1. ELITE CAVALRY IN HERODOTUS

Any study of the elite Achaemenid cavalry must inevitably begin with Herodotus, who devotes considerable space to describing the forces of the Great King Xerxes. One might imagine that some useful information on the composition of elite Persian cavalry units before this time might emanate from Xenophon's Cyropaedia. Yet it is generally understood that this work, in the main, constitutes a form of didactic historical romance.Footnote ¹⁰ As a result, it is of very limited utility for our purpose. It will be only be adduced where it has the potential to corroborate other snippets of evidential material, although considerable caution will nevertheless be warranted when using this text.

Persian cavalry are described at Marathon (490 b.c.) during the reign of Darius I, but there is little in Herodotus' description (6.112–13) that can be used to distinguish elite from regular cavalry. Furthermore, cavalry are not specifically mentioned in Herodotus' brief description of the battle, which seems to have been largely an infantry contest.Footnote ¹¹ Now to Xerxes' reign. In the oft-cited passage (Hdt. 7.40.1–41.2) describing the appearance of Xerxes' army before it made its way across the Hellespont on a pontoon bridge, some relevant information is provided. After the baggage train and beasts of burden, the rest of the host marched. But when more than half of this rather motley assemblage had passed, the truly Persian troops started to appear. First came ‘1,000 horsemen, selected out of all the Persians’ (ἱππόται χίλιοι ἐκ Περσέων πάντων ἀπολελεγμένοι, 7.40.2), then 1,000 spearman (αἰχμοφόροι, 7.40.2) carrying their spears turned downwards, with golden pomegranates on the butts of their spears. Next came a chariot carrying the chief Persian deity, and then Xerxes himself, followed by another 1,000 spearmen (αἰχμοφόροι again, 7.41.1) with their spears, bearing golden apples, carried in the customary fashion. After these, ‘1,000 picked Persian horsemen’ (ἵππος ἄλλη χιλίη ἐκ Περσέων ἀπολελεγμένη: 7.41.1) followed, and then 10,000 Persian infantrymen, with 1,000 bearing spears with golden pomegranates on the butts, and the rest bearing ones of silver. After these troops came ‘10,000 Persian horsemen’ (ἵππος Περσέων μυρίη, 7.41.2) and, after a space, the rest of the army marched.

What to make of this is problematic, though one might as well start with the infantry, since they have already been discussed elsewhere.Footnote ¹² In any case, identifying these units also has clear implications for the cavalry. Herodotus describes two units of 1,000 of spearmen or αἰχμοφόροι, and then 10,000 other infantry, which at least indicates that the Persian army was organized in decimal fashion.Footnote ¹³ These 10,000 seem to be the Immortals or ἀθάνατοι, described in more detail at Hdt. 7.83.1–2, with the different pomegranates (gold and silver) presumably denoting some sort of internal hierarchy. The Apple Bearers or μηλοφόροι described in later accounts are clearly indicated by their equipment (that is, golden apples), but we are left with a mystery 1,000-strong unit marching with spears turned downwards. Following the earlier speculation of Sir Frederick Maurice and Nick Sekunda, I recently tried to cut the Gordian knot, so to speak, by suggesting that the infantry units marching before and after the king's chariot are really both Apple Bearers, with Herodotus' accidentally doubling up the number (so 500 + 500, and not 1,000 + 1,000).Footnote ¹⁴ With no evidence to the contrary, it seems best to go along with my previous surmise. In sum, our interpretation of Herodotus presents us with a guard unit of 1,000 infantry (μηλοφόροι), plus a ‘palatine’ infantry grouping of 10,000 (ἀθάνατοι).

The Persian cavalry represent almost a mirror image of the infantry. Herodotus describes two lots of 1,000 picked horsemen, although the use of ἐκ Περσέων πάντων ἀπολελεγμένοι (7.40.2) and ἐκ Περσέων ἀπολελεγμένη (7.41.1) respectively does not allow us to determine which of these units is meant to be the senior.Footnote ¹⁵ There are also 10,000 other Persian cavalrymen, with nothing to suggest that they were particularly special, other than that they were ethnically Persian (or perhaps also Median).Footnote ¹⁶ These horsemen, as described here, could represent the cavalry equivalent of the Immortals, but that view is not entirely secure. Indeed, Ernst Obst has described them simply as cavalry ‘der Linie’.Footnote ¹⁷ Now, the notion of Herodotus muddling, or rather doubling, the number of the μηλοφόροι suggests that he could have done likewise with the picked cavalry; that is, instead of two units of 1,000 picked cavalry, we really have a 1,000-strong unit divided into two 500-strong groups.Footnote ¹⁸ Hence, 500 of these picked cavalry form the vanguard of the royal party (and indeed the whole ethnically Persian or Median contingent of Xerxes' army), while the other 500 form a rearguard. The 10,000 infantry and 10,000 cavalry are therefore not part of the group most intimately associated with the monarch; rather, they are meant to constitute the truly Persian army that travelled with the Great King when on campaign, or at least on this particular campaign. In short, we have a guard unit of 1,000 picked cavalry, together with a less-exalted cavalry force of 10,000.Footnote ¹⁹

The next occasion on which truly Persian troops are mentioned occurs at Hdt. 7.55.2–3, where the Persian army cross the Hellespont. This time, we look at the passage verbatim, for it is much shorter:

