ACTA Cf,A8SICA XXXVIIT (1995) 57-69 ISSN 0065·1141

THEMISTOCLES AND THE SUPPOSED SECOND

MESSAGE TO XERXES: THE ANATOMY OF A LEGEND.
by John Marr
(University of Exeter)

Themistocles' deceptive message to Xerxes on the eve of the battle of

Salamis was the most famous of all the stratagems attributed in anti­
quity to that wily operator. All the sources agree that it had the effe€t
of inducing Xerxes to undertake the naval engagement within the narrow
straits between Salamis and mainland Attica, and that this afforded the nu­
merically inferior Greek navy an advantage which they were able decisively
to exploit. 1
Our sources also mention another message, supposedly sent by Themis­
tocles to Xerxes after the battle of Salamis was over. However they differ
very considerably over the circumstances, purpose, and details of this se­
cond message. The purpose of this article is to examine the various versions
of thf' story, as it is presented in the sources, to explain how these versions
developed, and to assess the historicity of the story itself. 2

The evidence
1. Our earliest authority for the story (which, interestingly, is not found
in Aeschylus' Persians of 472 B.C.--the earliest literary source for the
battle of Salamis) is Herodotus, Histories 8. 108-110, a passage written
probably in the 430's or early 420's. This is a complex, and strongly anti­
Themistocles version, in which:
a. After the battle of Salamis, at a council of the commanders of the Greek
fleet, which had advanced as far as Andros, Themistocles urged them to
sail at once to the Hellespont to break the bridge of boats there, so cutting
off Xerxes' forces and stranding them in Greece.
b. He was strongly opposed by the Spartan Eurybiades, the commander-in­
chief, supported by the other Peloponnesian commanders, and over-ruled
on the grounds that this would make Xerxes desperate, and force him to
devastate mainland Greece.
c. Themistocles returned to the Athenian contingent, who most strongly
wished to sail to the Hellespont, alone if necessary, and, changing his po­
sition completely, made a public speech successfully dissuading them and

57
advising them to return home and to leave any expedition to the Hellespont
to the following spring.
d. But, in saying this he was not sincere, he had an ulterior motive, since
he now wished to turn the situation to his personal advantage, and be able
to make a claim on the King in the future, if ever he should get into trouble
with the Athenians and need a place of refuge, 'something which did in fact
occur' (a comment on the historian's part which should arouse suspicion).
e. He thereupon immediately sent a secret message to Xerxes via a trusted
group, who travelled in a boat to Attica. They included Sicinnu8, who had
also been the bearer of the pre-Salamis message. The message said that
Themistocles had done Xerxes a favour, by preventing the Greeks from
pursuing the Persian fleet to the Hellespont and destroying the bridges
there. So Xerxes could now return home at his leisure.
It should be noted that there is no mention of Aristides at any point in
this account.
2a. Diodorus, 11.19. 5-6 (the source for which was probably the fourth
century B.C. 'universal' historian Ephorus),3 says of Themistocles, after
the battle, but apparently still at Salamis, 'he devised another stratagem,
no less clever than this (previous) one .... He succeeded in greatly reducing
the numbers of the Persian army in the following way. He sent the tutor
of his sons to Xerxes to reveal to him that the Greeks intended to sail to
the span of boats and to destroy the bridge'. Xerxes believed this 'because
it was plausible' (cf. 11.17.2), and, fearing that he might be completely
prevented from getting back to Asia, since the Greeks already controlled
the sea, decided to return to Asia as quickly as possible, leaving behind
Mardonius with a picked force, still considerable in size, (400,000 men!),
but much reduced in comparison with the original.
In this account the message is presented as a successful patriotic
stratagem, and there is no mention of any expressed intention on Themis­
tocles' part to delay or prevent the Greek pursuit, let alone a claim to have
done so (as in Herodotus). The message is simply a (bogus) intelligence
tip-off about the Greeks' intentions. The messenger, though not named,
seems to be Sicinnus (as in Herodotus' version), since he was the tutor of
Themistocles' sons (d. Hdt.8.75).
2b. Diodorus 11.59.2. This passage comes within an interesting encomium
on Themistocles' life and achievements. As before, Diodorus follows the
tradition that the message concerning the Hellespont bridge was a suc­
cessful patriotic stratagem, which, we are this time told, reduced Xerxes'
forces in Greece by half: 'by the one plan concerning the bridge of boats
he reduced the land forces of the enemy by half, so it could be easily dealt
with by the Greeks'.

