P.Malon. inv. 538: Apollonius Rhodius, Argonautica V

Edited by Eumelos Malonius · 27 September 2026

Editio princeps of three papyrus fragments from Malona, Rhodes, attributed to a fifth book of the Argonautica that belonged to the poet’s Rhodian edition.

1. Discovery and provenance

The fragments came to light in September 2026 in a village house at Malona, on the east coast of Rhodes. A prospective buyer’s notary had required a structural survey of the north wall. Behind a later lime render the surveyor opened a small cavity, about 30 cm deep, and found in it a lidded jar of coarse local ware. The jar held a roll that had been folded rather than rolled, and it was broken along the folds into three pieces. To the editor’s knowledge, P.Malon. inv. 538 is the first literary papyrus in history brought to light by Greek conveyancing.

The owner of the house is the present editor, which he declares at once. He can therefore vouch for the find-spot, and he is aware that this vouches for nothing else.

Papyrus does not survive in the soil of Rhodes, and the editor does not pretend otherwise. The Derveni papyrus survived by being burnt. P.Malon. 538 appears to have survived by being forgotten: the jar was sealed with pitch, the cavity faced north, and the wall was never damp enough to matter. Anyone who has spent August in Malona will find the last point easier to believe than the first two.

The number follows that of the Ennian fragment already published from the same collection (P.Malon. inv. 537). The two pieces share a find-spot, an editor and, some may feel, a certain quality of evidence.

2. The Rhodian edition and the case for a fifth book

The obstacle is the end of Book 4. The poet takes leave of his heroes (4.1773–81) and lands them, gladly, on the beaches of Pagasae:

ἀσπασίως ἀκτὰς Παγασηίδας εἰσαπέβητε.

gladly you stepped ashore on the beaches of Pagasae. (4.1781)

After a line like that, one does not simply turn the page. Any fifth book has to explain how the poem continued after it had so plainly stopped.

The explanation offered here rests on three pieces of ancient evidence, all genuine.

1. The Lives. The two Vitae transmitted with the scholia say that Apollonius, after an unsuccessful recitation at Alexandria, withdrew in shame to Rhodes. There he revised the poem, was well received, and was given citizenship, which is why he is called “Rhodian”.
2. The proekdosis. At a handful of places in Book 1 (e.g. 1.285, 516, 801) the scholia quote readings from a προέκδοσις, an earlier edition. So the poem did exist in more than one form.
3. The unfinished story. Every earlier account took Jason home to Pelias. Pindar calls Medea “the death of Pelias” (Pyth. 4.250). Stesichorus wrote a whole poem on the funeral games (Athla), and those games were carved on the Chest of Cypselus (Paus. 5.17.9–11). The Argonautica as we have it leaves all of this out.

The hypothesis is this. The four-book poem is the Alexandrian text, which ends at Pagasae because a court poet does not end on regicide in front of a king. On Rhodes, among friends and far from the Museum, Apollonius went on: to Iolcus, to the cauldron, to the games, and to the rotting ship. The Alexandrian tradition kept the four-book text and let the fifth book go. It survived, if at all, only on the island where it was written.

The editor admits that this hypothesis is, in the technical sense, too good to be true. He offers it anyway, on the principle that a papyrus should be allowed to speak before it is cross-examined.

3. The papyrus

Three fragments of a single roll, written along the fibres. The back carries a list of oil measures in a later cursive, which may be the earliest record of olive production at Malona. It is at least the most reliable part of the find.

Hand. A practised, rounded book hand, bilinear except for φ and ψ, with no ligatures. It compares well with hands of the later third century BC, such as the Lille Stesichorus. On that dating the copy was made within the poet’s lifetime or soon after. The editor records this and draws no conclusion, since the conclusion would draw itself.

Layout. Scriptio continua, with no accents or breathings. Iota is written adscript (ΦΩΝΗΙ, fr. 2.2). One verse per line. The column held about 28 lines, going by the margins of fr. 1 and fr. 2 and the likely order of events.

Second hand. A later reader corrected fr. 1.1 and wrote ζη( ), i.e. ζή(τει), “query”, in the left margin of fr. 2.3. This is the note a scholar wrote when he doubted a reading. Here it stands against the line in which the Argo’s beam promises to carry Jason away. The ancient reader seems to have shared the modern editor’s doubts. Whether he doubted the reading or the whole book, the abbreviation does not say.

Order. The narrative order of events fixes the order of the fragments: the killing of Pelias (fr. 1), then Jason’s death (fr. 2). Fr. 3 cannot be placed. A Rhodian aition could come anywhere, and the editor would like it to come last.

4. Sigla and conventions

Each fragment is printed in two forms. The diplomatic transcript shows what the papyrus has, in capitals, with gaps. The edited text below it is accented, divided into words, and supplemented. The apparatus follows the edited text, then a translation, then the commentary.

