The Date of Kanishka

BY J.H. Marshall, C.I.E., Litt.D., F.S.A.
Journal of the Royal Asiatic Society of Great Britain and Ireland
1914, pp. 973-986



I p. 973 In the discussion on the date of Kanishka which took place last year much prominence was given to the so-called Vikrama era of 58 B.C., and arguments were put forward by those who favoured an early date for Kanishka to prove, not only that this era was founded by him, but that all the recorded dates of the Saka, Pahlava, and Kushan rulers of Northern India are referable to it. Some discoveries which I have recently made at Taxila throw, I think, fresh light on this question, and, though they do not settle precisely the date of Kanishka, appear to prove that he was not at any rate the founder of the era of 58 B.C. and could not have come to the throne until the close of the first century A.D. or later. One of these discoveries I made in a small chapel immediately west of the so-called "Chir" stupa. The chapel in question is built in a small diaper type of masonry, which came into vogue at Taxila about the middle of the first century A.D. and lasted for about a hundred years. Its entrance faces the main stupa, and near the back wall opposite this entrance, and about a foot below the floor,I found a deposit consisting of a steatite vessel with a silver vase inside, and in the vase an inscribed scroll and a small gold casket containing some minute bone relics. A heavy stone placed over the deposit had, unfortunately, been crushed down by the fall of the roof and had broken both the steatite vessel and the silver vase, but had left the gold casket uninjured and chipped only a few fragments from the edge of the scroll, nearly all of which I was, happily, able to recover p. 975 by carefully sifting and washing the earth in the vicinity. The cleaning and transcription of the record was a matter of exceptional difficulty, as the scroll, which is only 6 1/4 inches long by 1 3/8 inches wide and of very thin metal, had been rolled up tightly, face inwards, in order that it might be enclosed in the silver vase; moreover, the metal of which it is composed is silver alloyed with a small percentage of copper, which had formed an efforescence on the surface of the extremely brittle band, with the result that I could neither unroll it without breaking it nor subject it to the usual chemical treatment. By the use of strong acid, however, applied with a zinc pencil, I was able to remove the copper efforescence and expose, one by one, the punctured dots of the lettering on the back of the scroll, and then, having transcribed these with the aid of a mirror, to break off a section of the scroll and so continue the process of cleaning and transcription. In this way I succeeded in making a complete copy of the record from the back of the scroll, while the letters were yet intact. Afterwards I cleaned in like manner and copied the face of each of the broken sections, and was gratified to find that my second transcript was in accurate agreement with the first. I mention these details in order to explain why it is impracticable to present a purely mechanical reproduction of the record, and why the transcript given below, which was made section by section and necessarily without any reference to the meaning of the record, may be regarded as trustworthy. My reading of this inscription is as follows:-- Text Line 1. Sa 100.20.10.4.1.1 Ayasa Ashsdasa masasa divase 10.4.1 isa divase pradistavita Bhagavato dhatuo Dhurasa(?)- 1. 2. kena Dhitaphria,-putrana Bahaliena Noachae nagare Vastavena tena ime pradistavita Bhagavato dhatuo dhamara- p. 976 l. 3. ie Tachhasie Tanuae Bodhisatvagahami maha- rajasa rajatirajasa devaputrasa Khushanasa arogadachinae 1. 4. sarva-budhana puyae prachaga-budhana puyae arahana puyae sarvasa(tva)na puyae mata-pitu puyae mitra-macha-nati-sa- 1. 5, dhihona puyae atmano arogadachhinae nianae hotu a . de sa ma parichago Translation In the year 136 of Azes, on the 15th day of the month of Ashadha - on this day relics of the Holy One (Buddha) were enshrined by Dhurasakes (?), son of Dhitaphria, a Bactrian, resident at the town of Noacha. By him these relics of the Holy One were enshrined in the Bodhisattva chapel at Tanua(?) in Takhasila of the religious realm, for the bestowal of perfect health upon the great king, king of kings, the divine Khushana; for the veneration of all Buddhas for the veneration of individual Buddhas; for the veneration of the Saints; for the veneration of all sentient beings; for the veneration of (his) parents; for the veneration of (his) friends, advisers, kinsmen, and associates; for the bestowal of perfect health