58
3. Nepos Themistocles 5. 12. Xerxes was discomfited a second time by
the same man; Themistocles, fearing that Xerxes would prolong the war,
informed him (Nepos does not say how, via whom, or from where) that
there was a plan to destroy the bridge which he had made at the Hellespont,
and to cut off his return to Asia; and he persuaded him of this. Xerxes
therefore returned speedily (in less than 30 days) to Asia and considered
that he had not been defeated bu t saved by Themistocles.
Here again the message is presented as a succesful patriotic stratagem,
which got Xerxes and his forces out of Greece, and in it Themistocles
doec. not promise to do anything, or claim that he has done anything.
Xerxes' gratitude towards Themistocles is due simply to the warning that
the Greeks planned to destroy the bridge, a warning which he mistakenly
believes is sincere.
4. Plutarch Themistocles 16. 2-6. Themistocles proposes to his fellow
Athenian and political rival, Aristides, apparently while the Greeks are
still at. Salamis, that they should sail to the Hellespont and destroy the
bridge of boats. Aristides opposes this suggestion, offering the 'trapped
rat' theory which Herodotus assigned to Eurybiades (8.108). Themistocles
then changes his position. There is an implication in this account that his
original proposal was not serious, but no suggestion that he is acting out of
pique, or that he is exploiting the rebuff to further his own self-interest. He
suggests that everyone should concentrate on devising a plan to get Xerxes
out of Greece 0.1> lJ.uickly as possible. Anstides agrees. Themistocles then
sends Xerxes a message, via a captured royal eunuch, Arnaces, to say that
the Greeks have decided to sail to the Hellespont to destroy the bridge,
but that Themistocles, out of regard for the King, urges him to get back
across it quickly, while he (Themistocles) would cause some delays in, and
postponements of, the Greeks' pursuit. This Xerxes does, and so, thanks
to Themistocles and Aristicies, when the Greeks fought afterwards against
Mardonins in the decisive battle at Plataea, the numerical odds against
them were much reduced.
Here again, as in Diodorus and Nepos, the message is presented as a
patriotic stratagem, in this version planned jointly with Aristides the Just. 4
However, it is more than just a tip-off. Themistocles promises action on his
part. i.e. delaying tactics. This detail, like the change of position, recalls
the' version of Herodotus, where, however, Themistocles claims he already
has prevented a Greek pursuit. But, according to Plutarch, Themistocles'
promise to delay the Greeks' pursuit out of his supposed concern for Xerxes,
is not sincere. It is just part of the deception plan (though it does seem
to involve a certain lack of logic-why should Xerxes hurry to leave, if the
Greek pursuit is going to be delayed?).
On the name of the messenger, Plutarch has gone to a different, unknown,

59
source (cf. Polyaenus 1.30.4), presumably because of the credibility problem
involved in Herodotus' assignation of both deceptive messages to the same
bearer, Sicinnus.
5. Plutarch Aristides 9.5~6. This version is similar to that at Them'istocles
16.2--6, in that Themistocles' original suggestion is to sail to the Hellespont
at once, and it is made to Aristides, who opposes it. But here it is not fol­
lowed by a jointly devised patriotic stratagem to get Xerxes out of Greece
as quickly as possible. Instead Plutarch says simply 'So Themistocles sent a
eunuch prisoner of war, Arnaces, secretly, to tell the King that he, Themis­
tocles, had succeeded in turning back the Greeks after they had set out
intending to destroy the bridge, because he wished to save the King'. In
this version the message is a secret one, unknown to Aristides or the other
Greeks, and, though Xerxes and part of his army do subsequently withdraw
from Greece, there is no suggestion in this account that this wat> Themis­
tocles' patriotic objective, or that the claims in the message of concem for
the King were merely a deception. Themistocles reacts in an apparently
purely self-interested way to having his original proposal rejected, <1ud in
his message he claims that he ha" already done the King a favour by stop­
ping the Greek pursuit. This venlion is much closer to that of Herodotus,
though again as in Plut. Themistocles 16.2~6, there is a certain lack of logic
about Xerxes' haste to retreat, which it presents as a consequence of the
message.
6. Thucydides 1. 137.4. Themistocles, after being exiled by the Athenians,
spent some time in N.W Greece, and then travelled across the Aegean to
Asia, He reached Ephesus, and travelled inland with one of the Persians
who lived on the coast. He then wrote a letter to King Artaxerxes, the son
of Xerxes, 'who had recently come to the throne' (i.e. in late 465 or 464,
some 15 years after Salamis).5 Thucydides quotes some of the text of this
lett.pr, and summarises the rest (the clause in parenthesis):
EO~),OU oE ~ ypcrcp~ O'tl ~ 8£[llcr'tOX),iic; ~xw rccrpCt crt, OC; 4
xcrxCt [lEv rcAe:lcr'tcr 'EAAf;vwv dpycrcr[lcrl 'tov uf,le'I£pov olxov,
ocrov Xpovov 'IOV crov ltcr't£pcr b:lovecr E[lOt &v:iyxn ~[l\Jvo[l't)v,
rcoAu 0' En It),,dw &ycre&, eltElO~ EV 'Iiii acr(pcrAEl [lEv E[lot,
ExdV<jl OE EV E1tlXl vMv<jl TC&),lv ~ &ltOXOf,llO~ ey(yvno. xcrl (LOl
c::uc::pyecrtCt OcpdAEWl (yp&rpCtC; ef;v e<: EX I:a).,a[llvoc; TCpO&YY<:Acrlv
'Iii:; &VCtxwpf;crc::wc; xcrl 'I~V eWV YE(PUPWV, ~v rpc::UOWC; ltpocrc::­
TCotf;crCHO, 'toeC:: Ol' atnbv ou Ol&AUcrlV), XCtt viJv exwv crc:: f,l£A&ACt
&YCteCt opiicral n&pEl[ll OlWX0f,lEVOC; t.mo eWV 'EAA+,vwv OlCt e~v cr~v
(plHav, ~OUAO[lCtl 0' EVlCtUeOV Eltlcrxwv aUeoe; crot itEPl. i:)v ~xw
o't)Awcral.'
'I, Themistodes, have come to you, a man who more than any other Greek
did harm to your house, during the time when I was forced to defend myself

