James M. Campbell

Appendix IV.
Java. Java.An incident redeems the early history of GujarÁt from provincial narrowness and raises its ruling tribes to a place among the greater conquerors and colonisers. This incident is the tradition that during the sixth and seventh centuries fleets from the coasts of Sindh and GujarÁt formed settlements in Java and in Cambodia. The Java legend is that about a.d.603 Hindus led by BhruvijÁya SavelachÁla the son of Kasamachitra or BÁlya AchÁ king of KujrÁt or GujarÁt settled on the west coast of the island.1 The details of the settlement recorded by Sir Stamford Raffles2 are that Kasamachitra, ruler of GujarÁt, the tenth in descent from Arjun, was warned of the coming destruction of his kingdom. He accordingly started his son BhruvijÁya SavelachÁla with 5000 followers, among whom were cultivators artisans warriors physicians and writers, in six large and a hundred small vessels for Java. After a voyage of four months the fleet touched at an island they took to be Java. Finding their mistake the pilots put to sea and finally reached Matarem in the island of Java. The prince built the town of Mendang Kumulan. He sent to his father for more men. A reinforcement of 2000 arrived among them carvers in stone and in brass. An extensive commerce sprang up with GujarÁt and other countries. The bay of Matarem was filled with stranger vessels and temples were built both at the capital, afterwards known as Brambanum, and, during the reign of BhruvijÁya’s grandson ArdivijÁya that is about a.d.660, at Boro Buddor in Kedu.3 The remark that an ancestor of the immigrant prince had changed the name of his kingdom to GujarÁt is held by Lassen to prove that the tradition is modern. Instead of telling against the truth of the tradition this note is a strong argument in its favour. One of the earliest mentions of the name GujarÁt for south MÁrwÁr is Hiuen Tsiang’s (a.d.630) Kiu-che-lo or Gurjjara. As when Hiuen Tsiang wrote the Gurjjara chief of BhinmÁl, fifty miles west of Ábu, already ranked as a Kshatriya his family had probably been for some time established perhaps as far back as a.d.490 a date by which the Mihira or Gurjjara conquest of Valabhi and north GujarÁt was completed.4 The
Appendix IV.
Java. details of the help received from GujarÁt after the prince’s arrival show that the parent state had weathered the storm which threatened to destroy it. This agrees with the position of the BhinmÁl Gurjjaras at the opening of the seventh century, when, in spite of their defeat by PrabhÁkaravardhana (a.d.600–606) the father of SrÍ Harsha (a.d.606–641) of Magadha, they maintained their power at Broach and at Valabhi as well as at BhinmÁl.5 The close relations between the Gurjjaras and the great seafaring Mihiras or Meds make it likely that the captains and pilots who guided the fleets to Java belonged to the Med tribe. Perhaps it was in their honour that the new Java capital received the name Mendan, as, at a later period it was called Brambanum or the town of BrÁhmans. The fact that the Gurjjaras of Broach were sun-worshippers not Buddhists causes no difficulty since the BhilmÁl Gurjjaras whom Hiuen Tsiang visited in a.d.630 were Buddhists and since at Valabhi Buddhism Shaivism and sun-worship seem to have secured the equal patronage of the state.