upon himself. May this gift be... For Dhitaphria-putrana Mr. D. R. Bhandarkar suggests the reading dhitastria-putrana, which, in combination with bahaliena, he would translate "accompanied by his daughters, wife and sons". For the meaning of Bahaliena (="a man from Balkh") I am indebted to Professor Rapson and Professor Konow. To Mr. Bhandarkar I am also indebted for the suggestion that dhamaraia = Skt. dharmarajya. The absence of any royal titles attached to the name of Azes is exceptional, but will hardly occasion surprise when it is borne in mind that his era had been in use for more than a century, and that his dynasty had been p. 977 supplanted by that of the Kushans. When did this era of Azes commence? That it is one and the same as the era in which the Gondophernes and Panjtar records are dated will, I think, be admitted by everyone, and I shall not therefore pause to discuss the point. If, then, Dr. Fleet is correct in referring the dates of the latter records to the era of 58 B.C., it follows that it was Azes I and not Kanishka who founded that era. That Azes I came to the throne about that date is now, indeed, attested by the evidence of coins and other antiquities at Taxila, which indicate that he was reigning in the third quarter of the first century B.C., while the probability that he may have founded an era is also suggested by the abundance of his coins, which denote his pre-eminence among the Saka-Pahlava sovereigns. Notwithstanding, however, the very strong reasons which Dr. Fleet has adduced for referring the dates in the Gondophernes and Panjtar records to the era of 58 B.C., the identity of the era of Azes and the Vikrama era call hardly be regarded as fully established, and, to my mind, it is quite possible that the era of Azes will be found to have commenced a few years earlier or later than 58 B.C. Assuming that it started actually in that year, the date given in the new inscription will fall in the year A.D. 79; and the next important point is to determine which of the Kushan kings is referred to as reigning in that year. That he is identical with the nameless Kushan ruler mentioned in the Panjtar record of fourteen years earlier is probable; and here, again, I think Dr. Fleet may be correct in identifying the latter with Vima-Kadphises. On the other hand, it is also possible that Kujula-Kadphises may be meant. The monogram on the scroll is characteristic of coins of Vima- Kadphises, but it is also found on coins of his predecessor.(l) 1. Cf. Vincent Smith, Catalogue of Coins in the Indian Museum, p. 67, Nos. 17, 22, 24. p. 976 l. 3. ie Tachhasie Tanuae Bodhisatvagahami maha- rajasa rajatirajasa devaputrasa Khushanasa arogadachinae 1. 4. sarva-budhana puyae prachags-budhane puyae arahana puyae sarvasa(tva)na puyae mata-pitu puyae mitra-macha-nati-sa- 1. 5, dhihona puyae atmano arogadachhinae nianae hotu a . de se ma parichago Translation In the year 136 of Azes, on the 15th day of the month of Ashadha - on this day relics of the Holy One (Buddha) were enshrined by Dhurasakes (?), son of Dhitaphria, a Bactrian, resident at the town of Noacha. By him these relics of the Holy One were enshrined in the Bodhisattva chapel at Tanua(?) in Takhassila of the religious realm, for the bestowal of perfect health upon the great king, king of kings, the divine Khushana; for the veneration of all Buddhas for the veneration of individual Buddhas; for the veneration of the Saints; for the veneration of all sentient beings; for the veneration of (his) parents; for the veneration of (his) friends, advisers, kinsmen, and associates; for the bestowal of perfect health upon himself. May this gift be... For Dhitaphria-putrana Mr. D. R. Bhandarkar suggests the reading dhitastria-putrana, which, in combination with bahaliena, he would translate "accompanied by his daughters, wife and sons". For the meaning of Bahaliena (="a man from Balkh") I am indebted to Professor Rapson and Professor Konow. To Mr. Bhandarkar I am also indebted for the suggestion that dhamaraia = Skt. dharmarajya. The absence of any royal titles attached to the name of Azes is exceptional, but will hardly occasion surprise when it is borne in mind that his era had been in use for more than a century, and that his dynasty had been p. 977 supplanted by that of the Kushans. When did this era of Azes commence? That it is one and the same as the era in which the Gondophernes and Panjtar records are dated will, I think, be admitted by everyone, and I shall not therefore pause to discuss the point. If, then, Dr. Fleet is correct in referring the dates of the latter records to the era of 58 B.C., it follows that it was Azes I