The Xerxes Canal

The canal answers a named disaster. In 492 BCE the first Persian expedition against northern Greece, commanded by Mardonius, moved west along the Thracian coast with the fleet keeping pace offshore. Rounding the cape of Athos, the easternmost of the three prongs of the Chalkidike, the fleet met a north wind:

Crossing then over from Thasos they voyaged near the land as far as Acanthus, and putting out from thence they would have rounded Athos. But as they sailed, there brake upon them a north wind great and irresistible, and dealt very roughly with them, driving many of their ships upon Athos; three hundred, it is said, was the tale of the ships that perished, and more than twenty thousand men. For inasmuch as these coasts of Athos abounded in wild beasts, some were carried off by these and so perished; others were dashed against the rocks; and those of them that could not swim perished by reason of that, and others again by the cold.

Herodotus, Histories 6.44, trans. Godley^[1]

Three hundred ships and more than twenty thousand men are Herodotus's figures, reported at the distance of "it is said", and the wild beasts finishing off the swimmers belong to the storyteller as much as to the sea. The loss itself is credible: the eastern face of Athos is a lee shore under the northerly gales of the Aegean, without harbours, and a fleet of the period caught against it had nowhere to go. Athos had shown what it could do to ships, and the memory governed the planning of the next war. When Xerxes resolved on the great invasion, Herodotus opens his account of the preparations with the mountain: "First of all he had now for about three years been making all his preparations in regard of Athos", he writes, "inasmuch as they who first essayed to sail round it had suffered shipwreck".^[2]

When Herodotus reaches the finished work he passes a judgement on it that has shaped the canal's meaning ever since. Xerxes, he writes,

gave command for this digging out of pride, because he would display his power and leave memorials of it; for they might very easily have drawn their ships across the isthmus; yet he bade them dig a canal from sea to sea, wide enough to float two triremes rowed abreast.

Herodotus, Histories 7.24, trans. Godley^[2]

The judgement is announced as guesswork; Godley's rendering opens "As far as I judge by conjecture". The commentators have long noticed how much hangs on that admission. How and Wells observed that the motive of pride "is avowedly a suggestion of H.", adding that "the importance attached to the canal seems due to the historian".^[3]

More hangs on a single word. What Godley printed as "pride" is the Greek megalophrosunē. In Emma Bridges's rendering of the sentence, Herodotus reckons "that Xerxes ordered the digging only out of megalophrosunē (μεγαλοφροσύνης εἵνεκεν, 7.24.1)", "wishing to display his power and leave a monument to himself"; and, as she puts it, "The translation of the Greek megalophrosunē here is open to conjecture."^[4] The word's history in English is a history of verdicts. How and Wells (1912) and Godley (1922) chose "pride", carrying the Greek moralising tradition into the translation itself; de Sélincourt (1954) offered "mere ostentation"; Waterfield (1998) "a sense of grandiosity and arrogance"; Holland (2013) expanded the phrase into "the sheer scale of Xerxes' relish for magnificence". The literal sense is simply "greatness of mind", and Baragwanath has observed that the related quality of megalopsychia, greatness of soul, carries positive weight in Aristotle's ethics.^[4] Each rendering decides in advance what the canal was: "pride" convicts the king of hubris; "greatness of mind" describes a royal virtue of exactly the kind Achaemenid kings claimed for themselves. The Greek word, on its own, settles nothing.

A strategic reading needs no help from the lexicon. Leake, quoted with approval by How and Wells, held Xerxes "perfectly justified in cutting this canal" for the security it gave his fleet, and later students of the coast noted that with the canal open, ships caught by the south-west wind "might shelter from it in the bay of Acanthus, and from the North wind in the Singitic gulf".^[5] Hyland reads the work as one arm of a deliberate programme. Coming directly after the suppression of revolt in Egypt and Babylonia, the preparations had, he argues, "a dual purpose", "signaling the restoration of imperial unity through collective subject labor" while easing the movement of ships and supplies; the cut allowed triremes to avoid "the lee shore of Mount Athos itself" and proceed in shelter toward the Thermaic gulf. For Hyland the display and the engineering do not compete: the canal was "a display of Xerxes's capability to reshape natural landscapes to serve his will", exercised at the exact place where Persian forces had come to grief.^[6]

Herodotus describes the works at unusual length, and his chapters on the canal preserve more circumstantial engineering detail than he gives any other Persian project:

