ARCHIMEDES AT SYRACUSE: TWO NEW WITNESSES TO CASSIUS DIO'S ROMAN HISTORY 15 (TZETZES’ CARMINA ILIACA AND HYPOMNEMA IN S. LVCIAM)

Abstract

Cassius Dio's fragmentary Roman History 15 contains an account of Archimedes’ role in defending Syracuse during the Roman siege of 213–212 b.c., incorporating a legendary tale about a solar reflector Archimedes constructed to burn Roman warships, and including details of his death when the city fell. The textual basis of this famous episode depends on two derivative twelfth-century works: Zonaras’ Epitome of Histories (9.4–5) and Tzetzes’ Chiliades (2.35). After clarifying the present state of enquiry, this paper introduces two new witnesses, overlooked by editors of Dio and extensive scholarship on Archimedes, and assesses their value for reconstructing Dio's text. Comparative analysis of corresponding Dio-derived material in Tzetzes’ Carmina Iliaca and Hypomnema in S. Luciam, especially verbal correspondences with Zonaras’ Epitome, demonstrates that they are independent and, sometimes, superior witnesses to Dio's wording and content, reflecting Tzetzes’ selective use of the Roman History in different verse and prose compositions over several decades. The study considers editorial implications for this section of Dio's work and general characteristics of Tzetzes’ writings as repositories of testimonia and fragments.

A remarkable florescence of interest in Cassius Dio has recently illuminated historiographic and literary dimensions of his Roman History.¹ Editorial questions, perhaps understandably, have attracted less attention. The basis of enquiry remains Boissevain's edition (1895–1901), magnificent but inevitably defined by contemporary knowledge and methodologies, particularly regarding fragmentary books transmitted via excerpts, epitomes and indirect traditions.² Concurrently, transformative scholarship on Byzantine literature, investigating texts as literary-cultural artefacts rather than mere vehicles of classical transmission, is shedding light on responses to Dio's narrative of the Roman past. Although never a school text, selection for Constantine VII's project of historical excerption (940s/950s) implies esteem or even quasi-‘canonical’ status.³ Particular interest emerges in the eleventh century with John Xiphilinus’ partial Epitome (c.1071–1075), long utilized as a textual witness, now studied also in terms of compositional creativity.⁴ In the twelfth century, John Zonaras and John Tzetzes variously adapted Dio's work, including books since lost. While Zonaras’ Epitome of Histories has rarely lacked scholarly attention, recent studies progress beyond Quellenforschung to sophisticated analyses of literary and stylistic priorities.⁵ Similarly, innovative engagement with Tzetzes’ long-neglected writings, partly aligned with a fundamental rethinking of Byzantine poetry, is reappraising his aims, achievement and milieu, while affording opportunities to re-examine and discover testimonia and fragments embedded in his vast and diverse œuvre.⁶ Although Boissevain and his predecessors adduced Tzetzes’ Chiliades (Historiarum variarum Chiliades or correctly Histories), by far his best-known work, as an indirect witness to Dio's Roman History, corresponding material in Tzetzes’ Carmina Iliaca and Hypomnema in S. Luciam has been entirely overlooked. This paper introduces these two texts to classical researchers and assesses their value for reconstituting part of Roman History 15 that has not survived in a direct tradition, especially in light of verbal correspondences they exhibit with Zonaras’ Epitome. This study will demonstrate that Tzetzes’ scholia to his Carmina Iliaca and selected passages of the Hypomnema are independent and, sometimes, superior witnesses to Dio's wording and content, reflecting Tzetzes’ differing use of Dio's work in distinct verse and prose projects over several decades.

These discoveries are all the more surprising as the events in question have been intensively studied: the siege of Syracuse by Roman forces under Marcus Claudius Marcellus in 213–212 b.c., and particularly Archimedes’ involvement in defending his native city, the most famous historical episode in his biography.⁷ The siege has attracted inordinate interest largely owing to a fabulous tale of how Archimedes devised one or more mirrors to concentrate the sun's rays and set Roman vessels ablaze. Some brief remarks on the origin and evolution of this story are necessary to locate Dio's Roman History in this literary tradition.⁸ Unsurprisingly, near-contemporary sources and earlier derivative accounts, while reporting Archimedes’ other defensive contraptions, do not mention solar reflection or any such naval disaster.⁹ The myth of Archimedes’ burning-mirror (or mirrors) first emerges as incidental allusions in unconnected texts around the mid second century a.d., possibly indicating its genesis in Hellenic revivalist culture of the Second Sophistic.¹⁰ Against this recent background, Dio is the earliest historian known to have mentioned a burning-mirror at Syracuse. The legend essentially conflates unhistorical reports of technological inventions with actual scientific writings on catoptrics, a branch of geometrical optics concerned with reflective properties and applications of plane and concave/convex mirrors, including combustion.¹¹ Late antique authors investigated and embroidered this fictive episode.¹² The story drew renewed interest from eleventh- and twelfth-century Byzantine scholars, evincing broader intellectual currents in Constantinople, and continuing Archimedes’ transformation from mathematician-engineer to semi-divine miracle-worker.¹³ Prominent in this tradition are Dio-derived accounts by Tzetzes and Zonaras.

CASSIUS DIO ON THE SIEGE OF SYRACUSE: TZETZES’ CHILIADES AND ZONARAS’ EPITOME

Scholarship has long reconstituted Dio's lost account of the siege of Syracuse (15 F57.35) from indirect textual traditions transmitted in Tzetzes’ Chiliades and Zonaras’ Epitome.¹⁴ Both authors cite Dio as a source for this event or constituent episodes. As each witness drew on Dio's work independently, their concurrence, in wording and/or substance, provides mutually corroborative evidence for the form and content of the original text.¹⁵ The testimonies of Tzetzes and Zonaras particularly coincide regarding two aspects: first, mechanical devices that Archimedes constructed to repel Marcellus’ forces, including a solar reflector; second, how Archimedes met his death. Their texts are juxtaposed below with verbal parallels marked in bold.

