HANDBOOK OF GREEK & LATIN PALAEOGRAPHY by Sir E M Thompson. 3rd edition, published 1906. Kegan, Paul, Trench, Trübner & Co Ltd, London.

CHAPTER V. Forms of Books.

| HOME | contents | << | the roll | the codex or book | gatherings or quires | ruling | arrangement of text | punctuation: greek ; latin | breathings & accents & other signs: greek ; latin | palimpsests | systems of dating | additions & corrections | >> |

The Roll.

Among the Greeks the ordinary terms for a book (that is, a roll) were βίβλος and its diminutive βίβλίον. βιβλίον also meant a letter, and is used in this sense by Herodotus. Saidas in his Lexicon explains βιβλίον as ἐπιστολή. Earlier forms of these words were βύβλος and, more rarely, βυβλίον, which were clearly derived from the material, the βύβλος or papyrus, of which books were made. the corresponding word liber of the Latin people, in like manner, was adopted as a term for a book, primitively made of the bark or inner rind of the lime or other tree. Such bark-books, however, disappeared in presence of the more convenient and more plentiful papyrus imported from Egypt; but the old name was not unfitly transferred to a book made of the new substance, which in texture and general appear-ance was not unlike the old. ← For instances of confusion of material, see Wattenbach, Schriftw. 89.

A diminutive of the word liber was libellus, which, as a literary title, specially referred to a book of poems, a sense in which it is constantly used by the Roman poets, It came at lengih to be used as an equivalent of liber, and to express a book in general.

The old form of a book was the roll, the Latin volumen. The Greeks do not appear to have had any parallel expression at an early date; the word κύλινδρος being comparatively late. Another term was ένείλημα or ἐξείλημα ; more rare. were εἰλητάριον, εἴλητον. A mediæval Latin term is rotulus.

Again, a later Greek term was τόμος (originally a cutting of papyrus), applicable to a roll containing a portion of a collection or of a great work. Neither this term nor βιβλίον, nor liber nor libellus, could be applied in the singular number to more than a single roll or volume. A work consisting of many volumes, or several divisions, must be described by the plural forms βιβλία, τόμοι, libri, etc. On the other hand, the several books of a work, if written on one roll, counted only for one βιβλίον or liber. Thus Ulpian, Dige t. xxxii. 52, lays down : " Si cui centum libri sint legati, centum volumina ei dabimus. non centum quae quis ingenio suo metitus est. ... ut puta, cum haberet Homerum totum in uno volumine, non quadraginta octo libros computamos, sed unum Homeri volumen pro libro accipiendum est."

For subdivisions such terms as λόγος, σύγγραμμα, σύνταγμα also were used.

The word τεῦχος, in the sense of a literarv work in several volumes, was employed at a late period. Originally it seems to have been applied to the chest or vessel in which the several rolls of such work were kept, and came in course of time to refer to the contents. ← Birt, Ant. Buchw. 89. Xenophon, Anab. vii. 6, 14, mentions books ἐv ξυλἰνοις τεύχεσι. In like manner the terms ρandectes and bibliotheca, originally referring to a work in several rolls kept toget her in their chest, were afterwards used specially to mean a MS. of the entire Bible. Bibliotheca was used in this sense by St. Jerome. Others, as Cassiodorus, Bede, Alcuin, preferred Pandectes. Bibliotheca continued to bear this meaning down to the close of the fourteenth Century, if not later. ← See examples in Wattenbach, Schriftw. 126-129.

To distinguish a work contained in the compass of a single roll, there was the title μονόβιβλος or μονόβιβλον.

There can be no doubt that the convenience of sub-dividing the lengthy works of authors into rolls of moderate size must have been appreciated in the earliest period of the publication of Greek literature ; and, although the authors themselves may not originally have divided their writings into separate portions to suit the ordinary length of a conveniently-sized roll, yet the practice of the scribe would eventually react on the author. Thus we find the works of Homer divided into books of a length which could be contained in an ordinary roll ; and we know that in course of time authors did regularly adapt the divisions of their works to the customary length of the βιβλία and volumina.

The roll was rolled on a stick, ὀμφαλός or umbilicus, to which the last sheet of the papyrus, ἐσχατοκόλλιον, was attached. Many of the rolls found at Herculaueum had a mere central core of papyrus. A knob or button, usually of bone or wood, was affixed to each end of the stick, the name of which, ὀμφαλός, umbilicus, appears to have been also extended to these ornamental additions. Porphyrion, commenting on Horace, Epod. xiv. 8, says : " in fine libri umbilici ex ligno aut osse solent poni." Or, instead of the simple knob or button, there was a tip, κέρας, cornu, of ivory or some such ornamental material; and either might be plain or coloured. ← Tibullus, III. i. 13: " Atque inter geminas pingantur cornua frontes." Martial, iii. 2, 9, " picti umbilici " ; v. 6, 15, " nigri umbilici."
The edges, frontes, of the roll were cut down and smoothed with pumice, ← Ovid, Trist. I. i. 11, " Nec fragili geminæ poliantur pumice frontes " ; Catullus, xxii. 8, " pumice omnia æquata." and sometimes coloured. The wrapper of an ordinary roll might be of common papyrus, charta emporetica ; in case of a more valuable work, a vellum cover, stained with colour, ← See above, p. 39. was used as a protection—the φαινόλης or φαιλόνης, pænula (the travelling cloak), as it was commonly called. ← The " cloak (φαιλόνης) which St. Paul left at Troas (2 Tim. iv. 13), and which Timothy was to bring together with the books and parchments, may have been in fact a book-cover. See Birt, 65. Lucian, Adv. indoctum, 7, refers to an ornamental work thus: " ὁπόταν τὸ μὲν βιβλίον ἐv τῇ χειρὶ ἔχῃς πάγκαλον, πορφυρᾶν μὲν ἔχον τὴν διφθέραν, χρυσοῦν δὲ τὸν ὀμφάλόν " ; and Martial, i. 66, has the lines:—