and not Kanishka who founded that era. That Azes I came to the throne about that date is now, indeed, attested by the evidence of coins and other antiquities at Taxila, which indicate that he was reigning in the third quarter of the first century B.C., while the probability that he may have founded an era is also suggested by the abundance of his coins, which denote his pre-eminence among the Saka-Pahlava sovereigns. Notwithstanding, however, the very strong reasons which Dr. Fleet has adduced for referring the dates in the Gondophernes and Panjtar records to the era of 58 B.C., the identity of the era of Azes and the Vikrama era can hardly be regarded as fully established, and, to my mind, it is quite possible that the era of Azes will be found to have commenced a few years earlier or later than 58 B.C. Assuming that it started actually in that year, the date given in the new inscription will fall in the year A.D. 79; and the next important point is to determine which of the Kushan kings is referred to as reigning in that year. That he is identical with the nameless Kushan ruler mentioned in the Panjtar record of fourteen years earlier is probable; and here, again, I think Dr. Fleet may be correct in identifying the latter with Vima-Kadphises. On the other hand, it is also possible that Kujula-Kadphises may be meant. The monogram on the scroll is characteristic of coins of Vima- Kadphises, but it is also found on coins of his predecessor.(l) 1. Cf. Vincent Smith, Catalogue of Coins in the Indian Museum, p. 67, Nos. 17, 22, 24. p. 978 Again, the title maharajasa rajatirajasa also suggests Vima-Kadphises; indeed, it was stated by several speakers during the discussion on the date of Kanishka that Kujula-Kadphises was only a petty local chief (yavuga jabgou), never "a king of kings", like his successor. But this assertion is erroneous. On some of his coins Kujula- Kadphises styles himself maharaja rajadirajasa, (1) and, according to Cunningham, devaputrasa also.(2) That he ruled, moreover, at Taxila, and consequently over the north-west of the Punjab and Frontier generally, is abundantly clear from his coins, which are found there in larger numbers than those of any other kings except Azes I and Azes II.(3) Other considerations, too, favour the identification with Kujula- rather than Vima- Kadphises. For, in the first place, it would be natural for the first emperor of the dynasty to be styled "the Kushan Emperor" without any further appellation, while it would be equally natural for his successors to be distinguished from him by the addition of their individual names. Secondly, the stratification of coins at Taxila show that Kujula-Kadphises succeeded the Pahlava kings there, and consequently he can hardly have conquered the country before circa A.D. 50; and inasmuch as his coins betoken a fairly long reign there, and he is known from other sources to have lived to a great age, he may well have been ruling in the 122nd and 136th years of the era of Azes, i.e., approximately, in A.D. 65 and 79. For these reasons it would, in my opinion, be unsafe at present to regard as certain the identity of the emperor referred to in this record with Vima-Kadphises, notwithstanding other evidence which 1. Cf. R. B. Whitehead, Catalogue of Coins in the Punjab Museum, Lahore, vol. i, p. 180, Nos. 20, 22. 2. Cf. Num. Chron., vol. xii, p. 66, 1892. 3. Thus, within the walls of Sir-kap alone I estimate from my finds up to date that there are not less than 18.000 of his coins hidden within the soil. P. 979 undoubtedly exists for assigning an earlier date to Kujula-Kadphises. From the new discoveries at Taxila, coupled with already known facts, the succession of the Saka, Pahlava, and Kushan rulers in this part of India appears to have been as follows: Maues. Kujula-Kadphises and Azes I. Hermaeus. Azilises. Vima-Kadphises Azes II Aspavarma (Soter megas). Gondophernes Strategos. Kanishka. Abdagases, Sasan, Sapedanes, Huvishka. Vasudeva. The coins of Maues are relatively few, and this bears out the theory that Maues rose to power in Arachosia and did not extend his sway over Taxila until relatively late in his reign. Rare, too, are the coins of Azilises, who seems to have had a short reign and may have been represented at Taxila by local governors. The existence of Azes II, which was first postulated by Mr. Vincent Smith, is not generally admitted by other numismatists; but the following facts appear to me strongly to support Mr. Smith's view: (1) The coins which he assigns to Azes II are found generally nearer the surface than those of Azes I. (2) Aspavarma appears to have been strategos in the reign of Gondophernes(1) as well as in that of Azes, and it is impossible that this Azes can be Azes I, who came to the throne seventy-eight years before Gondophernes. (3) Coins of Azes II (with Aspavarma) are found in company with coins of Gondophernes.