I have previously observed, with respect to the infantry, that ‘[t]he details provided here are remarkably consistent with what is found at Hdt. 7.40–1, the main difference being the position of the units in the marching column’.Footnote ²⁰ The same observation could hold more or less true for the cavalry, but we clearly need to reconcile one lot of unnumbered ἱππόται and a group of a 1,000 horse (οἱ ἱππόται οἱ χίλιοι) with what Herodotus provides at 7.40–1. Let us see what can be done. Four distinct interpretations present themselves: (a) οἱ μύριοι Πέρσαι is shorthand for both the 10,000 infantry and the 10,000 cavalry, which mean that the two groups of ἱππόται at 7.55.2–3 are both 1,000 strong; (b) the first group of ἱππόται includes the 10,000 cavalry and one of the two units of 1,000 cavalry; (c) the 10,000 horse are missing, and the two cavalry groups comprise 1,000 men each, with the first lot going unnumbered; and (d) the first group of unnumbered ἱππόται is indeed the 10,000 cavalry, which means that one of the 1,000-strong units of ἱππόται found at 7.40–1 is missing.

The first option, originally proposed by W.W. How and Joseph Wells, seems rather arbitrary, and is not altogether convincing as a result.Footnote ²¹ It might be best to hold that οἱ μύριοι Πέρσαι refers only to the Immortals. The second solution is also difficult to defend. With regard to the third, suggested by De Sélincourt, one wonders why Herodotus would have omitted a whole myriad of cavalrymen.Footnote ²² The fourth option remains. Given that only one lot of 1,000 cavalry is clearly numbered as such, it is possible that we could well have been correct in assuming that Herodotus had made a mistake at 7.40–1; that is, he had inaccurately described two 1,000-strong picked cavalry groups, instead of only one 1,000-strong picked unit, which had been divided into two 500-strong squadrons. Such an inference provides further support for the supposition that there was only one 1,000-strong unit of picked infantry, even though we still find two units of αἰχμοφόροι at Hdt. 7.55.2–3. Yet no numbers are given here, so perhaps the unit, once again, had been split into two. To conclude, we have 12,000 cavalry in the first Herodotean passage and (arguably) only 11,000 in the second. A further avenue to describe the discrepancies between the passages is presented by Hdt. 7.55.3, where we discover that Herodotus had consulted at least two accounts for his description of the crossing. It is therefore possible that he had also consulted multiple accounts when he wrote 7.40–1, and had accidentally doubled up the 1,000-strong cavalry and infantry units as a result of misunderstanding his sources.

The next occasion that Herodotus refers to elite Achaemenid cavalry is in the aftermath of the Persian failure at Salamis (480 b.c.). Xerxes, about to leave Greece, handed over command to Mardonius, who then selected a force from the remaining host. Again, we look at the passage (Hdt. 8.113.2) verbatim:

The 10,000 Persian horse referred to in the first passage discussed above (Hdt. 7.40–1), and probably in the second (Hdt. 7.55.2–3), do not seem to have been required, while only one of the two 1,000-strong units of picked horse found at Hdt. 7.40.1–41.2 is mentioned explicitly.Footnote ²⁴ The θωρηκοφόροι, literally those ‘bearing cuirasses’, has sometimes been thought of as cavalry, and so could possibly represent one of the two 1,000-strong units referred to at Hdt. 7.40.1–41.2 – or indeed the 10,000 cavalry with some stretch of the imagination. But I prefer to follow my previous, if tentative, suggestion that this is meant to be the type of soldier that Herodotus refers to elsewhere as αἰχμοφόροι; that is, these are possibly the Apple Bearers or μηλοφόροι.Footnote ²⁵ With respect to the Persian units that Mardonius chooses, we find the Immortals, together with what might well be termed the infantry and cavalry guard units. If the θωρηκοφόροι were not guard infantry (and the term is used elsewhere to denote troops simply wearing cuirasses), we are left to ponder the status of those picked soldiers who fought so bravely with Mardonius at Plataea in 479 b.c. (Hdt. 9.63.1–2). Although some scholars have maintained a contrary view, these men were more likely to have been infantry (or at least dismounted cavalry) given the context of a hand-to-hand struggle with Lacedaemonian hoplites – even if Mardonius himself was mounted.Footnote ²⁶ In short, it would seem odd that these 1,000 picked men, described as the flower of the Persian infantry (περὶ ἑωυτὸν λογάδας Περσέων τοὺς ἀρίστους χιλίους), were not meant to be referred to at Hdt. 8.113.2.Footnote ²⁷ Thus we might infer from Hdt. 8.113.2, together with the context of the information presented to us about the later battle of Plataea, that there was indeed only one group of 1,000 elite horsemen present on the campaign. This means that the passage can effectively be reconciled with what we might well term the ‘corrected’ view of Hdt. 7.40–1 established previously.

