The relation of these two towns to each other is a problem. The Alexander historians make it clear that Alexandria was meant to be the capital of the Paropamisadae, and the references to it in the Milindapañha (p. 421 and n. 4) show that it was existing in the second century b.c., in the flourishing period of Greek rule, and probably in the first century also; and there is a Chinese mention of it round about 50 b.c. (p. 340). The literary evidence is then perfectly clear. But the evidence of the coins is equally clear that Kapisa was the Greek capital, for the coins of Pantaleon and Agathocles which show the Zeus of Kapisa holding Hecate τριοδῖτις on his hand prove beyond any doubt that (among other things) Kapisa was successively the seat of these two sub-kings (see on this p. 158). I need not enlarge here on the importance of Kapisa; this book should have sufficiently shown it, and Kapisa continued to be a capital for centuries after the name of Alexandria was forgotten. Now it is unthinkable that there should have been at the same time two Greek capitals of the Paropamisadae; and a solution of the difficulty must be attempted.

In the absence of excavation there can naturally be no certainty about the site of Alexandria; all the sites so far proposed—the most favoured has been the ruin-mounds at Opian near Charikar—are mere guess-work, and the French archaeological mission declined to locate it.

Ptolemy vii, 1, 6 calls the country between the Jhelum and the Ravi, with the cities of Bucephala, Iomousa, and Sāgala, ἡ Πανδοούων or Πανδαούων or χώρα. The readings vary; that of the best MS, X (Πανδόπων), is too corrupt to make anything of, and the others vary between Πανδόπων, Πανδόόπων (which Renou prints), and Πανδόόπων the word is certainly Pāndava, the form last mentioned being an exact transliteration. The Latin versions give Pandanorum, Pandonorum, Pandorum, and Pandiana regio.

The location of the name, which is part of Ptolemy's Hellenistic material (Chap, vi), shows that it can have nothing to do with the kingdom of the Pãndhya in the extreme south of the Indian peninsula, which moreover only really came within the purview of the West about the time of Augustus, though the two names have sometimes been confused, both in the Latin versions of Ptolemy and in modern writers.

Renou in his edition of Ptolemy's Indian books prints Εὐθυδημία. His note is: ‘Ἐὐθυδημία scripsimus: εὐθύδη X, εὐθημία Γ, εὐθυμηδία ω, -μέδια ἅ.’ His text is based on X (Vaticanus 191, thirteenth century) which all editors seem agreed is the best MS; Γ (Vatic. Palat. 388, fifteenth century) is said to exhibit some curious and erratic readings; ω is Renou's sign for the general body of MSS, and ἅ for the editio princeps of 1533 (Bâle) which Müller rated highly; it is based on Γ but has various divergences. The Latin versions all have -media. One MS of the group ω (Vat. Urbin. 82) is eleventh-century; -μηδία therefore is actually the oldest reading known, for what that may be worth.

It might be held that εὐθύδη of X warrants Εὐθυδημία, but Renou did not explicitly put it on that ground; there is little doubt from his note (scripsimus) that he adopted Bayer's old conjecture Εὐθυδημία because most writers since have done so. The historical considerations which make Εὐθυδημία impossible are dealt with in my text (pp. 247 sq.); here I am only considering Ptolemy's text.

The corruption of -μηδία into -δημία would be easy and obvious; Εὐθυδημία makes such good sense so long as one does not think about it, and the reasons which have affected the modern writer are precisely those which would have affected the ancient scribe.

Great changes had taken place in India since Alexander's day. He had found a number of disconnected states and peoples in the North-West, and had had no relations with, even if he had heard of, the most powerful of the Indian kingdoms, that of Magadha on the Ganges. Soon after his death the Maurya Chandragupta had seized the crown of Magadha, and, perhaps by 312, had extended his rule to embrace all India north of the line of the Vindhya mountains and the Nerbudda river. He was succeeded first by his son Bindusāra and then by his grandson Asoka, under whom the Mauryan empire was expanded to include a considerable part of peninsular India; but the southern conquests were only temporary and were apparently lost after Asoka died, and the empire was essentially a North Indian empire; the capital was Pātaliputra on the Ganges. The Seleucids and the Mauryas were always on friendly terms, and Greeks knew a good deal about the Mauryan empire as it had been under Chandragupta through the account of it given by Megasthenes, Seleucus' ambassador at his Court; probably they knew as much about it as they had known about the Persian empire in Xenophon's day, while Indians in turn knew a certain amount about the Greeks of the Seleucid East, whom they called Yavanas or Yonas (p. 417). It is however of some importance to the subsequent story to note that the Mauryan empire as most Greeks knew it was that of Chandragupta and not that of Asoka, that is, it was an empire of Northern India.

