30 852. saec. ix ex. ut vid. a) No /^-distinction: voca/ione, ui/iorum and tlhi. b) Rule observed, even Ille. Cf. Cat. Add. (1876—1881) p. 121; Facs. in Cat. of Anc. MSS of Brit. Mus. II, pi. 37. 31. *Paris Nouv. Acq. Lat. 2170 (Part I), saec. ix ut vid. a) No fi-distinction: etlam and cunc/is. b) Rule observed. Cf. Delisle, Melanges, p. 79: "peut remonter au X® siecle". 32. Escor. R II 18 (f. 95—95^). post a. 882. This folio contains the famous Oviedo catalogue, a) No /^-distinction: conla/ionum and can/?cum. Cf. Munoz, pi. IV; N. A. VI, 278; Becker, Catal. Bibl. Antiq., p. 59; Bibl. P. L. H., p. 135; Beer, p. 376 sqq. 33. Escor. P I 7. saec. ix ex. ut vid.^) a) No ^/-distinction: efiam, la^inum, ius^i^iam. b) Rule observed, even Ilia. Forked i-longa in alt, esalas. Cf. Exempla, pi. XIV; N. A. VI, 220, n. 4; Bibl. P. L. H., p. 101. 34. Escor. T II 25. saec. ix ex. ut vid. (Fr. Manero's photo.) a) No //-distinction: po/ius, mul//, iustitle. b) Rule observed, even Illis, prolnde. Forked i-longa in alt. ^) This and the following MS have the acrostic Adefonsi principis lihrum. It has generally been assumed that this referred to Alfonso II (795—843). As the writing of these two MSS resembles that of some dated MSS of about the year 900, I am inclined to believe that Alfonso III (848—912) is meant, especially as there is historical evidence for books having been presented by the latter as well as the former. Cf. Beer, p. 376 and 379. 64 12. Abhandlung: E. A. Loew 35. *Paris Nouv. Acq. Lat. 1298. saec. ix ufc vid. a) No ^i-distinction : etisLin and an^icam. b) Regular. Cf. Delisle, Mt^langes, p. 108: "minuscule nu'lee de cursive du xi^ siecle'\ Mixed minuscule and cursive is more in keeping with my date. 36. *Paris Nouv. Acq. Lat. 2167. saec. iX ut vid. a) No /i-distinction: pes/ilen^/a. b) Rule observed, even Ihs and Illis. Cf. Delisle, Melanges, p. 76: "du x*^ siecle". 37. *Paris Nouv. Acq. Lat. 260. saec. ix ut vid. a) No ^^-distinction: uitio and volupta/^s. b) Rule observed: Id, Ipse but illo. Cf. Delisle, Melanges, p. 114: "du xi® siecle". 38. *Paris Lat. 10877 (cf. Tours 615). saec. ix ex. ut vid. a) No ^^-distinction: totlus and grega^i. b) Not regular: incumbere, deinde (with short i). There is something foreign about this MS. Cf. Delisle, Melanges, p. 54: "probablement du X° siecle''. 39. *Paris Lat. 10876. saec. ix ex. ut vid. a) No ^i-distinction: conuersa^io and exconimunica^is. b) Not regular: inter, imperium, ista, proinde (all with short i) which is a transgression of the rule. This MS belongs to the same school as the preceding. Cf Delisle, Melanges, p. 54: "probablement du x^ siecle". 40. ^London Add. MS 30854. saec. ix ex. ut vid. a) No #i-distinction. b) Regular; even lUius. Cf. Cat. Add. (1876-1881), p. 121: "x^^ cent.'\ 41. Escor. I III 13. saec. ix/x ut vid. (Traube's photo.). a) No ^i-distinction. b) Regular. Cf. Bibl. P. L. H., p. 81: "saec. x". Studia palaeographica. 65 42. Madrid Tolet. 14. 22 (now 10029). saec. ix/x ut vid. a) Distinction made in some parts and not in others: e^iam, parenfi (no distinction); presena)a, na^ique (with distinction). The marginalia, apparently of the same time, observe the distinction : deprecaa)o. b) Regular. Cf. M. G. H. Auctt. Ant. T. Ill 2 (1879) pp. l & lit; ibid.facs; N. A. VI, 316 and 581: "saec. x"; Bibl P. L. H., p. 284 "saec. ix/x"; M. G. H. Auctt. Ant., T. XIV, p. xxxviii. 43. "^London Thompsonianus 97.^) a. 894. a) Distinction made: fora)a but ductile, a) Regular, even Illi. Cf. A descriptive catalogue of the second series of 50 MSS in the collection of H. Y. Thompson (1902) p. 304. 44. Madrid Tolet. 43. 5 (now 10 064). saec. ix/x ut vid. a) Distinction made: precedenq^um but ius/issime. b) Regular; but illi, also prolbendum. Cf. Exempla, pi. XVII: "s. ix si non antiquior"; Bibl. P. L. H., p. 299. Reasons for my date are given below, p. 83 sq. 45. Madrid Acad, de la Hist. 20 (F. 186),^) Hartel-Loewe no. 22. saec. ix/x ut vid. The Bible of San Millan. a) Distinction made in first part of MS: tribulaa)one, but angus^ia, can^kum. No distinction in last part of MS, which is by a different hand. The marginalia which are added make the distinction. ^) This excellently preserved MS (which I was privileged to examine in the library of its present owner to whom 1 here express my thanks) was purchased of Lord Ashburnham in 1897. The script is manifestly of the late 9*^ or early 10*^ century, and the subscription which dates it 894 (era 932) may be trusted. 2) The entire MS has been photographed for the Commission on the Vulgate. D. De Bruyne, one of its members, kindly allowed me to examine the photogi-aphs. Sitzgsb. d.philos.-philol.u. d, hist. Kl. Jahrg.1910. 12. Abb. 5 66 12. Abhandlung: E. A. Loew b) Regular, even lUis. Also slon, ebralce, with forked i-longa. Cf. Exempla, pi. XXV: ''saec. x"; N. A. VI, 332: "saec. ix"; Bibl. P. L. H., p. 500: '^saec. vm". Accor- ding to a subscription in the MS its date is 6G2! Berger, Hist, de la Vulg., p. 16. For discussion of the date see below, p. 84. 