Triremes were anchored off Elaeus in the Chersonese; with these for their headquarters, all sorts and conditions of men in the army were made to dig a canal under the lash, coming by turns to the work; and they that dwelt about Athos dug likewise. Bubares son of Megabazus and Artachaees son of Artaeus, Persians both, were the overseers of the workmen. This Athos is a mountain great and famous, running out into the sea; it is inhabited by men. At the mountain's landward end, it is in the form of a peninsula, and there is an isthmus of about twelve furlongs' width; here is a place of level ground or little hills, from the sea by Acanthus to the sea which is over against Torone. On this isthmus, which is at the end of Athos, there stands a Greek town, Sane; there are others too seaward of Sane and landward of Athos, which it was now the Persians' intent to make into island and not mainland towns; to wit, Dion, Olophyxus, Acrothoum, Thyssus, Cleonae.

Herodotus, Histories 7.22, trans. Godley^[2]

The passage fixes the geography. The isthmus of twelve stades, Godley's "furlongs", is about 2.2 km wide, running from the bay of Acanthus on the north to the Singitic gulf opposite Torone; the cut would turn the six towns of the peninsula into islanders. The workforce was the army's own labour, "all sorts and conditions of men", conscripted in relays and joined by the local population, with the operation supplied by sea from a base at Elaeus at the tip of the Chersonese, which Hyland notes lay more than 160 km from the dig.^[7] The lash is part of the picture Herodotus intends: Bridges points to the phrase hypo mastigōn, "under whips", as one thread of the slave-motif that runs through his whole narrative of the invasion, the king's subjects labouring as a master's property while Greeks, in the set speeches, fight to avoid the same condition.^[4] The conscription by peoples also belongs to a documented imperial practice of assembling multinational workforces for royal projects (see the peoples of the empire).

The method of excavation follows:

the foreigners dug as I shall show, dividing up the ground among their several nations. They drew a straight line near to the town of Sane; and when the channel had been digged to some depth, some stood at the bottom of it and dug, others took the stuff as it was digged out and delivered it to yet others that stood higher on stages, and they again to others as they received it, till they came to those that were highest; these carried it out and cast it away.

Herodotus, Histories 7.23, trans. Godley^[2]

A marked line, sectors allotted by nation, and spoil passed up the trench walls by relays standing on stages: the description is coherent and practical. Then comes the celebrated engineering lesson. In every national sector except one, the steep sides of the cut kept collapsing and doubled the labour:

for inasmuch as they made the span of the same breadth at its highest and its lowest, this could not but happen. But the Phoenicians showed therein the same skill as in all else that they do; having taken in hand the portion that fell to them, they so dug as to make the topmost span of the canal as wide again as the canal was to be, and narrowed it ever as they wrought lower, till at the bottom their work was of the same span as what the rest had wrought.

Herodotus, Histories 7.23, trans. Godley^[2]

The observation is sound: an unsupported trench in soft ground stands only if its sides are battered back, and the Phoenician sector was dug wide at the top and narrowed with depth. The ground itself confirms the problem, for the isthmus is composed of soft tertiary sands and marls, and Hyland notes Herodotus's emphasis on "repeated cave-ins" among the causes of delay.^[8][7] How and Wells doubted the neat national moral rather than the fact: battering was "the only possible way of making the canal", so the exclusive credit to Phoenician cleverness suggests that Herodotus "is here repeating a popular story".^[8] Around the labour grew a service economy: "There is a meadow hard by, where they made a place for buying and marketing", Herodotus adds, with ground grain arriving continually from Asia.^[2] Briant draws the implication out: a market at the dig means the workers bought their rations, and "It thus seems certain that the workers purchased their rations", perhaps from wages paid in silver or tokens, the administration selling its own stores back to its own conscripts, a mechanism he traces across the empire's military logistics.^[9]

At the head of the work Herodotus names two Persians, and neither is a cipher. Bubares son of Megabazus is identified by How and Wells with the Bubares who, a generation earlier, had married into the Macedonian royal house: Herodotus tells how Alexander I of Macedon ended a dangerous Persian inquiry "by the gift of a great sum and his own sister Gygaea to Bubares, a Persian".^[10][11] If the identification holds, one of the canal's two overseers was kin by marriage to the dynasty whose coast the fleet would pass. Hyland treats the pair as standing officials "tasked with the oversight of the wider region", the canal one item in a larger Thracian portfolio.^[12]

Of Artachaees son of Artaeus, Herodotus has more to tell. When the army reached Acanthus in 480 the king first honoured the town for the canal:

When Xerxes came to Acanthus, he declared the Acanthians his guests and friends, and gave them a Median dress, praising them for the zeal wherewith he saw them furthering his campaign, and for what he heard of the digging of the canal.