Besides of GujarÁt and its king the traditions of both Java and Cambodia contain references to Hastinagara or Hastinapura, to Taxila, and to Rumadesa.6 With regard to these names and also with regard to GandhÁra
Appendix IV.
Java. and to Cambodia, all of which places are in the north-west of India, the question arises whether the occurrence of these names implies an historical connection with KÁbul PeshÁwar and the west PanjÁb or whether they are mere local applications and assumptions by foreign settlers and converts of names known in the BrÁhman and Buddhist writings of India.7 That elaborate applications of names mentioned in the MahÁbhÁrata to places in Java have been made in the Java version of the MahÁbhÁrata is shown by Raffles.8 Still it is to be noticed that the places mentioned above, Kamboja or KÁbul, GandhÁra or PeshÁwar, Taxila or the west PanjÁb, and Rumadesa apparently the south PanjÁb are not, like Ayodhya the capital of Siam or like Intha-patha-puri that is Indraprastha or Dehli the later capital of Cambodia,9 the names of places which either by their special fame or by their geographical position would naturally be chosen as their original home by settlers or converts in Java and Cambodia. Fair ground can therefore be claimed for the presumption that the leading position given to Kamboja, GandhÁra, Taxila, and Rumadesa in Javan and Cambodian legends and place-names is a trace of an actual and direct historical connection between the north-west of India and the Malay Archipelago. This presumption gains probability by the argument from the architectural remains of the three countries which in certain peculiar features show so marked a resemblance both in design and in detail as in the judgment of Mr. Fergusson to establish a strong and direct connection.10 A third argument in favour of a GujarÁt strain in Java are the traditions of settlements and expeditions by the rulers of MÁlwa which are still current in south MÁrwÁr.11 Further a proverb
Appendix IV.
Java. still well known both in MÁrwÁr and in GujarÁt runs:

Je jae JÁve te kadi nahi Áve

Áve to sÁth pidhi baithke khÁve.

Who to Java roam ne’er come home.

If they return, through seven lives

Seated at ease their wealth survives.12

Once more the connection with GujarÁt is supported by the detail in the Java account which makes Laut Mira the starting point for the colonising fleet. This Sir S. Raffles supposed to be the Red Sea but the Mihiras’ or Meds’ sea may be suggested as it seems to correspond to the somewhat doubtful Arab name Baharimad (sea of the Meds?) for a town in western India sacked by Junaid. Against this evidence two considerations have been urged13: (a) The great length of the voyage from GujarÁt to Java compared with the passage to Java from the east coast of India; (b) That no people in India have known enough of navigation to send a fleet fit to make a conquest. As regards the length of the voyage it is to be remembered that though Sumatra is more favourably placed for being colonised from Bengal Orissa and the mouths of the GodÁvari and K?ish?a, in the case either of Java or of Cambodia the distance from the Sindh and KÁthiÁvÁ?a ports is not much greater and the navigation is in some respects both safer and simpler than from the coasts of Orissa and Bengal. In reply to the second objection that no class of Hindus have shown sufficient skill and enterprise at sea to justify the belief that they could transport armies of settlers from GujarÁt to Java, the answer is that the assumption is erroneous. Though the bulk of Hindus have at all times been averse from a seafaring life yet there are notable exceptions. During the last two thousand years the record of the GujarÁt coast shows a genius for seafaring fit to ensure the successful planting of north-west India in the Malay Archipelago.14