Tzetzes more clearly signals the provenance of this material. In a protracted, multi-stage process from the mid 1150s, Tzetzes composed the Chiliades as a verse commentary to his self-edited letter collection.¹⁶ As throughout his œuvre, he asserts ‘scientific’ learning in frequent, if fanciful, references to celebrated ‘mechanical writers’ (μηχανογράφοι) or ‘mechanicians’ (μηχανικοί).¹⁷ The figure most often mentioned is Archimedes, whose achievements exercised a peculiar fascination, none more so than his alleged construction of burning-mirrors.¹⁸ Tzetzes also implies familiarity with Archimedes’ writings on catoptrics.¹⁹ Tzetzes devotes Chil. 2.35 to ‘Archimedes and some of his machines’ (Περὶ Ἀρχιμήδους καὶ τινῶν αὐτοῦ μηχανῶν), which primarily concerns his role in defending Syracuse (verses 112–31) and the circumstances of his death (134–51). In a concluding ‘source notice’ (152–9), Tzetzes adduces Dio and Diodorus as historical authorities for Archimedes’ life and works (152 ὁ Δίων καὶ Διόδωρος γράφει τὴν ἱστορίαν), alongside four named—and other unnamed—mathematical authors: Anthemius, Heron, Philo and Pappus (153–9). The relevant sections of Dio and Diodorus are lost. Tzetzes apparently consulted both historians, as he indicates that their accounts differed regarding how Syracuse fell: either, as Diodorus, by treachery (134–5) or, as Dio, by a nocturnal attack as the Syracusans celebrated a festival (136–7), though Tzetzes seemingly conflates two stages of a single narrative: the seizure of the outer city during an all-night festival of Artemis and the subsequent capture of the citadel of Achradina by treachery (Polyb. 8.37; Livy 25.24–31; Plut. Vit. Marc. 18). Moreover, throughout the Chiliades, Tzetzes frequently cites Dio and Diodorus, in combination or individually.²⁰ Comparison with other witnesses to Dio's text, direct and indirect, suggests that Tzetzes had recent access to the Roman History, despite his claims to be writing ‘bookless’ (ἀβίβλης), a literary-rhetorical motif that simultaneously excuses inaccuracies and asserts prodigious mnemonic powers.²¹

Tzetzes drew the technical description of Archimedes’ complex hexagonal mirror (121–30) from a third source: a treatise on geometrical optics entitled On Paradoxical Mechanisms (Περὶ παραδόξων μηχανημάτων) by the sixth-century geometer, physicist and architect Anthemius of Tralles.²² Tzetzes seemingly acknowledges his indebtedness in the terminal source notice to Chil. 2.35: ‘many mention Archimedes, Anthemius the paradoxographer foremost’.²³ The surviving truncated text of On Paradoxical Mechanisms comprises three optical problems.²⁴ In the second, ‘How shall we cause combustion by means of the sun's rays in a given position that is not less than a bowshot away?’, Anthemius adduces Archimedes’ feat as a historical precedent, before conjecturing Archimedes’ likely method and offering his own improvement. Tzetzes’ precise debt to Anthemius is of no immediate concern; it suffices to note that his verses (121–30) summarize Anthemius’ text, preserving words and phrases, though Tzetzes’ selection, abridgement and transpositions betray misunderstanding or carelessness.²⁵ In particular, Tzetzes failed to appreciate that the multi-unit hexagonal reflector is explicitly Anthemius’ alternative geometrical solution, which Tzetzes mistakenly transforms into Archimedes’ invention.²⁶ Of broader relevance for reconstructing Dio's fragmentary work, Tzetzes’ effort to combine Dio and Anthemius offers a methodological template for studying other cases in which he sought to integrate historical narrative and technical exposition.²⁷

Zonaras’ testimony is more straightforward. Recent studies variously date the completion of his Epitome between c.1120 and c.1150.²⁸ Zonaras drew extensively and often exclusively on Dio's first 21 books to construct his account of Roman history from Aeneas to 146 b.c., which editorial convention numbers Epitome 7–9. There is no evidence that Zonaras consulted Diodorus’ Bibliotheca.²⁹ Zonaras explains that, writing in retirement in the island monastery of St Glyceria, he sometimes lacked access to necessary books, despite efforts to obtain them. Beyond conventional prefatory apologia, frequent reliance on a single-source narrative, interrupted by specific and substantial gaps, is consistent with a paucity of material.³⁰ Verbal and contentual correspondence with Tzetzes’ verses corroborates that Zonaras (9.4) also based his account of Archimedes’ defence of Syracuse on Dio's Roman History 15. Furthermore, in a later—often overlooked—passage of his Epitome, concerning an incendiary device used to save Constantinople from a naval assault in 515, Zonaras observes that Dio likewise recorded (ἱστόρησε) Archimedes’ use of a burning-mirror (or mirrors) to defend Syracuse.³¹

Accordingly, whereas Zonaras’ Epit. 9.4–5 derives from Dio alone, Tzetzes’ Chil. 2.35 cites Dio and Diodorus, neither extant, but also incorporates technical specifications from Anthemius. Up to and including Tzetzes’ report of a mirror (verse 122), close parallelism with Zonaras’ text suggests Tzetzes’ exclusive dependence on Dio—or, at least, one struggles to discern any element that Tzetzes might have alternatively taken from Diodorus. Given evident misunderstanding in one strand of modern scholarship, it is important to clarify that Zonaras (κάτοπτρον γάρ τι) and Tzetzes (τι κάτοπτρον) record, in identical terms, ‘a kind of mirror’, providing independent corroborative testimony to the presence of this phrase in their common source, Dio, notwithstanding Tzetzes’ decision to import additional details from Anthemius (verses 121–30).³² Only thereafter, from when the Romans enter Syracuse (verse 134), does Tzetzes expressly derive variant or supplementary content from Diodorus. Where Dio himself encountered information about a mirror can only be conjectured, though, as previously observed, the story is unattested before the mid second century.³³ The protracted composition of Dio's Roman History is variously located between the 190s and the 230s.³⁴ As Dio apparently mentioned one burning-mirror, whereas previous and subsequent sources, if they indicate, specify several, he possibly drew on a variant tradition.³⁵ In contrast, it seems most unlikely that the lost section of Diodorus’ Bibliotheca recorded a burning-mirror (or mirrors). In an extant earlier book (5.37.4), Diodorus promises discussion of Archimedes’ inventions when the narrative reaches his era. However, writing over a three-decade period from the late 60s to late 30s b.c., Diodorus predates the earliest witnesses to this legend by two centuries, during which others who wrote at length about Archimedes’ role in the siege, notably Livy and Plutarch, are silent on this matter.³⁶ Even so, almost the entire text of Tzetzes’ Chil. 2.35 (verses 106–52 of 106–59) long ago became a canonical ‘fragment’ of Diodorus, though nothing connects Diodorus to verses 106–33, of which one-third (121–30) derive from Anthemius, while the rest coincide with Zonaras’ Dio-derived Epitome. Tzetzes’ Chil. 2.35 (106–52) thus leads a double life as a fragment of both Dio's Roman History 15 and Diodorus’ Bibliotheca 26; this dual parentage is potentially justified only for verses 134–52.³⁷

TWO NEW WITNESSES: TZETZES’ SCHOLIA TO THE CARMINA ILIACA AND THE HYPOMNEMA IN S. LVCIAM

Unnoticed by editors of Dio and students of the Archimedean tradition, two other compositions by Tzetzes contain material very similar to Chil. 2.35 and Zonaras’ Epit. 9.4–5. This oversight is understandable: although critical texts of both exist, like much of his œuvre, they are obscurely published and virtually unknown outside ‘Tzetzean’ scholarship. This section introduces these works and examines their value as witnesses to the main episodes of Dio's text as preserved in this indirect tradition: Archimedes’ defence of Syracuse and his death when the city fell.

