Britain at the Edge of the Oikoumenē

Abstract

Diodorus Siculus 5.21–22 is the earliest substantial description of Britain to survive as a continuous ancient text. Written in the first century BC, it is centuries later than the voyage whose information it appears to preserve; yet its geography, ethnography and tin narrative can be traced with high probability to the late-fourth-century BC investigations of Pytheas of Massalia, most plausibly through Timaeus of Tauromenium. In that qualified but important sense, it is the nearest thing Britain possesses to an intact opening chapter of written history before Caesar: not Pytheas’ lost book itself, but an unusually coherent survival of the information it put into Mediterranean scholarship. (Cunliffe 2001/2002, pp. vii–ix, 93–115) (Roller 2006, pp. 57–91) (Shipley 2024, pp. 231–248)

The passage is also a philological object. Books I–V survive through a recoverable medieval tradition whose four principal witnesses are D (Naples, ex-Vind. gr. 04), C (Vat. gr. 130), V (Vat. gr. 996) and L (Laur. Plut. 70.1), with later descendants, restoration and contamination. The argument below therefore keeps three questions connected but distinct: what Greek Diodorus wrote; where his information came from; and how far archaeology and palaeogeography can test the world described by that text. (Bertrac 1993, pp. LXXVII–CXXIII) (Casevitz & Jacquemin 2015, pp. XXXIII–XXXVIII)

1. Where Britain appears in Diodorus’ universal history

Diodorus did not begin the Bibliotheca historica with a year-by-year annal. His forty-book project first establishes the peoples, landscapes, origin traditions and geographical theatre in which chronological history will unfold. The surviving opening books are therefore thematic and geographical as well as historical. Book V calls itself the Book of Islands (νησιωτική): after treating islands associated with Libya and the western Mediterranean, Diodorus turns to Europe, Sicily, the western islands and finally the great island he calls Πρεττανική. The transition at 5.21.1 is explicit: having discussed the Ocean and its islands on the Libyan side, he “transfers the account to Europe.” Diodorus 5.1–21, LacusCurtius ↗ (Jacquemin in Casevitz & Jacquemin 2015, pp. IX–XXXI)

Sicily, islands and the shape of the work

This organisation matters for 5.21–22. Britain is not an isolated excursus inserted into Roman campaign history; it belongs to Diodorus’ geographical preparation of the inhabited world. He even postpones the fuller account of British customs until the chronological narrative reaches Caesar, while retaining here the material needed to locate and characterise the island. The triangular comparison with Sicily is therefore particularly apt in the work of a Sicilian Greek writing a universal history from the Mediterranean outwards.

Why a Sicilian Greek historian stands behind Pytheas

Diodorus came from Agyrion in inland Sicily, but “Sicilian Greek” was the product of a much older colonial and cultural history. From the eighth century BC onward Greek apoikiai—independent overseas settlements traditionally described as colonies—were founded around the Sicilian coasts alongside Phoenician and indigenous communities. By the Classical and Hellenistic periods this had produced a distinctive western Greek intellectual world. Diodorus himself remarks that Greek settlements in Sicily occupied much of the coast and that long interaction led to cultural and linguistic mixture among island populations (Bibliotheca historica 5.6.5). Diodorus 5.6.5 ↗

Timaeus of Tauromenium belongs squarely in that Siceliot tradition. Born in fourth-century BC Tauromenium, he wrote a major history of Sicily and the western Mediterranean and became one of the principal authorities through whom later historians knew the Greek West. Diodorus names and criticises Timaeus at the opening of Book V (5.1.3), which is important evidence that Timaeus was inside this book’s source environment. That makes the route Pytheas → Timaeus → Diodorus historically intelligible as well as textually economical: a Massaliot explorer’s Atlantic observations entered the literature of the Greek West, were absorbed by a Sicilian historian, and later reached another Sicilian historian assembling a universal library. (Mette 1952, pp. 1–4) (Roller 2006, pp. 72–74)

2. Greek text: Diodorus 5.21–22

The displayed text now labels its editorial basis instead of silently blending editions. 5.21 follows Friedrich Vogel’s Teubner recension as openly encoded by Perseus; 5.22 follows Michel Casevitz’s 2015 Budé text as edited at Casevitz & Jacquemin 2015, pp. 31–32. The distinction is visible because several details matter for interpretation: Casevitz prints Πρεττανία / Πρεττανική, κασσίτερος, and in 5.22.2 πῶρον rather than Vogel’s πόρον. (Casevitz & Jacquemin 2015, pp. 31–32)

5.21 digital base: Vogel / Perseus. 5.22 critical control: Casevitz & Jacquemin 2015, pp. 31–32. Public comparison links: Scaife, Perseus, LacusCurtius and ToposText.