Appendix IV.
Java. That the Hindu settlement of Sumatra was almost entirely from the
Appendix IV.
Java. east coast of India and that Bengal Orissa and Masulipatam had a large
Appendix IV.
Java. share in colonising both Java and Cambodia cannot be doubted.26
Appendix IV.
Java. Reasons have been given in support of the settlement in Java of large bodies of men from the north-west coasts of India and evidence has been offered to show that the objections taken to such a migration have little practical force. It remains to consider the time and the conditions of the GujarÁt conquest and settlement of Java and Cambodia. The Javan date S. 525 that is a.d.603 may be accepted as marking some central event in a process which continued for at least half a century before and after the beginning of the seventh century. Reasons have been given for holding that neither the commercial nor the political ascendancy of Rome makes it probable that to Rome the RÚm of the legends refers. The notable Roman element in the architecture of Java and Cambodia may suggest that the memory of great Roman builders kept for Rome a place in the local legends. But the Roman element seems not to have come direct into the buildings of Java or Cambodia; as at AmrÁvati at the K?ish?a mouth, the classic characteristics came by way of the PanjÁb (TÁhia) only, in the case of Java, not by the personal taste and study of a prince, but as an incident of conquest and settlement.27 Who then was the ruler of RÚm near Taxila, who led a great settlement of Hindus from the PanjÁb to Java. Names in appearance like Rome, occur in north-west India. None are of enough importance to explain the prince’s title.28 There remains the word raum or rum applied to salt land in the south PanjÁb, in MÁrwÁr, and in north Sindh.29 The great battle of KÁrur, about sixty miles south-east of MultÁn, in which apparently about a.d.530 Yasodharmman of MÁlwa defeated the famous White HÚ?a conqueror Mihirakula (a.d.500–550) is described as fought in the land of RÚm.30 This great White HÚ?a defeat is apparently the origin of the legend of the prince of RÚm who retired by sea to Java. At the time of the battle of KÁrur the south PanjÁb, together with the north of Sindh, was under the SÁharÁis of Aror in north Sindh, whose coins show them to have been not only White HÚ?as, but of the same JÁvla family which the great conquerors ToramÁ?a
Appendix IV.
Java. and Mihirakula adorned. So close a connection with Mihirakula makes it probable that the chief in charge of the north of the Aror dominions shared in the defeat and disgrace of KÁrur. Seeing that the power of the SÁharÁis of Aror spread as far south as the KÁthiÁvÁ?a ports of SomnÁth and Diu, and probably also of Diul at the Indus mouth, if the defeated chief of the south PanjÁb was unable or unwilling to remain as a vassal to his conqueror, no serious difficulty would stand in the way of his passage to the seaboard of Aror or of his finding in Diu and other Sindh and GujarÁt ports sufficient transport to convey him and his followers by sea to Java.31 This then may be the chief whom the Cambodian story names Phra Tong or Thom apparently Great Lord that is MahÁrÁja.32

The success of the Javan enterprise would tempt others to follow especially as during the latter half of the sixth and almost the whole of the seventh centuries, the state of North India favoured migration. Their defeats by Sassanians and Turks between a.d.550 and 600 would close to the White HÚ?as the way of retreat northwards by either the Indus or the KÁbul valleys. If hard pressed the alternative was a retreat to Kashmir or an advance south or east to the sea. When, in the early years of the seventh century (a.d.600–606), PrabhÁkaravardhana the father of SrÍ Harsha of Magadha (a.d.610–642) defeated the king of GandhÁra, the HÚ?as, the king of Sindh, the Gurjjaras, the LÁ?as, and the king of Malava,33 and when, about twenty years later, further defeats were inflicted by SrÍ Harsha himself numbers of refugees would gather to the GujarÁt ports eager to escape further attack and to share the prosperity of Java. It is worthy of note that the details of PrabhÁkaravardhana’s conquests explain how GandhÁra and LÁ?a are both mentioned in the Java legends; how northerners from the PanjÁb were able to pass to the coast; how the MÁrwÁr stories give the king of MÁlwa a share in the migrations; how the fleets may have started from any Sindh or GujarÁt port; and how with emigrants may have sailed artists and sculptors acquainted both with the monasteries and stupas of the KÁbul valley and PeshÁwar and with the carvings of the Ajanta caves. During the second half of the seventh century the advance of the Turks from the north and of the Arabs both by sea (a.d.637) and through Persia (a.d.650–660);34 the conquering progress of a Chinese army from Magadha to Bamian in a.d.645–65035; the overthrow (a.d.642) of
Appendix IV.
Java. the Buddhist SÁharÁis by their usurping BrÁhmanist minister Chach and his persecution of the Jats must have resulted in a fairly constant movement of northern Indians southwards from the ports of Sindh and GujarÁt.36 In the leading migrations though fear may have moved the followers, enterprise and tidings of Java’s prosperity would stir the leaders. The same longing that tempted Alexander to put to sea from the Indus mouth; Trajan (a.d.116) from the mouth of the Tigris; and MahmÚd of Ghazni from SomnÁth must have drawn Saka HÚ?a and Gurjjara chiefs to lead their men south to the land of rubies and of gold.37

Of the appearance and condition of the Hindus who settled in Java during the seventh and eighth centuries the Arab travellers SulaimÁn a.d.850 and MasÚdi a.d.915 have left the following details. The people near the volcanoes have white skins pierced ears and shaved heads: their religion is both BrÁhmanic and Buddhist; their trade is in the costliest articles camphor aloes cloves and sandalwood.38

CAMBODIA.