60
against your father, who was attacking me; yet I did even more good at
the time of his withdrawal, when I was safe and he was in danger. I am
owed a favour in return ... and now I am here, an exile pursued by the
Greeks because of my friendship for you, with the power to do you much
good ... '
There has been much dispute over the correct translation and interpre­
tation of the parenthesis here. For the moment I leave it untranslated, but
I will ['('turn to this point in the next section of the paper.
7. Diodorus 11.56 5-8, records Themistocles' request for asylum from the
Persian King on the same occasion as that referred to by Thucydides,
though, following Ephorus (cf. Plutarch Themistocles. 27.1), he makes the
King Xerxes, not Artaxerxes, and refers not to a letter, but to speeches be­
fore UUl King, the first delivered on Themistocles) behalf by one Lysitheides,
acting as an intermediary, the second by Themistocles himself. However,
DiodorUf! does not report any of the arguments used by Thernistocles, or on
his behalf, apart from commenting rather vaguely that the King 'realised
that Themistocles had done him 110 wrong and so absolved him from pu­
nishment'. It has been thought that Diodorus' probable source, Ephorus,
did report the arguments. A second century A.D. papyrus (Pap. Oxy.1610
F1, 7 12 FGH 191.1) has AeJyoVO'L 0' 01. (lev O[1:l (me]/-lv'f]aEv ctU1:[ov 6)v]
T.:Epl "CE -cij<; v[ctV(lctJX!ct<; XO!L1:ij<; Y[EqJ1.JpctC; ltpol~YYElAEv.
'Some say that he reminded him of the warnings he had given
ctHlcerning both the sea battle and the bridge'. However, whoever
this author was, it was almost certainly not Ephorus. 6
8. J\epos Themistocles 9.1-4. Nepos is obviously following Thucydides,
1.137. 34, very closely here. III 9.1 he refers to the tradition that Themis­
toclu::l lllet Xerxes, but says that he prefers to believe Thucydides, who was
a fellow A thenian, and the authority closest in time to the actual events. Is
(i.e. Thucydides) autem ait Arta1:er".Een emIt venisse atque his verbis epistu­
lam m:i8i8se 'Themistocles veni ad te, qui plurima mala omnium Gmecorum
in dam:u.rn tuam intuli quamdiu mihi necesse fuif adversum patrem tuum
bel/are patriamque meam defendere. Idem multo plum bona feci postquam
in tnto 'ip,qe et ille in periculo esse coepit. Nam cum in Asiam reverti vel­
let, pmelio apud Salamina facto, litteris eum certiorem feci id agi, ut pons,
quem, in Hellesponto fecemt, dissolveretur atque ab hostibus circumiretur,
quo nuntio ille periculo est liberatus. Nunc autem confugi ad te exagitatus
a cnncta Graecia, tuam petens amicitiam .. ,
It i::l, surely, very significant that Nepos does not make Themistocles in
his letter mention, or claim any credit for, his famous pre-battle message;
he refers back only to his post-battle message, proelio ... facto. Since
Nepos is following Thucydides virtually word for word here, the obvious