3. A new Kassiteros translation

This translation tries to preserve Diodorus’ repeated reportative syntax—especially φασίν, infinitives in indirect discourse and participial chains—while remaining readable. It deliberately leaves several culturally loaded terms visible rather than smoothing them away: αὐτόχθονα is “autochthonous”; δυνάστας is “dynasts”; and ἐξημερωμένοι is rendered “tamed” or “civilized,” because the Greek metaphor matters to the ethnographic argument below. (cf. Casevitz & Jacquemin 2015, commentary pp. 113–357)

4. English translations and the modern Pytheas corpus

The passage has circulated through several different editorial objects: translations of Diodorus as a whole, editions of Book V, and reconstructions of Pytheas in which Diodorus is excerpted as indirect testimony. Those genres should not be conflated. The table therefore records what each work is actually doing rather than ranking translations on a single scale. (Oldfather 1933, pp. 153–159) (Scott 2022, F4–F5) (Shipley 2024, pp. 231–248, texts from p. 236)

Modern translations are cited selectively rather than reproduced in full. Scott, Shipley, Oldfather and Casevitz are therefore used for comparison and source criticism, while the continuous English text on this page is the Kassiteros translation; public-domain translations such as Booth remain useful for reception history.

5. The Greek text: manuscript tradition and Bertrac’s stemma

The textual problem begins well before any modern translation. For Books I–V, Pierre Bertrac’s reconstruction isolates four principal surviving witnesses conventionally labelled D, C, V and L. They do not simply constitute four equal copies: the early and restored strata of D differ, C is the only complete representative of its line, V was itself repaired from C, and L transmits an abbreviated, expurgated recension. Bertrac’s full treatment occupies pp. LXXVII–CXXIII of the 1993 CUF general introduction; the modern Book V edition by Casevitz works from this textual history. (Bertrac 1993, pp. LXXVII–CXXIII) (Casevitz & Jacquemin 2015, pp. XXXIII–XXXVIII)

A clickable reconstruction of the Bertrac stemma

The diagram below reconstructs the Books I–V relationships used in modern discussion after Bertrac. Horizontal rules now mark approximate manuscript dates; solid lines indicate normal descent, while dashed lines flag contamination or a less secure family association. The lines are guides to chronology, not claims that every copy can be dated to a single year. Click any manuscript box to jump to its full shelfmark and repository record.

The diagram is a functional reconstruction rather than a facsimile. Compare the published/Commons visualization derived from Bertrac: Diodorus Bibliotheca Stemma Codicum ↗ (Mellangoose, CC BY-SA 4.0). The underlying stemmatic argument remains Bertrac’s, not the graphic’s.

6. Manuscripts today: IIIF, reuse status and repository map

The article can now inspect three of the four principal transmission lines directly in-browser: L from the Biblioteca Medicea Laurenziana and C/V from DigiVatLib. The viewer uses each library’s own IIIF image service; no Vatican folios are bundled into this file. For C and V no exact 5.21–22 folio is asserted until it can be established independently; the tabs therefore expose the whole manuscript rather than inventing a target.

Are the principal manuscripts openly licensed?

Strict answer: among D/C/V/L, no complete principal witness is here documented with an complete primary witness carrying an explicit Creative Commons licence. L is openly accessible via IIIF and its Ictis crop is CC BY-SA; C and V are digitised but publication-restricted by the Vatican.

All 28 Books I–V witnesses

Where the stemma survives today

The map plots every repository location represented in the 28-witness stemma. Marker radius scales with the number of witnesses at each repository, so larger surviving concentrations are visible at a glance. Popups enumerate the manuscripts and link back to their records. The optional CARTO and Esri basemaps provide geographical context without altering the manuscript data.