Cambodia.The close connection between Java and Cambodia, the alternate supremacy of Cambodia in Java and of Java in Cambodia, the likelihood of settlers passing from Java to Cambodia explain, to a considerable extent, why the traditions and the buildings of Java and Cambodia should point to a common origin in north-west India. The question remains: Do the people and buildings of Cambodia contain a distinct north Hindu element which worked its way south and east not by sea but by land across the HimÁlayas and Tibet and down the valley of the Yang-tse-kiang to Yunnan and Angkor. Whether the name Cambodia39 proves an actual race or historical connection with Kamboja or the KÁbul valley is a point
Appendix IV.
Cambodia. on which authorities disagree. Sir H. Yule held that the connection was purely literary and that as in the case of Inthapatha-puri or Indraprastha (Dehli) the later capital of Cambodia and of Ayodhya or Oudh the capital of Assam no connection existed beyond the application to a new settlement of ancient worshipful Indian place-names. The objection to applying this rule to Cambodia is that except to immigrants from the KÁbul valley the name is of too distant and also of too scanty a reputation to be chosen in preference to places in the nearer and holier lands of Tirhut and Magadha. For this reason, and because the view is supported by the notable connection between the two styles of architecture, it seems advisable to accept Mr. Fergusson’s decision that the name Cambodia was given to a portion of Cochin-China by immigrants from Kamboja that is from the KÁbul valley. Traces remain of more than one migration from India to Indo-China. The earliest is the mythic account of the conversion of Indo-China to Buddhism before the time of Asoka (b.c.240). A migration in the first century a.d. of Yavanas or Sakas, from Tamluk or RatnÁvate on the Hugli, is in agreement with the large number of Indian place-names recorded by Ptolemy (a.d.160).40 Of this migration Hiuen Tsiang’s name Yavana (Yen-mo-na) for Cambodia may be a trace.41 A Saka invasion further explains Pausanias’ (a.d.170) name SakÆa for Cochin-China and his description of the people as Skythians mixed with Indians.42 During the fifth and sixth centuries a fresh migration seems to have set in. Cambodia was divided into shore and inland and the name Cambose applied to both.43 Chinese records notice an embassy from the king of Cambodia in a.d.617.44 Among the deciphered Cambodian inscriptions a considerable share belong to a BrÁhmanic dynasty whose local initial date is in the early years of the seventh century,45 and one of whose kings Somasarmman (a.d.610) is recorded to have held daily MahÁbhÁrata readings in the temples.46 Of a fresh wave of Buddhists, who seem to have belonged to the northern branch, the earliest deciphered inscription is a.d.953 (S. 875) that is about 350 years later.47 Meanwhile, though, so far as information goes, the new capital of Angkor on the north bank of lake Tale Sap about 200 miles up the Mekong river was not founded till a.d.1078 (S. 1000),48 the neighbourhood of the holy lake was already sacred and the series of temples of which the Nakhonwat or NÁga’s Shrine49 is one of the latest and finest examples, was begun at least as early as a.d.825 (S. 750), and
Appendix IV.
Cambodia. Nakhonwat itself seems to have been completed and was being embellished in a.d.950 (S. 875).50 During the ninth and tenth centuries by conquest and otherwise considerable interchange took place between Java and Cambodia.51 As many of the inscriptions are written in two Indian characters a northern and a southern52 two migrations by sea seem to have taken place one from the Orissa and Masulipatam coasts and the other, with the same legend of the prince of RÚm land, from the ports of Sindh and GujarÁt.53 The question remains how far there is trace of such a distinct migration as would explain the close resemblance noted by Fergusson between the architecture of Kashmir and Cambodia as well as the northern element which Fergusson recognises in the religion and art of Cambodia.54 The people by whom this PanjÁb and Kashmir influence may have been introduced from the north are the people who still call themselves Khmers to whose skill as builders the magnificence of Cambodian temples lakes and bridges is apparently due.55 Of these people, who, by the beginning of the eleventh century had already given their name to the whole of Cambodia, Alberuni (a.d.1031) says: The Kumairs are whitish of short stature and Turk-like build. They follow the religion of the Hindus and have the practice of piercing their ears.56 It will be noticed that so far as information is available the apparent holiness of the neighbourhood of Angkor had lasted for at least 250 years before a.d.1078 when it was chosen as a capital. This point is in agreement with Mr. Fergusson’s view that the