" Sed pumicata fronte si quis est nondum
Nec umbilicis cultus atque membrana,
Mercare : tales habeo."

For preservation against moths, etc., cedar oil was rubbed on the papyrus. ← "Ex cedro oleum, quod cedreum dicitur, nascitur, quo reliquæ res unctæ, uti etiam libri, a tineis et carie non læduntur."— Vitruvius, ii. 9, 13. A good poem was worthy of this protection : " cedro digna locutus " (Persius, i. 42) ; "cedro nunc licet ambules perunctus" (Martial, iii. 2, 7). But it imparted a yellow tint : " quod neque sum cedro flavus " (Ovid, Trist. III. i. 13).

The chest or box in which the rolls were kept was the κίστη, κιβωτός, capsa, cista, forulus, nidus, puteus, or scrinium. To tie bundles of rolls together was a destructive process, as the papyrus was injured ; so Petronius, Satyricon, cii. : " Chartæ alligatæ mutant figuram." Extensive works were arranged in their capsæ in decades, triads, or other sets, as we know from the examples of the works of Livy, Dio Cassias, Yarro, and others.

For convenience of reference when the roll was placed in a box or on a shelf, a vellum label, σίλλυβος or σίττυβος, ← Marquardt, Privatl. der Römer, 794. πιττάκιον, also γλῶσσα, γλωσσάριον, titulus, index, was attached to the edge of the roll and inscribed with the title of the work, ← See an engraving, copied from a sculpture, in Schwarz, De ornamentis librorum (1756), tab. ii., wherein are represented series of rolls placed on shelves, like bottles in a wine-bin, with the tituti depending in front ; also an engraving of a capsa, with rolls enclosed, on the title-page of Marini, Papiri Diplom. ; and Museo Borbonico, tav. xii. and, for distinction, was also coloured. ← See above, p. 39. Such tituli are perhaps the "lora rubra" of Catullus, xxii. 7. Cicero, writing to Atticus, iv. 4, gives both Greek and Latin names : " Etiam velim mihi mittas de tuis librariolis duos aliquos, quibus Tyrannio utatur glulinatoribus, ad cetera adininistris, iisque imperes ut sumant membranulam, ex qua indices fiant, quos vos Græci, ut opinor, σιλλύβους appellatis." And the lines of Tibullus, III. i. 9, may be quoted as describing the outward appearance of the roll:—

" Lutea sed niveum involvat membrana libellum,
Pumex cui canas tondeat ante comas ;
Summaqne prætexat tenuis fastigia chartæ,
Indicet ut nomen, littera facta, puer."

The text was written in columns, σελίδες, paginæ. The term σελίς (originally the gangway between the rowing benches of a ship) was first applied to the space between two columns, and then to the column itself. Other terms were the diminutive σελίδιον and καταβατόν. The lines of the columns ran parallel with the length of the roll ; ← Before the time of Julius Cæsar, official despatches appear to have been written "transversâ chartâ," that is, with the lines parallel with the breadth of the roll. Suetonius, Jul. Cæs. 56. and lead was used for drawing the ruled lines. Such ruling, however, was not always, and perhaps not generally, employed, for the horizontal fibre of the papyrus itself was a sufficient guide for the lines of writing; and the fact that the marginal line of the columns frequently trends away out of the perpendicular proves that in such instances there were no ruled lines to bound the columns laterally. These were generally narrow, at least in the texts which were written by skilled scribes for the market ; and occasionally we find the letters made smaller at the end of a line in order to accommodate words to the available space. An example of writing in wide columns is seen in the papyrus of Aristotle on the Constitution of Athens—a MS. which was written for private use and not for sale.

The title of the work was written at the end.

The reader unrolled the book with the right hand ; with the left hand he rolled up what he had read. ← See an engraving, from a sculptured sarcophagus, in Daremberg and Saglio's Dict. des Antiquites, s.v. "Bibliotheca," in wbish a man is represented reading from an open roll. To unroll a book was ἐξειλεῖν, ἀveiλeῖv, ἀveλισσειν, ἐλίσσειν, εἴλειν or εἰλεῖν, evolvere, revolvere, volvore, explicare. The book read to the end was " explicitus usque ad sua cornua" (Martial, xi. 107), or "ad umbilicum," as in Horace, Epod. xiv. 8 :—

" Deus nam me vetat
Inceptos, olim promissum carmen, iambos
Ad umbilicum adducere ;"

and in Martial, iv. 89 :—

" Ohe, jam satis est, ohe libelle,
Jam pervenimus usque ad umbilioos."