(2) After the death of Gondophernes his empire was split up into smaller principalities, and it was then that Hermaeus and Kadphises I appear to have made their successful invasion 1. Cf. R. B. Whitehead, Catalogue of Coins in the Punjab Museum, Lahore, vol. i, p.150, Nos. 35-8. 2. e.g. twenty-three coins of Gondophernes (with Sasan) in company with four of Azes II (with Aspavarma). p. 980 of Gandhara and Taxila. One of these principalities was ruled by Abdagases, another by Orthagnes, and others by princes whose coins I have now recovered for the first time at Taxila. Among them were Sasan, Sapedanes, and Satavastra(?). Coins of Gondophernes with the legend Sasasa have long been familiar to numismatists, and it has been a matter of dispute whether this word contained the name of a ruler or was merely an unexplained epithet of Gondophernes. But a new type of silver coin from Taxila, bearing on the obverse the legend Maharajasa Aspabhataputrasa(1) tratarasa Sasasa, seems to indicate that General Cunningham was right in interpreting it as the name of a ruler. I suggest that Sasan may have been a Viceroy of Gondophernes during the lifetime of the latter, and have made himself maharaja of his province on Gondophernes' death. The legends on the other coins referred to read respectively: maharajasa rajarajasa tratarasa dhramiasa Sapedanasa, (2) and maharajasa Satavastrasa.(3) The corrupt legends on the obverse are not clear,(4) but the symbol appears on all of them, and in other respects the coins are closely allied to those of Gondophernes. The titles of these potentates imply that they were independent at the time when these coins were struck; but there is nothing to prove that any of them was ruling in Taxila. Probably they were ruling in other parts of the country when Kujula-Kadphises and Hermaeus had already taken possession of Taxila. This supposition is supported by the fact that no copper coins of these later Pahlava princes have yet been found there, 1. Aspabhataputrasa may perhaps be read as Aspabhra- taputrasa, in which case Sasan may have been a nephew of Aspavarman. 2. The lower half of the second aksara of this name is somewhat doubtful. Perhaps it may be Sarpedanasa. 3. The reading Satavastrasa is clear, but it is difficult to believe that this is the name of a king. 4. On one of the coins of Sapedanes or Sarpedanes the Greek letters . . CAPHN . . . are visible. p. 981 and that the silver pieces alluded to above were all found together in one jar(1) in a stratum which has yielded many copper coins of Hermaeus and Kujula-Kadphises. In the absence of any silver mintage of Hermaes or Kadphises it is not, of course, surprising that silver coins should have found their way to Taxila from neighbouring Pahlava. principalities. If any of these princes succeeded Gondo phernes at Taxila and reigned for any length of time there, then the conquest of Kujula-Kadphises and Hermaeus can hardly have taken place before about A.D. 60, in which case there will be still more reason for identifying the former with the Kushan monarch referred to in the Panjtar record. Among the coins of Hermaeus and Kujula-Kadphises are a certain number struck in the name of Hermaeus alone,(2) but the vast majority are those of Hermaeus and Kadphises or of Kadphises alone, nor does there seem to be sufficient reason for supposing that Taxila was ever included within the kingdom of Hermaeus, prior to the conquest of the latter by Gondophernes. On the contrary there are good grounds for believing that Azes II was succeeded directly by Gondophernes, who afterwards proceeded to annex the Kabul kingdom of Hermaeus. Hermaeus, we may assume, formed an alliance with Kujula-Kadphises, recovered his own lost dominions, and after the death of Gondophernes took advantage of the break-up of the great Pahlava kingdom to invade Gandhara and Taxila.(3) 1. With them were a figure of a winged Aphrodite of gold repousse, a number of intaglio gems engraved with figures of Eros, Artemis, etc., and other pieces of gold jewellery. 2. Of two types, viz. B.M. Cat., pls. xv, 6, and xxxii, 8. 