details of Nakhonwat and other temples of that series show that the builders came neither by sea nor down the Ganges valley but by way of Kashmir and the back of the HimÁlayas.57 Though the evidence is incomplete and to some extent speculative the following considerations suggest a route and a medium through which the Roman and Greek elements in the early (a.d.100–500) architecture of the KÁbul valley and PeshÁwar may have been carried inland to Cambodia. It may perhaps be accepted that the Ephthalites or White HÚ?as and a share of the Kedarites, that is of the later Little Yuechi from GandhÁra the PeshÁwar country, retreated to Kashmir before the father of SrÍ Harsha (a.d.590–606) and afterwards (a.d.606–642) before SrÍ Harsha himself.58 Further it seems fair to assume that from
Appendix IV.
Cambodia. Kashmir they moved into Tibet and were the western Turks by whose aid in the second half of the seventh century Srongbtsan or Srongdzan-gambo (a.d.640–698), the founder of Tibetan power and civilization, overran the Tarim valley and western China.59 During the first years of the eighth century (a.d.703) a revolt in Nepal and the country of the BrÁhmans was crushed by Srongdzan’s successor Donsrong,60 and the supremacy of Tibet was so firmly established in Bengal that, for over 200 years, the Bay of Bengal was known as the sea of Tibet.61 In a.d.709 a Chinese advance across the Pamirs is said to have been checked by the great Arab soldier Kotieba the comrade of Muhammad Kasim of Sindh.62 But according to Chinese records this reverse was wiped out in a.d.713 by the defeat of the joint Arab and Tibet armies.63 In the following years, aided by disorders in China, Tibet conquered east to Hosi on the upper Hoangho and in a.d.729 ceased to acknowledge the overlordship of China. Though about a.d.750 he was for a time crippled by China’s allies the Shado Turks the chief of Tibet spread his power so far down the Yangtsekiang valley that in a.d.787 the emperor of China, the king of Yunnan to the east of Burma, certain Indian chiefs, and the Arabs joined in a treaty against Tibet. As under the great Thisrong (a.d.803–845) and his successor Thi-tsong-ti (a.d.878–901) the power of Tibet increased it seems probable that during the ninth century they overran and settled in Yunnan.64 That among the Tibetans who passed south-east into Yunnan were Kedarites and White HÚ?as is supported by the fact that about a.d.1290, according both to Marco Polo and to Rashid-ud-din, the common name of Yunnan was KÁrÁjang whose capital was Yachi and whose people spoke a special language.65 The name KÁrÁjang was Mongol meaning Black People and was used to distinguish the mass of the inhabitants from certain fair tribes who were known as Chaganjang or Whites. That the ruler of KÁrÁjang was of Hindu origin is shown by his title Mahara or MahÁrÁja. That the Hindu element came from the KÁbul valley is shown by its Hindu name of KandhÁr that is GandhÁra or PeshÁwar, a name still in use as Gand­Álarit (GandhÁra-rashtra) the Burmese for Yunnan.66 The strange confusion which Rashid-ud-din makes between the surroundings of Yunnan and of PeshÁwar is perhaps due to the fact that in his time the connection between the two places was still known and admitted.67 A further trace
Appendix IV.
Cambodia. of stranger whites like the Chaganjang of Yunnan occurs south-east in the Anin or Honli whose name suggests the HÚ?as and whose fondness for silver ornaments at once distinguishes them from their neighbours and connects them with India.68 Even though these traces may be accepted as confirming a possible migration of HÚ?as and Kedaras to Yunnan and Anin a considerable gap remains between Anin and Angkor. Three local Cambodian considerations go some way to fill this gap. The first is that unlike the Siamese and Cochin Chinese the Khmers are a strong well made race with very little trace of the Mongoloid, with a language devoid of the intonations of other Indo-Chinese dialects, and with the hair worn cropped except the top-knot. The second point is that the Khmers claim a northern origin; and the third that important architectural remains similar to Nakhonwat are found within Siam limits about sixty miles north of Angkor.69 One further point has to be considered: How far is an origin from White HÚ?as and KedÁras in agreement with the NÁga phase of Cambodian worship. Hiuen Tsiang’s details of the Tarim Oxus and SwÁt valleys contain nothing so remarkable as the apparent increase of Dragon worship. In those countries dragons are rarely mentioned by Fa Hian in a.d.400: dragons seem to have had somewhat more importance in the eyes of Sung-Yun in a.d.520; and to Hiuen Tsiang, the champion of the MahÁyÁna or Broadway, dragons are everywhere explaining all misfortunes earthquakes storms and diseases. Buddhism may be the state religion but the secret of luck lies in pleasing the Dragon.70