From the term " explicitus " came the mediæval " explicit," formed, no doubt, as a pendant to " incipit." The term to roll up a book was plicare. The beginning of the roll was held under the chin while the hands were employed in turning the umbilici. Hence Martial, i. 66, refers to " virginis .... chartæ, quæ trita duro non inhorruit mento" ; and again, x. 93, he has: " Sic nova nec mento sordida charta juvat."

The inconvenience of writing on the back of the roll is obvious, and this practice was probably very seldom, if ever, followed in the case of works intended for sale. Authors' copies, however, were often opisthograph, as in Juvenal, Sat. i. 4:—

" Impane diem consumpsevit ingens
Telephus, aut summi plena jam margine libri
Scriptus et in tergo necdum finitus Orestes ?"

The younger Pliny also, Epist. iii. 5, 17, in reference to his uncle's numerous works, uses the words : " Commentarios clx. mihi reliquit, opisthographos quidem et minutissime scriptos."

In the same manner worthless scribbling is referred to by Martial, viii. 62, as written on the back of the charta :—

" Scribit in aversa Picens epigrammata charta
Et dolet, averso quod facit illa deo."

Rough draughts or temporary pieces, or children's or scholars' exercises might also be so written. Martial, iv. 86, threatens his libellus with the fate of waste paper to be utilized for such purposes, if his verses fail to please :—

" Si damnavevit, ad salariorum
Curras scrinia protinus licebit,
Inversa pueris arande charta."

A most important instance of a scholar's exercise, written on the back of a papyrus, is found in the early copy of. the Epitaphios of Hyperides in the British Museum.

After the establishment of the book-shape in general use, the roll form was almost entirely abandoned for literary purposes in the middle ages. It survived, however, for some of the Greek liturgies, for mortuary rolls, for pedigrees, for certain brief chronicles in which historical genealogies form a principal feature, and in a few other instances, as in the " Exultet " rolls of Italy, in which it was found convenient. But in all these the writing was parallel with the breadth, not with the length, of the roll. For records, however, the roll form has been continued throughout the middle ages to our own days, particularly in England, where not only public documents relating to the business of the country, but also proeeedings of private manorial courts and bailiffs' accounts, were almost invariably entered on rolls.

The Codex or Book.

The earliest form of the book, in our modern sense of the word, that is, as a collection of leaves of vellum, paper, or other material, bound together, existed, as we have seen, ← See above, p. 20. in the case of waxen tablets, when two or more were fastened together and made a caudex or codex. Hence vellum books, following the same arrangement, were also called codices. Similarly, by usage the title liber, which had been transferred from the original bark roll to the papyrus roll, was also passed on to the vellum book. So too the Greek terms βίβλος, βιβλίον and other words, which had been employed to designate the earlier rolls, were transferred in the same way. The vellum codex came into general use when it was found how conveniently it could contain a large work in a much smaller space than could the papyrus roll. In the words of Isidore, Origg. vi. 13, 1 : " Codex multorum librorum est, liber unius voluminis."

That vellum MSS. existed in the classical period at Rome we know from Martial's Apophoreta. But these must have been few in number and articles of luxury. It was the requirements of the lawyers which necessitated the casting of the great law-books into a convenient form for reference ; and the vellum MS., more durable than papyrus and adapted for receiving writing on both sides of the leaves, satisfied those requirements in the most perfect manner. Hence the term σωμάτιον, a name for the vellum MS., expressive of the bulk of the contents; and hence, conversely, the title of codex which was given to great compilations, such as those of Theodosius and Justinian.

Again, the Bible, the book which before all others became the great work of reference in the hands of the early Christians, could only be consulted with convenience and despatch in the new form. From the writings of St. Jerome and others it is evident that Bibles in codex form existed at a very early date. When once this form of multiplying texts was adopted by the Church, its rapid diffusion became a matter of certainty through the medium of monastic institutions. The form adopted for the Bible would naturally becomo the model for theological and ecclesiastical books of all kinds. Thus the vellum codex was destined to be the recipient of Christian literature, as the papyrus roll had been that of the pagan world.

Still, however, for the older literature the papyrus continued to some extent to hold its ground; ← Birt, 100. although even in this department the codex began at once to make inroads. For, as regards the works of great standard authors, such as Homer in Greek and Cicero in Latin, there is evidence that even in the earliest centuries of our era the codex form was not uncommon. Ibid. 113. In St. Jerome's days vellum MSS. of the classics appear to have been in ordinary use, for his library of vellum codices included works of profane literature. Ibid. 115. In the end. the book form became so general that even papyrus was put together in leaves and quires in the same way as vellum. Several specimens of such papyrus books still exist, as has been already noticed. ← Above. p. 34.

Gatherings or Quires.