61
inference is that he did not interpret Thucydides' parenthesis at 1.137. 4
as containing any reference to Themistocles' pre-battle message.
9. Plutarch Themistocles 28. 1-5. Themistocles' letter to Artaxerxes in
465/4 has here become a speech, delivered in person (cf. Diodorus 11.56.8,
passage no 7). Plutarch, like Nepos, is obviously following Thucydides'
text at L 137.4 closely; he employs a number of very similar expressions.
Thus Themistocles refers to himself as a man iJ) rron& !ltv OrpdAOUcrt l1tpaCtl
XCtx& (i.e. his role in their defeat at Salamis), rrAdw oe ayCt9& xw)"'JcrC(V"Cl
't~v Otw~t v (by preventing the Greek pursuit), ace ,~c; 'EAACtOOC; EV acrrpCtAEl
YE\l0!lt\lT)C; 1tCtptaXE ,a OlXOl crw~6!lEVc( X{lpiO'(WeCtl 'tl x{ll V!llV. However, like
Nepos, Plutarch does not make Themistocles claim any credit for the pre­
battle message, and that, in my view, is a vitally important consideration
for a correct interpretation of the parenthesis at Thucydides 1.137.4, to
which we can now return.

The translation and interpretation of Thucydides 1.137.4

(passage no. 6)
Recent commentators (e.g. Gomme 7 and Hornblower 8 ), have taken the first
part of the parenthesis to mean 'citing the announcement (to Xerxes) of
the (intended) retreat (by the Greeks) from Salamis', and assumed that
Thucydides is saying that Themistocles in his letter of 465/4 made reference
to both his 'first' (pre-battle) and his 'second' (post-battle) messages in 480,
and claimed credit with Artaxerxes for both.
But the problem with this interpretation is that, as has been well pointed
out,9 the deliberately deceptive character of the notorious pre-battle mes­
sage was not something that could have been hidden from the Persian
King, or even plausibly contested, in 465/4. In Aeschylus' Persians, a
play performed less than eight years after Salamis, and seven years before
Thernistodes' flight to Asia, the true patriotic character of that message
had been very publicly commemorated and celebrated. There were many
influential Greek exiles at Artaxerxes' court, eager to ingratiate themselves
with a new King, who, through their contacts, would have known precisely
why Thernistocles had become famous throughout Greece for his cleverness.
It is inconceivable that Themistocles could have expected to be taken seri­
ously by Artaxerxes in a claim in 465/4 that his pre-Salamis message had
really been a favour conferred on the Persians.
There is a very clear division made in the quoted letter (signalled by fltv
... 0'), to indicate two different periods of time, one when Themistocles
claims he was under attack by Xerxes, and so forced to defend himself
(hence the 'great harm' which he did to Artaxerxes' house at that time
was, to some extent, pardonable); the other when he was 'safe' and Xerxes
was 'in danger', which was when he did the Persians a service in return