Herodotus, Histories 7.116, trans. Godley^[13]

Then he lost his overseer:

While Xerxes was at Acanthus, it so befel that Artachaees, overseer of the digging of the canal, died of a sickness. He was high in Xerxes' favour, an Achaemenid by lineage; he was the tallest man in Persia, lacking four finger-breadths of five royal cubits in stature, and his voice was the loudest on earth. Wherefore Xerxes mourned him greatly and gave him a funeral and burial of great pomp, and the whole army poured libations on his tomb. The Acanthians hold Artachaees a hero, and sacrifice to him, calling upon his name; this they do by the bidding of an oracle.

Herodotus, Histories 7.117, trans. Godley^[13]

Godley reckons the stature: "This would make Artachaees eight feet high", about 2.4 metres, a giant's height reported with the mock precision of a measurement, four finger-breadths short of five royal cubits.^[13] The chapter compresses a great deal into a small space: an Achaemenid noble dying on station at the empire's far edge; a royal funeral in the field, with the whole army pouring libations; and then a Greek city raising a hero-cult to a Persian overseer at the bidding of an oracle. Herodotus writes of the cult in the present tense, so the Acanthians were still sacrificing to Artachaees by name in his own day, two generations after the Persians had gone. A named Persian official worshipped as a hero by a Greek community is a rarity in the record, and the cult bound the canal's memory to the town that had watched it dug.

The canal was one node in a system, and Herodotus says so himself. The verdict-chapter ends with the workforce's next assignment: "The same men who were charged with the digging were also charged to join the banks of the river Strymon by a bridge."^[2] The chapter that follows widens into cordage and food:

Thus did Xerxes accomplish this work; and for the bridges he charged the Phoenicians and Egyptians with the making of ropes of papyrus and white flax, and storing of provision for his army, that neither it nor the beasts of burden in the march to Hellas should starve; in such places as enquiry showed to be the fittest he bade them store it, carrying it to the several places from all parts of Asia in vessels of merchandise and transports.

Herodotus, Histories 7.25, trans. Godley^[14]

The depots are then named: the grain went, as each shipment was assigned, "to the White Headland (as it is called) in Thrace, or Tyrodiza in the Perinthian country, or Doriscus, or Eion on the Strymon, or Macedonia".^[14] Even the cordage carries information: Godley notes that the "white flax" was "apparently not really flax" but Esparto grass, shipped in by the Phoenicians from Spain, the empire's procurement reaching the far western Mediterranean for a fibre.^[14]

Hyland places the Athos works inside a regional programme of roads and logistics extending from the Propontis to Macedonia, under the oversight of Artachaees and Bubares; Herodotus assigns the same labour pool to the Strymon bridge, and Hyland associates it with work on the Eion–Doriscus road (see the Royal Road).^[12] A Persepolis Treasury tablet, PT 15, records payment to Egyptian, Levantine and Greek kurtaš engaged in palace construction in Fars from autumn 483 to spring 482 (see the Persepolis archives).^[15] Its dates and multinational workforce make it a useful analogy for Achaemenid capacity to coordinate royal construction, but a tablet about palace workers in Fars does not directly corroborate Herodotus's account of the Thracian workforce or its sequence of tasks. Tamukkan on the Persian Gulf offers another comparison: officials under Cyrus and Darius directed labour into palaces, roads and at least one canal.^[15] These parallels establish an imperial repertoire; they do not independently document the Athos dig.

Herodotus is specific that the work was finished, and he includes a detail that presupposes completed hydraulic engineering:

When the bridges and the work at Athos were ready, and the moles at the canal's entrances, that were built to prevent the surf from silting up the entrances of the digged passage, and the canal itself was reported to be now perfectly made, the army then wintered, and at the beginning of spring was ready and set forth from Sardis to march to Abydos.

Herodotus, Histories 7.37, trans. Godley^[16]

Breakwaters at both mouths against silt-bearing surf are a working engineer's provision, and evidence that maintenance was built into the design. The canal was then used for the purpose it was cut for:

Now when the fleet had left Xerxes and sailed through the canal made in Athos (which canal reached to the gulf wherein stand the towns of Assa, Pilorus, Singus, and Sarte), thence taking on board troops from these cities also, it stood out to sea for the Thermaic gulf.