To conclude our discussion of Herodotus, it seems most likely that the truly Persian cavalry elite of Xerxes' day consisted of only one unit of 1,000-strong guard cavalry. If we follow Herodotus, there was also a 10,000-strong force of Persian cavalry. Yet the existence of these 10,000 cavalrymen, in the face of a lack of corroborating evidence from the period, aside from a possible aside in Xenophon's Cyropaedia (8.3.16), is not entirely secure. It is possible that the numbers presented for this group are part of a general exaggeration of Persian military forces on the part of Herodotus or his source/s, as is convincingly illustrated by the logistical analysis of Cuyler Young.Footnote ²⁸ While the Immortals, which we presume to be infantry, are indeed referred to outside Herodotus, for example by Curtius Rufus (3.3.13), Heracleides of Cumae (FGrHist 689 F 1 = Ath. 12.514b–c) and Pausanias (6.5.7),Footnote ²⁹ the myriad of cavalry, as a special grouping with a royal or at least ‘palatine’ association, is not mentioned by anyone else, although Hesychius, writing in the fifth century a.d., does refer to ἀθάνατοι (= Α 1531) as cavalry.Footnote ³⁰ But this is presumably in error, unless he is referring, here, to the Sassanian cavalry called Immortals by late-antique writers such as Procopius (1.14.31, 1.14.44).Footnote ³¹ In light of these problems, Cawkwell contends that the 10,000 ‘may have been Persian cavalry assembled from the satrapies’, and it is just possible that he might be right.Footnote ³² If the 10,000 cavalrymen were not, then, a special grouping, contrary to what Herodotus implies, one wonders whence its members might have come, and what their relationship to other cavalry units might have been. We will return to this vexing matter in Section 4 of this study.

2. BETWEEN XERXES' REIGN AND ALEXANDER'S CONQUEST

Between Xerxes' attempted invasion of Greece and the campaigns waged by Alexander the Great against Darius III, there is scant information available about the Persian military formations of the period. Internecine strife was wont to break out on occasion, with an array of rebellions, such as those led by Megabyzus, Artyphios (Megabyzus' son) and Pissouthnes, taking place. The most significant battle involving Persian troops arguably took place at Cunaxa in 401 b.c., between the forces of the royal pretender Cyrus the Younger, satrap of Lydia and Great Phrygia with Cappadocia (Xen. An. 1.9.1), and those of the ultimately victorious Artaxerxes II, previously installed officially as the Great King. For this battle, we have the eyewitness account of the Athenian mercenary leader Xenophon, who documents the engagement in his Anabasis. This is not to say, of course, that Xenophon, in his service with Cyrus, saw everything that he describes, but we can at least be sure of the general immediacy of his testimony. We can supplement Xenophon's account with fragments of Ctesias preserved by Plutarch in his Artaxerxes, though these provide little information directly of use.

Let us concentrate on the cavalry of Artaxerxes, for Cyrus' forces clearly do not represent the forces of a fully invested Great King. At An. 1.7.11, we read that there were 6,000 horsemen under the command of the noble Artagerses. These were stationed in front of the king, with this information coming from those who had deserted Artaxerxes.Footnote ³³ No mention is made of the ethnicity of these troops. This force is mentioned again at An. 1.8.24, in the context of (somewhat implausibly) being put to flight by the rampaging Cyrus and his 600 picked cavalry – a detail which hardly suggests that these 6,000 under Artagerses were what one could really describe as elite. There is, of course, the possibility that these 6,000 represented a mobilized portion of the 10,000 cavalry of the Herodotean model, but supporting this view is a fraught exercise given the paucity of evidence for such a formation existing outside the pages of Herodotus, at least in the very specific sense of a recognized grouping of 10,000. So, one might simply prefer to view these 6,000 men as cavalry of indeterminate origin in the Great King's service. That said, they were likely to have come, in this instance, from the Persian heartland, especially given the hasty circumstances under which Artaxerxes' army was assembled. It follows that there might indeed be a connection of sorts to the Herodotean 10,000, if this ever existed – but only with respect to origin. Barkworth even suggests that they were the only cavalry available.Footnote ³⁴ Whatever the case, brave and loyal they do not seem to have been, unlike just about any other description of elite Persian troops who, even if bested by their Greek or Macedonian counterparts, do not shirk from their duty in combat. The Great King was protected by what Xenophon describes at An. 1.8.26 as being a ‘compact body’ (witness στῖϕος). But even these were penetrated, for Cyrus was able to strike Artaxerxes through his cuirass, as is corroborated by Plutarch (Art. 11.2), who followed or indeed paraphrased Ctesias' now lost account. Horsemen are mentioned elsewhere throughout the narrative (for example, at An. 1.8.9, 1.8.18, 1.10.6–7, 1.10.12–13),Footnote ³⁵ but they do not seem to have been of elite status.Footnote ³⁶ In the end, Xenophon provides us with very little of importance. It is fortunate that the battles of Alexander and Darius provide us with further information relating to elite Persian cavalry.

