Books existed in Egypt long before they came into use in Greece. Systems of writing had been invented and developed for administrative purposes in both Egypt and Mesopotamia by c.3000 bc. While the Sumerians (see sumerian) and Babylonians used clay tablets for their cuneiform scripts, the Egyptians used papyrus. A blank sheet of papyrus was found in the tomb of the vizier Hemaka in Saqqara of c.3000 bc. The oldest surviving inscribed papyrus texts are the temple accounts of Abusir of c.2450 bc. A number of fine statues of seated scribes of the same period suggests that this profession was already well established and that writing had been practised for centuries, long enough for the ‘hieratic’ script to develop through the adaptation of hieroglyphs to the use of reed-brush and papyrus. The hieroglyph for ‘book-roll’ is first attested in the first dynasty (c.3000–2800 bc), and Egyptian literature was supposed to have begun with the writings of Imhotep, the architect of the first pyramid under King Djoser in the third dynasty (c.2600 bc). Religious books were kept in temples; although temple ‘libraries’, i.e. chambers designated for the storage of books, have survived only in Ptolemaic temples (Edfu, Philae, ed-Tod), literary references to books and libraries suggest their existence in the Middle Kingdom (thirteenth dynasty, c.1700 bc), and Diodorus (3) (1. 47, using the account of Hecataeus (2)) describes the library in the Ramesseum at Thebes (2).

The papyrus plant (Cyperus papyrus) grew mainly in the swamplands of lower Egypt and especially the Nile delta. It was used for many purposes: to make ropes, sandals, baskets, boats, and—most importantly—writing material. The Greeks called it βύβλος or βίβλος (see below), later πάπυρος (first attested in Theoph. Hist. pl. 4. 8. 2), believed to be derived from Egyptian ‘that of Pharaoh’, which suggests that its manufacture and marketing were a royal monopoly. The locus classicus describing its manufacture is Pliny (the Elder) Naturalis historia 13. 74–82 (translation and discussion in Lewis, 34–69, see bibliog. below). The triangular-sectioned stem is cut into segments c.30–40 cm. (12–16 in.) long, from which the outer hull is then removed; the white pith is sliced lengthwise into thin strips, which are placed vertically parallel to each other on a plane surface; over these, a second layer of strips is placed horizontally, and the ends squared off. They are then pressed in presses (prelis); the plant's natural juice glues the layers firmly together. After drying, the sheet is smoothed with a stone or a sea-shell. According to Pliny (1) the Elder, twenty sheets (κολλήματα, plagulae) were then glued together to form a roll (χάρτης, charta), on average 6–8 m. (20–26 ft.) long (in Pharaonic times, too, papyrus rolls of twenty sheets had been standard). If a longer roll was required, it had to be manufactured specially. The narrow edge (κόλλησις, ‘gluing’) of a sheet which, in Greek books, overlaps that of the sheet to its right is usually 20–25 mm. (¾–1 in.) wide. In rolls of Egyptian texts, however, the right-hand sheet overlaps the one to its left, so that the scribe, writing from right to left, had his reed brush travelling ‘downhill’ over the joins. If this was the way in which rolls were originally manufactured, Greek scribes turned them upside down, so that they, too, could write ‘downhill’ over the joins. The sheets on the inside of the roll show horizontal fibres; in this way the joins are better protected from being pulled apart. Only the first sheet of the roll has its vertical fibres on the inside, at right angles to those of the following sheets; this is the ‘protocol’ (πρωτόκολλον) which protected the outside of the book when it was rolled up; it was sometimes a parchment sheet (cf. [Tib.] 3. 1. 9), or reinforced with parchment. Whether the ends of rolls were similarly reinforced is not clear; the surviving ends of book-rolls show no sign of an eschatocollion (Mart. 2. 6. 3). Rollers (umbilici, Hor. Epist. 14. 8; Mart. 4. 89) with decorative knobs (cornua, Mart. 11. 107) attached to the last sheet survive in some rolls from Herculaneum but none have been found in Egypt.

