# Reading collection: texts, editions and limits

This is a modern, curated reading collection for a reconstruction set around 250 BCE. It is **not an inventory recovered from the Library of Alexandria**. A work’s survival today, its probable date, and evidence for an ancient Alexandrian copy are separate questions. No individual shelf placement is documented.

The collection contains **42 surviving works or explicitly named collections**, accessible through **40 English reading texts and two complete scanned editions in five PDF volumes**, together with **28 records for missing works: 25 fragmentary and three lost**. The five-book Greek Pentateuch is one catalogue record. The four Theocritean Idylls and the individual plays are separate records, each containing a complete poem or play—not a few lines disguised as a full work.

“Complete” means complete within the identified surviving text and modern edition. It does not mean an untouched ancient autograph, every work ever written by an author, or the recovery of material now lost. Thucydides is unfinished; the Bacchae has manuscript gaps; the transmitted History of Animals includes disputed material; the surviving Poetics is not the whole of Aristotle’s projected treatment. These distinctions are also stated in the individual records.

## What belongs to the date of the reconstruction?

Most selections clearly predate 250 BCE. Early Hellenistic works require more care. Callimachus’s Hymn to Zeus, selected Idylls of Theocritus, and Aratus’s Phaenomena are individually identified instead of treating a modern anthology as a securely dated ancient collection. The earlier Aetia material is distinguished from Books III–IV and the Lock of Berenice, which include material after 246 BCE. The dating of Hecale is explicitly uncertain.

The **Greek Pentateuch is a borderline chronological inclusion**, not a claim that all five Greek books were certainly finished by 250 BCE. Modern scholarship commonly places their translation in the middle or later third century BCE in Ptolemaic Egypt; exact location, circumstances and chronology are disputed. The much later Letter of Aristeas is a source for the royal-commission tradition, not a contemporary Library accession record. Later books of the Septuagint are excluded from the reader. See [Greenspoon and Ross, “Septuagint,” Oxford Bibliographies](https://academic.oup.com/reference/62341/reference-article-abstract/561720572) and [James K. Aitken, “The Ptolemaic Setting for the Translation of the Greek Pentateuch” (2020)](https://www.repository.cam.ac.uk/items/b34fdcd3-1095-483f-8460-af288e6a626c). Aitken’s proposal emphasizes Egyptian scribal settings, including the Fayum, rather than establishing an Alexandrian royal project.

Theophrastus’s Enquiry into Plants belongs to the late fourth and early third centuries BCE. It was repeatedly revised and appears unfinished; its nine surviving books must not be confused with the separate Plant Explanations. The other works bound in Hort’s second modern volume are identified as separate texts, not additional books of the Enquiry. See the [Stanford Encyclopedia of Philosophy’s Theophrastus entry](https://plato.stanford.edu/entries/theophrastus/).

No writings of Hipparchus, Apollonius of Perga, Hero, Ptolemy or Hypatia are staged here. Eratosthenes’s mature geographical work and the later treatises of Archimedes are likewise not silently backdated into 250 BCE.

## Downloaded sources and rights

The downloadable editions are bundled locally. The machine-readable [source manifest](/texts/source-manifest.json) records each source URL, catalogue URL where applicable, byte length, SHA-256 hash, transformations and scan-verified corrections. The acquisition script is `scripts/fetch-texts.mjs`. Running it checks source identities and boundaries and regenerates selected reading copies and their navigation index. Existing source files are reused unless `--refresh` is supplied.

### Project Gutenberg editions

Twenty-nine original Gutenberg source files supply 37 of the 40 text records. Their original headers, end markers and full licence notices remain in the downloadable editions. Public-domain status here follows the identified editions’ United States publication history and Gutenberg’s distribution notices; readers elsewhere should check their own jurisdiction. Display-only removal of boilerplate does not remove the notices from the downloadable source.