46. Madrid Tolet. 10. 25 (now 10 007). a. 902. a) Distinction made by first scribe: sena)unt but celes- t'lum. Often ^ is used: exeun§uni. No distinction by second scribe. Here the work of the corrector can be watched ; he adds the tail to i where t is assibilated. On f. 47"^ eogam seems to be by second scribe. The scribe toward the end of the book uses q) lor assibilated tl. Likewise a later entry on f. 147"" makes the distinction. These valuable details I have from W. M. Lindsay. b) Regular, but illut, illo. The second scribe has Itaque occasionally with forked i-longa. The clubbed shafts of tall letters tend to become angular. Cf. Exempla, pi. XIX; Monaci, Facs. di antichi MSS, pi. 88; Bibl. P. L. H., p. 265. 47. Madrid Tolet. 35. 1 (now 10 001). saec. ix/x ut vid. a) No fi-distinction: ter^ia, tihi. b) Regular. Forked i-longa in alt, efralm. Cf. Exempla, pi. XX VIP; Bibl. P. L. H., p. 296: "saec. ix/x'\ 48. Leon Eccl. Cathedr. 14. saec. x in. (Clark's photo.) a) No ^i-distinction: tihi and ratio. b) Regular. Shafts of tall letters have angular tops. Cf. Beer-Diaz Jimenez, p. 15. 4r». Barcelona Eivipullensis 49. a. 911. a) No ^i-distinction: letitia, ahstinentie. But ^ is used for soft ti at the end of a line: senten^a. b) Regular: Ipsa, Ihu, even Illis. Studia palaeographica. 67 Cf. Beer, Die Handschriften des Klosters Santa Maria de Ripoll, I 34 and pi. 2 and 3 (see above no. 18); Steffens, Lat. Pal.^, pi. 66 b (= 54 of 1«* ed.). 50. Escor. a I 13. saec. x in. a) Distinction made: Ius^iq)as, diligena)a. b) Regular, even Illi. Cf. Munoz, pi. Y: "a. 912"; Exempla, pi. XV: *'fortasse a. 812"; N. A. VI, 226: "saec. ix"; BibL P. L. H., p. 10: ''a. 912", where the note on p. 13 contains Ewald's discussion of the date. Beer (p. 383 note and p. 384 note 3) favors 812; Traube, Text- geschichte der Regula S. Benedicti, p. 64 (= 662). The reasons for my date are given below, p. 82 sq. 51. Manchester John Rylands Library MS Lat. 93. a. 914. Written at Cardena by Gomiz. (Lindsay's photo.) a) No //-distinction by original scribes: scien/iam, potes- ta/ibus. But a contemporary corrector makes the dis- tinction: acogo (f. 58), acogonibus (f. 292). b) Rule observed, but ille, ihs (also Ihs). Forked i-longa in alt, hiems, ludalca. The subscription which dates the MS will be pub- lished by Dr. M. R. James in his catalogue of the John Rylands MSS. 52. Escor. T II 24 (formerly Q II 24). saec. x ut vid. See pi. 5. a) Distinction made: alqpus but latino, quaesiqpo but quaestio. b) Regular. Cf. Exempla, pi. VIII (older literature given) ; Munoz, pi. 3; N. A. VI, 272; Bibl. P. L. H., p. 112; Beer, Praefatio to Tolet. 15. 8, p. XXIV; Steffens, Lat. Pal.^, pi. 36 (= SuppL, pi. 17). In these works the MS is dated saec. vm, saec. vni/ix, a. 733 or 743. The grounds on which my date is based are given below, p. 81 sq. My facsimile I owe to the courtesy of Dr. Franz Steffens to whom I here express my thanks. 5* 68 12. Abhandlung: E. A. Loew 53. Madrid Tolet. 15. 12 (now 10 067). a. 915. a) No fi-distinctioii by one scribe: e^^am, perfecfionis. Distinction made by another: ecgani, ])iit per^^mescit. See plate 4 containing a facsimile of both hands. b) llegiilar. One hand writes invariably illius; another has Illo.'' Also alt with forked i-longa. The up-strokes of the scribe who makes the fi-distinction are strongly clubbed and often tend to end in an angle — a feature of the early 10^^ century. Of. Exempla, pi. XX; Bibl. P. L. H., p. 293. 54. Madrid Acad, de la Hist. 24 (F 188). Hartel-Loewe, no. 25. a. 917? a) Distinction made by first scribe:^) districaq)one. No distinction at end of MS: e^iam, ra^io. b) Regular. In first part even Ille. Forked i-longa in Igne. The script is not the compact sort of the 9<^ century. Cf. Exempla, pi. XXI; N. A. VI, 332; Bibl. P. L. H., p. 503. The subscription which furnishes the date seems to have been tampered wnth. Cf. pi. in Exempla. 55. Madrid P 21 (now 1872). saec. x in. ut vid. a) Distinction made: graq)as but fdulo. b) Regular. Cf. Exempla, pi. XXVIII: "saec. x/xi". The script is plainly against this date. 56. Escor. S I 16. saec. x in. ut vid. a) No //-distinction: trisfi/ia. b) Regular, illius. The script presents a strange ap- pearance. Cf. Exempla, pi. XXXVII: "saec.xi utvid"; Eguren, p. 82. For my date see below, p. 84 sq. 57. *Paris Nouv. Acq. Lat. 238 (tly-leaf). saec. x ut vid. a) Distinction made: posioQonem but mar/ires. b) Regular. Cf. reference cited to no. 22. 1) These facts 1 learn from W. M. Lindsay. The plate in the Exempla reproduces the portion where no distinction is made. Studia palaeographica. 