5. Fr. 1: Pelias’s daughters and the ram

Transcript

  αρνειον (Π², above the line)
1 [[ΚΡΙΟΝ]]ΔΟΤΕΔΗΠΡΕΣΒΥΝΜΕΛΕΙΣΤΙΤΑΜΟΥΣΑ
  ΧΑΛΚΕΙΩΙΚΑΤΕΘΗΚΕΛΕΒΗΤΕΝΙΠΑΜΦΑΝΟΩΝΤΙ
  ΕΚΘΟΡΕΝΑΡΝΕΙΟΙΟΝΕΟΝΤΕΚΟΣΑΙΔΕΣΙΔΟΥΣΑΙ
4 ΚΟΥΡΑΙΘΑΜΒΗΣΑΝΜΕΓΑΔΗ̣[

Text

Apparatus. 1 ἀρνειόν Π² s.l.: κριόν Π^ac (unmetrical) ‖ 4 ἤ̣[πυσαν ed. pr.: ἴ[αχον, the Homeric formula, is possible, but the trace suits η better than ι

Translation. When she had cut the old ram limb from limb and set it in the gleaming bronze cauldron, out leapt the ram’s own new offspring. The girls, seeing it, were astonished, and loudly they cried…

Commentary

1 ἀρνειόν | κριόν. The first scribe wrote the word he knew from prose handbooks of myth, where the animal is always a κριός (so Apollod. 1.9.27). The corrector restored the epic word, which the metre requires. The mistake says something about the copyist’s reading. It says nothing about the poet, who knew his Homer (ἀρνειός, e.g. Od. 10.527).

πρέσβυν. Used of an animal, the adjective has no parallel. It looks ahead: the ram is an old male of the flock, and the next old male into the pot will be Pelias. The whole scene is a rehearsal, and the adjective gives the game away to everyone except the daughters.

μελεϊστὶ ταμοῦσα. The adverb comes from the Cyclops’s kitchen: τοὺς δὲ διὰ μελεϊστὶ ταμὼν ὡπλίσσατο δόρπον, “he cut them limb from limb and made his supper” (Od. 9.291). Medea cooks like Polyphemus but calls it medicine. Apollonius’s readers were meant to notice, and Pelias’s daughters were not.

2 λέβητ’ ἔνι παμφανόωντι. In epic, a gleaming cauldron is above all a prize at funeral games (Il. 23.259, 267). If the fifth book went on to describe Pelias’s games, as Stesichorus did, the cauldron may well have turned up again as a prize, scoured. The editor offers this in a spirit of hope.

3 ἀρνειοῖο νέον τέκος. The ram comes out of the pot as its own offspring: father and son in one animal, and the generation gap closed by boiling. Ovid gives the same moment more space, and a lamb that skips off looking for an udder (Met. 7.312–21). Apollonius gives it three words and a genitive.

4 κοῦραι. The daughters are not named, and the reason is metrical: Πελιάδες starts with three short syllables and cannot stand in a hexameter. The anonymity nevertheless suits them. The poem will not name women who are about to be tricked into killing their father. Later tradition excuses one of them, Alcestis, who is said to have held back, but this fragment offers no exceptions.

6. Fr. 2: The death of Jason

Transcript

1  ΑΛΛΟΤΕΔΗΠΡΥΜΝΗΣΥΠΟΔΟΥΡΑΤΙΚΕΚΛΙΤΙΗΣΩΝ
   ΔΗΤΟΤΕΔΩΔΩΝΙΣΦΗΓΟΣΠΑΛΙΝΕΚΛΑΓΕΦΩΝΗΙ
   —— (paragraphos)
   ΣΧΕΤΛΙΕΚΑΙΝΥΝΑΥΤΙΣΕΓΩ̣[          ζη( ) in left margin, Π²
4  ]Υ̣ΣΤΑΤΟΝΗΜΑΡ
   (lower margin)

Text

Apparatus. 2 paragraphos below the line, marking the start of speech ‖ 3 ζη( ) in the left margin, Π² ‖ 3–4 σὲ κομίσσομαι … ἤλυθεν e.g. ed. pr. (space for about 30 letters) ‖ 4 ἤλυθεν: ἤνυσεν also possible, but see the commentary

Translation. But when Jason lay beneath the timber of the stern, then the Dodonian oak cried out once more: “Wretch, even now I myself will carry you again, [to the house of Hades, since] your last day [has come].”

Commentary

1 πρύμνης ὑπὸ δούρατι. In the version found in later handbooks, Jason dedicated the Argo at the Isthmus and was lying asleep beneath it when a piece of the stern fell and killed him. Medea predicts exactly this at Eur. Med. 1386–7: Ἀργοῦς κάρα σὸν λειψάνῳ πεπληγμένος, “your head struck by a relic of the Argo”. The fragment keeps the setting and gives the relic a voice.

κέκλιτ(ο). The pluperfect of κλίνω means “lay”. It is the ordinary word for resting, and also for the laying-out of the dead. Here Jason is doing the first and about to need the second.

2 Δωδωνὶς φηγός … πάλιν. Athena built into the keel a beam from the oak of Dodona (1.526–7). The ship cried out once at the launch (1.524–5) and spoke with a human voice in Book 4 (4.580–91). πάλιν, “once more”, points back to those occasions. The poem’s only oracle is also its only piece of furniture that talks, and it gets the last word on the hero.