62
for which he now claims to deserve some favour. But what else can these
two periods of time be but a) the period up to and including the battle of
Salamis and b) the period after the Greeks' victory there? And it seems
clear from the context that it is only to events within the latter period
that the phrase X::Xl [lOt EUl::pYl::ry[::x Oq>dAE"tcn, and the following explanatory
parenthesis apply.
A close analysis of the word order of the phrase ypa~::x<; "t~'J "tl:: Ex
E::X),::X[lL'JO<; npo:XYYI::Acrt'J "t~<; &'Jctxwp~ryEW<; also suggests that the transla­
tion of Gomme and Hornblower, 'citing the announcement (to Xerxes) of
the (intended) retreat (of the Greeks) from Salamis', cannot be the right
one. The position of tx EctAct[lL'JO<;, coming between and framed by the ar­
ticle 7~'J and its noun c:po:xyye:AOL'J, suggests that the four words go closely
together in sense, 'the warning message sent from Salamis'. This is surely a
much lllore natural interpretation of the words than to take EX E::XAct[ll'JO<;
with ;;1i<; &'JctXWp~ryEW<; 'the retreat from Salamis'. Thus if the more natural
interpretation is right, then "t~'J • •• rcpOCtYYEAOL 'J "t~<; &'JctXWp~ryEW<; should
mean 'warning message for his retreat' i.e. to retreat from Greece. This
does Hot seem an inherently impossible sense for the word rcpoaYYEAcrt<; to
bear, especially as it is found only here in Greek literature. lO
Both Nepos, Themistocles 9.2, and Plutarch, Themistocles 28 1-5, are
obviously following this passage of Thucydides very closely, with clear
similarities in their language. Yet they do not present Themistocles as
making any claims at all in his letter/speech about his first, pre-Salamis
message. The 'good things' which he alleges the Persians owe to him are
solely things which he did for them after the battle. The clear implication
is that neither Nepos nor Plutarch interpreted Thucydides' parenthesis as
involving any reference to Themistodes' pre-Salamis message.
The conclusion must surely be that there is no implication in Thucydides'
parenthesis that Themistocles, in his letter to Artaxerxes in 465/4, made
any reference at all to his pre-Salamis message to Xerxes.
Fmtlwrmore, Thucydides does not mean to suggest in the second half
of his parenthesis, xcti ... OU OlCtAUOL'J, that Themistocles' claim to have
prevpnt,pcl the immediate destruction of the Hellespont bridges (presumably
by delaying the Greek pursuit) was a claim or a promise originally made in
his post-Salamis message. It is, surely, a fresh claim, made now for the first
tim.l~ in his letter of 465/4. It presumably helped to explain to a sceptic
why, despite the supposed aggressive intentions of the Greeks in late 480,
they did not in fact venture as far as the Hellespont till late 479 (Herodotus
9.114).
Some manuscripts omit the -Cl:: between "t~'J and EX EctA::X[ll'JO<;, but it
seems preferable to retain it, and to translate the parenthesis in full as
follows:

63
 'citing both the forewarning message sent (by him) from Salamis (to
Xerxes) to withdraw (from Greece), and the (subsequent) Greek fail­
ure to destroy the bridges at that time, which he falsely claimed (in
his letter) had been due to him'.
There is a slight discrepancy of detail between Herodotus and Thucydides
(which has perhaps misled commentators) about the place from which the
'second message' was sent--Herodotus says Andros, after the Greek fleet
had advanced there from Salamis. But there appear to have been two an­
cient traditions about this. Diodorus 11.19.5-6, and Plutarch, Themistocles
16.2-6, seem to be following the alternative tradition found in Thucydides,
that it was sent from Salamis.
Thucydides' compressed and rather obscure expression in this parenthe­
sis seems to have been misunderstood by the unknown authors quoted by
the unknown author of the fragment preserved in a second century A.D.
papyrus (see passage no. 7), as well as by many modern scholars. However,
as we have seen, it was not misunderstood either by Nepos or by Plutarch.
To sum up. The conclusion must be that Thucydides' summarising
parenthesis refers to only one message sent by Themistocles; this was the
one supposedly sent after the battle of Salamis (the 'second message'), and
the implication is that it said simply: 'You must retreat immediately, as the
Greeks are intending to sail to the Hellespont and destroy the bridges'. For
Thucydides this message had contained no promises to delay the Greeks'
pursuit, or claims to have done so. It was purely an intelligence tip-off (as
it is in Diodorus and Nepos).
Subsequently, in his letter to Artaxerxes in 465/4, Themistocles, accor­
ding to Thucydides' informants, referred back to that message, and alleged
that it had in fact been a service done on his part to the Persians, for which
he now deserved a favour in return, i.e. political asylum. In addition he
made the (untrue) further claim, that the fact that the Greeks had not
immediately sailed to the Hellespont and destroyed the bridges was his
(Themistocles') doing. He had delayed their pursuit.

The origins and development of the 'second message' tradition.

If we accept, following a unanimous ancient tradition going back to Aeschy­
lus' Persians, that the famous pre-Salamis deceptive message sent by
Themistocles to Xerxes is historical,11 then it is in the highest degree un­
likely that, just a few days later, in the bitter aftermath of the Persians'
stunning defeat, which was at least partly the result of their misguided
belief in that message, either a) Xerxes and his high command would have
been prepared to believe any further supposedly confidential messages em­
anating from anyone on the Greek side, let alone Themistocles again, or b)
the Greeks could have thought that they would.