As the pen runs more smoothly along the fibres than across, the scribes normally used the inside of the roll first where the glued join (κόλλησις) runs at right angles to the fibres; this is conventionally called ‘recto’, its back ‘verso’, but these terms should only be used with regard to codices (‘recto’ = right-hand page), whereas for rolls the terms ‘front’/‘back’, or ‘inside’/‘outside’ are preferable. In some rolls, however, the writing runs across the fibres on the inside, parallel to the joins, in long lines from edge to edge. Such rolls (rotuli) were employed only for Greek and Latin documents, never for literary texts which were designed for continuous reading. Documents, on the other hand, were often glued together to form rolls, the blank backs of which could then be used for literary texts (for private use, not for sale). As the inside of the roll was nearly always used first, a dated document may provide a terminus post quem for the literary text; the reverse case (document on the back of a book roll, cf. Mart. 8. 62) is relatively rare. Papyrus rolls were dressed with cedar-oil (κεδρία, cedrium) to protect them against worms (Vitr. De arch. 2. 9. 13).

A papyrus roll would take a book of Thucydides (2), or a play of c.1500 lines, or two to three books of Homer. The length of the books of Apollonius (1)'s Argonautica, or the books of Polybius (1), Strabo, or Diodorus (3), may have been determined by this standard format. The text is arranged in columns (σελίδες, paginae); the number of lines per column, usually between 25 and 45, varies with the height of the column and the size of the writing. The length of the lines also varied; in hexameter and trimeter poetry it was determined by the verse, but lyric poetry and prose were (at any rate, from the 2nd cent. bc onwards) written in shorter lines of between 5 and 10 cm. (2–4 in.), with an average of between 18 and 25 letters per line. Some book rolls have wide upper and lower margins as well as generous spaces between columns. Sometimes a line accidentally omitted from the text is added in the upper or lower margin, usually with a reference mark.

Apart from papyrus, leather was also used in Egypt to produce rolls for literary texts. The Annals of Thutmosis III were written on leather rolls and deposited in the temple of Amūn at Karnak. The inscription in the library of the Edfu temple also lists leather rolls. In the ancient near east, leather was widely used as a writing material. Ctesias (in Diod. Sic. 2. 32. 4) describes the ‘royal hides’ (βασιλικαὶ διφθέραι) which the Persians used for their chronicles; from Persia, leather rolls may have come into use in Ionia and then in other parts of Greece. In Italy, too, animal hides were used at an early stage (Dion. Hal. 4. 58). While leather had to be tanned, vellum or parchment was manufactured by placing the hide in slaked lime for some days; then flesh and hairs were scraped off and the hide was rubbed with calcium oxide, stretched in a frame, and finally smoothed with pumice. The name ‘parchment’ (περγαμηνή, pergamena) is derived from Pergamum, which became a centre of production and export in the 2nd cent. bc when papyrus from Egypt was in short supply. The earliest extant Greek documents on vellum (parchment), however, prove that the manufacturing process had been known earlier.

In the production of books parchment played only a minor role compared with papyrus, which remained the dominant writing material throughout Greek and Roman antiquity. In Crete, papyrus may have been used as early as the second millennium bc, as strands of papyrus have been found baked into Minoan clay sealings. Egyptian papyrus was also exported to Phoenicia (see phoenicians) and beyond; the story of Wen-Amūn (11th cent. bc) mentions 500 rolls of finished papyrus sent to Byblos in return for timber. The fact that in Greek (Aesch. Supp. 947, Hdt. 2. 100, etc.) byblos means ‘papyrus roll’ suggests that originally it may have been imported from Byblos, Phoenician Gubla. So it is at least possible that papyrus as a writing material was known in the Greek world in Mycenaean times, but there is as yet no firm evidence for its use as a vehicle for Greek literature before c.500 bc when book-rolls of papyrus first appear on Attic vases. Herodotus' remark (5. 58) that the Ionians had used leather rolls at a time when papyrus was scarce, and hence kept the word διφθέρα (‘hide’) for ‘book’ (βύβλος), does not rule out early acquaintance with papyrus. In the 8th cent. bc, papyrus is mentioned in cuneiform texts in Assyrian accounts as ‘reeds from Egypt’, and an 8th-cent. palimpsest papyrus written in Hebrew has come to light near Murabba᾽at on the Dead Sea (PMur. 17 = Discoveries in the Judaean Desert (1960) 2 pp. 93–100).