ἔκλαγε φωνῇ. κλάζω is used of birds, dogs and prophets: of anyone, in fact, who makes a sharp sound that the listener would prefer not to hear. Iota adscript in ΦΩΝΗΙ is correct and is one of the few things in this edition that the editor will defend without reservation.

3 σχέτλιε. A rebuke: “wretch”, “you stubborn man”. The beam blames Jason as it kills him. He broke his oaths to Medea, and in Book 4 the same beam spoke of the wrath of Zeus.

κομίσσομαι. Read as a short-vowel aorist subjunctive with future sense, a Homeric usage (cf. ἐρύσσομεν, Il. 1.141). The verb means both “carry” and “carry off”. The ship that carried Jason out now carries him off. This line is the one that bears the marginal query. The ancient reader’s doubt could concern the form, the pun or the whole book, and the editor shares it at every level.

4 εἰς Ἀΐδαο δόμους, ἐπεὶ ἤλυθεν. The supplement is exempli gratia. It fills the space, scans as a line of five dactyls (quick, like a falling timber), and supplies the sense the line needs. ἤνυσεν, “has run its course”, would also fit. ἤλυθεν is preferred because the day comes to Jason as he sleeps; he does nothing to bring it on.

ὕστατον ἦμαρ. The column ends here, and on the evidence so far so does the hero. The lower margin is blank.

7. Fr. 3: Lindos and Malona

Transcript

1  ]ΡΗΣΙΝΕΠΑΚΤΗΙ[
   ]ΝΛΙΝΔΟΙΟΚΟΛΩΝ[
   ]ΑΙΜΑΛΩ .[
4  ]ΑΡΗΙΣΙΒΟΟΣ[

Text

Apparatus. 1 πέτ]ρῃσιν e.g. ed. pr. ‖ 2 κολων[αῖς vel κολών[ης ‖ 3 Μαλώ[νη- ed. pr.: μᾶλ’ ὤ[ (Doric μᾶλα, “apples”) a sceptic: κ]αὶ μάλ’ ὤ[ another sceptic, with more reason ‖ 4 ἀρῇσι: ΑΡΗΙΣΙ Π, with iota adscript

Translation. … [rock]s on the shore … / … of the hill of Lindos … / … and Malona … / … with curses … of an ox …

Commentary

2 Λίνδοιο κολων[. Lindos, with its acropolis on a hill above the sea. Together with the curses and the ox in line 4, the name points to only one story. Heracles came ashore at Lindos, took an ox from a ploughman, and ate it while the owner stood by and cursed him. The Lindians afterwards sacrificed to Heracles Bouthoinas (“ox-feaster”) with curses instead of prayers (Callim. Aetia frr. 22–3 Pf.; Apollod. 2.5.11). Callimachus told this story in the Aetia. If Apollonius retold it in an edition revised on Rhodes, he was retelling his rival’s Rhodian story on Rhodes. That would be either an act of homage or an act of war. With Apollonius and Callimachus there is seldom any third possibility.

Heracles left the Argonauts in Book 1, so the story cannot be part of the voyage. It is best taken as the poet speaking in his own person, perhaps in a closing sphragis, where he signs the poem from his new home. This is why the editor would like the fragment to stand last.

3 Μαλώ[νη-. If this is right, it is the first attestation of the village of Malona, some 2,200 years before the next one. The ancient name of the place is otherwise unknown. The reading therefore cannot be checked, and the editor finds this a great convenience.

Two sceptics have come forward. The first reads μᾶλ’ ὤ[, with Doric μᾶλα, “apples” (epic μῆλα). This gives Malona an etymology that nobody has proposed before. It also has a certain local truth, since μῆλον came to cover the citron (μῆλον Μηδικόν), and Malona lives by its citrus groves. The second sceptic reads κ]αὶ μάλ’ ὤ[, “and indeed, O …”, the dullest reading available. It is also the most probable. The editor rejects it on grounds of sentiment and says so.

4 ἀρῇσι βοός. ἀρή is the epic word for a curse. The curses of the Lindian ploughman became, by a characteristically Greek piece of logic, the correct liturgy. Anyone who has tried to buy or sell a house on Rhodes will recognise the principle.

Works cited

• Davies, M. & Finglass, P. J., Stesichorus: The Poems (Cambridge 2014)
• Fränkel, H. (ed.), Apollonii Rhodii Argonautica (Oxford 1961)
• Hunter, R., Apollonius of Rhodes: Argonautica Book IV (Cambridge 2015)
• Malonius, E. (ed.), P.Malon. inv. 537: a new fragment of Ennius (Malona 2026)
• Mastronarde, D. J., Euripides: Medea (Cambridge 2002)
• Pfeiffer, R. (ed.), Callimachus, vol. 1 (Oxford 1949)
• Vian, F. & Delage, É., Apollonios de Rhodes: Argonautiques, 3 vols (Paris 1974–81)
• Wendel, C. (ed.), Scholia in Apollonium Rhodium vetera (Berlin 1935)