Appendix IV.
Cambodia. This apparent increased importance of dragon or NÁga worship in north-west India during the fifth and sixth centuries may have been due partly to the decline of the earlier Buddhism partly to the genial wonder-loving temper of Hiuen Tsiang. Still so marked an increase makes it probable that with some of the great fifth and sixth century conquerors of Baktria KÁbul and the PanjÁb, of whom a trace may remain in the snake-worshipping
Appendix IV.
Cambodia. NÁgas and Takkas of the Kamaon and Garhwal hills, the Dragon was the chief object of worship. Temple remains show that the seventh and eighth century rulers of Kashmir, with a knowledge of classic architecture probably brought from beyond the Indus, were NÁga worshippers.72 The fact that the ninth century revision of religion in Tibet came mainly from Kashmir and that among the eighteen chief gods of the reformed faith the great Serpent had a place favours the view that through Tibet passed the scheme and the classic details of the Kashmir NÁga temples which in greater wealth and splendour are repeated in the Nakhonwat of Angkor in Cambodia.73 It is true that the dedication of the great temple to NÁga worship before the Siamese priests filled it with statues of Buddha is questioned both by Lieut. Garnier and by Sir H. Yule.74 In spite of this objection and though some of the series have been Buddhist from the first, it is difficult to refuse acceptance to Mr. Fergusson’s conclusions that in the great NÁkhon, all traces of Buddhism are additions. The local conditions and the worshipful Tale Sap lake favour this conclusion. What holier dragon site can be imagined than the great lake Tale Sap, 100 miles by 30, joined to the river Mekong by a huge natural channel which of itself empties the lake in the dry season and refills it during the rains giving a water harvest of fish as well as a land harvest of grain. What more typical work of the dragon as guardian water lord. Again not far off between Angkor and YunnÁn was the head-quarters of the dragon as the unsquared fiend. In Carrajan ten days west of the city of Yachi Marco Polo (a.d.1290) found a land of snakes and great serpents ten paces in length with very great heads, eyes bigger than a loaf of bread, mouths garnished with pointed teeth able to swallow a man whole, two fore-legs with claws for feet and bodies equal in bulk to a great cask. He adds: ‘These serpents devour the cubs of lions and bears without the sire and dam being able to prevent it. Indeed if they catch the big ones they devour them too: no one can make any resistance. Every man and beast stands in fear and trembling of them.’ Even in these fiend dragons was the sacramental guardian element. The gall from their inside healed the bite of a mad dog, delivered a woman in hard labour, and cured itch or it might be worse. Moreover, he concludes, the flesh of these serpents is excellent eating and toothsome.75