What kind of people now were these Greeks whom the Seleucid settlement scattered throughout Asia? For the third and much of the second centuries b.c. the answer is simple: just Greeks, with all that that implies. Most certainly they were not, as it was once the fashion to suppose, a people of Eurasians and Levantines. I have said elsewhere what I have to say about Greek ‘decadence’ in the Hellenistic period; there is small sign of it down to about the middle of the second century b.c., and I trust that, as regards the Farther East, readers of this book will come to the same conclusion. But the later period may need some special consideration, for in the latter part of the second century much of the Greek and Greek-speaking world lost its political independence. In 168 Macedonia fell to Rome. Between 163 and 141 Iran and Babylonia passed out of Seleucid hands into those of the Parthians. In 146 Greece itself became a Roman province, and in 133 Rome took over the Attalid kingdom of western Asia Minor. About 130 the Graeco-Bactrian kingdom fell to the Yuehchi, and by about 110 or even earlier the Saca invasion of Greek India had begun. We might therefore expect to find some alteration in the first century b.c., and to a certain extent we do. Leaving the Roman provinces aside, there was a change in Egypt in the direction of mixture of blood and of social ideas; but Egypt was not a land of Greek cities and offers no analogy with Asia.

That the sign þ which appears in Greek legends on the Kushan coinage and on that of two earlier rulers has the value sh was remarked in 1872 by Cunningham, who thought that the sign was a peculiar form of Rho. In 1887 Sir A. Stein put forward the theory that it was a revival of the obsolete Greek letter San (which had the value sh), since the oldest minuscule form of San resembles, though apparently it is not identical with, the þ sign on the Kushan coinage. This has raised a good deal of discussion; on the one hand, various other origins have been suggested for the þ sign, including the Aramaean Tsade, and on the other, much learning has been expended in an attempt to show that this sign had not the value sh at all but the value r (Rho), while Professor F. W. Thomas has taken the view, which resembles Cunningham's, that the sh sign þ was derived from the r sign Á (Rho), and has also said that after the fifth century San only survived as a numeral (and as a numeral it cannot come in question, for it was written with a totally different sign). But I have never seen the one enlightening Greek text on the subject quoted.

In order to save much repetition in the text, the principles followed in this book on these two matters are set out here.

It has been widely believed that the monograms on the Greek coins from Bactria and India, or most of them, denoted mint-cities; and even to-day, I understand, it is thought that some of them must be mints, though one numismatist has stated that they may sometimes ‘denote the name of the local magistrate under whose authority the coin was struck’. Yet Cunningham's laborious effort to work out the mint-cities from these numerous monograms was a complete failure, and it is admitted that, after many years of study, no single monogram of any mint has been identified, while on the other hand the types of at least two mint-cities, the ‘Zeus enthroned’ of Alexandria-Kapisa and the humped bull of Pushkalāvati, are perfectly certain.

Why it was ever supposed that the Greek kings in the East would make such a radical breach with Seleucid custom I cannot imagine; the continuity between the eastern Greek kingdoms and the Seleucid realm is as marked as that of other Seleucid Succession states, indeed in many ways more so; this book has, I trust, shown the trouble taken by both houses, that of Euthydemus and that of Eucratides, to prove that they were Seleucids. No one seems ever to have doubted that the Seleucid monograms represent moneyers, and the Seleucid system of monograms at the Antioch mint has been elucidated by Mr E. T. Newell; the monograms are those of continuing mint-masters and changing city magistrates.