In several respects, the later of the two works is more easily integrated into current scholarship. As his only known endeavour in hagiography, Tzetzes wrote the Hypomnema in S. Luciam, a short prose ‘memorial’ to St Lucia of Syracuse, probably in 1154 or possibly 1158, in connection with one or other of two Sicilian embassies to Constantinople.³⁸ The text is thus contemporary with the Chiliades, though neither work derives from the other. In this inventive literary fantasy, Tzetzes incongruously makes St Lucia a proud descendant of Archimedes. She acclaims her forebear's achievements, which become vehicles for strained Christianizing allegories. Her remarks on Archimedes’ defensive contraptions in two passages (3, 11) are essentially abridged prose versions of material versified in Chil. 2.35 (112–30), employing the same or similar vocabulary, and likewise combining Dio-derived historical description with technical specifications lifted from Anthemius’ treatise:³⁹

οὐδὲ Μαρκέλλου πολέμων τὰς συρραγὰς καὶ Ἀρχιμήδους ἐκείνου τὰς μηχανάς, αἷς τριήρεις μὲν αὔτανδροι ἀνασπώμεναι πρὸς τὸ τεῖχος τὸ Συρακούσιον, εἶτα μετέωροι ἐκριπτούμεναι, αἱ δὲ λίθοις ἄνωθεν ἑκάστῳ ἀνδρὶ ἁμαξιαίοις βαλλόμεναι, αἱ δ᾽ ἔτι πορρώτερον ἐνορμίζουσαι ἀεροτόνοις σφενδόναις, αἱ δ᾽ ἐπὶ πλέον ἀφεστηκυῖαι ἐν ἑξαγώνοις κατόπτροις ἡλιακαῖς ἀντανακλάσεσιν ἐκπυρούμεναι αὔτανδροι τῷ βυθῷ παρεπέμποντο.

Nor Marcellus’ clash of war and the machines of that Archimedes, by which triremes, men and all, are drawn up to the Syracusan wall, then, once aloft, cast forth, while some are hit from above by each man with wagon-sized stones, some moored still further off [hit] with air-driven catapults, others at a greater distance, set alight by solar reflections from hexagonal mirrors, are sent into the depths, men and all.

… ὃς Μάρκελλον διά τε γῆς καὶ θαλάσσης τῇ Συρακούσῃ προσβάλλοντα ταῖς μηχαναῖς ἀπεκρούσατο, πῇ μὲν ὁλκάδας ἐκείνου πρὸς τὸ τεῖχος μετεωρότερον ἀναρτῶν μηχαναῖς, καὶ οὕτω βυθῷ παραπέμπων, πῇ δὲ ἀκοντίζειν ποιῶν ἄνδρα ἕκαστον λίθους ἁμαξιαίους, καὶ ταύτας ὑποβρυχίους ποιῶν, ἐξορμησάσας δ᾽ αὐτὰς ἔτι πορρώτερον ταῖς ἀεροτόνοις τῶν μηχανῶν, εἰ δὲ καὶ τόξου βολὴν ἀποσταίησαν τῶν τειχῶν πρὸς τὸ πέλαγος, ἐν ἑξαγώνοις κατόπτροις τετραπλούμενοις λαμπάδας ἡλίου δεχόμενος καὶ ταύτας φλέγων ἀντανακλάσεσιν.

… [Archimedes] who, when Marcellus was attacking Syracuse by both land and sea, beat him off with machines, here suspending that man's ships aloft above the wall from machines and thence sending them into the depths, there enabling each man to shoot wagon-sized stones and sinking them below the surface, while using the air-driven types of machine for those moored still further off, but, if they were also out to sea a bowshot from the walls, then by receiving the sun's beams into fourfold hexagonal mirrors and setting them ablaze by reflections.

With regard to Archimedes’ death, however, Tzetzes’ Hypomnema (12) has greater critical value, as it exhibits verbal correspondences with Zonaras’ Epitome that are not found in Tzetzes’ Chil. 2.35. This concurrence and its implications will be examined below.

The testimony of the other new witness is more complicated and significant. Tzetzes obliquely touches on Archimedes in his earliest Homeric composition, conventionally titled Carmina Iliaca or Μικρομεγάλη Ἰλιάς.⁴⁰ Forced to eke a living by writing and teaching, in c.1133–1140 Tzetzes composed this condensed narration of the Trojan War in hexameters as a literary-educational ‘showpiece’ designed to secure pupils and/or patrons.⁴¹ An episode in Il. 5.1–8, where Athena invests Diomedes with a flame-like aura, prompts Tzetzes (2.44–8) to draw a contrastive analogy with historically distant events at Syracuse:

ὣς τῶν μαρναμένων Διομήδης ἔνθορε μέσσον,

τεύχεσι λαμπόμενος πυρὶ τεχνήεντι ἀφλέκτῳ, (45)