69 58. *London Add. MS 25600. a. 919. a) Distinction made: pudicia)a, ius^iogae, but fniieant. b) Regular, even Illis. Cf. Cat. Add. (1854-75) p. 208; Facs. Pal. Soc, pi. 95; Arndt-Tangl II, pi. 36; Cat. Anc. MSS Brit. Mus. II, pi. 38. The shafts of the letters h, d, h, i-longa and I have a prefix (or serif) at the top consisting of a small stroke made obliquely from left to right and upwards. In some MSS it is made at a right angle with the main shaft and often extends beyond it thus giving it the form of a mallet-head (cf. pi. 5, 6, 7). This graphic feature is noteworthy, as it is lacking in MSS of the preceding periods. 59. Leon Eccl. Cathedr. 6. a. 920. (Clark's photo.) a) Distinction made: edia)onem but legeri^is. b) Regular. Cf. Beer-Diaz Jimenez, p. 5; Berger, Hist, de la Vulg., p. 17. 60. Madrid Tolet. 11. 3. a. 945. (Kept in Vitrina 2% Sala l\) (Haseloff's photo.) a) Distinction regularly made by one scribe: inia)um, uia)is but extitit. Yet another scribe (to judge from the facsimile in Munoz) seems unsteady in his use, for he makes the distinction in some words and not in others: silena}um (1. 1) but silen/mm (1. 6); contemplaa)onis (1. 7) but contempla^^onum (1. 4). The examples are from Muiioz' facsimile. b) Regular. The tops of tall letters have a prefix. Cf. no. 58. Cf. Munoz, pi. VI and p. 117. 61. *London Add. MS 30844. saec. x ut vid. a) Distinction made: precgum. b) Regular, even Ilia. Cf. Cat. Add. (1876-1881), p. 119. 70 12. Abhandlunsr : E. A. Loew 62. Madrid Acad, de la Hist. 25 (F 194), Hartel-Loewe, no. 8. a. 946. a) Distinction made: pif:jriq)am but timore, celesfia. b) Regular, even Ille; forked i-longa in hebralca. Cf. Exempla, pi. XXII; N. A. VI, 331; Bibl. V. L. IL, p. 493. 63. Manchester John Rylands Library MS Lat. 99. a. 949. Written at Cardena. (Lindsay's photo.) a) Distinction made: poenitenqpani, tribulaogo but sa- lufis, ^imore. b) Regular. The tops of tall letters have a prefix. Cf. no. 58. The subscription which dates and places this MS will be given by Dr. M. R. James in his forth- coming catalogue of the John Rylands MSS. 64. *Paris 2855 (part II). ca. a. 951. a) Distinction made: acq)onem, but deser^^ and moles- ^iarum. b) Regular, yet ihm, ilium. The tops of the tall letters have a prefix. Cf. no. 58. Cf. Delisle, Melanges, p. 53, where older liter- ature is cited; Facs. see Sylvestre, Paleog. Univ. Ill, pi. 206; Facs. de I'ecole des chartes, pi. 277. 65. Escor. a II 9. a. 954. a) Distinction made: profanaa)onibus but cuncfe b) Regular. Script not compact. Tlie tall shafts tbicken at the top in a triangular form. Cf. Exempla, pi. XXIII; Bibl. P. L. H., p. 19. 66. *Paris Nouv. Acq. Lat. 239. saec. x. a) Distinction made: tris^ia}e but celes/ia. Cf. Delisle, Melanges, p. 78. 67. Leon Eccl. Cathedr. 21 (additions on a page left blank). saec. x. (D. De Bruyne's photo.) a) Distinction made. Studia palaeographica. 71 The script may even be more recent. It shows foreign influence, e. g. p = prae; p with superior o = pro ; m with apostrophe = mus, etc. The Catal- ogue by Beer- Jimenez does not describe these additions. 68. *Floren. Laur. Ashburnh. 17. saec. x ex. ut vid. a) Distinction made: generaor)onem but tibi. b) Regular, even Ilia, Illius, Illi. The tops of the tall letters have a prefix. Cf. no. 58. Cf. Facs. in Collez. Fiorent., pi. 33; Rivista delle Bibl. e degli Archivi XIX (1908) p. 5. See above p. 52, n. 1. 69. Madrid Acad, de la Hist. F 212. Hartel-Loewe, no. 44. saec. X ex. ut vid. a) Distinction made: spaq)um but complec^^tur. The tops of the tall letters have a prefix. Cf. no. 58. Cf. Exempla, pi. XXIV: "a. 964"; N. A. VI, 334: "saec. x"; Bibl. P. L. H., p. 514: "saec. xi". 70. *Paris Nouv. Acq. Lat. 2170 (last 22 leaves). saec. x ut vid. a) Distinction made: ins^^tua)onis, oraogone. Cf. Delisle, Melanges, p. 79. 71. *London Add. MS 30846. saec. x ut vid. a) Distinction made: supplicaa)one but pecca^is. Cf. Cat. Add. (1876-1881) p. 120. 72. *London Add. MS 30845. saec. x ut vid. a) Distinction made: cessaa}one but pecca^is. Cf. Cat. Add., p. 120; Facs. in The Musical No- tation of the Middle Ages (London 1890) pi. I. 73. Escor. d I 2. a. 976. (Traube's photo.) a) Distinction made: raq)one, sacerdo/^bus. b) Regular. Forked i-longa in laici. Tops of tall letters have prefixes. Cf. N. A. VI, 238; Bibl. P. L. H., p. 43; Facs. in N. A. VIII, 357, containing a line of script and one of arabic numerals, perhaps the earliest example in a western MS. ii 72 12. Abhandlung: E. A. Loew 74. *Paris Nouv. Acq. Lat. 2180. ante a. 992. a) Distinction made: eqjam, ius/ia)a, but iuventu^^. b) Regular; Ibi but illi. Cf. Delisle, Melanges, p. 101. 75. Escor. d I 1. a. 992. a) Distinction made: oblaq)ones but re/inent. b) Regular. The tops of tall letters have a prefix. Cf.no. o8. Cf. Exempla, pi. XXVII b; N. A. VI, 236; Bibl. P. L. H., p. 43. 76. *Paris Nouv. Acq. Lat. 1296. saec. x ut vid. a) Distinction made: aucqjo but es/hno, congesfio. This is perhaps the oldest Latin MS on paper; sheets of vellum are interspersed. Cf. Delisle, Melanges, p. 109: "du xif siecle". 