The invasion of India had been the work of three men, whose far-reaching plan had come within a very little of complete success; that success was prevented, at the last moment, by two other men, who were also working on a far-reaching plan of their own. Dimly as we discern the outlines of their several schemes and actions, the magnitude of them gives us the feeling that the age of giants had come again and that we are back among the men who fought for the heritage of Alexander. For though Greeks could change their sky they could not change their souls. The gods had given them every gift save one, the gift of combination; and they tore each other to pieces beneath the shadow of the Hindu Kush with the same enthusiasm which Greek city-states and Macedonian generals had always put into the business round the Aegean home-sea. This chapter is concerned with the story of how and why Demetrius failed to secure the Mauryan empire, the story of the Seleucid Antiochus IV, called Epiphanes, and his cousin Eucratides.

Antiochus IV has often had hard measure from his historians. Some have repeated the Hellenistic gossip which made of him half a fool—vain, silly, theatrical; it is worth precisely what any Hellenistic gossip is worth, and the less that serious history has to do with it the better. To others, he is little but the king who persecuted the Jews; that story can be read in many books, and I need only say here that, whatever he did to the Jews, they have had an ample revenge.

The two main Chinese sources for the conquest of Greek Bactria are chapter 123 of the Shi-ki of Ssu-ma Ch'ien, who is said by Hirth to have finished his history about 99 b.c. (some put it rather later) and who reproduces Chang-k'ien's Report, more or less interwoven with his own narrative but in places apparently given verbatim; and chapters 96 (both parts) and 61 fols. 1-6 of the Ch'ien-han-shu (Annals of the Former Han) of Pan-ku, who died in a.d. 92 and whose history, left incomplete at his death and finished by his sister, runs from 206 b.c. to a.d. 24; chapter 61 fols. 1–6 contains among other things a Life of Chang-k'ien, largely drawn from himself, and chapter 96 is Pan-ku's own account of the Western Countries, based on Chang-k'ien's Report which is sometimes apparently quoted verbatim, but incorporating later material. Ssu-ma Ch'ien is supposed to be the more valuable source for what Chang-k'ien actually wrote, but it is not always easy to say what is Chang-k'ien and what is Ssu-ma Ch'ien; the latter writer brings in later material just as Pan-ku does, and for historical purposes chapter 123 of the Shi-ki requires the same kind of critical analysis as has been applied to many Greek historians; I have done what little I can from a translation (see for example p. 281) but I do not pretend that it can be satisfactory. Pan-ku (who of course needs a similar analysis) had much new information at his disposal which Ssu-ma Ch'ien had not possessed, and his occasional corrections of the latter on matters like geography can be valuable.

The Gārgī Samhitā is an astrological work of uncertain date (it has been dated anywhere from the Christian Era to the third century a.d.), one of whose chapters, the Yuga-purāna, contains an historical account of (among other matters) the Greek advance to Pātaliputra, written as usual in the form of a prophecy. The existing texts of the Yuga-purāna are written in Sanscrit with (it is said) traces of Prakritisms; in the opinion of the late Dr Jayaswal, who devoted special attention to this work, the extant account must go back to a historical chronicle, written either in Prakrit or in mixed Sanscrit-Prakrit, which he dates in the latter half of the first century b.c. on the ground that it mentions no dynasties later than the Sacas. Historians of India have usually considered the historical account of the Yavanas in the Yuga-purāna as valuable, an opinion shared by Jayaswal, who regards the work as the earliest known Purāna and as exhibiting an independent tradition; occasionally someone has dissented from this view, but the manner in which the accounts of the Greek Apollodorus and of the Yuga-purāna complement each other (Chap. iv) ought to be conclusive for the Yavana sections, as the two are presumably independent.

The history of the Yuga-purāna in modern times is peculiar. H. Kern in 1865 first brought it to notice in the preface to his edition of the Brihat Samhitā; he possessed a single MS, apparently rather broken, and he gave a translation of the greater part (not all) of what are §§ 5 and 7 in Jayaswal's translation (p. 453). Of §6 he merely said that it contains complaints against heretics, presumably Buddhist monks.