Codex Alexandrinus

The earliest MSS. on vellum are usually of the broad quarto size, in which the width equals, or nearly equals, the height. The quires of which they are composed consist, in most instances, of eight leaves, that is, of four folded sheets, τετράς or τετράδιον, quaternio (whence our word quire ), and this number continued in general favour for all sizes of volumes throughout the middle ages. Quires of three sheets or six leaves, of five sheets or ten leaves, and of six sheets or twelve leaves, are also met with. For example, the famous Codex Vaticanus of the Greek Bible is made up of ten-leaved quires. Each quire was actually numbered or signed, to use the technical word, either at the beginning, in the upper margin, or, more generally at the end, in the lower inner corner. In the Codex Alexandrinus the signatures are at the beginnings of the quires, in the centre of the upper margin. The numbers were frequently, in Latin MSS., accompanied with the letter Q (for quaternio ). The practice of numbering the leaves of the quires, e.g. A. i., A. ii., A. iii., etc., dates from the fourteenth Century. Catch-words, exclamantes, to connect the quires together, first appear, but rarely, in. the eleventh Century; from the twelfth Century they become common.

In putting together the sheets for the quire, care was generally taken to lay them in such a way that hair-side faced hair-side, and flesh- (or inner) side faced flesh-side. Thus, when the book was opened, the two pages before th reader had the same appearance, either the yellow tinge of the hair-side or the whiter surface of the flesh-side. In Greek MSS. the arrangement of the sheets was afterwards reduced to a system : the first or lowest sheet being laid with the flesh-side downwards so that when the sheets were folded that side always formed the first page of the quire. ← C. R. Gregory, Les Cahiers des MSS. Grecs, in the Comptes Rendus of the Acad. des Inscriptions, 1885, p. 261.
In the Codex Alexandrinus, however, the first page of a quire is the hair-side of the skin, In Latin MSS. also the hair'-side appears to have generally begun the quire.

To the folded sheet was given the title diploma ; a barbarous mediæval name for it was arcus. ← Wattenbach, Schriftw. 153. The leaf was χapτίov, φύλλον, folium. The line of writing was στίχος, versus, linea, and riga.


In the earlier centuries of the middle ages, the ruled lines of vellum MSS. were drawn with a hard-pointed instrument, a blunt bodkin or stilus, on one side of the leaf, the lines being impressed with sufficient force to cause them to stand out in relief on the other side. The ruling was almost in variably on the hair- (or outer) side of the skin. Marginal lines were drawn to bound the text laterally. The distances of the horizontal lines from one another were marked off with pricks of the compass in vertical order down the page. In earlier MSS. these prickings are often found near the middle of the leaf, or at least within the space occupied by the text, and the lines are drawn right across the sheet and not confined within the vertical boundaries. It was afterwards the custom to prick off the spaces close to the margin and to keep the ruled lines within limits; and eventually the prickings often disappeared when the edges were shorn by the binder. Each sheet should be ruled separately ; but two or more sheets were not infrequently laid and ruled together, the lines being so deeply drawn on the upper sheet that the lower sheets also received the impressions. In rare instances lines are found ruled on both sides of the leaf, as in some parts of the Codex Alexandrinus. In this MS. also, and in some other early codices, ruling was not drawn for every line of writing, but was occasionally spaced so tliafc some lines of the text lay in the spaces while others stood on the ruled lines. Ruling with the lead point or plummet came into ordinary use in the twelfth Century ; coloured ink was also used for ruled lines in the fifteenth Century.

Arrangement of the Text.

Codex Vaticanus

The text, which in early MSS. was written continuously without separation of words, might be written across the face of the page ; and in some cases, as in poetical works, no other arrangement could well be followed. But, continuing the system observed in the papyrus rolls, the arrangement in columns was usual. The superior convenience of the column over the long line is obvious, particularly when a small character was the type of writing, The number of columns in a page was ordinarily two ; but three and even four were also allowed. Codex Sinaiticus. The Codex Sinaiticus of the Greek bible has four columns in a page; so that the open book presents a series of eight columns to the reader, which, it has been observed, would forcibly recall the long row of paginæ of the papyrus roll. ← The phrase of Eusebius, Vita Const, iv. 37, " ἐν πολυτελῶς ἠσκημένοις τεύχεσι τρισσὰ καὶ τετρασσά, " probabiy refers to the number of columns. See Wittenbach, Schriftw. 119. the Codex Vaticanus has three columns in a page in the portion containing the Old Testament ; and other early MSS. or fragments of MSS. exhibit the same arrangement, e.g. the Vatican fragments of Sulhist, the Latin Pentateuch of Lyons, and others in the libraries of Rome, Milan, etc. ← See Wattenbach, Schriftw. 149. It may also be noted that the most aneient dated MS. in existence, the Syriac MS. of A.D. 411, contaiining the Recognitions of Clement of Rome (Brit. Mus. Add. MS. 12,150), is written in triple columns. But the tri-columnar system appears to have been generally abandoned after the sixth century. The Utrecht Psalter, written at the beginning of the 9th Century, in triple columns, is not an instance which counts for late usage, the MS. being only an exact copy of an older codex. ← The later copies of this Psalter also maintain the same arrangement. Usually the later examples are the result of necessity, as in the case of Psalters in parallel versions or languages. ← A Psalter in four parallel columns (the Greek and. the three Latin versions), A.D. 1105, is in the Bibl. Nationale, MS. Lat. 2195, See Pal. Soc. i. 156. A late instance, however, of a text written in this fashion, without any compelling causes, occurs in the Latin Bible of the 9th Century, Add. MS. 24,142, in the British Museum.