64
 Furthermore, although there may have been some argument on the Greek
side, in the days after Salamis, about how soon and how far to press the
naval offensive in the light of their unexpected victory, it is surely incredible
that any Athenians, however bold, would have suggested that the Greek
fleet should do something as risky as to sail 150 miles north to the Helle­
spont, so late in the year (it was now October), while the territory of Athens
and Attica was still under Persian occupation, and when they were still in
the dark about the exact condition and morale of Xerxes' fleet. It seems
virtually certain that, in the immediate aftermath of Salamis, Themistocles
did not urge the other Greek commanders (or Aristides, for that matter),
to sail the fleet to the Hellespont, and hence that there was no argument
about this, where he was opposed by Eurybiades (or Aristides), and as a re­
sult of which he was outvoted (or persuaded to change his mind). Thus the
context provided by our sources for Themistocles' supposed second mes­
sage is a fictitious one, whether the message is to be viewed as a patriotic
stratagem or as a self-serving piece of opportunism.
It is also surely significant that there is nothing in our earliest source for
these events, Aeschylus' Persians (472 B.C.), about a message, deceptive
or otherwise, sent to Xerxes after Salamis, in the section of the Persian
messenger's speech where it might have been expected (lines 480-514).
In fact the Hellespont bridge is not mentioned at all there, and Xerxes'
post-Salamis withdrawal is presented (admittedly with much patriotic ex­
aggeration for a Greek audience) as a disorderly and disastrous rout, which
did not require any further spurring on after the naval defeat.
After the victories at Plataea and Mycale in 479, and the siege and
capture of Sestos, on the Hellespont, by the Athenians during winter 479/8,
now that, for the first time since the invasion began, the security at Greece
was properly assured, there may have been some argument over whether
they should have sailed to the Hellespont earlier, and attacked the bridges
sooner, and so won an even more crushing victory. Greeks being Greeks,
it is quite likely that some national commanders claimed that they had
wanted to do so immediately after Salamis, but had been opposed and out­
voted by the others. Hence the notion, that there had been some argument
about this issue immediately after the battle, may have got into circulation
quite early on. But this was an erroneous notion, and, as we have seen,
all the indications are that the alleged message, which is presented as a
consequence of that argument, is unhistorical.
15 years after Salamis, in 465/4, Themistocles now an Athenian exile,12
was given sanctuary by the Persian King Artaxerxes. It seems to have been
customary for those who sought favours from the King to claim they had
done him or his family some services (cf. Pausanias' remark in his letter to
Xerxes, reported at Thucydides 1. 128.7). It is quite possible that in his
initial approach, probably by letter, supported verbally by an intermediary,

65
Themistocles made the conveniently uncheckable claim that, at a council
of the Greek commanders after Salamis, he had successfully opposed a
suggestion that the Greeks should sail immediately to the Hellespont to
destroy the bridges, and so had been intentionally instrumental in allowing
Xerxes to get back safely to Asia. In this approach to Artaxerxes in 465/4
Themistocles made no claims about his fateful pre-Salamis message (how
could he?), nor did he claim credit for what he had said in any post-Salamis
message to Xerxes either, since no snch message had been received or sent.
However, for reasons which will soon become apparent, in later reports
of his letter to Artaxerxes, such as the one at Thucydides 1.137.4, he is
presented as referring to a post-Salamis message, and as claiming credit for
it.
Any claim by Themistocles to have done the Persians some service wa.'i
not very plausible, but doubtless what really counted with Artaxerxes,
in acceding to Themistocles' request at the time, was not any dubious
claim of past services, or optimistic promises for the future, but the sheer
propaganda value of the defection of the famous Themistocles at a time
when Persian morale in the Ea.'lt Mediterranean was IOW. I3
After Themistocles' sons and family returned to Athens, probably in the
early 450'S,14 there seems to have ensued a fierce and long-lasting propa­
ganda battle between what may perhaps be called 'conservatives' (men such
as Cimon and the Alcmaeonids, who were virulently anti-Themistocles) and
'radicals' (e.g. Ephialtes and Pericles, pro-Themistocles), and their respec­
tive political supporters. IS
It wa.'3 of course common knowledge in Athens in the 450's that Themis­
tocles had been given sanctuary and honours by the Persian King. But why
was it that the victor of Salamis had been so favourably treated by the son
of his bitterest foe? The conservatives were ready with an answer. They
claimed that the reason Artaxerxes had responded so positively was that
Themistocles-~who) after all, had allegedly been involved in the treasonable
activities of Pausanias in the late 470's, and had consequently been con­
demned in absentia by an Athenian court in c.469 16 --had actually started
his treasonable career very much earlier, immediately after the battle of
Salamis. That was when he had, on his own authority, without consultillg
the other Greek commanders, sent a secret message to Xerxes, to the ef­
fect that Xerxes could get back to Asia safely and at his leisure, because
Themistocles 'Would hinder and delay the Greek pur8uit (and he did indeed
do so, when others wanted to press on to the Hellespont); and they claimed
that this message had been sent with a view to the future possibility of his
defection to the Persian side.
This hostile version is substantially the one which appears in Herodotus
and at Plutarch Aristides 9.6. Doubtless it came to Herodotus mainly from