The Milindapañha or Questions of Milinda is the one extant work professedly dealing with any of the Greek monarchs in the Far East; for Milinda, beyond any question, is the king Menander. It exists in a Pali version and, in part, in a Chinese translation of the fourth century A.D. of which two recensions are extant. The Pali work falls into two well-marked divisions; the first comprises pp. 1–89 in Trenckner's edition of the Pali text, being books I–III inclusive; the second and longer part comprises all that follows. It is now generally agreed that Part II is later than Part I and the work of a different hand, and it is also generally agreed that Part I (or perhaps I should say the original of Part I) cannot be placed too long after Menander's death; but I need not quote the datings suggested, for none of those who have professedly dealt with the work have investigated Menander's chronology and have usually put him near the end of the second century b.c. or even in the first century. The Chinese translation includes Part I and a few pages of Part II.

The work is cast in the form of a dialogue between Menander and a Buddhist sage Nāgasena, with an introduction in which Menander, at his capital Sāgala, appears as a great king fond of learned disputations, together with his 500 Yonakas, four of whom play a part in setting the scene for the dialogue proper. In the first part Menander's professed object is not the pursuit of knowledge but a dialectical victory over Nāgasena, though he does not in fact keep his end up very well.

Hermaeus' Greek legend on his own coins, both silver and copper, which were struck in the A!lexandria-Kapisa mint, is always βασιλέως σωτῆρος Ἑρμαίου, ‘Of king Hermaeus, saviour’, and the Kharoshthi legend is the corresponding Maharajasa tradatasa Heramayasa, with the ‘Zeus enthroned’ of Alexandria-Kapisa. The Kadphises coins, which are of inferior style, resembling the debased copies of Hermaeus' money issued after his death, fall into two classes. The first class has: obverse, bust of Hermaeus diademed and Greek legend βασιλέως στηρος συ Ἑρμαίου (often corrupted); reverse, Heracles facing with club and lion's skin, and a Kharoshthi legend signifying ‘Kujula Kadphises, Kushan, yavuga (chief)’. This class has the square omicron but no other square letters. The second class has: obverse, bust of Hermaeus diademed, and a Greek legend, usually mutilated, reading ‘Kujula Kadphises, Kushan’; reverse, the same as the first class. This class, beside the square omicron, has the square sigma and phi which one associates with the coins of Gondophares; as both classes must be near in time, it is clear that the period is getting too late for letter forms to mean much chronologically.

Menander was the most famous of the Yavana kings, and his legend attests the impression he made upon the world about him; and a sketch must now be attempted of the kingdom of the man who for a little while had held Asoka's capital and whose conquests were exalted by a Greek historian even above those of Alexander. The deaths of Demetrius and Apollodotus and the return of Eucratides to Bactria left him master of the position in India, and thenceforth to his death he, the one man who had successfully resisted Eucratides, ruled the whole of the territory still remaining to the Greeks in that country, excluding the Paropamisadae; if he had not the royal title before (p. 167) he must have taken it, presumably by a vote of his army, when Demetrius was killed. The growth of his legend, and the establishment of his coinage in Barygaza, postulate for him a reign of reasonable length; at the same time the fact that his son, Strato I, was too young to rule alone when he died sets a definite limit to that length. He legitimised his rule by marrying Demetrius' daughter Agathocleia; the evidence that she was his queen seems conclusive. If cadets of the house of Euthydemus still survived, they must have accepted his rule as the only security against Eucratides and his line; he never had any civil war—at least in all his vast coinage no coin seems known which has been overstruck by anyone else or upon anyone else's money.

It is not possible to get an accurate chronology for Menander's reign, but one must approximate as nearly as possible.

The three lectures here published are the Lees-Knowles Lectures in Military History for 1929–30, delivered at Trinity College, Cambridge, in the spring of this year; they were somewhat shortened in actual delivery. The two matters which may perhaps claim some novelty, the evolution of the great war-horse and of the Hellenistic great ships, are now rather more fully treated, chiefly by means of notes and appendices; indeed, the principal justification for the appearance of these lectures in book form must be the section on the ships, about which comparatively little has hitherto been written to much purpose. Professor F. E. Adcock very kindly read through my typescript before it went to the publishers and made several suggestions, more particularly in the early part of the first lecture, which have much improved the book. My best thanks are due to him, and also to the Syndics of the Cambridge University Press for undertaking the publication of these lectures.