οὐχ οἷον φλογέεσκε Συρακοσίοιο κατόπτρου,

τῷ ῥ’ Ἀρχιμήδης φλόγα εἵλκυσεν ἠελίοιο,

πρῆσε δὲ Μάρκου Μαρκέλλου νέας Αἰνεάδαο

As Diomedes leapt into the midst of those doing battle,

from his arms shining with artificial fire without flame,

not such as blazed from a Syracusan mirror,

by which that Archimedes drew flame from the sun

and burned ships of Aenean Marcus Marcellus

Tzetzes distinguishes the fiery glare of Diomedes’ equipment from the burning flame of Archimedes’ mirror. These verses attest Tzetzes’ interest in this topic perhaps two decades before he wrote his Chiliades.⁴² More intriguing than this brief allusion, however, are two scholia that Tzetzes wrote on his own verses (45–6). These scholia were published as early as 1770, and the first was re-edited in 1851, but both publications seemingly passed unnoticed by scholars of Archimedes. Even Leone's critical edition (1995) remains a bibliographical rarity outside Italian libraries.⁴³ The first scholium elucidates πυρὶ τεχνήεντι (2.45) with a rationalizing exegesis of Diomedes’ divinely bestowed radiance: κάτοπτρον μηχανητὸν εἶχεν ἐπὶ τῆς ἀσπίδος ὁ Διομήδης δόκησιν πυρὸς ἀποστίλβον, … πρὸς ἐκπτόησιν τῶν πολεμίων καθ᾽ ἡλιακῆς ἀκτῖνος φερόμενον καὶ τῇ ἀντανακλάσει προσεπιτεῖνον τοῖς ἀπειροτέροις τὴν δόκησιν, ‘Diomedes had a mirror devised on his shield, shining forth an apparition of fire, … causing alarm among the enemy through a solar ray and, by reflection, intensifying the apparition for the more inexperienced’. Typical of his claims to scientific authority, Tzetzes appends a fictitious catalogue of ancient ‘mechanicians’ (μηχανικοί) who recommend this combat application of mirrors in optical/catoptrical writings.⁴⁴ Around the early/mid 1140s, Tzetzes reprised this techno-allegorical interpretation of supernatural brilliance at Il. 5.1–8 in his Allegories of the Iliad. Now in verse, he again differentiates Diomedes’ ‘flameless fire’ (πῦρ ἀφλεγές) from Archimedes’ ‘burning mirror’ (καυστικὸν κάτοπτρον), substantiating his scientific knowledge with a similar but augmented list of ‘mechanical writers’ (μηχανογράφοι).⁴⁵

Tzetzes’ second and much longer scholium to Συρακοσίοιο κατόπτρου (2.46a) is more complex and intriguing.⁴⁶ Its structure comprises four basic units, defined by content and/or source(s). The first three broadly mirror the sequence of Chil. 2.35, though each text contains material not found in the other. The first section concerns Archimedes’ defensive devices, including a burning-mirror. Again Tzetzes combines Dio's historical narrative with Anthemius’ technical exposition (166.13–168.2 = Chil. 2.35.106–31). The second briefly digresses on Archimedes’ extraordinary accomplishments with pulleys, as reported in multiple sources drawing on variant traditions (168.2–6 = Chil. 2.35.132–3).⁴⁷ The third returns to the siege, particularly Archimedes’ death, at least partly based on Dio, as indirectly witnessed by Zonaras and Tzetzes’ Chiliades (168.6–18 = Chil. 2.35.133–51). The fourth quotes a tetrastichic inscription on a statue of Marcellus, as recorded by Plutarch (168.18–169.6 = Plut. Vit. Marc. 30.5).⁴⁸ It will be demonstrated below that this scholium draws independently on the same source as Zonaras’ Epit. 9.4–5 and Tzetzes’ Chil. 2.35.106–31, namely Dio's Roman History. To this end, the first and third sections are quoted here in full:⁴⁹

Συρακοσίοιο κατόπτρου· ὁ Ἀρχιμήδης οὗτος ἐπὶ Ἱέρωνος ἦν καὶ Πυθαγόρου, γέρων γεωμέτρης μηχανικὸς Συρακούσιος τὴν ἡλικίαν ἐγγὺς ὀγδοήκοντα ἐτῶν, ὃς μυρία μὲν πεποίηκεν ἕτερα, ἀλλὰ καὶ Μάρκελλον τὸν στρατηγὸν τῶν Ῥωμαίων τῇ Συρακούσῃ πολέμῳ βαρεῖ κατὰ γῆς καὶ θαλάσσης ἐπιτεθέντα μεγάλοις συνέθραυσε μηχανήμασι. πρῶτα μὲν γὰρ μηχανὰς ἐκ τοῦ τείχους καθιείς, αὐτάνδρους ἀνέσπα τὰς τριήρεις ἐκείνου καὶ μετεωρίζων εἰς ὕψος καὶ κατασπῶν εἰς τὴν θάλασσαν ὑποβρυχίους ἐποίει. τοῦ Μαρκέλλου δὲ τὰς λοιπὰς ὀλκάδας πορρώτερον τοῦ τείχους πελαγιώτερον καθελκύσαντος, μηχανήμασι πάλιν ὁ αὐτὸς Ἀρχιμήδης τὸν καθένα τῶν Συρακουσίων ἐποίει λίθους σφενδονᾶν ἁμαξήρεις καὶ πάλιν τὰς ὀλκάδας βυθίζειν. ἀπογνοὺς δὲ ὁ Μάρκελλος τὸν τείχους ἐπίπλουν τόξου βολὴν πόρρωθεν εἵλκυσε ταύτας· ὁ δὲ κάτοπτρον ἑξάγωνον ποιήσας ποδιαῖον τὸ μέγεθος καὶ ἀνατείνας αὐτό, ὡς ἀπέχειν ἐξ ἴσου διαστήματος μεσημβρινῆς, θερινῆς καὶ χειμερινῆς ἡλίου ἀκτῖνος καὶ θεὶς παρ’ ἑκάστην ἐκείνου γωνίαν ἕτερα τέτταρα μικρότερα κάτοπτρα, τῇ ἀνακλάσει τῆς ἀκτῖνος τὰς ὁλκάδας ἐνέπρησε τοσοῦτον ἀφεστηκυίας. ὁ δὲ Μάρκελλος ἐπὶ τοῖς συμβᾶσι δεινοπαθῶν ἔλεγεν· ‘εἰ δυνατὸν ἐξεγένετο πᾶσαν τὴν περὶ τὸν κόσμον ὑπάρχουσαν στρατιὰν εἰς ταὐτὸν συναχθῆναι καὶ τῷ ἐμῷ στρατεύματι ἀντιτάξασθαι, ἔστησα ἄν, εὖ οἶδα, νικητικώτατον τρόπαιον. νῦν δέ—ἀβάλε μοι τῆς αἰσχύνης—ἕν τι δαιμόνιον γηρανδράριον γενναίως καὶ ἀσυγκρίτως οὕτω καταγωνίζεται, ὡς μηδὲ τὸ σύνολον ἀνταίρειν τολμᾶν.’