With regard to the breaking up of the text into paragraphs, and more particularly into the short sentences known as στιχοί, the reader is referred to what is said below under the heads of Punctuation and Stichometry.

As already noticed, the text of early MSS. was generally written continuously without separation of the words ; and this practice continued as a rule down to about the ninth Century. But even when the scribes had begun to break up their lines into words, it still continued to be the fashion to attach short words, e.g. prepositions, to those which immediately followed them. It was hardly before the eleventh Century that a perfect system of separately-written words was established in Latin MSS. In Greek MSS. it may be said that the system was at no time perfectly followed, for, even when the words were distinguished, there was always a tendency to separate them inaccurately.

The first lines of the main divisions of the text, as for example the several books of the Bible, were often written in red for distinction.

In order to save space, and to get as much as possible into a line, or to avoid division of a word, the letters were often written smaller towards the end of the line ; and in Latin MSS., with the same object, two or more letters were linked or combined in a monogrammatic form.

At first, in uncial Latin MSS., there was no enlargement of letters in any part of the text to mark the beginnings of sections or chapters; yet, in some of the earliest examples, the first letter of the page, without regard to its position in relation to the text, is made larger than the rest.

Rubrics and titles and colophons (that is, titles, etc., written at the ends of books) were at first; written in the same charaeters as the text; afterwards it was found convenient, as a distinction, to employ different characters. Thus in later uncial Latin MSS. titles might be in capitals or rustic capitals; in minuscule MSS. they might be written in capitals or uncials. The convenience of having the title at the beginning of a MS., instead of only in colophon-form at the end, was soon recognized ; but the use of the colophon still continued, the designation of a work being frequently recorded in both title and colophon down to the latest period.

Running titles or head-lines appear in even some of the earliest MSS., in the same characters as the text, but of smaller size.

In the division of words at the end of a line, it was the ancient practice to break off with a complete syllable. In Greek, however, in the case of compound words, the last consonant of the prefix was carried on to the next syllable, if this was a vowel or began with a vowel, as κα-τεῖ-δον ; and the same method was ob-served with a preposition and the following word, as κα-τέ-μοῦ. With such a system in vogue it is not surprising to find it extended occasionally to other cases, as ταῦ-τοῦχ. In simple words the sigma was not uncommonly carried on to a following consonant, as μέγι-στος.

In Latin MSS., while the observance of the true syllabic division was maintained according to ancient usage, and, when two consonants came together, they were properly assigned to their several syllables, as dic-tus, prop-ter, ig-navus, pris-cus, hos-pes, hos-tis, yet in some ancient texts the first consonant is drawn over to the second, as di-ctas, ho-stis, etc., in accordance with the Greek practice noticed above; and in some MSS. we find the older style altered to suit the later, as in the Fulda MS. of the Gospels, corrected in the sixth century by Victor of Capua, ← Zangemeister and. Wattenbach, Exempla, Codd. Lat., tab. xxxiv. and the Harley Gospels of about the year 600. ← Brit. Mus. Cat. Anc. MSS., pt. ii. p. 14. The coupling stroke or hyphen, to indicate connection of the two parts of the divided word, appears to have been unknown in the early centuries. A point performs this duty in early instances. In the eleventh century the hyphen at the end of the line shows itself on a few occasions ; in the twelfth century it becomes more systematic, and is also repeated at the beginning of the next line.


The earliest form in which a system of punctuation appears is that found in ancient inscriptions, wherein the several words are divided from one another by single, double, or treble dots or points. This, however, is not punctuation in the sense in which we use the term—the system whereby sentences are marked out, and the sense of the text is made clear.

The ancient practice of writing literary texts continuously, without distinction of words, was not, indeed, quite universal ; for the astronomicai treatise known as the Ἐυδόξου τεχνή, earlier than 154 B.C., at Paris, is an instance to the contrary. But it was certainly by far the more ordinary method, and in the uncial vellum MSS. of the earlier middle ages it may be said to have been the only method that was followed. In the documents of ordinary life the distinction of words was, from early times, more frequently, though still only partially, observed. When the minuscule writing came into use us the literary hand, separation of the words from one another gradually followed ; but never was this system fully perfected. For example, prepositions were still attached to the following words, and there was always a tendency to detach a final letter, and to attach it to the next following word.

The inconvenience which we experience in reading a continuously written text could not have been so greatly felt by the scholars of the old Greek world ; otherwise separation of words, and a perfect system of punctuation, would have been established long before was actually the case. Still the distinction of paragraphs was found a necessity at an ancient period. Hence arose the dividing stroke, the παράγραφος, known, at all events, as early as Aristotle's time, separating paragraphs by being inserted between them at the beginnings of lines ; but, it should be remembered, the stroke really belonged to the concluding paragraph, and marked its termination, and did not form an initial sign for the new paragraph which followed. The paragraph-mark was not, however, uniformly the horizontal stroke ; the wedge > (διπλῆ), the mark which is also often found at the end of a work, 7 (κορωνίς), and similar forms were employed. This system of distinguishing paragraphs appears in use in the early papyri, and analogously the dividing stroke marks off the speeches of the different characters in the surviving papyrus fragments of the tragedians, as, for example, in the very ancient remains of the Antiope of Euripides.