3. The prevalent view taken by historians and numismatists is that Kadphises I conquered Hermaeus circa A.D. 20 or even earlier (cf. Vincent Smith, The Early History of India, 3rd ed., 1914, p. 236; Rapson, Indian Coins, p. 16, par. 65). In that case Kadphises I must have been driven back from Taxila and Kabul by Gondophernes. I find nothing to support this supposition. p. 982 Of the nameless king, Soter Megas, all that can be said at present is that his coins are not found in Sir-kap; and, as they are common enough on the sites round about, it may be inferred that he was certainly later than Kujula- Kadphises, but how much later yet remains to be seen. To revert, however, to Kanishka. We have seen that he was not the founder of the era commencing in 58 B.C., or thereabouts, and that there is no place for him and his immediate successors among the Saka and Pahlava kings, who were ruling at Taxila in the first centuries before and after Christ. I turn now to more positive evidence regarding his date. That he followed and did not precede the two Kadphises is abundantly clear from my excavations both in the city of Sir-kap and at the Chir stupa. Sir-kap was built during the Greek domination and, apparently, remained in occupation as a city until the reign of Vima-Kadphises. In it I have now cleared a reasonably representative area, measuring some 3 3/4 acres, and including part of a main street, several side streets, and a number of large edifices. I have unearthed buildings of the Greek, Saka, Pahlava, and Kushan epochs, and I have discovered, buried in small hoards beneath their floors or dropped singly in the chambers. alleys, and roads, coins of the following kings:-- Greek Saka and Pahlava 1. Agathocles 11. Maues. 2. Lysias. 12. Vonones (with Spalahora). 3. Eucratides. 13. Azes I. 4. Antialcidas. 14. Azilises. 5. Apollodotus. 15. AzesII. 6. Heliocles. 16. Gondophernes. 7. Hippostratus. 17. Abdagases. 8. Philoxenus. 18. Sasan. 9. Telephus. 19. Sapedanes. 10. Hermaeus. 20. Satavastra(?). Kushan 21. Kujula-Kadphises. 22. Vima-Kadphises. p. 983 Yet I have not come across a single coin of Kanishka, Huvishka, or Vasudeva. How can this entire absence of their coins be explained, except on the hypothesis that these three emperors came later than Vima-Kadphises, during whose reign the city appears to have been deserted --particularly when it is remembered that their coins remained in circulation long after their deaths? At the Chir stupa, on the other hand, I have brought to light a series of buiidings covering a much longer period-- namely, from the middle of the first century B.C. to the fourth or fifth century A.D. These buildings are characterized by four clear and distinct types of masonry. Those in the lowest stratum are of rubble, often faced with finely cut kanjur stone; built over them are structures of small "diaper" masonry; above these, again, are buildings of a larger and more massive type of "diaper"; and uppermost of all come stupas and chapels of semi-ashlar, semi-diaper masonry. Now, coins of the Saka and Pahlava kings are found associated with the first of these four classes, and coins of the two Kadphises are found in buildings of the second class; but not a single coin of Kanishka, Huvishka, or Vasudeva has been found in any building earlier than those of the third class. Nor is this evidence derived only from coins found in the debris of these buildings. In one case the relics in a small stupa associated with a building of the third class, and certainly not older than the second century A.D., were accompanied by coins of Huvishka and Vasudeva only. Thus in Sir-kap we have, represented by their coins, a succession of rulers from the second century B.C. until the latter part of the first century A.D., but not a trace among them of Kanishka, Huvishka, or Vasudeva; and at the Chir stupa we have a succession of Saka, and Pahlava kings followed by the two Kadphises, with Kanishka, Huvishka, and Vasudeva coming later. And if we look at other monuments associated with p. 984 Kanishka and Huvishka, we are forced to the same conclusion regarding their date. Thus the original masonry of the Kanishka Stupa at Peshawar is of a type which at Taxila, at any rate, was unknown in the Saka-Pahlava period, but is paralleled there in buildings of the second century A.D. True, Peshawar is at some distance from Taxila, and it is possible, though not likely, that a local style may have been independently evolved there. But at Manikyala, which is within 40 miles of Taxila, no such explanation will avail. There the great stupa erected during Huvishka's reign is similar in all its details--in its dwarfed pilasters, degenerate Corinthian capitals, bevelled torus mouldings, notched Indian brackets, and the like(1) -- to monuments of the second and third centuries A.D. at Taxila, but markedly different from those of the first century B.C. In concluding this brief note let me add that the new information which I have