Although writing may have been employed early on in the composition of Greek poetry (and the complex structure of both Iliad and Odyssey is hardly conceivable without it), the performance of poetry continued to be oral throughout the Archaic and Classical periods. When Archilochus describes himself as a ‘woeful messenger-stick’ (ἀχνυμένη σκυτάλη, fr. 185 W), he does not necessarily imply that his epode was transmitted on a piece of papyrus or leather, wrapped round a stick (see skytale). On the other hand, much of early epic poetry is reflected in both lyric poetry and black-figure vase-painting, Corinthian and Attic, and it seems doubtful whether this can be accounted for by oral transmission (by itinerant rhapsodes and choirs) alone. It is therefore reasonable to assume that book-rolls played a part in the transmission of Greek poetry in the 7th and 6th cent., if only as aides-mémoire to the performers. In the 6th cent., the tyrants Polycrates (1) of Samos and Pisistratus of Athens are said to have been admired for their collections of books (Athen. 1. 3a). Pisistratus is credited with a revision of the texts of Homer which until then had been ‘confused’ (Cic. De or. 3. 137); he is also said to have inserted lines about Salamis (1) and Theseus into the texts (Il. 2. 558, Od. 11. 631). On the other hand, [Pl.] Hipparch. 228b says that Pisistratus' son Hipparchus (1) brought the poems of Homer to Attica and forced the singers at the Panathenaea to perform them—in preference, perhaps, to other parts of the Epic Cycle. In any case, there clearly was, in the later 6th cent., an authoritative text of Homer which served as a basis for rhapsodic recitals at the Panathenaea.

From c.500 onwards, book-rolls (evidently of papyrus) appear on Attic vases; as far as the writing can be identified, they all contain poetry. The Duris cup in Berlin of c.485 bc (Beazley, ARV² 431. 48) illustrates the use of book-rolls in schools, and Ar. Nub. 961–72 describes the ‘ancient education’ (ἀρχαία παιδεία), i.e. in the schools of c.500 bc, where the children were made to memorize epic poetry and to sing it in the traditional mode. Reading books for pleasure is mentioned in Euripides Erechtheus (fr. 369 N, of 422 bc?), and in Aristophanes Ranae 52 f. (405 bc) Dionysus says he read Euripides' Andromeda on board ship. The earliest references to booksellers are in Eupolis (PCG fr. 327) and in Plato (1) (Ap. 26d), where Socrates says that a copy of Anaxagoras could be bought ‘from the orchestra’ (in the Agora of Athens?) for one drachma ‘at most’. The term ‘bookseller’ (βιβλιοπώλης) is first attested in Theopompus (2) the comic poet (fr. 79 KA). Xenophon (1) (An. 7. 5. 12–14) refers to ‘many written books’ being exported on ships from Athens to the Black Sea (see euxine), and in De memoria 4. 2. 10 Socrates asks Euthydemus whether he wants to become a rhapsode, having bought the complete works of Homer.

The intellectual revolution of the sophists and the interest in dithyrambs and tragedy boosted demand for books in Athens, where book production and the book trade flourished in the 4th cent. bc. It made the vast collecting activities of Aristotle and his pupils possible and led to the formation of libraries, notably that of Aristotle himself (Strabo 13. 608).