But to write every paragraph distinct by itself would have entailed a certain loss of space. If the last line were short, there would remain a vacant space after it, unoccupied by writing. In the earliest specimens therefore we find this space occupied by the first words of the next paragraph, a slight break being left to mark :its commencement, t hus :—


The next step was to draw back the first letter of the first full line of the new paragraph, and leave it slightly projecting into the margin ; and then lastly to enlarge it. The letter made thus prominent being a sufficient indication of the commencement of the new paragraph, the stroke or wedge between the lines was no longer necessary and ordinarily disappeared, Thus the two lines given above would, in this last stage of development, be written thus :—


Of course, if the paragraph commenced at the beginning of a line, the large letter took its natural place as the initial; but, arranged as above, any letter, even one in the middle of a word, might be enlarged.

Codex Alexandrinus

This system is found in action in the Codex Alexandrinus, attributed to the 5th Century, and continued to be practised throughout the middle ages. But it should be noted that, although rendered unnecessary by the introduction of the large initial, the paragraph mark also appears in this MS., but generally in anomalous positions, particularly above the initial letters of the different books—an indication that the scribes of the day had already begun to forget the meaning and proper use of the mark.

We next have to consider punctuation by points. As already stated, these were used in ancient inscriptions. The earliest instance of their employment in a Greek MS. occurs in the very ancient fragment known as the Artemisia papyrus, at Vienna, wherein the double point (:) occasionally closes a sentence. Again, in the fragments of the Phædo of Plato, found at Gurob, the satme double point appears as a mark of punctuation ; and it may also be here added that a short horizontal stroke or dash also serves the purpose of separating the different speeches in the same fragments. The double point also, in addition to the παράγραφος, occasionally marks the close of the paragraphs in the Paris papyrus 49, a letter of about 160 B.C. But such isolated instances merely show that there was a knowledge of the value of such marks of punctuation, which, however, in practice were not systematically employed.

A more regular system was developed in the schools of Alexandria, its invention being ascribed to Aristophanes of Byzantium (260 B.C.). This was the use of the full point with certain values in certain positions (θέσεις) : the high point (στιγμὴ τελεία), équivalent to a full stop ; the point on the line (ὑποστιγμή), a shorter pause, equivalent to our somicolon; and the point in a middle position (στιγμὴ μέση), an ordinary pause, equivalent to our comma. In the Codex Alexandrinus the middle and high points are pretty generally used. But the middle point eventually disappeared ; and about the ninth Century the comma was introduced. It also became a common practice to mark the conclusion of a paragraph or chapter with a more emphatic sign, such as two or more dots with or without a horizontal dash, : , :- , . The mark of interrogation also first appears about the 8th or 9th Century.


The punctuation of Latin MSS. followed in some respects the systems of the Greeks. In the poem on the Battle of Actium, found at Herculaneum, points are used to mark off the words, a practice borrowed from inscriptions ; and in the early MSS. of Virgil in the Vatican Library points are found employed for the same purpose, although they appear to be due to a second, but still early, hand. From the Latin grammarians we knowthat they adopted the Greek system of punctuation by points (θέσεις, positurae ), to which they gave the titles of "distinctio finalis," " subdistinctio," and " distinctio media" ; but in practice we find that the scribes used the points without consistently adhering to their meaning.

In some of the more ancient MSS. marks of punctuation are entirely wanting, only a short space being left blank in the line to indicate the conclusion of a passage or paragraph, as in Greek MSS., but without the accompanying dividing line (παράγραφος) or the enlarged letter at the beginning of the first full line, which the Greek scribes employed. Yet the paragraph mark was used to separate paragraphs or divisions of the text (as, for example, in the poem on the Battle of Actium) when the new paragraph began a line ; and its eventual conversion from a mere siga of separation between two paragraphs into a sign belonging to the head of the new paragraph was a natural development. Our modern ¶ if is directly derived from the simple ancient form T.

In early uncial MSS. it is not uncommon to find tho point, more often in the middle position, used as an ordinary stop ; and at the end of a paragraph or chapter, a colon, or colon and dash, or a number of points, occasionally indicate a final stop. In the seventh century the high point is used with the force of a comma, the semi-colon with its modern value, and a point and virgule, ˙7, or other combinations of points, as a full stop. In the Carlovingian period and the next centuries we have the inverted semicolon, holding a position between our comma and semicolon, and the comma itself. The origin of the former of these is uncertain. It appears first with some regularity in MSS. of the eighth century ; but it is noticeable that a mark which resembles it occurs in the Actium poem, being there formed by the addition of an oblique stroke to an ordinary point. Along with these later signs also appears the mark of interrogation in common use.

Breathings and Accents and Other Signs.—Greek.

Breathings and accents, like the Greek system of punctuation by points noticed above, are also attributed to Aristophanes of Byzantium, as part of the δέκα προσῳδίαι, of which he is cailed the inventor.

