To the north, still conscious of their nomad origins, the Kushans sought to restore contact with the Chinese borderlands in which their wanderings had begun. No doubt their great resources in animal transport as a nomadic people gave them the means to set the Chinese trade flowing, and to form a bridge between the civilizations of India and China. At the same time the fact that in Bactria they were the successors of a Hellenistic civilization gave them special points of contact with the Mediterranean world. The staple item in the westbound trade was of course silk, which could now avoid Parthian territory and be diverted southwards to the Indus delta, to finish its journey to the Roman empire by sea. In return, besides gold coin, Rome sent manufactured goods of many kinds - woollen tapestries, engraved gems and cameos, figurines and metalware; but perhaps most important of all, the magnificent glassware of Alexandria, since China in the 1st century A.D. had not developed the manufacture of glass. At the same time the Indian territories of the Kushans exported both towards China and to Italy their exquisite ivories. For all this trade the French excavations at Begram in Afghanistan have proved the most revealing source of information, but minor finds are known from many sites.
The complexity of the migration in which the Kushans took part and the variety of the tribes which accompanied them naturally raise many difficult problems of race and language. So far as the ruling group were concerned, their drooping moustaches and bulbous features, as frequently depicted in the sculpture of Gandhara, or on the coins, have caused many commentators to doubt, perhaps with insufficient reason, that they could have been Indo-Europeans. Kennedy argued eloquently that they represented a Turkish, or at any rate non-Iranian physical type.1 The medieval Arab writer al-Birunt regarded them as Tibetans ;2 whilst their contemporary, the Syrian Bardesanes in his Book of the Laws of the Countries? reports the existence of matriarchal tendencies in their society. At the same time their dress and equipment - the buckled cloak, long shirt, and the baggy trousers of the horseman, depicted in many works of Gandhara sculpture, and especially on the statue of the emperor Kanishka from Mathura (pi. 79) - were no different from that of the other Iranians of the steppe. Like the Sakas they wore scale-armour, and their weapons included a straight sword over three feet long.
It is not known whether the Kushans possessed a special language, distinct from that of their associated tribes. However, it is now clear that the official language of the Kushan empire, an Iranian dialect written in Greek script, is in fact Bactrian - that is, the local eastern Iranian dialect of the province in which the Kushans had settled after crossing the Oxus. This language is now known from the inscriptions of a large part of the Kushan coinage, and especially from the twenty-five-line lapidary "Inscription of Nokonzoko" discovered by the Delegation Archeologique Franchise en Afghanistan at Surkh Kotal, and interpreted by the late Professor W. B. Henning.4 It seems likely that this adaptation of the Greek script for the writing of the Bactrian language had already been carried out by the Greeks in Bactria before the coming of the Kushans, since the devices employed for the rendering of the specifically Iranian sounds are entirely of the type that would occur to the Greek mind; e.g., the rendering of the aspirate by upsilon (since in Greek initial upsilon invariably carries the rough breathing), and of the /sibilant by san. It is also curious that the "Inscription of Nokonzoko" contains one word in common with the Khotanese texts of Central Asia. This is the Bactrian xsono, Khotanese ksuna - " regnal year"-but it remains to be determined in what dialect this word originated.
The next group in the tribal hierarchy after the Kushans were the Tocharians, but of their language no trace survives in Bactria. Some scholars are inclined to attribute to the Tocharians the two IndoEuropean centum dialects found in manuscript fragments from Kucha and Qarashahr. However, though there is a possibility that one or both of these dialects may have been connected with that of the migrating Tocharians, it seems clearer until positive proof comes to hand to designate them Agnaean and Kuchaean, rather than to forejudge the question by terming them "Tocharian A and B". A rather similar problem arises in connection with the dialect of the Sakas in India. Words and names attributable to them are found on the Indian coins and inscriptions, and have been variously connected. Thus the name of the Saka satrap Castana has been compared with the modern Pashto word tsaxtan "master", suggesting a connection between Pashto and the Indo-Scythian dialect. On the other hand such IndoScythian words as horamurta " supervisor of donations " and bakanapati "priest" have been compared with Khotanese. Yet since Pashto and Khotanese are not closely related, the linguistic situation in the Kushan empire must have been more complex than at first appears.
Over the chronology of the Kushan empire much controversy prevails, but a definitive solution may now be within reach. The relative chronology of the rulers can be deduced from their coins, and is fairly well agreed. Starting in the first decades of the Christian era, there ruled Kujula Kadphises, then a "nameless king", designated on coins merely by his title, Soter Megas "The Great Saviour", and Vima Kadphises. The next series of coins provides the names of three further rulers, Kanishka (Kaniska) I, Huvishka (Huviska) and Vasudeva. However, the absolute chronology of these reigns, and our knowledge of several minor members of the dynasty who did not strike coins, has to be deduced by assembling the evidence of a large number of dated inscriptions, some in Kharosthi and others in Brahmi script, from various sites in the subcontinent. To these must now be added the two dated Bactrian inscriptions from Surkh Kotal in Afghanistan. Complications arise in the study of the inscriptions from the fact that several different eras are used for dating. To devise a chronological scheme, the eras have first to be identified. In fact, the best hypothesis assumes the existence of four quite different eras, all probably originating from the accession years of dynastic founders, on the following lines: (a) An Indo-Bactrian era of c. 15 5 B.C. (often called the Old Saka era). (b) The era of Azes commencing in 57 B.C. (c) The Saka era, used originally by the western satraps of Ujjain, commencing in A.D. 78. (d) The era of Kanishka, c. A.D. 128. The epigraphical evidence on which this scheme is based is given in Appendix iv.
The Nomad Invasions