… this Archimedes lived at the time of Hieron and Pythagoras, an aged geometer [and] mechanician of Syracuse, close to eighty years of age, who had made countless other things, but also used great machines to crush Marcellus, the Roman general, as he beset Syracuse with heavy fighting on land and sea. For first, letting down machines from the wall, he drew up [Marcellus’] triremes, men and all, and raising them aloft to a height and dragging them down into the sea, he sank them below the surface. When Marcellus pulled back the rest of his ships from the wall further out to sea, again with machines the same Archimedes made it possible for each one of the Syracusans to sling wagon-sized stones and again to sink the ships. Giving up on sailing against the wall, Marcellus pulled them back as far as a bowshot, but [Archimedes], having made a hexagonal mirror, one foot in dimension, and having tilted it so that it was at an equal distance from the midday rays of the sun, in summer or midwinter, and having set at each angle of that [mirror] four other smaller mirrors, by the reflection of the rays set the ships on fire when they were so far away. Bemoaning these events, Marcellus said: ‘if it were possible that all the forces existing around the world were gathered in this place and arrayed opposite my army, I would have raised, I know well, a trophy to a very great victory. But now—oh, the shame of it—one withered old man with miraculous power so nobly and incomparably resists, that the entire [army] does not dare oppose [him].’

τελευτᾷ ὁ γέρων ὧδέ πως. ἁλούσης τῆς Συρακούσης ὑπὸ Μαρκέλλου ἢ παννυχιζόντων Συρακουσίων Ἀρτέμιδι ἢ προδεδομένης παρά τινων ἐπέστη τῷ γεωμέτρῃ Ῥωμαῖος ξιφήρης μηχανικόν τι διάγραμμα γράφοντι καὶ καθεῖλκεν αὐτόν· ὁ δέ ‘ἀπόστηθι, ἄνθρωπε, τῆς γραμμῆς’ αὐτῷ ἔλεγεν. ὡς δ᾽ ὁ Ῥωμαῖος ἐπέκειτο καὶ σφοδροτέρως εἷλκεν αὐτόν, ὁ Ἀρχιμήδης στραφεὶς καὶ συνεὶς τὸ πραχθὲν ἐβόα· ‘δότω μοί τίς τι τῶν ἐμῶν μηχανῶν’. καὶ περιδεὴς γεγονὼς ὁ Ῥωμαῖος ἀνεῖλεν αὐτόν. ὁ δὲ βασιλεὺς Ῥωμαίων ὁ Μάρκελλος ἐθρήνησε τοῦτο ἀκηκοὼς καίπερ πολέμιος ὢν καὶ τὴν κόμην ἐσπάραξεν ἐπὶ τούτῳ καὶ πολεμίῳ πεφυκότι καὶ γέροντι καὶ μεγαλοπρεπῶς αὐτὸν μετὰ πάντων Ῥωμαίων καὶ ἀρίστων Συρακουσίων ἐνεταφίασε, τὸν δὲ τούτου οἶμαι φονέα πελέκει κατέκρινεν.

The old man met his end in this way. When Syracuse was taken by Marcellus, either while the Syracusans were celebrating an all-night festival of Artemis or because betrayed by certain people, a Roman swordsman came upon the geometer as he was drawing some mechanical diagram and began to drag him away, but [Archimedes] said to him ‘Stand away, man, from my drawing’. As the Roman persisted and dragged him more vigorously, Archimedes, having turned around and perceived what was happening, cried out ‘Let someone give me one of my machines’. And becoming fearful, the Roman slew him. The king of the Romans, Marcellus, mourned on hearing this, even though he was an enemy, and tore his hair for this person who had been both an enemy and an old man, and with all the Romans and noblest Syracusans he gave him a magnificent burial, while his murderer, I believe, he condemned to the axe.

Before proceeding to a closer textual examination, a question of dating must be addressed. While Tzetzes’ composition of the Carmina Iliaca is securely anchored in the 1130s, the date of his scholia has not been demonstrated. Recent studies assume or imply the contemporaneity and integrity of Tzetzes’ verses and scholia, but it remains uncertain whether this erudite ‘showpiece’, combining virtuoso hexametric display and sample teaching text, was originally furnished with a full pedagogic apparatus or later modified in light of classroom practice.⁵⁰ In particular cases, evidence permits tentative inferences about relative chronology. Arguably in favour of an early dating of scholium 2.46a is the high incidence of errors or conflicting details, all of which were omitted from or corrected in Chil. 2.35, and are thus explicable as mistakes of a younger Tzetzes.⁵¹ Elsewhere, at least, Tzetzes seized opportunities to correct errors committed in his youth.⁵² If this interpretation is correct, Tzetzes’ interest in this subject and access to relevant historical and technical sources go back to his earliest writings. This chronology would also mean that scholium 2.46a was written around the same time or possibly before Zonaras completed his Epitome, while Tzetzes would not begin composing his Chiliades and Hypomnema for another fifteen to twenty years.⁵³

Whenever Tzetzes wrote scholium 2.46a, comparative analysis indicates that he did not merely rework the same material found in Chil. 2.35 but consulted Dio's text directly; that is to say, scholium 2.46a and the Chiliades are discrete projects. Hitherto scholarship has recognized Tzetzes’ Chiliades and Zonaras’ Epitome as independent witnesses to Dio's account of the siege of Syracuse. When scholium 2.46a is adduced as a third independent witness, the prospect of constituting Dio's text is correspondingly enhanced. Furthermore, in Tzetzes’ Chiliades, composed in iambic decapentasyllabic verse, vocabulary and syntax are variously affected by metrical considerations that do not constrain the prose of his scholium. Most importantly, although linguistic parallels between Tzetzes’ two texts, even if separated by two decades, may be explained as a consequence of their common authorship, verbal correspondence between Tzetzes’ scholium and Zonaras alone must originate in their common source. This is apparent in the first section of the scholium, concerning Archimedes’ military devices:

The independent descent of scholium 2.46a from Dio's text is best illustrated by the sequence καθιείς, αὐτάνδρους ἀνέσπα, where καθίει and ἀνέσπα are found sequentially in Zonar. 9.4 but not in Tzetzes’ Chil. 2.35, while αὐτάνδρους is found in Chil. 2.35 (116) but not in Zonar. 9.4. Within this intricate intertextual pattern, while bilateral correspondence between Tzetzes’ two works could—but does not necessarily—indicate that he found αὐτάνδρους in Dio, agreement between Tzetzes’ scholium and Zonaras guarantees the presence of καθιεί(ς) and ἀνέσπα in Dio's text.⁵⁴ Significantly, the phrase ἀνατείνας αὐτό in Tzetzes’ scholium finds a parallel in Zonar. 9.4 (πρὸς τὸν ἥλιον ἀνατείνας) but not in Tzetzes’ Chil. 2.35, indicating Dio's use of this participle in reference to Archimedes’ mirror. This concurrence again underlines the presence of the κάτοπτρον in Dio's account and provides a rare splinter of verbal certainty at precisely the point where the testimony of Zonaras and Tzetzes’ Chiliades diverges, owing to Tzetzes’ insertion of technical content from Anthemius’ treatise. Incidentally, where the wording of Chil. 2.35 has proved obscure, scholium 2.46a can now also clarify how Tzetzes (mis)understood Anthemius’ specifications for a complex reflector.⁵⁵