77. *Londoii Add. MS 30851. saec. x/xi ut vid. a) Distinction made: s^ilancga. b) Regular, even lUud. The tops of the tall letters have a prefix. Cf. no. 58. Cf. Cat. Add. (1876-1881) p. 120. 78. *London Add. MS 30847. saec. xi ut vid. a) Distinction made. Cf. Cat. Add. (1876-1881) p. 120. 79. *Paris Nouv. Acq. Lat. 2179. saec. xi ut vid. a) Distinction made :Indignaq)o but quaesfionarii, vestigia. Cf. Delisle, Melanges, p. 95. 80. Escor. e I 13. saec. xi ut vid. a) Distinction made: geronq)us but ualen^aius. b) Regular, even lUud. Tall letters are very long and have a prefix at the top. Cf. no 58. Cf. Exempla, pi. XXIX. *'saec. x/xi." 81. *Londoii Add. MS 30850. saec. xi ut vid. a) Distinction made: oraa}one but uolupta^i. Cf. Cat. Add. (1876-1881) p. 120; Facs. in The Musical Notation of the Middle Ages, pi. IV. Studia palaeographica. 73 82. *Paris Nouv. Acq. Lat. 2178. saec. xi ut vid. a) Distinction made: paa)en^?s. Cf. Delisle, Melanges, p. 85; Facs. pi. II in catal- ogue of sale (1878). 83. Escor. & II 5. saec. xi ut vid. (Clark's photo.) a) Distinction made: paqr)enq)a but odis^i. b) Regular. Cf. Bibl. P. L. H., p. 75. 84. Madrid Tolet. 35. 2 (now 10110). saec. xi. a) Distinction made: Insurgena)um. Cf. Exempla, pi. XXX. The date "a. 1006'' is ffiven in index on the authority of Merino. But there is much uncertainty in connection with this date. The script is very ill-formed and may be older than saec. xr. 85. *Paris Nouv. Acq. Lat. 235. saec. xi ut vid. a) Distinction made: aedificaoc)0 but ])ro^inus, modesfiam. The tops of the tall letters have a prefix. Cf. no. 58. Cf. Delisle, Melanges, p. 75. 86. *Paris Nouv. Acq. Lat. 2176. saec. xi ut vid. a) Distinction made: raogone but multi. Cf. Delisle, Melanges, p. 70; Facs. pi. IV in catal- ogue of sale (1878). 87. *Paris Nouv. Acq. Lat. 2177. saec. xi ut vid. a) Distinction made: Iusfiq)ae, paq)enfi. I noted lustida (p. 473). The use of ci for soft ti begins to creep into MSS during the 11"* century, and is often found after that time. — The tops of the tall letters have a prefix. Cf. no. 58. Cf. Delisle, Melanges, p. 71. 88. Escor. & I 3. a. 1047. (Clark's photo.) a) Distinction made: iaa)onem but con^inet. The tops of tall letters have a prefix. Cf. no. 58. Cf. Munoz, pi. XI, p. 121; Beer, p. 218. 74 12. Abhcindlun*,': E. A. Loew 89. *London Add. MS 30855. saec. xi iit vid. a) Distinction made. Cf. Cat. Add. (187G-1881) ]). 122. 90. Madrid Nacion. (Beatus super Apocalypsim.) a. 1037 1065. Now kept in Yitrina 1% Sala P. a) Distinction made: ea)am but al/^'tudo. b) Regular: alt with forked i-longa. Cf. Munoz, pi. XII (where no press-mark is given). 91. Madrid Nacion (Forum judicum from Loon.) a. 1058. Now kej)t in Vitrina 4*, Sabi I\ a) Distinction made: preq)o but faculta^ibus. b) The tops of tall letters have a prefix. Cf. no. 58. Cf. Munoz, pi. XIII (no press-mark). 9± Madrid Acad, de la Hist. F 211. ITartel-Loewe, no. 47. saec. XI ut vid. a) Distinction made: quaesia}0. b) TJegular, but illius. The tops of tall letters have a prefix. Cf. no. 58. Cf. Exempla, pi. XXXVI. 93. Madrid Eoyal Private Library 2 J 5. a. 1059. a) Distinction made: graoja but salu^is. b) Regular, but illo. The tops of tall letters have a prefix. Cf. no. 58. Cf. Exempla. pi. XXXII. 94. Madrid A 115 (now 112). saec. xi (a. 1063?) a) Distinction made: negoq)is. b) Not regular: in often with short L Sign of decay of script. The tall letters have a prefix occasionally, as a rule they thicken at the top in the form of a triangle. Cf. Exempla, pi. XXXlll whence Arndt-Tangl*, pi. 8d. 95. Madrid A 2 (now 2). saec. xi ut vid. (D. DeBruyne's photo.) a) Distinction made. Cf. Berger, Hist, de la Vulg., p. 20. Studia palaeogjraphica. '^ 96. *Paris Nouv. Acq. Lat. 2171. ante a. 1067. a) Distinction made: lus^iogam, forogores. b) Regular, but ilium. Cf. Delisle, Melanges, p. 68: "premiere moitie du xi^siecle"; Ferotin, Le liber ordinum, p. xni. 97. Leon Eccl. Cathedr. 2. a. 1071. (Clark's photo.) a) Distinction made: iusf/q)am. b) Regular. Cf. Beer-Diaz Jimenez, p. 2. 98. '^Taris Nouv. Acq. Lat. 2169. completed a. 1072. a) Distinction made: raq)one l)ut imitit ques^io. b) Regular. The tops of tall letters have a prefix. Cf. no. 58. Cf Delisle, Melanges, p. 107 ; Ferotin, Le liber ordinum, p. XXXIII. 99. *London Add. MS 30848. saec. xi ut vid. a) Distinction made. b) Regular: Ilia and illuc. Cf. Cat. Add. (1876-1881) p. 120. 100. Madrid Acad, de la Hist. F 192. Hartel-Loewe, no. 29. a. 1073. a) Distinction made: lecq)o but nocfis. b) Regular, but ilia. The shafts of the tall letters have a prefix. Cf. no. 58. Cf. Exempla, pi. XXXV; N. A. VI, 332. 101. Madrid R 216 (now 6367). a. 1105. a) Distinction made: fornicaq)onem. b) Regular, but ilia. Cf. Exempla, pi. XXXVIII. 102. *London Add. MS 11695. a. 1109 (or 1091).^) a) Distinction made: condiogone but consfituta. b) Regular: Ipsius, even Ilia. The tops of tall letters have a prefix. Cf. no. 58. Cf. Delisle, Melanges, p. 60; Facs. Pal. Soc, pi. 48, 49; Arndt-TangP, pi. 37; Facs. de Fecole des chartes. no. 353. Colored facs. in Westwood's Pal. Sacra Pict. 1) The subscription which gives us the date is not quite clear. Cf. Prou, Manuel de Paleogr.3 (1910) p. 101, note 4. 