1 Sir Stamford Raffles’ Java, II. 83. From Java Hindus passed to near Banjar Massin in Borneo probably the most eastern of Hindu settlements (Jour. R. A. Soc. IV. 185). Temples of superior workmanship with Hindu figures also occur at Waahoo 400 miles from the coast. Dalton’s Diaks of Borneo Jour. Asiatique (N. S.) VII. 153. An instance may be quoted from the extreme west of Hindu influence. In 1873 an Indian architect was found building a palace at Gondar in Abyssinia. Keith Johnson’s Africa, 269.?

2 Raffles’ Java, II. 65–85. Compare Lassen’s Indische Alterthumskunde, II. 10, 40; IV. 460.?

3 Raffles’ Java, II. 87.?

4 Compare Tod’s Annals of RÁjasthÁn (Third Reprint), I. 87. The thirty-nine ChohÁn successions, working back from about a.d.1200 with an average reign of eighteen years, lead to a.d.498.?

5 Compare Note on BhinmÁl page 467.?

6 According to Cunningham (Ancient Geography, 43 and Beal’s Buddhist Records, I. 109 note 92) the site of Hastinagara or the eight cities is on the SwÁt river eighteen miles north of PeshÁwar. In Vedic and early MahÁbhÁrata times Hastinapura was the capital of GandhÁra (Hewitt Jour. Roy. As. Soc. XXI. 217). In the seventh century it was called PushkalÁvatÍ. (Beal’s Buddhist Records, I. 109.) Taxila, the capital of the country east of the Indus, was situated about forty miles east of Attok at ShÁhderi near KÁlaka-sarai (Cunningham’s Ancient Geography, 105). According to Cunningham (Ditto 109), Taxila continued a great city from the time of Alexander till the fifth century after Christ. It was then laid waste apparently by the great White HÚ?a conqueror Mihirakula (a.d.500–550). A hundred years later when Hiuen Tsiang visited it the country was under Kashmir, the royal family were extinct, and the nobles were struggling for power (Beal’s Buddhist Records, I. 136). Rumadesa. References to Rumadesa occur in the traditions of Siam and Cambodia as well as in those of Java. Fleets of RÚm are also noted in the traditions of Bengal and Orissa as attacking the coast (Fergusson’s Architecture, III. 640). Coupling the mention of RÚm with the tradition that the Cambodian temples were the work of Alexander the Great Colonel Yule (Ency. Brit. Article Cambodia) takes RÚm in its MusalmÁn sense of Greece or Asia Minor. The variety of references suggested to Fergusson (Architecture, III. 640) that these exploits are a vague memory of Roman commerce in the Bay of Bengal. But the Roman rule was that no fleet should pass east of Ceylon (Reinaud Jour. As. Ser. VI. Tom. I. page 322). This rule may occasionally have been departed from as in a.d.166 when the emperor Marcus Aurelius sent an ambassador by sea to China. Still it seems unlikely that Roman commerce in the Bay of Bengal was ever active enough to gain a place as settler and coloniser in the traditions of Java and Cambodia. It was with the west not with the east of India that the relations of Rome were close and important. From the time of Mark Antony to the time of Justinian, that is from about b.c.30 to a.d.550, their political importance as allies against the Parthians and Sassanians and their commercial importance as controllers of one of the main trade routes between the east and the west made the friendship of the KushÁns or Sakas who held the Indus valley and Baktria a matter of the highest importance to Rome. How close was the friendship is shown in a.d.60 by the Roman General Corbulo escorting the Hyrkanian ambassadors up the Indus and through the territories of the KushÁns or Indo-Skythians on their return from their embassy to Rome. (Compare Rawlinson’s Parthia, 271.) The close connection is shown by the accurate details of the Indus valley and Baktria recorded by Ptolemy (a.d.166) and about a hundred years later (a.d.247) by the author of the Periplus and by the special value of the gifts which the Periplus notices were set apart for the rulers of Sindh. One result of this long continued alliance was the gaining by the KushÁn and other rulers of PeshÁwar and the PanjÁb of a knowledge of Roman coinage astronomy and architecture. Certain AfghÁn or Baktrian coins bear the word Roma apparently the name of some AfghÁn city. In spite of this there seems no reason to suppose that Rome attempted to overlord the north-west of India still less that any local ruler was permitted to make use of the great name of Rome. It seems possible that certain notices of the fleets of RÚm in the Bay of Bengal refer to the fleets of the Arab Al-Rami that is Lambri or north-west Sumatra apparently the Romania of the Chaldean breviary of the MalabÁr Coast. (Yule’s Cathay, I. lxxxix. note and Marco Polo, II. 243.)?