66
his Akmaeonid informants, themselves suspected of having sent a treason­
able signal to the Persians at Marathon in 490, (cf. Hdt. 6.115,121-4), and
so, for them, a particularly satisfying piece of revenge on their hated enemy.
Herodotus and Plutarch in the Aristides actually make Themistocles claim
in his message that he already has prevented any Greek pursuit. This par­
ticular variation of/improvement on the original story may perhaps have
been influenced by an originally separate tradition that Themistocles made
a public speech to the Athenian sailors at the very end of the campaigning
season in 480 (the substance of which is correctly reported by Herodotus,
but which he misleadingly presents as preceding the second message)17 say­
ing, in essence, 'Let's be sensible and go home for the winter sowing. We
can push on to the Hellespont and final victory next spring' (cf. Hdt. 8.109).
The 'radicals' ingeniously countered this allegation, not by denying that
any IllPssage was sent by Themistocles at all after Salamis--they accepted
thp 'tradition' that there had been a message----but by asserting that the
message had had a fine patriotic motive. It was another piece of classic
Themistoclean deception of the enemy-just like his pre-Salamis message,
which everyone knew about. It had been agreed to by the Greek high
command, and approved even by his political rival, the paragon of virtue,
Aristides the Just. It contained a bogus warning, and its purpose was to get
Xerxes and a large chunk of his army to withdraw from Greece as quickly
as possible, which it brilliantly succeeded in doing. Naturally, to ensure
credibility, Themistocles had had to claim in it that he was acting out of
friendship for Xerxes, but he had made no promises, or claims of action
on his part. The message had said simply, 'Retreat immediately, as the
Greeks are about to sail to the Hellespont, and cut you off by destroying
the hridges. Your friend, Themistocles'.
According to the radicals Themistocles had originally been all in favour
of a hold strike at the Hellespont bridges immediately after Salamis, but
he had been opposed and thwarted on this by the over-cautious Pelopon­
nesians. (This is the one pro-Themistocles detail which has been incor­
pora.ted into Herodotus' otherwise hostile account. Hence Themistocles'
treasonable message is implausibly presented by Herodotus as motivated
primarily by pique at having his first plan rehuffed.)
Thernistocles loyally accepted this decision, they argued, and then came
up with an alternative stratagem, designed to get a large part of Xerxes'
occupying forces out of Greece as quickly as possible. So there had indeed
been a post-Salamis deceptive message, but it had been just as patriotic
and successful as his pre-Salamis one.
Some years later, in 465/4, when Themistocles, in desperation, had had
to approach Arta.x:erxes for sanctuary, he had (according to this sympa­
thetic version) referred back in his letter to that post-Salamis message, and

67
had claimed that it had actually been a genuine warning to Xerxes, moti­
vated by genuine concern; and he had bolstered his appeal to Artaxerxes
by a further claim, that he had subsequently delayed the Greek pursuit,
thus explaining why the Greeks had not in fact reached the Hellespont till
late 479. All of this, of course, had been merely artful deception, to fool
Artaxerxes into offering him asylum.
This sympathetic version is the OIle which is implicit in the parenthesis at
Thucydides 1. 137.4, and is substantially the one we find in Diodorus, Nepos
and Plutarch's Themistocles, except that at Themistocles 16.5 Plutarch
makes Themistocles add to his message, with a certain lack of logic, a
promise that he would delay the Greek pursuit. This element was not a
part of the original pro-Themistocles version of the post-Salamis message
(it belonged to the hostile version), but it is easy to see how it came to be
attached to it subsequently, either by Plutarch or his source, since it was
accepted even by the radicals that Themistocles had retrospectively, i.e. in
his letter to Artaxerxes in 465/4, claimed that he had delayed the Greek
pursuit after Salamis (though that was an untrue claim).
Thus, for the pro-Themistocleans from the late 450's onwards, the reason
Themistocles was given sanctuary by Artaxerxes in 465/4 was that the
King wa..'l induced by their hero to believe mistakenly a) that Themistocles'
post-Salamis deceptive message of warning about the Hellespont bridge had
actually been a service to Xerxes, motivated by genuine concern for him.
and b) that Themistocles had also deliberately delayed the Greek pursuit
of his father, to allow him ample time to get back to Asia.
The tradition of a second message thus became a fable conver1:Ue, 'an
agn>ed fable' accepted by both sides, and these two rival versions of it, both
of whic:h originated in the 450's, were enthusiastically propounded by the
pro- and anti-Themisto cles camps in the battle for public opinion at Athens.
At first they were oral traditions, the debate being conducted by argument
in the assembly and other public fom at Athens. But within a generation,
when accounts of, and details from, earlier fifth century Athenian history
started to be written down in the works of e,g, Herodotus, Thucydides
and Hellanicus of Lesbos, the two rival versions were recorded in a more
permanent form. They thus had a strong influence OIl the subsequent
character of the Themistocles tradition in literature, though, inevitably,
over time they became, to some extent, confused with and contaminated
by each other. The co-existence of two rival original versions probably
also accounts for some minor variations of detail found in the accounts in
our sources i.e. the message was sent from Salamis/Andros; the messenger
was Sicinnus/ Arnaces/the tutor of Themistocles' children; he was the same
man as/different from the bearer of the pre-Salamis message.
The only trouble for historians was, and is, that they were rival versions
of what was almost certainly a wholly fictitious event~the sending of a