Herodotus, Histories 7.122, trans. Godley^[16]

On Herodotus's own account, then, the fleet passed through into the Singitic gulf, levying troops from the gulf towns as it went, and continued toward the battles of the campaign (see Thermopylae).

Whether the canal was worth its cost is a separate question, and the modern excavation sharpened it. Benedikt Isserlin, who directed the investigations of the site, proposed that "a workforce of about 2,000 men could finish the canal in 60 to 80 days", against Herodotus's three years.^[17][18] Hyland replies that the numbers actually allocated are unknown, that the ground caused real delays, and that the three years belong to the region-wide programme rather than to the ditch alone; "perhaps the canal-builders finished their task early in order to turn to other assignments", such as the Strymon bridge.^[18] A minor dating dispute hangs on the same point: Diodorus synchronises the canal with the bridging of the Hellespont on the eve of the crossing, but Hyland holds that "Diodorus misses the larger context of region-wide preparations described by Herodotus", and keeps the start around the summer of 483.^[19][18] The engineering arithmetic cannot decide the older argument about motive. Herodotus gives neither the number assigned to the canal nor the duration of the digging itself, while Isserlin's estimate depends on a hypothetical workforce and Hyland's suggestion that labourers moved to other assignments remains a possibility rather than a record. Strategy and display may have coincided, but the surviving figures do not prove that the canal was over-provisioned.

Within a few centuries the canal's existence had become a test case for the credibility of Greek historiography. Juvenal, writing around 600 years after the event, made it the first item in a catalogue of Greek audacities:

People believe that ships once sailed over Athos, and all the lies that Greece has the nerve to tell in her histories: that the sea was covered with boats, and the ocean provided a solid surface for wheels. We believe deep rivers failed, that streams were all drunk dry by the Persians at lunch…

Juvenal, Satire 10.173–178, trans. Rudd^[20]

At line 174 the Latin is Graecia mendax, "lying Greece", and Bridges reads the sneer as knowing: for Juvenal the invasion stories "had been repeatedly reinvented, embellished and exaggerated until it had become impossible to distinguish historical truth from literary fiction".^[21] The mockery garbles its own target, ships sailing over Athos, the mountain standing in for the cut, and the garble is itself evidence of how far the story had drifted from the ditch in the ground. A more precise doubt was older and more serious. The Hellenistic scholar Demetrius of Skepsis, reported by Strabo, "declared that after passing through ten stades of earth the canal must have been stopped by a bank of rock a stade broad": a claim that the channel was never carried from sea to sea, and that ships must have been hauled across the gap.^[22][23] Lucian mocked the story too, and in the nineteenth century Stein and Wecklein revived a version of the doubt, which How and Wells pronounced "unwarranted": Herodotus, they observed, "distinctly affirms both the completion (ch. 37) and the use of the canal (ch. 122)".^[23]

The strongest ancient witness is not Herodotus at all. Thucydides, describing Brasidas's campaign of 424 BCE, introduces the peninsula as "a promontory running out from the king's dike with an inward curve", and places the town of Sane "close to the canal, and facing the sea in the direction of Euboea".^[24] Within living memory of men whose fathers had watched the digging, "the king's dike" was simply a landmark from which geography was measured, a matter-of-fact reference with no argument to make, which is exactly what gives it weight.

The ground kept its own record. Captain Spratt in 1847 "found distinct traces of the ancient cutting almost across the whole isthmus", failing only near the shores; the line survived, in Rawlinson's summary, as "a line of ponds from two to eight feet deep and from sixty to ninety broad"; and a stretch of about 200 yards at the centre where the ground showed no disturbance kept the partial-cut theory breathing, though it was already conjectured that the middle had been filled in later to restore a road into the peninsula.^[25] Godley's Loeb of 1922 could therefore assert, despite "the incredulity of antiquity", that the canal was "no doubt actually made and used", with traces "said to exist": an editor's confidence still running ahead of proof.^[26] Lazenby, writing in 1993 as the new fieldwork began, reported that "Traces of the Athos canal have been discovered" and were under study, and judged the canal "probably completed, as Herodotos says, despite the doubts expressed by Demetrios of Skepsis".^[27]

The investigations directed by Isserlin between 1991 and 2001 settled the question. In Hyland's summary, "Modern excavations have established its course between Nea Roda and Tripiti, with an estimated depth of 9 feet (2.7 m) and top width of 90 feet (27 m)".^[17][28] A course established from shore to shore leaves no room for Demetrius's stade-wide bank of rock, and none for the haulage it would have required; the undisturbed ground of the Victorian surveys overlay silted channel. There is even a small vindication inside the old doubt: the 100-foot breadth Demetrius reported agrees closely with the roughly 90 feet the travellers traced and the 27 m the survey measured, so the sceptic was carrying accurate information about a canal he argued had never been finished.^[28][25] Herodotus's twelve-stade isthmus, his completed cut, his moles and his fleet passage now stand corroborated at every point the ground can test.