76 12. Abhandlung: K. A. Loew Studia palaeographica. 77 103. Madrid Archiv. Hist. Nacion. 989-B. Vitrina 40. a. 1 1 10. a) palado: ci is used for assibilated ti. The spelling on the whole is that of an ignorant notary. Cf. Facs. in Mufioz, pi. XIV (where no press mark is given). 104. *Rome Corsinian. 369 (formerly 40 E 6). saec. xii. a) Distinction made in Visigothic portion: cogniogo, persecua}onis. The non-Yisigothic hand often writes ci for assibilated ti. In Visigothic script are ff. 144 — 156 and additions on f. 106.^) The rest of the MS is in ordinary minuscule by contemporaneous hand. This is the sixth example known to me of a Spanish MS in Italy. It has been correctly described by Zacarias Garcia: Un nuevo manuscritto del comentario sobre el apocalipsis de San Beato de Liebana, in Razon y Fe XII (August 1905) p. 478-493. The MS is palaeographically very instructive. The Visigothic script in it is impure, showing a mixture of ancient and foreign elements, especially in the abbreviations. The tops of tall letters as in other recent MSS have a prefix. Cf. pi. 7. The above evidence is instructively supplemented by a consideration of the following corrections and additions, and by the testimony of notarial documents. In Escorial T II 24 (formerly Q II 24) on line 6 of folio 73 (cf. Exempla, pi. VIII) the scribe originally wrote qiiesitio with the assibilated form of ti. The word however should have been qnestio. The corrector who crossed out the superfluous i also changed the form of the second i. 1) The additions it seems escaped the notice of Garcia. As they occur in the non-Visigotliic portion of the MS they furnish further evid- ence for his contention that the whole MS was written in Spain. \\ One of the scribes of Madrid Tolet. 10. 25, a. 902 does not make the ^i-distinction. In this part of the MS the activity of the corrector is plainly noticeable: he adds the tail to the i where ti has the soft sound. The scribe or scribes of Manchester John Uylands Library MS 93 make no distinction, but contemporary additions have it (f. 58, 292) and a later corrector changes the ordinary form of ti to o) where it is assibilated, e. g. on f. 129. The MS Madrid Acad, de la Hist. F 186 shows a wavering in the matter of the ^^-distinction. The marginalia, which seem to me by a later hand, invariably observe it. The same indecision with regard to the ti-usage is found in Madrid Tolet. 10. 25. The later entry on f. 147^ makes the distinction. The documents which I have been able to study in the facsimiles of Munoz furnish data which may fairly be regarded as confirming the evidence of the MSS.^) In a document of 857 (Mufioz, pi. 16) § is used for assi- bilated ti, but not o).^) In a document of 898 — 929 (Munoz, pi. 17) no distinction is made, ci doing service for assibilated ti. But in a document of 904 (Munoz, pi. 18) we have the distinction: preq:)o but dedisfis. It is needless to enumerate the later documents. As a rule the distinction is made as in MSS. Occasionally it happens that o) is used indiscriminately (cf. Munoz, pi. 22 and 41). In the more recent documents ci is used for assibilated ti. Yet in a document of 1137 (Munoz, pi. 42) the two forms of ti are still strictly differentiated: uendiq}ones but tih\. The earliest examples of Visigothic cursive show no a'-distinction, as we learn from the cursive pages of Autun 27 (cf. p. 52, n. 1). There is likewise no distinction in the Escorialensis of Augustine (Camarin de las reliquias) in the cursive part containing the Benedictio cerei. But this writing, as Traube has pointed out (Nomina Sacra, p. 191, note 1), must not be regarded as Spanish. 2) In the cursive portion of Escor. R II 18 (ante a. 779) assibilated ii is regularly represented by ^. The same is true of the additions in cursive found in many MSS posterior to the 8t^» century. 78 12. Abhandlungr: E. A. Loew Studia palaeographica. 79 A study of the usage illustrated by the foregoing data gives us the folloAving facts with regard to f^-forms in Visi- gothic MSS. 1. The distinction is never found in MSS which are in- disputably of the 8*^*^ or early 9^'* century. 2. The distinction is invariably made in the more recent MSS, beginning (to use the safest limits) with the second half of the 10'*' century and extending to the 12^'S i. e. as long as the script lasts. 