The oldest surviving specimens of literary papyrus rolls date from the second half of the 4th cent. The Timotheus (1) papyrus (P Berol. 9875, Persians), found at Abusir north of Memphis, written in long lines in large, clumsy letters, may antedate Alexander (3) the Great's conquest of Egypt. The carbonized papyrus roll found at Derveni, near Thessalonica, a commentary on an Orphic cosmogony, is written in small letters (c.2 mm. (1 10 in.) high) in a careful, skilled hand which makes the columns look almost like stoichedon inscriptions (i.e. exactly aligned vertically). Given the regularity and elegance of Attic writing on vases and also of private letters on lead tablets, such as the letter of Mnesiergus in Berlin (Syll.³ 1259), the Derveni papyrus has a far stronger claim to being a typical representative of a 4th-cent. book. From the beginning of the Ptolemaic period through to the 8th cent. ad, an uninterrupted series of book-rolls and, later, codices chiefly from Egypt illustrates the development of Greek (and, to a lesser extent, Latin) books and their scripts.

The museum and the library at Alexandria (1), founded by Ptolemy (1) I Soter, became the most important centre of scholarship, and literary criticism in particular, for centuries to come. The work of the Alexandrian scholars led to a standardization in the formats of Greek books and in the layout of the texts. While in early 3rd-cent. bc texts of poetry, verse (other than hexameters and trimeters) is not written metrically, books from the 2nd cent. bc onwards show lyric passages arranged in short metrical units or kōla; their colometry is preserved in the medieval manuscripts of Pindar's epinikia and the choruses of Attic drama (see tragedy; comedy). Prose texts are usually set out in fairly narrow columns in lines of equal length. Aids to the reader, such as spaces between sentences, punctuation, or horizontal dashes (paragraphoi) in the left-hand margin or between line-beginnings, very rare at first, become more frequent in the 2nd and 1st cent. bc and increasingly common during the Roman period, as do accents and breathings. Aristarchus (2) devised a system of critical reference signs to alert the reader to textual problems and their discussions in commentaries; most of these appear in book-rolls, though not always with the functions assigned them by Aristarchus. Commentaries (ὑπομνήματα) were written in rolls separate from the texts on which they comment, but the rolls carrying texts often have notes or excerpts from commentaries in the margins, although they are rarely as copious as in the Louvre papyrus of Alcman's Partheneion.

Greek book-rolls continued the Egyptian tradition of book illumination; the surviving illustrated papyri are mostly scientific or mathematical books, or magical papyri; fragments of illustrated papyrus rolls of Homer or drama are very rare. Book illumination becomes more common only in the later Roman period, with the victory of the codex (see below). Titles, too, were added, either in the left-hand margin against the opening line, or at the end of a book under its last line. Title-tags (sillyboi) were sometimes appended to book-rolls. Book titles first came into use with Attic drama, because the titles of plays had to be entered for the competition. While the older lyric poems are usually referred to by their opening lines, Aristophanes (1) (Nub. 553 f.) quotes Eupolis' Marikas by title and then his own play, the Knights (τοὺς ἡμετέρους Ἱππέας). Dithyrambs, too, may have been entered for competitions by their titles, perhaps as early as the 6th cent., if Herodotus' statement that Arion (2) ‘gave his dithyrambs names’ (ὀνομάσαντα, 1. 23) is to be trusted. Prose works in the 5th and 4th cents. bc generally do not seem to have had titles, even though Plato (Plt. 284b) refers to his dialogue Sophist by this title (ἐν τῳ̑ Σοφιστῇ). Normally, prose works, too, are referred to by quoting their opening words, a practice which may have been promoted by Callimachus (3)'s catalogue of the Alexandrian library, the Pinakes (‘Panels’). The division of longer works into books must be attributed to the Alexandrian scholars who edited them.

Rolls were kept on shelves, or in boxes or buckets (τεῦχος, capsa, scrinium). Sometimes rolls were kept in a vellum cover (διφθέρα, paenula) with a coloured label. The vellum label with Πίνδαρος ὅλος (‘The Complete Pindar’, P. Antinoop. Papyri 1. 21) may have been attached to a box or bookcase containing the seventeen books of the Alexandrian edition. While private libraries had boxes or movable cases, as did the library of Eumenes (2) II of Pergamum, the great public libraries of early imperial Rome had bookcases inserted in niches (θυρίς, fenestra) in the walls. This is a feature of both Egyptian temples (Edfu, Philae, ed-Tod: see above) and Coptic monasteries; it seems probable that Caesar saw them in Alexandria (in the Serapeum library?) and hence planned to introduce them in Rome's first public library, a plan which Asinius Pollio carried out after Caesar's death in the Atrium Libertatis, thus creating the model for Augustus' library on the Palatine.