Following this description of Archimedes’ contraptions, scholium 2.46a records Marcellus’ reaction in direct speech (167.13–168.2 ὁ δὲ Μάρκελλος ἐπὶ τοῖς συμβᾶσι δεινοπαθῶν ἔλεγεν· ‘εἰ δυνατὸν … τολμᾶν’), in which he bemoans his forces’ defeat by ‘one withered old man with miraculous power’ (ἕν τι δαιμόνιον γηρανδράριον). Unlike preceding material, the provenance of this quotation cannot be traced. No corresponding speech episode occurs in Tzetzes’ Chiliades and Zonaras’ Epitome, though this absence is not crucial. Ordinarily, Tzetzes’ fifteen-syllable verses accommodate only short declarations, apophthegms or bons mots.⁵⁶ Zonaras, in accordance with generic convention and authorial taste, generally omits or condenses extended speech episodes in his sources.⁵⁷ Comparative analysis of Zonaras’ text and Dio's extant books or excerpts reveals that he pursued several abridging methodologies, either excluding the content of speeches, debates and dialogues entirely or, where they serve causal or explicative functions, summarizing selected arguments or drastically reducing the general tenor to a terse résumé. However, even when extended direct discourse is completely excised, Zonaras often retains its surrounding narrative framework by recording that a speech occurred (typically using λέγειν, διαλέγεσθαι, δημηγορεῖν) and by acknowledging its performative impact on the audience and the course of events. Zonaras thus ‘de-rhetoricizes’ Dio's text, transforming direct discourse into descriptive report.⁵⁸ In contrast to extended orations, Zonaras more often reproduces short direct-speech episodes such as pithy utterances and ripostes essential to a story, which, it seems, were frequent across Dio's work.⁵⁹ The direct speech in Tzetzes’ scholium may not meet these criteria, inasmuch as Marcellus’ frustrated outburst is not deliberative, epideictic or hortatory oratory, but rhetorical interjection without historical consequence. However, with some imagination, one might discern a parallel narrative development in Zonaras’ text at this point, immediately following the destruction of the Roman fleet: ἀπογνοὺς οὖν ὁ Μάρκελλος τὴν πόλιν αἱρήσειν διὰ τὸ τοῦ Ἀρχιμήδους εὐμήχανον, ‘Marcellus, then, despairing of taking the city on account of Archimedes’ mechanical ingenuity …’.⁶⁰ Zonaras’ wording does not signal an excised speech episode using his usual verbal markers, but his report of Marcellus’ aporia precisely coincides with the positioning of the direct speech in Tzetzes’ scholium and encapsulates its content and sentiment in a manner consistent with Zonaras’ method of abridgement. Curiously, the closest extant counterpart occurs in Plutarch's Life of Marcellus: directly after Archimedes’ machines shatter the Roman naval assault (15.1–16.2), Plutarch likewise quotes the general's response, but here Marcellus’ words differ entirely—a ‘jesting’ (σκώπτων) remark on the futility of fighting ‘this geometrical Briareus’ (17.1). This structural correspondence implies the existence of a prior historiographic template—or, at least, an established episode in one historical tradition—in which, at this point in the narrative sequence, Marcellus voices admiration of or exasperation at Archimedes’ supernatural/technological powers. Tzetzes would thus have found another example of this model in an alternative source, which could have been Dio. More generally, the quotation in scholium 2.46a evokes a persistent face-saving motif of Roman historiography that credits initial reverses at Syracuse to the singular scientific genius of ‘one old man’ rather than collective Syracusan martial superiority.⁶¹ Tzetzes rehearses this Archimedean topos in other works, likewise employing the adjective γηρανδράριος, unique to Tzetzes’ œuvre and possibly his own coinage. The presence of such ‘Tzetzean’ vocabulary in scholium 2.46a is unsurprising and need not exclude a classical antecedent, given the place of adaptive epideictic erudition in Tzetzes’ artistry.⁶² Ultimately, nothing in its language or style requires that this specimen of direct speech derives from Dio (or any other ancient author). Yet, unless we infer that Tzetzes himself, perhaps loosely inspired by Plutarch, put these words into Marcellus’ mouth, Dio remains the most likely source. The question must remain open.

Finally, the third section of scholium 2.46a, concerning the fall of Syracuse and Archimedes’ death, although, overall, less noteworthy than the first section, evinces the same compositional history and pattern of intertextuality with the other indirect witnesses, while modestly augmenting our understanding of this lost passage in Dio's Roman History. Below, the text of the scholium is juxtaposed to that of Tzetzes’ Chiliades (with corresponding wording marked in bold) and Zonaras’ Epitome (correspondences underlined):

Tzetz. Schol. Carm. Il. 2.46a (168.6–18):

τελευτᾷ ὁ γέρων ὧδέ πως. ἁλούσης τῆς Συρακούσης ὑπὸ Μαρκέλλου ἢ παννυχιζόντων Συρακουσίων Ἀρτέμιδι ἢ προδεδομένης παρά τινων ἐπέστη τῷ γεωμέτρῃ Ῥωμαῖος ξιφήρης μηχανικόν τι διάγραμμα γράφοντι καὶ καθεῖλκεν αὐτόν· ὁ δέ “ἀπόστηθι, ἄνθρωπε, τῆς γραμμῆς” αὐτῷ ἔλεγεν. ὡς δ᾽ ὁ Ῥωμαῖος ἐπέκειτο καὶ σφοδροτέρως εἷλκεν αὐτόν, ὁ Ἀρχιμήδης στραφεὶς καὶ συνεὶς τὸ πραχθὲν ἐβόα· “δότω μοί τίς τι τῶν ἐμῶν μηχανῶν”. καὶ περιδεὴς γεγονὼς ὁ Ῥωμαῖος ἀνεῖλεν αὐτόν. ὁ δὲ βασιλεὺς Ῥωμαίων ὁ Μάρκελλος ἐθρήνησε τοῦτο ἀκηκοὼς καίπερ πολέμιος ὢν καὶ τὴν κόμην ἐσπάραξεν ἐπὶ τούτῳ καὶ πολεμίῳ πεφυκότι καὶ γέροντι καὶ μεγαλοπρεπῶς αὐτὸν μετὰ πάντων Ῥωμαίων καὶ ἀρίστων Συρακουσίων ἐνεταφίασε, τὸν δὲ τούτου οἶμαι φονέα πελέκει κατέκρινεν.