3. Certain MSS, written between the two periods indicated show a wavering in usage, one scribe making the distinction and another not; or one scribe making it in some cases and not in others. There can be but one interpretation of these facts. The custom of making the ^i-distinction in book-script was con- sciously introduced. This graj)liic innovation, which on the face of it has something formal and conventional (since the ligature § which did service for assibilated ti in cursive was rejected as unsuitable in book-hand), was in all probability in- troduced in connection with liturgical books, where a need was felt of facilitating the reading aloud. The form o) was to tell the reader at once that he should give the soft sound of t. As such scribal changes, however, are adopted slowly, and reach some schools much sooner than others, it need not surprise us that scribes of one school should continue in the old way long after those of another had adopted the new one. The absence of the ^i-distinction may therefore say less to us than its presence. Its presence is at once a hint that the MS is not of the oldest kind. But there are MSS in which one scribe makes the distinction and another does not.^) These are manifestly MSS of the transition period, in which the struggle between the old and the new can l)e witnessed, the younger scribe adopting the innovation, the older persisting in his old-fashioned way as he had been taught. The fact that these MSS were written, as the dated as well as the undated MSS show, precisely in the interval between two periods the first of which displays the invariable absence, the second the invariable presence of the distinction, is the best possible proof that the custom of making the distinction was then in the actual process of adoption by the various schools of Spain. The question as to which centre was first to practice the distinction and which were the centres more backward about doing so must be left for further investigation. What are the more precise limits of the transition period? The earliest dated example known to me of a MS with the ^i-distinction is Thompsonianus 97, written, according to a subscription, in the year 894. As the form of the letters cor- responds to that of other dated MSS of the same time, there is no reason for questioning the originality of the subscription. The latest dated example know^n to me of a MS in which the scribe shows insecurity in his usage is of the year 945.^) As several dated MSS which fall between 894 and 945 show the /i-distinction (at least by one hand), it is fair to consider these two dates as the extreme limits of the transition period. From all this it must follow that a MS w^ithout the dis- tinction is in all probability older than 894 (as many MSS of the type of Thompsonianus 97 still ignore the distinction); that on the other hand a MS with the ^^-distinction is hardly older than 894, and in most cases much younger. The MSS which may be pointed out as disputing the cri- terion just formulated are, I believe, so few in number that they could fairly be regarded as mere exceptions to a rule. But such MSS remain exceptions only if we accept their ^) Cf. in my list the numbers 42, 45. 4G, 5o, 54 and GO. ^) Cf. no. 60 of list. It is only fair to note that this statement is based on a facsimile of Munoz which is less trustworthy than a photo- graph. The photographs which I had of this MS showed the distinction regularly. I 80 12. Abhandlung: E. A. Loew traditional dates.^) If we can sliow those dates to be un- tenable or improbable on |)alaeograpliical grounds the validity of the ^i-criterion will thus at once be both tested and con- firmed. This T shall attempt to do. I preface my argument with a few remarks on the script as such. Briefly, we may distinguish four stages of develo|)ment: a) The first stage is exemplified in the oldest MSS, saec. viii— IX. The script has striking compactness. The pen-stroke is not fine. The shaftless letters are rather broad, the arcs of m, n and h are low; their last stroke turns in. The separation of words is imperfect. The })oint of interrogation is usually a later addition. The suspensions Ims and que are generally denoted by a semi-colon placed above h and q (cf. pi. 8). b) The second stage is illustrated by the MSS of the end of the 9*^ and the beginning of the lO**' century. The script is looser and larger; the shafts of tall letters are club- shaped; the shaftless letters have more height than breadth; the final stroke of m, w, h often turns out. The separation of words is more distinct; the interrogation point is used. The suspensions hus and que are represented now by means of the semi-colon, now by means of an s-like flourish (cf. pi. 4). c) The third stage is seen in MSS of the 10*'' and 1 1*'> centuries. The letters are better spaced; the pen-stroke is often fine. The body of the letters is rather tall and narrow. The final stroke of m, w, h etc. regularly turns out. Particu- larly characteristic are the shafts of tall letters, which end in a little hook or mallet-head. The suspensions hus and que are denoted by an s-like flourish placed above h and (/, i. e. the semi-colon of the first stage is here made in one convention- alized stroke (cf. plates 5 and 6). 