In the Greek imagination the canal never stood alone. Bridges observes that "The paired actions of the carving of the Athos canal and the crossing of the Hellespont serve to demonstrate both the king's transgression and alteration of natural boundaries" and his role as "would-be enslaver"; the neat antithesis that Xerxes turned land into sea and sea into land "is not expressed by Herodotus, but would later become a literary commonplace".^[29] By 330 BCE the pairing was already rhetorical shorthand for the whole invasion. Aeschines, rousing a jury against the background of Alexander's war on Persia, asked:

Is not the king of the Persians, who dug through Athos, who yoked the Hellespont, who demanded earth and water from the Greeks, and who dared to write in his letters that he was master of all men from the rising of the sun until its setting – is he not fighting now not merely for power over others, but for his life?

Aeschines, Against Ctesiphon 3.132^[30]

The Persian side of the ledger is silence. No inscription of Xerxes mentions the canal, or indeed the campaign: as Bridges puts it, "there is no place for that campaign in the view left to posterity by the official Persian sources".^[31] Yet the act itself has an Achaemenid grammar. Darius I had cut a canal from the Nile to the Red Sea and commemorated it on stelae in the royal voice, the ships sailing through "to the sea which goes to Persia", "as was my desire".^[32] Hyland sets Athos beside it: the Greek cut "falls short in comparison with the length of Darius's more impressive canal near Suez", but "served a comparable purpose in removing a barrier to the passage of ships wheresoever the king wished to send them". Behind both stands an older Near Eastern idiom of royal mastery over waters and mountains; Hyland notes that "Sennacherib had corralled both loyal and rehabilitated Levantines, alongside Greeks of uncertain origin, into riverine shipbuilding projects on the far side of his empire".^[33] Read in that tradition, the canal is a routine assertion of the king's power over landscape, of a piece with roads, bridges and paradises; read in the Greek tradition, the same ditch is a crime against nature. The two readings describe one object.

The hubris-reading itself has a datable history. How and Wells noticed that Aeschylus never mentions the canal, although the *Persae* returns again and again to the bridged Hellespont; "Only later did the turning of land into sea as well as sea into land by Xerxes become a rhetorical commonplace".^[34] The oldest Greek indictment of Xerxes's engineering fixed on the bridge alone. The canal was recruited into the topos afterwards, which is a reminder that the moral reading of the work is reception, with its own growth over time, and cannot be quoted as evidence for what the digging meant in 483.

The moles at the mouths were built against silting, and silting won in the end. The canal is recorded in use only by the fleet it was dug for; no source records later passage, maintenance or the date at which the channel closed. By 424 the cut was a boundary-mark in Thucydides's geography; his Acte is a small polyglot world, its towns "inhabited by mixed barbarian races" speaking two languages, Tyrrheno-Pelasgian, Bisaltian, Crestonian and Edonian communities alongside a Chalcidian element, with Sane on the canal itself.^[24] The cut had been driven through the neck of a frontier district where Greek and Thracian settlement interleaved, and its human geography outlived its water.

In literature the canal had a second, longer career as a commonplace. Propertius, declining epic themes, waves away "two seas joined at Xerxes' command"; Manilius declines to sing how "a channel was let into the land, and a road made on the waves of the sea"; in Seneca's declamation exercises the works prove that arrogant good fortune, superba felicitas, "has no firm foundation"; and as late as the fifth century CE Sidonius was still listing Xerxes's crossing of the Hellespont and the Athos canal among the clichés a poet should avoid.^[21] For six hundred years writers who had never seen the isthmus, and who half-disbelieved the story, reached for the canal as the fixed emblem of engineering pride; Juvenal's "lying Greece" is the point at which the emblem consumed the fact.

The fact, however, outlasted the emblem. A silted channel about 2.2 km long still crosses the isthmus between Nea Roda and Tripiti, measurable in depth and breadth two and a half millennia after the Phoenician gangs battered back its walls.^[17][28] Of the invasion's two great engineering works, the bridges were cables and boats, gone within the year; the canal is the one work of Xerxes's war that can still be measured, and the one piece of evidence for it that no Greek wrote.