Tzetz. Chil. 2.35.134–51:

οὗτος, κατὰ Διόδωρον, τῆς Συρακούσης ταύτης,

προδότου πρὸς τὸν Μάρκελλον ἁθρόως γενομένης (135)

εἴτε, κατὰ τὸν Δίωνα, Ῥωμαίοις πορθηθείσης,

Ἀρτέμιδι τῶν πολιτῶν τότε παννυχιζόντων,

τοιουτοτρόπως τέθνηκεν ὑπό τινος Ῥωμαίου.

ἦν κεκυφὼς διάγραμμα μηχανικόν τι γράφων,

τὶς δὲ Ῥωμαῖος ἐπιστὰς εἷλκεν αἰχμαλωτίζων. (140)

ὁ δὲ τοῦ διαγράμματος ὅλος ὑπάρχων τότε,

τίς ὁ καθέλκων οὐκ εἰδώς, ἔλεγε πρὸς ἐκεῖνον⋅

“ἀπόστηθι, ὦ ἄνθρωπε, τοῦ διαγράμματός μου”.

ὡς δ’ εἷλκε τοῦτον συστραφεὶς καὶ γνοὺς Ῥωμαῖον εἶναι,

ἐβόα⋅ “τὶ μηχάνημα τὶς τῶν ἐμῶν μοι δότω”. (145)

ὁ δὲ Ῥωμαῖος πτοηθεὶς εὐθὺς ἐκεῖνον κτείνει,

ἄνδρα σαθρὸν καὶ γέροντα δαιμόνιον τοῖς ἔργοις.

ἐθρήνησε δὲ Μάρκελλος τοῦτο μαθὼν εὐθέως,

λαμπρῶς τε τοῦτον ἔκρυψεν ἐν τάφοις τοῖς πατρῴοις

σὺν τοῖς ἀρίστοις πολιτῶν καὶ τοῖς Ῥωμαίοις πᾶσι, (150)

τὸν δὲ φονέα τοῦ ἀνδρὸς οἶμαι πελέκει κτείνει.

Zonaras, Epit. 9.5 (2.264.14–15, 264.27–265.2):

[ὁ Μάρκελλος] τηρήσας οὖν τοὺς Συρακουσίους παννυχίδα τῇ Ἀρτέμιδι ἄγοντας πανδημεί, … καὶ τῶν λοιπῶν τῆς Συρακούσης ἐκράτησεν. ἐγκρατεῖς δὲ τούτων οἱ Ῥωμαῖοι γενόμενοι ἄλλους τε πολλοὺς καὶ τὸν Ἀρχιμήδην ἀπέκτειναν. διάγραμμα γάρ τι διαγράφων, καὶ ἀκούσας τοὺς πολεμίους ἐφίστασθαι, “πὰρ κεφαλάν” ἔφη “καὶ μὴ παρὰ γραμμάν.” ἐπιστάντος δὲ αὐτῷ πολεμίου βραχύ τε ἐφρόντισε καὶ εἰπών “ἀπόστηθι, ἄνθρωπε, ἀπὸ τῆς γραμμῆς” παρώξυνέ τε αὐτὸν καὶ κατεκόπη.

Even if, despite a gap of possibly two decades, the wording and the content of Tzetzes’ scholium are more closely aligned with Chil. 2.35 than with Zonaras’ Epitome, this triple comparison again demonstrates that the scholium is an independent witness to Dio's text. Evidently, in both of Tzetzes’ works—and explicitly in the Chiliades—he combined information drawn from Dio and Diodorus. Scholium 2.46a likewise reports variant accounts of the city's capture, ἢ παννυχιζόντων Συρακουσίων Ἀρτέμιδι ἢ προδεδομένης παρά τινων, ‘either while the Syracusans were celebrating an all-night festival of Artemis or because betrayed by certain people’, two versions that in Chil. 134–6 Tzetzes respectively assigns to Dio and Diodorus.⁶³ In contrast, as noted above, there is no evidence that Zonaras used Diodorus’ Bibliotheca as a source for any part of his Epitome. Consequently, given significant differences between Tzetzes’ and Zonaras’ accounts of this episode, with each author providing details not mentioned by the other, Zonaras must be regarded as the more authoritative guide to Dio's text, while it may be presumed that divergent and additional content in Tzetzes’ scholium and/or Chiliades is likely to derive from Diodorus. Conversely, and of greater interest for present purposes, where the wording of Tzetzes’ scholium corresponds to Zonaras’ text (but diverges from Tzetzes’ Chiliades), that concurrence can only be explained by their collateral descent from Dio. This common-source relationship is most apparent in Archimedes’ direct speech: Tzetzes’ scholium (168.10) “ἀπόστηθι, ἄνθρωπε, τῆς γραμμῆς” and Zonar. 9.5 (2.265.1–2) “ἀπόστηθι, ἄνθρωπε, ἀπὸ τῆς γραμμῆς”, distinct from Tzetzes’ Chiliades (2.35.143) “ἀπόστηθι, ὦ ἄνθρωπε, τοῦ διαγράμματός μου”.

In this context, however, the other neglected work under discussion, Tzetzes’ Hypomnema, acquires unexpected textual significance. In its antepenultimate section, where St Lucia is condemned to be beheaded, Tzetzes places in her mouth an inspiring account of Archimedes’ last moments, where she quotes his final words:⁶⁴

πάλιν τάδ᾽ ἀνεφθέγξατο· ‘Ἀρχιμήδης ἐκεῖνος ὁ πρόγονος ὁ ἐμὸς καὶ μέχρι τῆσδε τῆς τελευταίας ἀναπνοῆς μοι τὸν λόγον χαρίζεται. μηχανικὸν γάρ οἱ διάγραμμα γράφοντι, προδοσίᾳ τῆς Συρακούσης ἁλούσης ἐπιστάς τις Ῥωμαῖος πολέμιος εἷλκεν αὐτὸν τῆς χειρὸς ἐπικεκυφότα τῷ διαγράμματι· ὁ δὲ Συρακουσίων τινὰ εἶναι οἰόμενος “ἀπόστηθι”, ἔλεγεν, “ἄνθρωπε, τῆς γραμμῆς”. ὡς δ᾽ “εἴπερ μὴ ἕπῃ, ἀφελῶ σου”, εἶπε, “τὴν κεφαλήν”. Ἀρχιμήδης· “ἑλεῖς”, ἔφη, “ἄνθρωπε, τό, κλῦθι ἐμοῦ, ἐγὼ δέ τοι ἐξεροῦμαι, ναί, τώς σοι, ὡς Ἀρχιμήδης, βέλτερον ἔντι παρὰ κεφαλὰν εἶναι ἢ παρὰ γραμμάν”· κἀμοὶ δέ … παρὰ κεφαλὰν βέλτερον εἶναι ἢ παρὰ τὴν εἰς Θεὸν ἄγουσαν εὐθεῖαν γραμμήν.’