The most important innovation in the shape of the book was Roman in origin. The codex was created when the wooden panels of writing-tablets fastened together with thongs were replaced by parchment (membrana). At first used as notebooks (Hor. Sat. 2. 3. 2; Quint. Inst. 10. 3. 31), parchment codices had come into use for classical literature by the 1st cent. ad (Mart. 14. 184–92 advertises them as cheap pocket editions), while the normal form of the book was, in the Latin west as in the Greek east, the papyrus roll. What eventually established the codex was its adoption by the Christians; the vast majority of biblical and NT texts from the early 2nd cent. onwards are in codex form. Pagan classical authors appear in parchment and papyrus codices from the 2nd cent. ad and more frequently in the 3rd; by the 4th cent., three out of four literary texts were in codex form.

Codices were assembled from quires consisting of wide sheets folded vertically in the middle and then stitched together along the fold. Quires of four or five sheets (quaterniones, quiniones = 8 or 10 leaves) are common, thinner or thicker quires are exceptional and may be early experiments. Sheets of papyrus were often placed so that on facing pages the direction of the fibres was the same (recto pages with horizontal and vertical fibres alternating), but the alternative arrangement (i.e. vertical facing horizontal fibres, and vice versa) is not uncommon. Vellum sheets were usually arranged with hair sides facing and flesh sides facing. While papyrus sheets in Greek book-rolls, and in codices of the 2nd and 3rd cents., rarely exceed 35 cm. (13½ in.) in height and 23 cm. (9 in.) in width, wider sheets (30–35 cm.: 12–13½ in.) were often used in Greek and Coptic codices of the 4th, 5th, and 6th cents., when large parchment codices were also common. Small formats (even miniature codices like the Cologne parchment codex of the life of Mani with pages of 38×45 mm. (1½×1¾ in.)) are also found in this period.

The text was written before the sheets were stitched together, usually in one column per page, prose texts sometimes in two narrow columns like those in rolls (more frequently in parchment than in papyrus codices). Pagination is frequent; sometimes quires are also numbered on their first pages. Codices were bound between wooden panels covered in leather. Papyrus codices remained common in Egypt until the late 6th cent.; elsewhere, parchment codices begin to prevail from the later 3rd cent. onwards. For codices, the advantages of parchment over papyrus are obvious: less fragile folds, greater durability, greater capacity, and they are easier to use. Constantine ordered 50 parchment copies of the Scriptures for the churches in Constantinople, and Jerome records that the papyrus manuscripts in the library of Caesarea (2), having become worn by use, were replaced by parchment codices.

Little is known about the book trade in the Greek world, probably because the private copy always remained the commonest form of book production. In the 4th cent. there seem to have been itinerant booksellers who ‘carried around’ bundles of Isocrates' speeches (Dion. Hal. Isoc. 18). More information is available on the book trade in Rome from the 1st cent. bc onwards, and some publishers like Cicero's friend Atticus, or the Sosii, Horace's publishers, or Martial's and Quintilian's publisher, Tryphon, are well attested; the younger Pliny (2) (Ep. 11. 11. 2) was surprised to learn that booksellers in Lyons (Lugdunum (1)) were selling his books. The evidence for book prices is scanty and contradictory; on the whole, it seems that in Rome books were not expensive: Martial quotes 5 denarii as the price of a luxury edition of his poems (1. 117), while a cheap one would cost 6 to 10 sesterces (1. 66). Palimpsests, i.e. texts written on reused papyrus or parchment after the original writing had been rubbed or washed off, were cheaper still (Catull. 22).