gleaned about the Saka-Pahlava rulers, coupled with the opening words of the inscription given above, have suggested to me a more satisfactory solution of the difficulties connected with the Patika copper-plate, King Moga, and the chronology of the local Satraps of Mathura. The most orthodox view, at present, is to identify the King Moga referred to in the Patika copper-plate with King Moa (= Maues) of the coins (circa 120 B.C., according to the generally accepted chronology), to regard Patika as a contemporary of Moga, and to place Rajuvula and Sodasa slightly later. On the other hand, Dr. Fleet 1. Cf. Cunningham, ASR., vol. v, pl. xxiv, which, however, is not entirely accurate. Fergusson's woodcut (ed. 1910, p. 98, fig. 27) is a mere travesty of the original. The attribution of this monument, as itment, as it now stands, to the eighth century A.D. or thereabouts is one of the most amazing blunders ever made by Fergusson, as amazing as his attribution of the Dhamekh stupa at Sarnath to the eleventh century. The style of the architectural decorations around the plinth and base of the super- structure is precisely that which prevailed at Taxila in the second century A.D., but was completely transformed during the three succeeding centuries. p. 985 differentiates between King Moa of the coins and Moga of the copper-plate, as well as between Patika of the Taxila plate and Patika of the Mathura lion-capital; he refers the year 78 of the Taxila inscription, as well as the year 72 of the Amohini record of Sodasa's reign, to the Vikrama era of 58 B.C.; and he places Rajuvula and Sodasa, together with King Moga and Patika, in the early part of the first century A.D. Neither of these solutions can be considered satisfactory: the first, because it ignores the fact that the style of the sculptures of Sodasa's reign at Mathura entirely precludes their being ascribed to so early a date as the second quarter of the first century B.C.; the second, because there is no reason for supposing that there was another Patika, and because it is prima facie improbable that King Moga was ruling in the first century A.D., in the same year as Gondophernes. Now, let us consider what these three records-the Patika copper-plate, the Mathura lion-capital, and the Amohini ayagapata slab--have to tell us regarding the succession of these Satraps. From the first we learn that Patika was not yet invested with satrapal powers in the year 78, when his father, the Satrap Liaka-Kusulaka, was still alive. From the Mathura lion-capital we learn that, at the time it was inscribed, Patika had become "Great Satrap", that Rajuvula was also a "Great Satrap", and his son Sodasa only "Satrap". And from the Amohini slab we learn that in the year 72, in which it is dated, Sodasa had become "Great Satrap" in succession to his father Rajuvula. Thus we have the following order of succession indicated in these records:--- Liaka-Kusulaka | Patika -- approximately contemporary with -- Rajuvula _______|_______ | | Arta = daughter Sodasa Kharahostes(1) 1. See Fleet, JRAS, October, 1913, p. 1009. p. 986 From this it follows that if Liaka-Kusulaka was Satrap in the year 78 of the era of 58 B.C., Sodasa could not have been Great Satrap in the year 72 of the same era. In older to get over this difficulty Dr. Fleet, as stated above, assumes the existence of two Patikas. But is it necessary to refer these dates to one and the same era, or to refer the year 78 of the Patika plate to an unspecified era? I hold that it is not. The new Taxila inscription proves that the year 136 of that record is dated, not in an unspecified era and during the reign of Azes, but in the era founded by Azes himself; and, if we compare this inscription with the Patika plate, we find that the words maharayasa mahamtasa Mogasa occupy the same position as the word Ayasa in the new record. Let us see, therefore, what the result will be if we refer the year 78 of the Patika plate to the reign of Maues and the year 72 of the Amohini slab to the era of Azes. According to the numismatic and other evidence from Taxila, Maues or Moa immediately preceded Azes, and, as he must have enjoyed a fairly long reign, I place his accession about 95 B.C. The year 78 of his reign, therefore, when Liaka-Kusulaka was Satrap, will fall about 17 B.C. The reigns of Patika and Rajuvula we shall place roughly between 10 B.C. and A.D. 10, that of Sodasa after A.D. 10 (the year 72 of the Amohini record falling in A.D. 15), and that of Kharahostes, say, A.D. 30-45. This chronology seems to me to accord satisfactorily with the numismatic evidence and all else that we know about the lineage of these Satraps.