Repository coordinates are approximate institutional/city points and are not intended as shelf-level geolocation. CARTO Positron base map; © OpenStreetMap contributors © CARTO.

7. Printed Greek: from Stephanus to Vogel–Fischer and Casevitz

The printed history of Book V begins later than the history of Diodorus in print. Opsopoeus’ Basel Greek edition of 1539 contained Books XVI–XX, not Book V. Henri Estienne (Henricus Stephanus), Geneva 1559, is the editio princeps for Books I–V. From there the text passed through Rhodoman, Wesseling, the nineteenth-century recensions of Dindorf and Bekker, the Teubner of Vogel/Fischer, and finally the modern Budé.

Page-link policy. A scan is linked directly to 5.21 only where the host exposes a durable page anchor. For several early books the viewer URL does not preserve a stable leaf/page target; those links deliberately open the digitised volume or catalogue rather than inventing a fragile locator. The text itself can always be located from Book V, chapter 21.

Why a new critical edition in 2015?

The surviving reviews and published paratext do not give us a personal manifesto in which Casevitz says that he set out to replace Vogel for one single defect. The documented explanation is more prosaic and more important. The Collection des Universités de France was producing a modern Book V; Bertrac’s 1993 reconstruction had sharpened the stemma of Books I–V; the four fundamental manuscripts could be weighted genealogically rather than simply counted; and Poggio’s Latin could be used more systematically as an indirect witness. Casevitz’s edition is therefore not merely a reprint with new punctuation but a new critical recension within the modern Budé programme. (Bertrac 1993, pp. LXXVII–CXXIII) (Casevitz & Jacquemin 2015, pp. XXXIII–XLVIII) (Blanc 2017, pp. 293–295)

What Vogel actually inherited and changed

Vogel’s title-page formula—editionem primam curavit Imm. Bekker, alteram Ludovicus Dindorf, recognovit Fridericus Vogel—is itself an editorial genealogy. His text revised the Bekker–Dindorf tradition through manuscript evidence and published collations rather than copying a single codex. He was also alert to contamination: Vat. gr. 995, later central to the Poggio problem, was already treated as a hybrid witness. The principal difference after Bertrac is methodological weighting: agreement between manuscripts matters less than whether those manuscripts represent independent branches. (Marcotte & Cohen-Skalli 2015, pp. 82–85) (Vogel 1888–1893, Praef. pp. XIV, XXII)

8. Pytheas behind Diodorus: a source chain that can be tested

The most economical reconstruction is Pytheas → Timaeus → Diodorus. It does not require Diodorus to have copied On the Ocean directly, and it explains why the passage contains information characteristic of Pytheas while Diodorus never names him here. Scott consequently treats Diodorus 5.21.3–4 and 5.22.1–4 as evidence for Pytheas; Roller likewise identifies the Massalian orientation of the Rhône route, the interest in tides and the treatment of strangers as Pythean signatures, while allowing Timaeus as Diodorus’ immediate source. (Scott 2022, F4–F5) (Roller 2006, pp. 72–73)

Why Timaeus is not a hypothetical name inserted for convenience

Diodorus explicitly names Timaeus in the opening of Book V (Diodorus 5.1.3). More decisively, Pliny explicitly attributes the British tin-island notice to “the historian Timaeus” at Naturalis historia 4.104. Pliny’s immediately surrounding geographical material also preserves Pythean Britain and Thule. In the amber dossier, Timaeus is again linked with material whose earlier authority is Pytheas. These are independent reasons to place Timaeus between the explorer and the first-century compilers. Pliny, NH 4.104 ↗ (Mette 1952, pp. 1–4)

Ictis / Mictis: one tin-island dossier, two transmissions

Diodorus 5.22 and Pliny NH 4.104 should not be treated as two unrelated reports that merely happen to resemble one another. They are two transmissions of the same Pythean–Timaean tin-island dossier. Diodorus preserves the more extended operational narrative: production around Belerion, smelting and casting, wagon transport at ebb tide, sale to merchants and the Gaul–Rhône route. Pliny compresses the notice under Timaeus, gives the transmitted island-name as Mictis, mentions hide-covered wicker boats, and attaches a six-day sailing figure.