3. THE ELITE CAVALRY OF DARIUS III

Persian cavalry fought in three engagements against Alexander the Great. Our sources, in the main, are (a) Arrian, who based his account on Ptolemy and Aristobulus, who were both present on the expedition and therefore saw many of the events described, and (b) Diodorus Siculus and Quintus Curtius, who, for the most part, followed the sometimes unreliable Cleitarchus, who seems to have been somewhat biased against Alexander. This Cleitarchean tradition underpins what has been commonly referred to, albeit somewhat too simplistically, as the Alexander ‘Vulgate’.Footnote ³⁷ In general, the Ptolemy/Aristobulus tradition presented by Arrian has been regarded as preferable where there are discrepancies between these two broad traditions – as often there are. Yet the Cleitarchean tradition cannot be ignored, for it records incidents not mentioned by Arrian, and describes actions not witnessed directly by his sources, as will be seen below in the case of the Granicus.Footnote ³⁸

At the Granicus (334 b.c.), where the Persian forces operating in north-west Anatolia were commanded by a group of nobles (Arr. An. 1.12.8), with Darius being absent, more than 10,000 cavalry were supposedly present (Diod. Sic. 17.19.4). It is necessary to determine which of these might be regarded as in some way elite, for it is surely too facile to assert that these are the special grouping of 10,000 Persian cavalry referred to by Herodotus, the more so when Arrian (An. 1.14.4) more or less doubles their number. Despite our general misgivings about the Cleitarchean source tradition that he follows, Diodorus presents more detail of interest to us than does either Arrian or Curtius, even if his account of the battle tends to dwell on the heroic and the downright sensational. At 17.20.2, he records that the satrap of Ionia, one Spithrobates, a son-in-law of the Great King, hurled himself towards the Macedonian lines ‘with a large force of cavalry’ (μετὰ μεγάλης δυνάμεως ἱππέων), and had with him forty fellow combatants described as ‘royal relatives’ (συγγενεῖς).Footnote ³⁹ When Spithrobates and his brother Rhosaces fell, these men, again referred to as συγγενεῖς, surrounded the stricken nobles and flung their javelins at Alexander (Diod. Sic. 17.21.1).Footnote ⁴⁰ These συγγενεῖς were eventually either killed, or put to flight (Diod. Sic. 17.21.4).

Some reflection on Diodorus' use of συγγενεῖς is warranted here. We are left with two options: (a) that συγγενεῖς is being used to describe members of an elite cavalry unit, as is suggested by Diodorus' later use of the word at 17.59.2: συγγενεῖς ἱππεῖς; or (b) that the word is being used to describe members of a particular class within Persian society. The latter could mean that the 40 men described as συγγενεῖς were not necessarily a detachment of what one might suppose was a larger elite unit, perhaps 1,000-strong, as per the Herodotean model. Yet the way in which Diodorus describes their involvement suggests that they acted as a kind of ‘guard’ to the noble Persian commanders. What their parent unit might have been, and how it relates to the Herodotean accounts of Persian elite cavalry, will become clearer when we tackle the later battle of Gaugamela. In essence, Diodorus seems to be telling us that elements of an elite grouping operated at the Granicus, even though the Great King was not present. That one of the nobles commanding the Persian forces was a relative of the king by marriage adds to the picture of these cavalrymen acting as a kind of ‘royal guard’, regardless of the Great King's presence or absence. These men seem to have come from the Persian heartland, at least if the royal association holds, instead of being satrapal troops. It is therefore possible that other elements of the large cavalry contingent at the Granicus might also have been of Iranian origin and had accompanied the συγγενεῖς in their march to Anatolia. Of this, however, we cannot be certain.

For Issus (333 b.c.), we unfortunately have only very rough accounts of the Persian forces. Since Darius, on this occasion, was present, one might presume that elements – if not all – of the elite cavalry were available. Yet the descriptions provided by the historians of Alexander's campaign, such as Arrian (An. 2.8.5–8) and Diodorus (17.31.2), generally fail to elucidate which units were ranged against the Macedonians and their allies. Only Curtius provides details of the guard cavalry, for he writes, at 3.9.4, that Darius was followed by 3,000 elite horsemen (tria milia delectorum equitum), described as ‘his customary bodyguard’ (assueta corporis custodia).Footnote ⁴¹ A strength of 3,000, though problematic, has been accepted by a number of scholars.Footnote ⁴² Note, too, that Diodorus (17.31.1) tells us that Darius, immediately before the battle, summoned his ‘friends’ (φίλοι) and ‘relatives’ (συγγενεῖς), with some of these being given commands, and others being ordered to fight by his side.Footnote ⁴³ This could provide a neat tie-in to the material that he presents on the Granicus, but it is not certain that Diodorus, here, is using the word συγγενεῖς in the same way as he does at 17.20.2 – that is, to refer to some kind of ‘guard’ cavalry.Footnote ⁴⁴

Curtius' relatively lengthy description of Darius' army in the build-up to Issus, which borders on the fanciful given the continued appearance of the Immortals, needs to be adduced at this point, for it has a bearing on his claim that there were 3,000 cavalry guardsmen at Issus.Footnote ⁴⁵ As I have argued elsewhere, the descriptions appear to be based ‘to some degree on Herodotus, or else sources themselves based on Herodotus, but the description of the king's appearance at 3.3.17–19 is clearly not from Hdt. 7.40–1’.Footnote ⁴⁶ I added that, ‘together with the comparable parade in the Cyropaedia (8.3.1–34, and especially 8.3.9–18), it is perhaps advisable to view all three instances as essentially independent pieces of evidence to demonstrate that military processions were an important feature of Achaemenid behaviour’.Footnote ⁴⁷ As for material relevant to us, one looks in vain to find anything of real worth, for no troops in the whole procession (3.3.9–25) are specifically described as horsemen, much less a royal cavalry bodyguard, aside from ‘the horsemen of twelve nations with varying arms and customs’ (equitatus duodecim gentium variis armis et moribus, 3.3.13) – surely not a reference to elite cavalry. 15,000 cognati regis (‘kinsmen of the king’) are found at 3.3.14, but we get no inkling that there were cavalry in a strictly military sense.Footnote ⁴⁸ Likewise, there were 10,000 hastati following the king's chariot (3.3.20), but the reference to hastae implies infantry.Footnote ⁴⁹ The only other notable information is that 200 of the king's propinqui or relatives were present (3.3.21), but it is uncertain whether these men can be equated with the συγγενεῖς, or whether they were meant to be drawn from the aforementioned cognati regis.Footnote ⁵⁰ It might not be too pessimistic to dismiss the entire passage as evidence.