| Works presented | Translator / identified edition | Original source |
| --- | --- | --- |
| Iliad, all 24 books | Samuel Butler, 1898 | [Gutenberg 2199](https://www.gutenberg.org/ebooks/2199) |
| Odyssey, all 24 books | Samuel Butler, 1900 | [Gutenberg 1727](https://www.gutenberg.org/ebooks/1727) |
| Works and Days; Theogony | Hugh G. Evelyn-White, Loeb, 1914 | [Gutenberg 348](https://www.gutenberg.org/ebooks/348) |
| Theocritus, Idylls I, VII, XV, XVII | C. S. Calverley, third edition | [Gutenberg 11533](https://www.gutenberg.org/ebooks/11533) |
| Herodotus, Histories, all nine books | G. C. Macaulay, 1890, both volumes | [2707: Books I–IV](https://www.gutenberg.org/ebooks/2707), [2456: Books V–IX](https://www.gutenberg.org/ebooks/2456) |
| Thucydides, all eight surviving books | Richard Crawley, 1874 | [Gutenberg 7142](https://www.gutenberg.org/ebooks/7142) |
| Xenophon, Anabasis; Hellenica | H. G. Dakyns | [1170](https://www.gutenberg.org/ebooks/1170), [1174](https://www.gutenberg.org/ebooks/1174) |
| Agamemnon; Libation Bearers; Eumenides | E. D. A. Morshead, The House of Atreus | [Gutenberg 8604](https://www.gutenberg.org/ebooks/8604) |
| Oedipus the King; Oedipus at Colonus; Antigone | Francis Storr, Loeb, 1912 | [Gutenberg 31](https://www.gutenberg.org/ebooks/31) |
| Medea; Bacchae; Trojan Women; Electra; Alcestis | Gilbert Murray, historical editions identified in each record | [35451](https://www.gutenberg.org/ebooks/35451), [35173](https://www.gutenberg.org/ebooks/35173), [10096](https://www.gutenberg.org/ebooks/10096), [14322](https://www.gutenberg.org/ebooks/14322), [10523](https://www.gutenberg.org/ebooks/10523) |
| Clouds; Frogs | William James Hickie | [2562](https://www.gutenberg.org/ebooks/2562), [7998](https://www.gutenberg.org/ebooks/7998) |
| Plato, Republic | Benjamin Jowett, third edition, 1888 | [Gutenberg 55201](https://www.gutenberg.org/ebooks/55201) |
| Plato, Apology; Phaedo; Symposium; Phaedrus; Timaeus | Benjamin Jowett | [1656](https://www.gutenberg.org/ebooks/1656), [1658](https://www.gutenberg.org/ebooks/1658), [1600](https://www.gutenberg.org/ebooks/1600), [1636](https://www.gutenberg.org/ebooks/1636), [1572](https://www.gutenberg.org/ebooks/1572) |
| Aristotle, Politics | William Ellis, translation first published 1776 | [Gutenberg 6762](https://www.gutenberg.org/ebooks/6762) |
| Aristotle, Poetics | S. H. Butcher | [Gutenberg 1974](https://www.gutenberg.org/ebooks/1974) |
| Aristotle, Nicomachean Ethics | D. P. Chase; introduction by J. A. Smith | [Gutenberg 8438](https://www.gutenberg.org/ebooks/8438) |
| History of Animals | Richard Cresswell, 1862 translation in the 1887 edition | [Gutenberg 59058](https://www.gutenberg.org/ebooks/59058) |
| On Ancient Medicine; On Airs, Waters, and Places | Francis Adams, 1849 translation reprinted in 1886 | [Gutenberg 72583](https://www.gutenberg.org/ebooks/72583) |

Hesiod, the four Idylls, the six plays selected from anthologies, and the two Hippocratic treatises are complete individual works selected by explicit headings in their source files. The complete source anthologies remain in `/texts/editions/`. The Hesiod reading copies retain the anthology’s endnotes to preserve the source numbering, even where a note concerns another work. The Hippocratic selections omit Adams’s prefatory “Argument”; his source volume remains available. The modern “Theban plays” binding is not represented as an ancient connected trilogy.

Herodotus’s two complete translation bodies are combined under one outer reading boundary so that a reader cannot silently stop after Book IV. Both original notices and licences are retained; the unchanged separate source volumes remain downloadable. The nine book headings are checked before the first end marker.

The plain-text Nicomachean Ethics header does not identify the translator. The [Gutenberg HTML edition](https://www.gutenberg.org/files/8438/8438-h/8438-h.htm) credits D. P. Chase; J. A. Smith wrote the introduction. These roles are not interchangeable.

### Mair’s Callimachus and Aratus, 1921

[Callimachus, Hymn I to Zeus](https://www.theoi.com/Text/CallimachusHymns1.html) is A. W. Mair’s complete translation of the poem corresponding to 96 Greek lines. [Aratus, Phaenomena](https://www.theoi.com/Text/AratusPhaenomena.html) is G. R. Mair’s complete translation corresponding to 1,154 Greek lines. Both derive from Loeb Classical Library 129, *Callimachus and Lycophron; Aratus* (William Heinemann / G. P. Putnam’s Sons, 1921), now public domain in the United States. Only the historical translations are reproduced; the modern website’s navigation, advertisements and layout are not bundled. Callimachus’s explanatory footnotes are omitted and this is disclosed.