1) Although with great hesitation, I have ventured to disagree with the date given by Delisle in the case of nos. 35 and 37 of my list. If his dates are correct, I should be at a loss to explain the // usage in these MSS. Studia palaeographica. 81 d) The last stage of the script is characterized by the decay and awkwardness of the old forms and the employment of foreign elements (cf. plate 7). We are now in a position to test the fi-criterion. I select first the most important exception. The MS Escor. T II 24 (formerly QII24)^) containing the Etymologies of Isidore has long enjoyed the distinction of being the oldest dated MS in the script (see plate 5). The traditional date is 733 or 743. A computal note in the text (f. 68) says: ^'usque in hanc pre- sentem erani que est BCCLXXl" which is the year 733. A few lines below occurs: '^iisque in hanc praefatam BCCLXXXI crani^ which is the year 743. One of these dates is plainly wrong. From the calculation in the text it appears that 743 is the correct year. In the judgment of Eguren, Munoz y Rivero, Ewald and Loewe, Beer and Stefi'ens, not to mention older authorities, the script did not seem to belie the date established by the computal note. Stefi'ens gives 743 as the date of his facsimile, but he is cautious enough to add: "unter der Voraussetzung, dais jene Eintragung ein Original ist und nicht etwa eine Abschrift aus einem anderen Codex". R. Beer, in his learned Praefatio to the reproduction of the Toletanus 15. 8 compared that MS with Escor. T II 24, thus trying to determine the age of the undated MS by the aid of the pre- sumably dated one. He says of our MS: "litterae sunt ali- quanto altiores ductusque magis tenues", thus pointing out essential difl'erences. But when he continues and says ''sed utriusque libri scriptura, ut ex Exempl. Scr. Visig. tab. VIII et ex tab. 17 supplementi Steff'ensiani perspicere licet, in uni- versum non est dispar", he seems to me to be withdrawing his earlier judgment just quoted. It is also plain that a certain calligraphic difference escaped Beer's notice: one MS uses only one form for ti, the other two distinct forms. But indeed a careful examination of the script of the Escorialensis will disclose other traits foreign to the oldest type of Visigothic ^) For literature see no. 52 of the list. Sitzgsb.d. pliilos.-philol. u. d. bist.Kl. Jahrg. 1910, 12. Abh. 82 12. Abhandlunsr: E. A. Loew Studia pulaeo<]fraphica. 83 writing. Foremost is the general impression already noted by Beer: the proportions of the letters, their relation to one an- other. It is plainly not the old, compact, broadly-flowing writing. In the oldest MSS the m and 7i and the arch of h all turn in. In the Escorialensis and the more recent MSS these strokes thicken at the end and turn out. In the older type the letter g has often a rather short and curved down- stroke, in the Escorialensis and the more recent type of MSS it is very long. But the unfailing ear-mark of the recent type is the hook or mallet-shaped end of the shafts of ft, d, h, i-lon(ja and L which is unknown in the oldest MSS. The Escorialensis has such shafts.^) The abbreviation sign over h and q for hus and que has the form of an uncial s as in the more recent type of MSS (cf. plate 5). — In short, purely graphic considerations are against the traditional date of 743. I may state my conviction that the computal note is merely a copied one, and that Escor. T II 24 may be fairly held to confirm the value of tl as a criterion for dating. The MS Escor. a I 13^) furnishes an excellent instance of the caution with which the inscriptions and subscriptions of Spanish MSS must be used.^) According to a note in cursive on f. 186^ the MS was written "regnante adefonso principe in era DCCCCL" i. e. in 912. Ewald has pointed out that in 912 there was no reigning Alphonse, as Alphonse III had died in 910. By assuming that the scribe inserted a superfluous C he gets era DCCCL corresponding to 812, which agrees with the reign of Alphonse II (795 — 843) and thus 812 was (pre- sumably) the date of the MS. Munoz has 912. The des- cription in the Exempla is "fortasse 812'', the reservation being ^) More precisely one of the scribes of this MS whose writing is seen in our plate. The facsimile in the Exempla shows another hand which does not make this type of shaft. 