So, whence did Curtius (3.9.4) get the idea that the guard cavalry constituted 3,000 horsemen, at least for Issus? It is impossible to draw any definite conclusions, but it is prima facie odd that such a unit does not appear in the parade described above. Curtius is clearly being inconsistent, and it is quite likely that he is mixing and matching from different sources. But the reference to 3,000 guard cavalry does not seem to come directly from the same principal source used by Diodorus, usually thought of as Cleitarchus, for Diodorus clearly states that the guard cavalry was 1,000-strong, even though this, admittedly, was at the later battle of Gaugamela. Either Curtius has misinterpreted his source, or else he is using one of even lesser repute than that used by Diodorus for the strength of the συγγενεῖς. A 1,000-strong guard cavalry unit accords with other evidence, not least of which is (a) Herodotus' reference to an elite 1,000-strong cavalry unit at multiple loci, particularly Hdt. 7.55.3 and 8.113.2, where only a single unit of 1,000 horse is mentioned; and (b) the likelihood that the guard infantry unit, the μηλοφόροι, was also 1,000 strong. Moreover, Arrian (An. 3.21.1, with 3.16.1) also seems to assert that the leader of the συγγενεῖς was a chiliarch, so one would envisage a 1,000-strong force, just as the infantry chiliarch commanded the 1,000 μηλοφόροι.Footnote ⁵¹ A chiliarch, after all, commands 1,000 men. If Darius did indeed have 3,000 guard cavalry at Issus – bearing in mind that no other source tells us their strength for the battle – this was clearly not the king's assueta corporis custodia; rather, it was an unusual formation, with the 1,000 συγγενεῖς presumably being part of this makeshift force.

For Gaugamela (331 b.c.), we are fortunate to have the reasonably detailed account of Arrian, who records the Persian order of battle provided by Aristobulus.Footnote ⁵² Arrian (3.11.5) writes that the μηλοφόροι, the presumably 1,000-strong guard infantry, were posted together with what are described as ‘the king's kinsmen’ (οἵ ... συγγενεῖς οἱ βασιλέως) in the centre. These troops appear to be cavalry, with their close association with the king surely denoting an exalted status. Curtius (4.14.8) provides similar information, for we read that Darius was with ‘the elite of his infantry and cavalry’ (delectis equitum peditumque), though the words used provide us with little detail about the type or composition of this cavalry and infantry guard, or indeed their number – he does not mention a 3,000-strong cavalry force with the king on this occasion, as he did for Issus.Footnote ⁵³ Diodorus (17.59.2) provides more useful detail. He writes that the cavalry were the first to engage the enemy and that Darius, on the left of the field, Footnote ⁵⁴ led against the Macedonians his ‘kinsmen cavalry (συγγενεῖς ἱππεῖς), chosen for their bravery and loyalty, the 1,000 being drawn together into one formation (χιλίους ἐν μιᾷ περιειλημμένους εἴλῃ)’.Footnote ⁵⁵ From this, it seems reasonably obvious that the συγγενεῖς were indeed cavalry, and that the whole unit was 1,000-strong. Yet it is implied by Diodorus, with his reference to the whole unit being marshalled together, that they sometimes operated in smaller groups of less than 1,000 men – as we possibly saw earlier at the Granicus (Diod. Sic. 17.20.2).Footnote ⁵⁶ In the face of this evidence, Briant's contention that ‘[à] l’époque de Darius III, ils [sc. the συγγενεῖς] forment un corps de cavaliers d'élite, au nombre de 10 000' appears to be difficult to support, the more so since he cites the Diodoran locus in question.Footnote ⁵⁷