The web transcription had errors, including one omitted sentence. Thirteen Callimachus repairs and ten Aratus repairs were visually checked against the [1921 scanned edition](https://archive.org/details/callimachuslycop00calluoft). The manifest gives the exact before/after string and scan URL for every repair; the acquisition script asserts one match for each to catch upstream changes.

For Callimachus, all six English pages of Hymn I were compared (printed pages 37, 39, 41, 43, 45, 47; scan leaves 48, 50, 52, 54, 56, 58). Among the repairs are *watery Iaon*, *thy birthpangs*, *Thenae was nigh*, *signs* for *sings*, and restoration of the sentence “The ancient poets spake not altogether truly.” The final prayer is present through the poem’s end.

Aratus received targeted opening/end and suspected-error checks, **not a complete line-by-line critical collation**. Corrections include *midst*, *his coil circles*, *nor on what task*, *windless calm*, *though urged*, *burst*, and *setting of a star*, plus two omitted articles. The unusual phrase *teaming mastich* was checked against the print and deliberately left unchanged. The final paragraph includes the ending through line 1,154.

### Brenton’s Greek Pentateuch

The source is [eBible.org’s Brenton English Septuagint](https://ebible.org/find/details.php?all=1&id=eng-Brenton), explicitly identified there as the 1851 translation by Sir Lancelot Charles Lee Brenton and public domain. Its complete verse-per-line ZIP is preserved locally, including its original rights notice. This is the Greek Septuagint in English—not a Hebrew-text English Bible substituted for it, and not a modernized derivative translation.

The reading copy selects all 5,803 source verse rows for Genesis, Exodus, Leviticus, Numbers and Deuteronomy. It retains all 187 chapters: 50, 40, 27, 36 and 34 respectively. No later biblical book is placed in this reader. Verse-per-line source formatting omits introductions, footnotes and marginal variants; that limitation is disclosed. Chapter and verse divisions are later reading aids, not an Alexandrian roll arrangement. Brenton’s text descends from later Greek transmission and does not recover the exact wording of an original third-century BCE copy.

### Complete scanned scientific editions

The scans preserve title pages, commentary, original pagination and diagrams. OCR is not substituted for a mathematical edition. Archive metadata identifies these editions as `NOT_IN_COPYRIGHT`; the printed dates and translators were checked.

| Work / volume | Edition and contents | Bytes | PDF pages |
| --- | --- | ---: | ---: |
| [Euclid I](/texts/euclid-volume-1.pdf) | Thomas L. Heath, 1908: introduction and Books I–II | 16,777,247 | 449 |
| [Euclid II](/texts/euclid-volume-2.pdf) | Heath, 1908: Books III–IX | 8,442,899 | 456 |
| [Euclid III](/texts/euclid-volume-3.pdf) | Heath, 1908: Books X–XIII and later appendices | 8,998,878 | 575 |
| [Theophrastus I](/texts/theophrastus-volume-1.pdf) | Arthur Hort, 1916: Enquiry into Plants I–V | 21,374,540 | 522 |
| [Theophrastus II](/texts/theophrastus-volume-2.pdf) | Hort, 1916: VI–IX, then separate minor works | 29,539,492 | 520 |

Original Internet Archive items: [Heath I](https://archive.org/details/thirteenbookseu02heibgoog), [II](https://archive.org/details/thirteenbookseu00heibgoog), [III](https://archive.org/details/thirteenbookseu01heibgoog); [Hort I](https://archive.org/details/enquiryintoplant01theouoft), [II](https://archive.org/details/enquiryintoplant02theouoft). The confusing Heath item identifiers were checked against the actual title pages and contents, not assumed to match volume order.

Heath’s treatment of Books XIV–XV in volume III is a later appendix; these are not part of Euclid’s thirteen books and are not backdated into the reconstruction. Theophrastus’s English opening is at PDF page 39 of volume I. Volume II’s botanical text and its later appended minor works are distinct. The two complete botanical volumes total about 51 MB; loading them can take time on a slow connection.

All five PDFs opened without encryption. Representative pages from every volume were rendered and visually inspected, including Euclid’s geometric diagrams and Hort’s botanical English text. Review images are retained under `qa/text-corpus/`. This verifies legibility of sampled pages, not every scanned page’s perfect condition.

## Missing books: witnesses, not invented translations

The 28 missing records do not offer fabricated full texts. Each names a surviving witness or scholarly edition. “Fragmentary” includes quotations and papyri; some collections also preserve complete individual poems. Sappho’s collection is therefore not labeled wholly lost. “Lost” is used here for Euclid’s Porisms, Conics and Pseudaria, whose complete original texts are unavailable; later discussion or mathematical reconstruction is not presented as those works themselves.