The six-day sailing figure should not automatically be removed from the tin-island notice merely because Pliny places it beside his discussion of Thule. Mette’s influential reconstruction treats the number as a displacement from the neighbouring Thule material, and Cunliffe regarded such an interpolation as possible; that remains an important explanation, but it is not a manuscript fact. Pliny’s wording deserves to be tested on its own terms: Timaeus historicus a Britannia introrsus sex dierum navigatione abesse dicit insulam Ictim—“Timaeus the historian says that the island Ictis is six days’ sailing introrsus from Britain” (Pliny, Naturalis historia 4.104 ↗). Introrsus does not literally mean “along the coast”: Lewis & Short’s basic sense is “towards the inside, inwards, into”. But neither can it sensibly mean “inland” here, because Pliny explicitly defines the distance by navigatione. In a striking maritime parallel at NH 2.219 Pliny contrasts the Pontus flowing extra into the Propontis with its never flowing introrsus in Pontum; there the opposition is between an outward and an inward direction within connected seas, not between sea and dry land. In geographical descriptions the same adverb can mark the landward/interior side of a coast (e.g. NH 6.169). The most economical sense in 4.104 is therefore directional: sailing on the inward or landward side of Britain’s oceanic frontage, rather than outward into the open Atlantic. A coastwise or Channel-facing voyage is compatible with that sense, although “alongshore” is an interpretation of the geography, not the dictionary meaning of the word. (Lewis & Short 1879, s.v. introrsum/introrsus)

This leaves open a possibility that deserves to be retained rather than emended away: the six days may have belonged to Pytheas’ actual tin itinerary. If the fuller source behind Timaeus began not from an abstract point called “Britain” but from the south-western tin-producing zone represented by Belerion—Cornwall and Devon in the Diodoran account—the number could describe a coastal voyage from the region of production to a more easterly region of sale, including the Wight–Solent zone. Pliny has compressed the point of departure to a Britannia, so this cannot be demonstrated from his sentence alone; nor should a precise route be reverse-engineered from an ancient “day’s sail” without assumptions about vessel, weather and stopping practice. Yet the hypothesis has two advantages: it preserves the six-day datum instead of declaring it corrupt, and it gives introrsus a coherent landward/coastwise orientation within the same production–exchange system that Diodorus describes. The older Wight historiography had already noticed that Pliny could be read as a six-day voyage “along the British coast”; the present argument is more cautious, treating that as a testable geographical interpretation rather than a lexical translation. (Mette 1952, pp. 3–4, 31–32) (Cunliffe 2001/2002, pp. 74–76)

What about Posidonius?

Chronology sharply limits what Posidonius can explain. Pytheas’ voyage and On the Ocean belong most plausibly to the 320s BC; Timaeus of Tauromenium (c.350–260 BC) was already using Pythean northern material in the early Hellenistic period; Posidonius of Apamea lived only c.135–c.51 BC—roughly two centuries after Pytheas and more than a century after Timaeus. Pliny explicitly names Timaeus as the authority for the Ictis/Mictis tin-island notice at NH 4.104. Whatever later reshaping Diodorus may have received, Posidonius therefore cannot be the origin of information that had already entered Timaeus’ work. (Baron 2013, esp. pp. 1–16) (Kidd 2016, s.v. Posidonius)

This chronological distinction also clarifies the ethnography. The classic Posidonian reconstruction is strongest for Diodorus’ extended account of the continental Gauls in 5.25–32, where close parallels with separately attested Posidonian fragments make substantial dependence probable, although modern scholarship rightly resists treating the whole block as a verbatim fragment. Britain in 5.21–22 is different: its measurements, northern scientific observations and the Timaean tin-island tradition point behind Timaeus to Pytheas. A useful working distinction is therefore Pytheas for the earliest British ethnology; Posidonius as a major, later source for the ethnography of the continental Gauls, while allowing Diodorus to combine and rewrite more than one source. (Tierney 1960, pp. 189–275, esp. 198–207) (Lampinen 2014, pp. 229–259, esp. 243–244) (Nash 1976, pp. 111–126)