By way of contrast, Sekunda, referring to skirmishes and screening operations on the eve of Gaugamela, holds that ‘it seems that there were at least three regiments of élite cavalry, of which at least one was composed of Kinsmen [that is, the συγγενεῖς]’, while he adds: ‘It may be that all three squadrons were composed of Kinsmen, out of which one was designated the Royal Regiment.’Footnote ⁵⁸ First, Sekunda adduces the somewhat inaccurately titled ‘master of the horse’ Satropates and his ‘1,000 élite cavalry’ (citing Curt. 4.9.7).Footnote ⁵⁹ These horsemen were sent to watch the Euphrates crossing, although Satropates was killed in the process of preventing such a passage (Curt. 4.9.25). Second, just before the battle proper, the senior Persian officer Mazaeus is mentioned as ‘commanding 3,000 élite’ cavalry' (citing Curt. 4.12.1, 4.12.18).Footnote ⁶⁰ Third, Sekunda adduces the ‘squadron of 1,000 Kinsmen, chosen for their courage and loyalty’, as per Diod. Sic. 17.59.2. Now, from this information, there is no reason to suppose that any group other than those described by Diodorus are meant to represent kinsmen cavalry or συγγενεῖς. Sekunda keeps referring to ‘elite’ horsemen, presumably in the modern sense of a unit with an exalted or ‘guard’ standing (Curtius, in fact, uses the phrase cum mille delectis at 4.9.7).Footnote ⁶¹ The locus at 4.12.1 mentions nothing about an elite status. Indeed, it refers to a screening activity, or securing roads that Alexander's forces were likely to use, while Curt. 4.12.18 only mentions Mazaeus having taken a position on a hill ‘with his picked cavalry’ (cum delectis equitum). Like Sekunda, Tarn also sees these men as ‘guard cavalry’.Footnote ⁶² But Mazaeus had 6,000 cavalry under his overall command, as per Curt. 4.9.7.Footnote ⁶³ This could mean that those introduced at Curt. 4.12.18 could simply be a more manageable group deemed of superior quality, or even a detachment of appropriate size for the task at hand. While the formations attested by Curtius could have been ‘elite’ in some sense, they do not seem to be a ‘guard’ unit. We do not even know whether they were formally recognized as a specific unit – the language does not allow us to draw any firm conclusions.

Finally, it is worthwhile to reflect on the fate of the elite Persian cavalry after Alexander's victory. Arrian (An. 7.6.3) writes that Alexander, approaching the end of his reign, either incorporated, or else was proposing to incorporate, as Sekunda suggests, Bactrian, Sogdian, Arachotian, Zarangian, Areian and Parthyaean horsemen into his Companion cavalry, together with ‘the horsemen called the Euacae’ (οἱ Εὐάκαι καλούμενοι ἱππεῖς).Footnote ⁶⁴ According to Sekunda, ‘[i]t seems reasonable to identify this regiment with the “Kinsmen”’, the more so since a precise identification of this group has proved troublesome.Footnote ⁶⁵ Now, we do not know what the Old Persian is for ‘kinsmen’, but Sekunda, reflecting on the Sanskrit word svaka, proposes that it could be *huvaka, or just possibly *xvaka, if we follow Hinz.Footnote ⁶⁶ Whatever the real form, and Sekunda eventually plumps for *huvaka on account of comparative linguistic grounds that we need not explore here, it could be transliterated into Greek as Εὐάκαι. This would explain Arrian's use of the word. Of course, we have no other evidence to support this conjecture. But it is not improbable that some of the συγγενεῖς, the Great King's premier cavalry unit, or what remained of it, were drafted into an elite formation in the Macedonian army, especially since the exalted μηλοφόροι – their infantry equivalent – were not, it seems, disbanded. Rather, they were incorporated into a joint Persian and Macedonian guard infantry formation, at least if Phylarchus' testimony (FGrHist 81 F 41 = Ath. 12.539e) is given credence.Footnote ⁶⁷

4. BROADER IMPLICATIONS

From the literary material presented above, there appears to be only one group of 1,000 cavalry that can be consistently identified throughout the Achaemenid period as being closely associated with the Great King, even if this identification can only be connected to two isolated periods in Achaemenid history, these being (a) Xerxes' preparation for his invasion of Greece in 480 b.c., and (b) during Darius III's attempted repulse of Alexander the Great's expeditionary army in 333–331 b.c. Who these men were, and who commanded them, therefore warrants some consideration, together with their relationship to other cavalry units in Persian service.

Let us first consider the unit's composition. The very name of the συγγενεῖς, literally ‘the [king's] kinsmen’, would suggest that there were drawn from the upper echelon of Persian society. But the nomenclature used in the Greek sources has occasioned some debate. Hammond describes the συγγενεῖς as ‘nobles’, perhaps reflecting Xenophon's use of the word at Cyr. 8.3.13 to refer to a social class – and not a cavalry unit.Footnote ⁶⁸ Strack once suggested, not unreasonably, that Greco-Roman writers confused this social class, as described by Xenophon, with the Persian guard cavalry.Footnote ⁶⁹ He contends that loci such as Diod. Sic. 17.59.2, 17.20.2 and Arr. An. 3.11.5, with their use of συγγενεῖς, do indeed refer to an elite cavalry unit, while the word's appearance at Diod. Sic. 17.35.2–3 and Arr. An. 7.11.1 refers to those ‘in der Nähe des Königs’ – which does not necessarily mean members of a military unit. Perhaps Sekunda has the best means to deal with the problem.Footnote ⁷⁰ He suggests that the συγγενεῖς ἱππεῖς were drawn from the aforementioned social group, which was allegedly 15,000-strong, as per Ath. 4.146c and Curt. 3.3.14, and was known by Greek writers as the συγγενεῖς.Footnote ⁷¹ The unit was therefore comprised of the sons of the Persian social elite, with service in the συγγενεῖς perhaps constituting a kind of military apprenticeship before its members moved on to other military or civil positions – as might service in the μηλοφόροι, their infantry equivalent.Footnote ⁷² This view, of course, must remain conjecture, but it does at least seem to fit the available evidence.