The principal evidence groups are:

- **Callimachus:** [Susan A. Stephens’s scholarly introduction](https://dcc.dickinson.edu/callimachus-aetia/callimachus), [the Aetia introduction](https://dcc.dickinson.edu/callimachus-aetia/introduction), and [Suda kappa 227](https://cs.uky.edu/~raphael/sol/sol-cgi-bin/search.cgi?db=REAL&enlogin=&field=adlerhw_gr&login=&searchstr=kappa%2C227). The tenth-century Suda is a late witness. Pinakes is not asserted to be a recovered shelf catalogue.
- **Archaic lyric and epic:** [CHS’s Sappho selection](https://www-current.chs.harvard.edu/primary-source/sappho-sb/), D. A. Campbell’s *Greek Lyric I: Sappho and Alcaeus* (Loeb 142, 1982), Evelyn-White’s Hesiodic fragments, and [Proclus’s Chrestomathy in Durham’s Living Poets project](https://livingpoets.dur.ac.uk/w/index.php/Proclus%2C_Chrestomathy). Summaries of Epic Cycle plots do not preserve the full poems. Attributions are uncertain.
- **Fragmentary tragedy:** [P. J. Finglass’s Sophocles survey](https://www.cambridge.org/core/journals/new-surveys-in-the-classics/article/introduction/B74A6CC866CDD4D3B94D9ED23E3FF223), [David Fitzpatrick on Tereus](https://university.open.ac.uk/arts/research/pvcrs/2007/fitzpatrick), and Collard and Cropp’s *Euripides: Fragments* (Loeb 504 and 506, 2008–2009). Trackers, Hypsipyle and Phaethon are fragmentary, not wholly absent.
- **Philosophy:** [Hutchinson and Johnson’s Protrepticus research project](https://www.protrepticus.info/), [Theophrastus in SEP](https://plato.stanford.edu/entries/theophrastus/), and [Diogenes Laertius 9.46–49 on Democritus](https://www.perseus.tufts.edu/hopper/text?doc=Perseus%3Atext%3A1999.01.0258%3Abook%3D9%3Achapter%3D7). Scholarly reconstruction remains distinguishable from transmitted text.
- **Medicine:** Heinrich von Staden’s [Herophilus: The Art of Medicine in Early Alexandria](https://books.google.com/books/about/Herophilus_The_Art_of_Medicine_in_Early.html?id=rGhlIfJZkVoC) and [Galen’s report concerning Erasistratus’s On Fevers](https://www.aroonchande.com/files/am/Erasistratus.pdf), *On the Doctrines of Hippocrates and Plato* 396.3–33 De Lacy. Galen reports a dispute about interpolation; his testimony does not recover the treatise intact.
- **History and geography:** [Josephus, Against Apion I](https://penelope.uchicago.edu/josephus/apion-1.html), especially §§73–105 for Manetho and §§128–160 for Berossus; Brill’s scholarly *Jacoby* collection for Hecataeus **of Miletus**, not Hecataeus of Abdera.
- **Mathematics:** [Ivor Bulmer-Thomas’s Euclid article](https://mathshistory.st-andrews.ac.uk/DSB/Euclid.pdf), discussing later testimony from Pappus and Proclus. It does not establish exact physical holdings in 250 BCE.

Modern copyrighted scholarly editions are linked for verification, not copied into the reading collection. A text’s eventual disappearance is not assigned to a particular destruction of the Library.

## Navigation and verification conventions

`readingStart`, `readingEnd` and section offsets are UTF-16 character positions in the **raw UTF-8-decoded file**, before newline normalization or removal of boilerplate. Each anchor string is supplied alongside its offset. Reader “pages” are application pagination, not ancient roll divisions or printed-edition page numbers. Word counts are approximate modern translation counts, sometimes including editorial matter; they are not estimates of Library roll lengths.

The acquisition routine verifies 40 downloaded source files and produces 16 selected or combined reading files. Every source and generated asset has a byte/hash receipt in the manifest. Validation checks include unique catalogue IDs, survival and evidence enums, actual file existence, raw anchor offsets, expected book divisions, all five Pentateuch chapter totals, preserved Gutenberg notices, and the combined Herodotus boundary. The local test receipt is `qa/text-corpus/validation.json`; reader interaction checks are recorded separately in `qa/reader-playtest.md`.

The public-domain editions are useful reading access, not substitutes for current critical editions. Their vocabulary, classifications and interpretations can be outdated. Ancient medical, astronomical, ethnic and geographical claims are presented as historical texts, not current factual or practical advice. Wider historical evidence and the limits of reconstructing the Library are set out in [Historical basis](./historical-basis.md).