2) For literature see no. 50 of list. 3) Other examples are not wanting. Of. nos. 3.S, 34, 45, 52, 84 and 102 of list. doubtless a concession by Ewald to Loewe. For according to the latter's notes as edited by Hartel the date of the MS was 912 and not 812. Ewald's explanation did not seem thor- oughly convincing to Traube. But Beer's date is 812. In connection with one of the Codices Ovetenses mentioned in the inventory of 882 he notes: "es ist zweifellos der heutige Escorialensis a I 13 'de la yglesia de Oviedo' (vgl. Hartel-Loewe, p. 10 ff.), dessen Beschreibung in alien wesentlichen Stiicken mit der vorliegenden uberein- stimmt. Durch diese Identifikation wird auch die Datierung (des ersten Teiles des Codex) 812 (Jahr der Alphonsischen Schenkung, nicht 912) gestutzt". But can not the Escorialensis be a copy of a MS which Avas presented in 812 and catalogued in 882? Whereas against this early date is the script of the MS, which is not of the old type. The letters are somewhat irregular and awkward, which lends the script an appearance of antiquity. The shafts of tall letters thicken at the end. The upright strokes of m and n thicken below and turn out. The abbreviation sign over h and q is an s-like flourish. Judged by purely graphic standards the MS should belong at the beginning of the 10^^ century. As for the subscription the very nature of the error in it hints that it was copied from an original having DCCCL. The scribe unconsciously inserted the extra C because he was accustomed to writing DCCCC — a type of mistake we commit every January. Thus though the year 912 need not be the exact date when the MS was copied, it is more than likely that it was Avritten after era 900, which would fully account for the presence of the ^i-distinction, not found in the MSS of the beginning of the 9^^ century. The MS Madrid Tolet. 43. 5 1) shows a cruder and less calligraphic type of writing than the MS just considered and that perhaps lends it an impression of antiquity. But it lacks ij Cf. no. 44 of list. 84 12. Ahlinndlunjx: E. A. liOow all resemblance to the earliest kind of Visi 30847 72 — Archiv. Hist. 989-B 76 30848 75 — Royal Private Library — 30850 72 2 J 5 74 — — . — 30851 72 — Univ. 31 61 ___ — - 30852 63 — — 32 62 __ 30854 64 Manchester John Rylands MS ._ _— . 30855 74 lat. 93 67, 77 — — 31031 34 ManQhesterJohn Rylands MS 99 70 — Thompsonianus 8 250 — — — — 116 62 97 65 ,79 Milan Ambros. Josephus (pa- Lucca 490 30, 44 pyrus) 11, 39 Lyon 523 32, 20 , 30 Milan Ambros. B 31 sup. 40 C 98 inf. 28, 39 Madrid Acad. His b. 20 (F 186) — C 105 inf. 39 65 , 77, 78,, 82. , 84 - D 268 inf. • 39 — — Hist. 24 (F 188) ; — — L 99 sup. 40 68 . 78, — 210 sup. 4I2 Hist. 25 (F 194) 70 — — S 45 sup. 30,40, — — — F211 74 Milan Trivulziana 688 44 F212 71 Modena I N 11 44 — — F192 75 Monte Cassino 4 58, 86 Madrid (Bibl . Nacion.) A 2 i — — 5 46 (now 2) 74 19 59 — A 115 (now 112) 74 — 150 32 — P 21 (now 1872) 68 — — 187 47 — R 216 (now 6367) 75 289 15 90 12. Abhamllwnij: E. A. liOew Stadia palaeo^'niphica. 91 Monte Cassino 295 — — 303 — — 332 — — 753 Munich (Hof- iind Staatsbil.l.) — hit. 4542 — - 4547 — — 4549 — — 4564 — — 4623 — — 4719'" — — 6277 — — 6402 — — 6437 — — 14102 — — 14421 — — 29033 — — 29158 Naples (Bibl. Naz.) JV A 8 - — VI B 12 46 15 15 46 48 47 48 48 10 48 48 48 13 32 48 32 33, 39 46 No vara Capitol. 84 Oxford IJodl. Canon. Class. lat. 41 — — Douce f. 1 — — - 140 — — Laud. 108 ¥i\Y\?^ (Bibl. Nations le) 30, 44 15 36 50 50, lat. 653 5„ 43, 86 — 1732 -io 2855 70 — 299 1 59 3836 37 4667 59 750& 178 7530 263, 46 8093 59 8913 30 8921 24,. 37 9427 31 and 11. 1 10837 50 — 10^76 9,, 64 Paris lat. 10877 9i, 64 — — 11529 38 — 11530 38 — — 11627 37 1 — 116S1 37 — 12097 20 — 12134 37 ~ 12135 38 — 12155 38 — 12168 24„ 33 — — 12217 38 — 12254 60 - 12 598 35 — 13048 38 ~ 13246 17,, 30 — 13440 38 — — 14086 35 — — 17451 38 - 17 655 31 Noiiv. Acq. lat. 235 73 — — — 238 (;i, 68 — — ~ 239 70 — — - 260 64, 80, — — — 1296 72 — — — 1298 64, 80, — — - 1628 38, 47, 53, 57 "— — — — 1629 47, 53, 57, - — — 2167 64 — — _ — 2168 62 — — _ _ 2169 75 — — — — 2170 63, 71 - — 2171 — — 2176 73 — ~ 2177 73 2178 73 2179 72 — — — _ 2180 72 -- r.alu/A' 270 44 Rome Basilicanus D 182 3-2 Casanat. 641 1 252 Rome Corsinian. 369 76, 87 Home Sessor. 40 (Vittorio Emanuele 1258) 45 Rome Sessor. 41 (V. E. 1479) 45 — — 55(V.E.2099) 30,44 — — 63(V.E.2102) — — 66 (V. E. 2098) — — 94 (V. E. 1524) — — 96(V. E. 1565) Rome Vallicell. D 5 Rome Vatie. lat. 491 — — — 595 — - — 3320 — — — 3342 — — — 3375 — — — 3973 ~ — — 5007 — — — 5763 Borgian Pal. lat — Vatic. — Reirin. 45 45 45 45 46 50 46 46 25, 32 46 30 40 46 50 50, 50 36 St. Gall 44 70 185 214 238 348 722 731 914 lat. 339 68 202 235 lat. 316 — 317 30, 32, — 1024 30 49 49 49 34, 49, 49 49 30, 49 49 St. Paul in Carinthia XXV * Sigiienza Decretale 150 Toledo Capitol. 99. Tours 615 30 49 31 61 62 64 Troussures Nov. Test. 34 ! Turin (Bibl. Nazion.) **A II 2 40 - - D V 3 37 I — - F IV 1 fasc. 6 50 ! - — G V 26 40 — G VII 15 15 Vercelli Capitol. CXLVIII 44 — - CLVIII 432 — - CLXXXIIT 12, 27, 28, 43, 86 — — CLXXXVIII 30 — — ecu 43 Verona Capitol. 1 40, 41 — — 11 41 — ~ III 41 — IV 41 — — XV 42 — -- XXXIll 41 — - XXXVIl 42 — XXXVIII 42 — XL 31 — XLII 41 — — LV 42 — — LXI 42 — — LXII 42 — — LXXXIX 153, 47, 51 2, 53, 56 — — CLXIII 42 Vienna (Hof-Bibl.) lat. 16 50 — lat. 17 39 — - 847 33 — — 1616 35 Wolfenbiittel Weissenb. 64 40 — — 99 32 Wiii-zburg Mp. Theol. Fol. 64* 33 Ziirirh Cantonsbibl. CXL 49 — — (Rheinau)30 49 i I m to a a m (U <^ W X X o 01 (U (L> I X o CO CO aj to 03 a. -3 w i- ti^. i^. fO fl^ a a to :. . &rtf>&t'