Who commanded the συγγενεῖς? Alexander Meeus has recently contended that the command of the συγγενεῖς was one of two important chiliarchies. These were (a) command of the μηλοφόροι (that is, the office of *hazarapatiš), this being the official credited by Greek writers with controlling access to the Great King, and (b) command of the συγγενεῖς ἱππεῖς.Footnote ⁷³ He adds that, at Gaugamela, Nabarzanes was the ‘equestrian chiliarch’, a view supported by Welles and Lewis.Footnote ⁷⁴ But Collins has recently refuted this argument by adducing evidence suggesting that the commander of the μηλοφόροι was indeed the chiliarch par excellence, and that the commander of the συγγενεῖς was merely one of many chiliarchs, or ‘commanders of a thousand’.Footnote ⁷⁵ One might still wonder, however, why it is the infantry guard commander rather than his cavalry equivalent who is normally assumed to enjoy the pre-eminent role, the more so given that it is often thought that cavalry rather than infantry was the pre-eminent Achaemenid military arm.Footnote ⁷⁶

The answer to this conundrum might very well lie in Tuplin's recent treatment of the Achaemenid cavalry, and particularly his discovery, gleaned from an exhaustive exposition of various sources from both within and outside the Persian empire, that cavalry are not given any particular prominence in the kinds of sources that might matter most, such as iconography.Footnote ⁷⁷ Moreover, Tuplin asserts that, despite an increase in the use of cavalry – both Iranian and allied – as the empire evolved, the status of cavalry vis-à-vis infantry is unlikely to have changed during the period of importance to us; that is, from Xerxes through to Darius III. As Collins admits, the συγγενεῖς ‘were no doubt commanded by a chiliarch of some sort’, being a unit composed of a 1,000 men, but there is no certainty that this commander was necessarily equivalent in importance to the exalted commander of the μηλοφόροι.Footnote ⁷⁸ This matter surely warrants a more detailed treatment elsewhere. Suffice it to say here that, on the face of it, there is no reason to assume that the συγγενεῖς and their commander enjoyed greater or equivalent seniority to the μηλοφόροι and their respective commander. In fact, if one extrapolates from the conclusions advanced by Tuplin regarding Achaemenid cavalry in an admittedly general sense, there is some cause to believe that the συγγενεῖς could even have been the inferior of the two guard units, the second among equals.

This leads us to the issue of the 10,000 Persian cavalry found in Herodotus, and how these relate to the συγγενεῖς, or indeed other cavalry in Persian service, such as those in satrapal service. Determining who these men were is a difficult exercise, and largely depends on whether we accept them as cavalry troopers from the Persian heartland, or from elsewhere, including satrapies outside this area, as per Cawkwell's view.Footnote ⁷⁹ If one considers all of Herodotus' lengthy exposition of the forces of Xerxes immediately before his Greek campaign, it should become readily apparent that Herodotus makes a clear division between those forces whom he understands as being truly Persian, or at least Iranian, and those who emanated from other parts of the Great King's empire. Hence, we are meant to understand the troops described at Hdt. 7.40–1 as Persian cavalry, or at least Iranian, rather than an amalgam of peoples from various parts of the empire. If so, the Herodotean 10,000 seem to be regarded, on the surface, as being analogous to the 10,000 infantry, whom we generally refer to as the Immortals or ἀθάνατοι. But do two groupings of 10,000 standing infantry and 10,000 standing cavalry really make sense? By striving for symmetry, Herodotus could possibly be privileging the trope that Persian cavalry were of especial importance compared to the infantry, a view which also presumably influenced Xenophon's reference to ‘the first-formed 10,000 cavalry’ (ἰππεῖς οἱ πρῶτοι γενόμενοι μύριοι) in the historical romance of the Cyropaedia (8.3.16) – a numbering which could possibly stem, in any case, from the Herodotean tradition. This is especially so given that Xenophon, as we saw in the rather more factually based Anabasis, did not seem to recognize an extant formation of 10,000 ‘palatine’ cavalry in the Great King's service on the eve of the fourth century b.c. Yet maintaining the standing cavalry of an empire at precisely the same strength as its standing infantry is an unusual circumstance at any point in history. This is particularly so given the greater logistical needs associated with maintaining cavalry, especially in the field, where forage might not always be available owing to climactic conditions or enemy harassment.Footnote ⁸⁰

In his survey of the evidence pertaining to Achaemenid cavalry, Tuplin points to a general view among our sources that the Great King's cavalry were either from the Persian heartland (Elam, Persia and Media), or ‘points east’, the other possibilities being Armenia and Cappadocia, at least by the era of Darius III.Footnote ⁸¹ This brings us to cavalry sourced from the satrapies. If the amount of truly Iranian cavalry was not as large as some have previously imagined, it is likely that a substantial portion of those cavalrymen serving in Persian armies not directly under the Great King's command hailed from the satrapies, particularly the eastern ones. The degree to which any of these men could be considered elite is difficult to ascertain; that said, it is clear that some cavalry formations emanating from outside the Persian homeland were regarded as superior to others. For example, at Gaugamela, Mazaeus, who commanded the Persian right, including ‘the best of the cavalry’ (witness μετὰ τῶν ἀρίστων ἱππέων), is said to have entrusted 2,000 Cadusii and ‘a 1,000 picked Scythian horsemen’ (χιλίους Σκυθῶν ἱππεῖς ἐπιλέκτους) to undertake the critical manoeuvre of sacking Alexander's heavily guarded baggage (Diod. Sic. 17.59.5).Footnote ⁸² Moreover, it would not be entirely unwise to posit that each satrap had a small permanent cavalry force that acted as a kind of ‘satrapal guard’.Footnote ⁸³ Perhaps the 600 trusted cavalry under Cyrus the Younger's immediate command at Cunaxa was an example of such a unit (Xen. An. 1.8.24).Footnote ⁸⁴ If such a force did indeed exist, it would clearly have formed a very important part of satrapal forces, either singularly, or combined, as was the case at the Granicus, and presumably would have been counted among the Great King's more useful units in the great battles of the era.