The rough (⊢) and the smooth (⊣) breathings (πνεύματα) at first represented the left and the right half of the letter H, which itself was originally the aspirate. They were soon worn down to └ and ┘, in which shapes they are found in early MSS. ; and eventually those square forms became the rounded ‘ and ’, the period at which they definitely arrived at this last stage being the 12th Century. Only occasionally are marks of breathing found in the more ancient MSS., and then it is generally the rough breathing that is distinguished.

The accents (τόνοι) are : the grave ` (βαρεῖα), or ordinary tone ; the acute ´ (ὀξεῖα), marking a rise in the voice ; and the circumflex ῀ (ὀξυβαρεῖα or περισπωμένη), combining the other two, and indicating a rise and fall or slide of the voice. Originally, in theory, all syllables which were not marked with the acute accent or circumflex received the grave accent, as Θὲόδὼρὸς ; and several examples of this actually occur in the Harris Homer. In the same MS., and occasionally in the Bankes Homer, we also see instances of the indication of normally oxytone words (in which the acute accent falls on the last syllable) by placing a grave accent on the penultimate, as ὲλων. In later MSS. a double accent marks emphatically με̏ν and δε̏.

Breathings and accents were not systematically applied to Greek texts before the seventh Century.

The rest of the ten signs attributed to Aristophanes of Byzantium, to assist in the correct reading of texts, ure as follows :—

The χρόνοι, or marks to distinguish a long (   ̅ ) and a short (   ̆ ) syllable, instances of their employment occurring in the Harris Homer and in some other early documents on papyrus.

The διαστολή or ὑποδιαστολή, a virgule or comma inserted between words where the distinction inigbt be ambiguous, as εστι, voυς, not εστιν, ους.

The hyphen (ὑφέν), a curve or line drawn under the letters to indicate connection, as, for example, to indicate compound words. In the Harris Homer the hyphen, in the form of a long straight line, is used for this pnrpose.

The apostrophe (ἀπόστροφος), which, besides marking elision, was used for other purposes, and whose form varied from a curve to a straight accent or even a mere dot. It was very generally placed in early MSS. after a foreign name, or a name not having a Greek termination, as, for example, Ἀβρααμ', and alter a word ending in a hard consonant, as κ, χ, ξ, ψ, and also in ρ. When a double consonant occurred in the middle of a word, an apostrophe was placed above the first or between t.he two letters. In a papyrus of A.D. 542 (Pal. Soc. ii. 128), a dot represents the apostrophe in this position; and in a MS. of the 8th or 9th Century (Pal, Soc. ii. 126), a double apostrophe is employed. The apostrophe is also used to distinguish two concurrent vowels, as ιματια' αυτων. In some instances it is even placed between two different consonants, as e.g. αριθ' μος, in the Vienna MS. of Dioscorides.

In addition to the marks and signs already noticed, there are some others which occur in Greek MSS.

Marks of diæresis, placed over ι and ν when at the beginning of a word or when they do not form a diphthong with a foregoing vowel, occur in papyri, being either a single or double dot or short stroke, or, in some instances, a short accent; in later MSS. the form is usually a double dot.

Quotations are indicated by marks in the margin, the most common being the arrow-head, > or <, and tbe cross, horizontal stroke, or waved stroke being also used. More rarely, quoted passages are indented, that is, written within the marginal line of the text.

To distinguish words consisting of a single letter, a short acute accent or similar mark is found in use, as, in the Codex Alexandrinus, to mark η in its various meanings as a word. Apparently from ignorance or confusion the scribes of this MS. even placed a mark on η when merely a letter in a word. The article is found similarly distinguished in a papyrus of A.D. 595 (Pal. Soc. ii. 124).

To fill small spaces left vacant at the end of a line, an arrow-head or tick was employed ; as, for example, in the papyrus of Hyperides (Lycophron), and in the Codex Sinaiticus.

Arbitrary signs, or signs composed of dots or strokes, are used as reference marks to marginal scholia, or to indicate insertion of omitted words or passages. In the papyrus of Hyperides (Lycophron) the place for insertion of an omitted line is marked, and has the word ἄνω, while the line itself, written in the margin above, has κάτω. In the papyrus of Aristotle on the Constitution of Athens, a letter or word inserted between the lines has sometimes a dot on each side.

In the same manner various signs are employed to indicate transposition, such as numerical letters, or (as in the papyrus of Aristotle) slanting strokes and dots (/·) placed above the words.

To distinguish words or other combinations of letters from the rest of the text, a line was drawn above them ; thus the grammatical forms in the papyrus attributed to Tryphon, in the British Museum, and the reference letters in the Oxford Euclid of A.D. 888 are so marked.

Besides actually striking out a letter or word or passage with a pen-stroke, the ancient scribes indicated erasure by including the word or passage between inverted commas or brackets or dots, one at the beginning and one at the end; sometimes by accents above, as e.g. των́ (to erase the ν ), τ́ά and π́αντά (to cover the whole word), as seen in the Codex Alexandrinus; sometimes by a line above, as και ; sometimes by a dot above, rarely below, each letter.