Pytheas, Aristotle’s Lyceum and the strategic horizon of Alexander

Roller places Pytheas inside the scientific world of Aristotle’s generation rather than treating him as an isolated merchant-adventurer. The connection is circumstantial but unusually suggestive. Aristotle’s Lyceum drew investigators from across the Greek world; Aristotle wrote a lost Constitution of the Massalians (attested by Athenaeus 13.576) and discusses Massalian institutions in Politics 5.5.2 and 6.4.5. Demosthenes 32, a case involving Massalian merchants and an Athenian maritime loan, independently shows regular Massalia–Athens contact in the middle fourth century. Roller therefore suggests that Pytheas could have visited Athens or belonged to the wider information network around the Peripatos, and even asks whether he may have supplied Massalian information to Aristotle. His scientific vocabulary and interests also sit naturally beside Eudoxan and Peripatetic work. The important negative evidence is that Aristotle’s surviving discussions of Arctic phenomena do not appear to know Pytheas’ later discoveries; if there was contact, Roller argues, it was probably before the voyage and before On the Ocean circulated. (Roller 2006, pp. 63–65) (Shipley 2024, pp. 233–234)

Roller also reconstructs a political background that may have made Atlantic reconnaissance newly urgent. In the 330s and 320s the western Mediterranean watched Alexander III’s expansion with uncertainty. Arrian says that Alexander contemplated further expeditions, including—in one formulation—Britain (Anabasis 7.1.4), and western embassies came to him in 323 BC; Roller asks whether Massalia had reason to discover how the Atlantic and northern routes fitted into the strategic geography of a possible western campaign. He even raises the possibility that reconnaissance might have helped answer whether Alexander could approach the western Mediterranean by an unexpected route. This should not be turned into a claim that Pytheas was an Alexandrian agent: Roller himself stresses that the journey’s broad, leisurely scientific programme does not look like urgent military intelligence, while Shipley calls the proposal that Alexander sent him “excessively speculative”. The defensible conclusion is contextual: fears about Alexander may have supplied one catalyst for a voyage whose surviving results are overwhelmingly scientific and geographical. (Roller 2006, pp. 58–60, 63–65) (Shipley 2024, p. 234 n.20)

9. Measuring Britain: Pytheas as observational scientist

Pytheas matters not only because he travelled farther north than earlier Greek geographical writers but because he attempted to turn travel into measured, communicable knowledge. Even Strabo, one of his most hostile readers, concedes at 4.5.5 (C201) that Pytheas appears to have handled celestial phenomena and mathematical theory competently. The surviving fragments associate him with the position of the north celestial pole, gnomon observations, longest-day and winter-sun measurements, tides, sailing times and high-latitude atmospheric or marine phenomena. (Shipley 2024, pp. 233–248)

Scientific language at the edge of experience

Roller draws attention to Pytheas’ effort to describe phenomena for which ordinary Mediterranean vocabulary was inadequate. The northern summer night is transmitted with unusually precise wording; rare or innovative expressions include προσάρκτιον (“toward the Bear”, northerly) and παρωκεανῖτις (“along the ocean coast”). Pliny’s Naturalis historia 2.186 preserves the striking Latin phrase angusto lucis ambitu, a “narrow circuit of light”. Roller notes that this could reflect auroral light, but continuous or near-continuous high-latitude summer twilight is another possibility: the source is evidence for acute observation, not a licence to identify the aurora with certainty. (Roller 2006, pp. 71–72)

The same scientific profile appears in other fragments. Hipparchus, Commentary on Aratus and Eudoxus 1.4.1, credits Pytheas with a geometrically described north-pole position. Aëtius 3.17.3 attributes to Pytheas the connection of ebb and flood with the waxing and waning Moon. Strabo 2.1.18 (C75) preserves northern solar-height and longest-day data through Hipparchus. These are not the observations of a passive collector of travellers’ tales; they show a writer operating in the scientific vocabulary and measurement culture of later fourth-century Greek geography. (Mette 1952, pp. 13–15) (Shipley 2024, texts 1–4, 25)

Where did the 42,500 stadia come from?