So, it appears that there might not have been a particularly large number of truly Persian or Iranian cavalry available to the Great King, particularly given that, as Evans points out, it cannot be assumed that each cavalryman only required one horse.Footnote ⁸⁵ Herodotus' cavalry myriad, instead of being a clearly defined formation composed exclusively of ethnic Persians, as he asserts, might therefore have been a force cobbled together from a number of areas with cognate fighting styles for the purpose of the campaign in question. If this was the case, the view that the Herodotean 10,000 cavalry must correspond exactly to the ‘palatine’ nature of the 10,000 infantry, a force whose strength seems to be less disputable, could well be mistaken. All this, of course, must remain speculation, but does at least help to explain why the cavalry myriad is never referred to again in our sources (aside from the rather dubious Cyropaedia), at least as a defined unit or formation. Moreover, it explains why the formation never receives special descriptive attention from Herodotus, unlike the ἀθάνατοι, who are named at several points in his narrative.Footnote ⁸⁶ Indeed, it is an infantry formation that was always kept to a precise strength of 10,000 men – not a cavalry unit (Hdt. 7.83.1). A possible implication is that the Herodotean cavalry myriad was a rather more ad hoc formation, and one assembled for a specific purpose, this being Xerxes' Greek expedition. Yet Xenophon (Oec. 4.5), who we might presume was not entirely ignorant of Persian institutional arrangements on account of his friendship with Cyrus the Younger, suggests that there was at least some permanent cavalry, these being paid for by those nations accustomed to give tribute to the Great King, though no numbers are advanced.Footnote ⁸⁷ Regardless, we seem to be left with a ‘palatine’ ethnic Persian force of cavalry and infantry that is rather smaller than what has previously been envisaged, with the συγγενεῖς being the only clearly defined standing cavalry unit of royal significance. This unit continued intact through to the era of Darius III, in a similar way to the infantry μηλοφόροι.

5. CONCLUDING REMARKS

In sum, our discussion of the Persian elite cavalry provides additional evidence to support the claim that the truly Persian army of Xerxes was not a particularly large force.Footnote ⁸⁸ If we accept Herodotus' figures, the 10,000 cavalry and 1,000 guard cavalry or συγγενεῖς supposedly formed a kind of combined Persian (or at least Iranian) force that accompanied the Great King in the field, just as the 10,000 ἀθάνατοι and the 1,000 μηλοφόροι constituted his central infantry force. If one follows my earlier lead with the elite infantry, it is possible to argue that Herodotus' 10,000 Persian cavalry, regardless of whether they were a standing force or otherwise, were not really elite or ‘select’ in the modern sense of the word, much less of ‘guard’ status. Rather, they constituted the combined Persian or Iranian cavalry for Xerxes' Greek campaign. The numbers alone surely require this conclusion, for the existence of a ‘guard’ cavalry unit of 10,000 is unheard of in any period of antiquity.Footnote ⁸⁹ To compare with the Roman empire, the equites singulares Augusti, the elite cavalry of the Caesars, numbered somewhere in the vicinity of 1,000 in the late first century a.d., rising to 2,000 in the early third century a.d., while their Eastern Roman empire equivalent in the fourth to sixth century a.d., the scholae, had a paper strength of 3,500, being comprised of seven 500-strong units – and this at time when cavalry was becoming the main force on the battlefield.Footnote ⁹⁰ If one accepts an ‘elite’, or even ‘palatine’, force of 10,000 cavalry, a considerably larger non-elite or satrapal force must be posited. This, however, is something which, on the face of it, looks rather implausible, the more so in view of recent work suggesting that the infantry constituted the core of Achaemenid armies throughout the period in question.Footnote ⁹¹ Herodotus' 10,000 Persian cavalry, if they did number 10,000 for the Greek expedition, were elite only in the sense that they were more closely associated with the Great King than any other cavalry forces brought along for the campaign, save the guard cavalry proper – that is, the συγγενεῖς. Despite Herodotus' statement (7.40.1–41.2) that there were two 1,000-strong units of picked cavalry, a detail which he himself seems to refute later in his narrative, there seems to have been only one unit of royal guard cavalry in our period. Despite some thoughts to the contrary, largely prompted by Curtius' largely inexplicable reference (at 3.9.4) to a 3,000-strong cavalry force accompanying the king, their nominal strength was 1,000. These men, evidently of high social standing, and presumably well trained, therefore constituted the premier cavalry unit of Achaemenid Persia.Footnote ⁹² They could presumably operate in small groups, especially when the Great King was absent, perhaps so as to provide a ‘royal’ presence and thereby instil confidence in lesser formations, but they could also be massed together, as reportedly occurred at Gaugamela.Footnote ⁹³