Accents and other Signs.—Latin.

Accents were seldom used by Latin scribes. Occasionally they mark a monosyllabic word, as the exclamation ό, or a preposition, as á ; and sometimes they are employed to emphasize a syllable.

As in Greek MSS., quotations are indicated by marks in the margin or by indentation ; and arbitrary signs are used to mark the place of insertion of omissions. Common reference marks are hd hs = hic deest, hoc supra or hic scribas, etc. Transposition of words might be indicated in various ways, as by letters or numbers, and very commonly by oblique strokes above the line, as m̋ea m̋ater = mater mea.

Finally, for correction, the simple method of striking out with the pen and interlining or adding in the margin was followed, as well as that of marking words or letters for deletion with dots above or below them.

Besides the above, other marks and signs are found in both Greek and Latin MSS., such as the private marks of correctors or readers. There are also critical symbols, such as the diple and the asterisk employed by Aristarchus in the texts of Homer, and the obelus and asterisk used by St. Jerome to distinguish certain passages in versions of the Latin Psalter. But the consideration of these is beyond the scope of the present work
↑top .


Codex Ephraemi

A palimpsest MS. is one from which the first writing has been rubbed off in order to make the leaves ready to receive fresh writing. Sometimes this process was repeated, and the leaves finally received a third text, the MS. being in such a case doubly palimpsest. This method of obtaining writing material was practised in early times. The term "palimpsest" is used by Catullus, ← Carm. xxii. 5. apparently with reference to papyrus ; also by Cicero in a passage Ad Fam. vii. 18. wherein he is evidently speaking of waxen tablets ; and by Plutarch, who narrates Cum princip, pliilosoph., ad fin. that Plato compared Dionysius to a βιβλίον παλίμψηστον, his tyrannical nature, δυσέκπλυτος, showing through like the imperfectly erased writing of a palimpsest MS., that is, a papyrus roll from which the first writing had been washed. The word, however, indicating, as it does, the action of scraping or rubbing, could originally have only been strictly applied to material strong enough to bear such treatment, as vellum or waxen tablets. Papyrus could only be washed, not scraped or rubbed, and the application of the term to a twice-written papyrus or waxen tablet or vellum MS. indifferently, proves that the term had become so current as to have passed beyond its strict meaning.

If the first writing were thoroughly removed from the surface of vellum, none of it, of course, could ever be recovered. But, as a matter of fact, it appears to have been often very imperfectly effaced ; and even if, to all appearance, the vellum was restored to its original condition of an unwritten surface, yet slight traces of the text might remain which chemical re-agents, or even the action of the atmosphere, might again intensify and make legible. Thus many capital and uncial texts have been recovered from palimpsest MSS. Of modern chemical re-agents used in the restoration of such texts the most harmless is probably hydro-sulphuret of ammonia.

Great destruction of vellum MSS. of the early centuries of our era must have followed the fall of the Roman empire. Political and social changes would interfere with the market, and writing material would become scarce and might be supplied from MSS. which had become useless and were considered idle encumbrances of the shelves. In the case of Greek MSS., so great was their consumption that a synodal decree of the year 691 forbade the destruction of MSS. of the Scriptures or of the fathers, imperfect or injured voiumes excepted. It has been remarked that no entire work has in any instance been found in the original text of a palimpsest, but that portions of différent MSS. were taken to make up a volume for a second text.

The most valuable Latin texts are found in the volumes which were re-written from the seventh to the ninth centuries. In many instances the works of classical writers have been obliterated to make room for patristic literature or grammatical works. On the other hand, there are instances of classical texts having been written over Biblical MSS. ; but these are of late date.

In the great Syriac collection of MSS. which were obtained from the monastery in the Nitrian Desert of Egypt and are now in the British Museum, many important texts have been recovered. A volume containing a work of Severus of Antioch, of the beginning of the 9th century, is written on palimpsest leaves taken from MSS. of the Iliad of Homer and the Gospel of St. Luke of the 6th century (Cat. Anc. MSS. i. pls. 9, 10) and of the Elements of Euclid of the 7th or 8th century. Another volume of the same collection is doubly palimpsest, a Syriac text of St. Chrysostom, of the 9th or 10th century, covering a Latin grammatical work of the 6th century, which again has displaced the annals of the Latin historian Licinianus of the 5th century (Cat. Anc. MSS. ii. pls. 1, 2). At Paris is the Codex Ephraemi, containing portions of the Old and New Testaments in Greek, of the 5th century, which are rewritten with works of Ephraem Syrus in a hand of the 12th century; and some fragments of the Phaeton of Euripides are found in the Codex Claromontanus. At the Vatican are portions of the De Republica of Cicero, of the 4th century, under the work of St. Augustine on the Psalms of the 7th century; and an Arian fragment of the 5th century. At Verona is the famous palimpsest which contains the MS. of Gaius of the 5th century, as well as the Fasti Consulares of A.D. 486. At Milan are the fragments of Plautus, in rustic capitals of the 4th or 5th century, covered by Biblical text of the 9th century. Facsimiles of many of these MSS. are given by Zangemeister and Wattenbach in their Exempla Codicum Latinorum.