Diodorus 5.21.4 gives Britain three sides of 7,500, 15,000 and 20,000 stadia, totalling 42,500 stadia. Strabo 2.4.1 (C104), while transmitting Polybius’ attack on Pytheas, says that Pytheas reported the accessible circuit of Britain as “more than 40,000 stadia”; Pliny NH 4.102 gives a Pythean circuit in Roman miles. The numerical family is close enough to make common Pythean ancestry probable, while the discrepancies warn us that later conversion, rounding and epitome intervened. (Cunliffe 2001/2002, pp. 95–99)

No surviving fragment says that Pytheas physically measured every bend of the coastline. A more realistic reconstruction combines coastal sailing times, route distances, bearings and anchorages, subsequently regularised into stadia by Pytheas or by geographers using him. His astronomical measurements constrain latitude; they cannot directly yield the length of a winding coast or longitude. This distinction matters when Diodorus’ geometrical triangle is compared with a modern map.

What the gnomon evidence actually says

A gnomon is a vertical staff of known height g on level ground. At local solar noon its shadow has horizontal length s. The local right triangle gives the Sun’s altitude α = arctan(g/s); its zenith distance is z = 90° − α. On a spherical Earth this angle is related to geographical latitude φ by the Sun’s declination δ: at meridian transit, z = |φ − δ|. At the equinox δ ≈ 0°, so φ ≈ z; near the northern summer solstice, for Massalia, φ = z + ε, where ε is the obliquity of the ecliptic. This is a latitude measurement. “Meridian passage” here means the Sun crossing the observer’s local north–south plane; it does not provide longitude. (Dicks 1960, pp. 178–183) (Cunliffe 2001/2002, pp. 60–62)

Massalia as Pytheas’ latitude datum

What the sources securely preserve is not a Pythean longitude system but a Massaliot latitude observation. Strabo 1.4.4 (C63), reporting Hipparchus, says that the parallel through Byzantium was taken to be the same as the parallel through Massalia because the gnomon-to-shadow ratio stated by Pytheas at Massalia agreed with the ratio Hipparchus assigned to Byzantium at the same season. Strabo 2.1.12 (C71–72) then shows Hipparchus using that parallel as one element in a much later geographical grid. The historical sequence matters: Pytheas supplied an observational latitude datum; Hipparchus integrated it into mathematical geography. (Shipley 2024, pp. 238–239, texts 2–3) (Dicks 1960, pp. 178–183)

The celebrated numerical ratio 120 : 41⅘ should also be stated with care. Strabo gives it at Geography 2.5.41 for the summer-solstice shadow on the Byzantium/Massalia parallel; at 1.4.4 he says that Hipparchus’ Byzantium ratio matched the ratio Pytheas reported at Massalia. Thus the number belongs to the later transmitted/calculated geographical apparatus, even though it is tied to a Pythean observation. Using it illustratively, α = arctan(120/41.8) ≈ 70.795°, so z ≈ 19.205°; adding a fourth-century-BC summer-solstice declination of about 23.75° gives φ ≈ 42.95° N, close to Massalia’s true latitude. The agreement explains why the datum was scientifically valuable, without pretending that every decimal is a direct field measurement by Pytheas. (Strabo 1.4.4; 2.5.41) (Shipley 2024, p. 238 n.37)

Winter sun heights and longest days

Strabo 2.1.18 (C75) preserves, through Hipparchus, a northward sequence in which the midwinter Sun rises to about 9, 6, 4 and fewer than 3 cubits, together with longest-day values. Shipley explains the astronomical convention and the way such observations define latitude. The figures cannot all automatically be assigned to direct autopsy by Pytheas; some may have been transformed, interpolated or calculated by later mathematical geographers. Their importance is that Pytheas’ northern itinerary entered precisely the quantitative framework in which Hellenistic geography constructed parallels of latitude. (Shipley 2024, pp. 238–241)

10. Philological pressure points

11. Interactive Leipzig-style interlinear gloss

The interlinear is organised to follow the Leipzig Glossing Rules while respecting the fact that Ancient Greek is highly fusional. Each orthographic Greek word is aligned with an academic romanisation, a direct English word-gloss, and a supplementary morphology stack. The latter expands grammatical information without inventing morpheme boundaries that are not visible in the Greek. Leipzig Glossing Rules ↗

Morphological key