68
supposed 'second message' to Xerxes, after the battle of Salamis.

NOTES

1. E.g. Aeschylus Persians 353-373, Herodotus 8.75-6, 79-80, Thucydides 1.74.1,

Diodorus 11.17.1,4, Cornelius Nepos Themistocles 4.3-4, Plutarch Themistodes
12.4-5. The additional detail in Herodotus' account, which is found also in Diodoru8
and Nepos, that this message was also kept a secret from the other Greeks, and had
the effect of forcing them to fight against their will, is extremely unlikely (see e.g.
F.J. Frost, Plutarch's Themistocles [1980] pp. 139··143). It is, however, consistent
wit.h the strongly anti-Themistocles character of Herodotus' narrative throughout
Book 8 of the Histories (for which see e.g. A.,J. Podlecki, The Life of Themistocles
[1975] pp. 67-72).
2. As one would expect, a very large number of secondary works and commentaries,
which deal inter alia with the battle of Salamis and the role of Themistocles, refer
to the story of the second message. Some scholars accept it as historical, e.g. Frost
op.cit. p.164; some are sceptical of some of the details, e.g. A.R. Burn, Persia
and the Greeks (2nd edition, 1984). pp. 468-470 and note 52, How and Wells, A
Commentary on Herodotus (1912), vol.2 p.272; some reject it completely e.g. ,J.E.
Powell, Herodotus Book VIII (1956), p.135, C. Hignett, Xerxes' Invasion of Greece
(1963) pp. 241-·2. However there does not seem to have been a comprehensive
examination and comparison of the different versions of the story which are to be
found in the sources.
3. Set" e.g. Podlecki, op.cit. p.92.
4. This detail may possibly have been familiar to Ctesias, a Greek doctor at the Persian
court, who wrote a Persica, probably in the late 390's, d. FGH 688 F13 (30).
5. For the date see M.E. White, 'Some Agiad dates: Pausani&<; and his sons', lHS 84
(1964), p.142 and note 13.
6. See F.,J. Frost, op.cit. p.162 note 51, referring to the article by T.W. Africa, AlP
88 (1962), 86·~89.
7. A.W. Gomme, A Historical Commentary on Thucydides voU (1945), pp. 440-441.
8. S. Hornblower, A Commentary on Thucydides, voU: Books I-III (1991), p.222.
9. e.g. by Frost, op.cit. pp. 261~7.

10. See Liddell and Scott, Greek··English Lexicon, 1tpo&YYE:Amc;.

11. Of recent authors only Hignett, op. cit. pp. 403-408, rejects the story entirely.
12. Thucydides 1.135.2~3, 138.6, Nepos Themistocles 8.2-3, Plut. Them.23.1.
13. d. Pluto Them. 28.6, Frost op. cit. p.217.
14. See J.K. Davies, Athenian Propc,·tied Families 600-300 B.C. (1971) 6669 VI p.218.
15. For Cimon and the AJemaeonids &'> Themistocles' enemies cf. Plut Them. 20.4, 23.1,
Aristides 25.10, Cimon 10.7. For Ephialtes and Pericles as his successors and
political heirs cf. Ath. Pol. 25. 3.. 4, 26.1, 28.2, Pluto Cimon 15. 1-3.
16. For the chronology see e.g. Podlecki, op.cit. pp.197-8.
17. See Burn, op.cit. p.470.

69