Again she spoke aloud as follows: ‘Archimedes, that forebear of mine, even up to my last breath has pleasing words for me. For as he was drawing a mechanical diagram, when Syracuse was captured by treachery, a certain Roman enemy, having come upon him, began dragging him off by the hand as he was bent over his diagram. He, supposing that it was some Syracusan, said, “Stand away, man, from my drawing”. Then [the Roman] said, “If indeed you do not follow me, I shall take off your head.” “Take it, man”, said Archimedes, “Hear me, I assuredly declare, yes, that thus, for Archimedes, it is better for you to be at my head than at my drawing”. And for me [St Lucia] too … it is better at the head than at the line that leads straight to God.’

There are, unsurprisingly, obvious parallels, in language and substance, between this passage of Tzetzes’ Hypomnema and both Chil. 2.35 and scholium 2.46a. In particular, Archimedes’ reported exclamation “ἀπόστηθι, ἄνθρωπε, τῆς γραμμῆς” again coincides with the scholium and with Zonaras, against the wording in Chil. 2.35, reinforcing the authority of the—now three—prose witnesses over the metrically constrained diction of Tzetzes’ verses. There are also omissions from the Hypomnema compared to Tzetzes’ other two versions. However, the most striking feature is the inclusion of Archimedes’ remark ‘βέλτερον … παρὰ κεφαλὰν εἶναι ἢ παρὰ γραμμάν’. This utterance, recorded with Doric orthography, finds no parallel in Tzetzes’ Chiliades or in scholium 2.46a, but it does correspond, almost exactly, to Zonaras’ Dio-derived account: ‘πὰρ κεφαλάν’ ἔφη ‘καὶ μὴ παρὰ γραμμάν’. Hitherto, Zonaras was the unique witness to this quotation. As there is no suspicion or likelihood that Tzetzes consulted this passage of Zonaras’ Epitome when composing his Hypomnema, the only explanation for this phraseological coincidence is a common source. Here again, one of Tzetzes’ three treatments of Archimedes at Syracuse transmits Dio-derived material not transmitted in the other two. In this case, although certainty about the precise setting of this quotation in Dio's text remains elusive, Tzetzes’ Hypomnema appears, at least, to provide a narrative and oratorical context for Archimedes’ words, whereas Zonaras’ severe abridgement offers, at best, a disjointed skeletal outline of what was undoubtedly a more developed story in Dio's Roman History.

CONCLUSIONS

The current constitution of Dio's account of Archimedes’ role in the siege of Syracuse in Roman History 15, depending on Zonaras’ Epit. 9.4–5 and Tzetzes’ Chil. 2.35, is essentially a product of nineteenth-century editorial knowledge and method. While subsequent studies have more clearly differentiated Tzetzes’ use of three cited sources (Dio, Diodorus and Anthemius), the wider bibliography continues to harbour misconceptions and inconsistencies. This investigation has drawn attention to corresponding Dio-derived content in a scholium to Tzetzes’ Carmina Iliaca (2.46a), his earliest engagement with Homeric verse (c.1133–1140), and in his Hypomnema (1154/1158), a hagiographical opusculum, contemporary with the composition of his Chiliades (mid/late 1150s). Comparative analysis of these two texts identifies fluctuating verbal correspondences with Zonaras’ Epitome that have no parallels in Chil. 2.35. Correspondingly, each of the new witnesses contains details not found in the other and/or in the Chiliades. This shifting pattern of convergence and divergence shows that both the scholium and the Hypomnema are independent witnesses to Dio's Roman History, reflecting Tzetzes’ selective exploitation of Dio's work in three distinct projects over two decades. If Tzetzes’ scholia to his Carmina Iliaca are correctly dated fifteen to twenty years before his Chiliades and Hypomnema, this Homeric exegesis charts his first traceable encounter with Dio's work. The possibility that the scholia may also predate Zonaras’ Epitome, if its dating to the 1140s is preferred, would make scholium 2.46a the earliest witness to this section of Roman History 15. Consequent advances in comprehending Dio's text relate primarily to language, but the foregoing investigation also sheds modest light on content. The ability of scholium 2.46a to confirm wording used by Zonaras specifically in connection with Archimedes’ ‘burning-mirror’ substantiates the occurrence of this device in Dio's account, should such verification be needed, if only to counter lingering misapprehensions in some studies of ancient mathematics. Whatever Dio's source, his retention or insertion of this apparently recent legendary element, even if briefly mentioned, may have implications for understanding the historian and the sort of work he wished to write. The scholium also raises the possibility that this part of Book 15 contained direct discourse, in the form of a rhetorical interjection by Marcellus, though this inference cannot advance beyond conjecture. In addition, the Hypomnema, uniquely of Tzetzes’ three texts, supplies corroborative evidence for a quotation of Archimedes that was previously known only from Zonaras (9.5).

Researchers must now take account of all three of Tzetzes’ writings, whose collective and individual testimony, particularly where verified by concurrence with Zonaras’ Epitome, affords a better prospect of arriving at Dio's original text via several collateral lines of textual tradition. Furthermore, it is clear that Tzetzes’ scholium and/or Hypomnema are sometimes more faithful witnesses, especially to Dio's wording, in part because the two prose texts are not constrained by metrical considerations that govern lexis and syntax in the Chiliades, and despite the scholium containing factual lapses attributable to youthful inattentiveness. This new editorial scenario poses complications beyond a novel arrangement and numbering of multiple ‘fragments’. For example, future editors will be obliged to determine the significance of verbal correspondence between the scholium and the Hypomnema and/or the Chiliades even in the absence of parallels in Zonaras’ text, itself a selective abridgement; in short, does shared vocabulary across Tzetzes’ works, separated by fifteen to twenty years, go back to Dio or does it result from currently obscure intertextual links or from mere coincidence?⁶⁵ Similarly, lexical eccentricity in the scholium alone, relative to the rest of Tzetzes’ œuvre, may also originate in Dio.⁶⁶ Ultimately, given the growth of interest in both authors, it is hoped that the increasingly nuanced enquiries into literary and textual aspects of Dio's Roman History will extend investigation to Tzetzes’ lesser-known writings and a re-evaluation of their traits as sources for Dio-derived material. Certainly, scholars have not yet fully mined this seam of research.⁶⁷