7 Compare Fergusson’s Architecture, III. 640; Yule in Ency. Brit. Cambodia.?

8 Java, I. 411. Compare Fergusson’s Architecture, III. 640.?

9 See Yule in Jour. Roy. As. Soc. (N. S.), I. 356; Fergusson’s Architecture, III. 631.?

10 Of the Java remains Mr. Fergusson writes (Architecture, III. 644–648): The style and character of the sculptures of the great temple of Boro Buddor are nearly identical with those of the later caves of Ajanta, on the Western GhÁts, and in SÁlsette. The resemblance in style is almost equally close with the buildings of Takht-i-Bahi in GandhÁra (Ditto, 647). Again (page 637) he says: The Hindu immigrants into Java came from the west coast of India. They came from the valley of the Indus not from the valley of the Ganges. Once more, in describing No. XXVI. of the Ajanta caves Messrs. Fergusson and Burgess (Rock-cut Temples, 345 note 1) write: The execution of these figures is so nearly the same as in the Boro Buddor temple in Java that both must have been the work of the same artists during the latter half of the seventh century or somewhat later. The Buddhists were not in Java in the fifth century. They must have begun to go soon after since there is a considerable local element in the Boro Buddor.?

11 Traditions of expeditions by sea to Java remain in MÁrwÁr. In April 1895 a bard at BhinmÁl related how BhojrÁja of Ujjain in anger with his son Chandrabau drove him away. The son went to a GujarÁt or KÁthiÁvÁ?a port obtained ships and sailed to Java. He took with him as his BrÁhman the son of a Magh Pandit. A second tale tells how Vikram the redresser of evils in a dream saw a Javanese woman weeping, because by an enemy’s curse her son had been turned into stone. Vikram sailed to Java found the woman and removed the curse. According to a third legend ChandrawÁn the grandson of Vir PramÁr saw a beautiful woman in a dream. He travelled everywhere in search of her. At last a Rishi told him the girl lived in Java. He started by sea and after many dangers and wonders found the dream-girl in Java. The people of BhinmÁl are familiar with the GujarÁti proverb referred to below; Who goes to Java comes not back. MS. Notes, March 1895.?

12 Another version is:

?

13 Compare Crawford (a.d.1820) in As. Res. XIII. 157 and Lassen Ind. Alt. II. 1046.?

14 The following details summarise the available evidence of GujarÁt Hindu enterprise by sea. According to the Greek writers, though it is difficult to accept their statements as free from exaggeration, when, in b.c.325, Alexander passed down the Indus the river showed no trace of any trade by sea. If at that time sea trade at the mouth of the Indus was so scanty as to escape notice it seems fair to suppose that Alexander’s ship-building and fleet gave a start to deep-sea sailing which the constant succession of strong and vigorous northern tribes which entered and ruled Western India during the centuries before and after the Christian era continued to develope.15 According to Vincent (Periplus, I. 25, 35, 254) in the time of Agatharcides (b.c.200) the ports of Arabia and Ceylon were entirely in the hands of the people of GujarÁt. During the second century after Christ, when, under the great RudradÁman (a.d.143–158), the Sinh or Kshatrapa dynasty of KÁthiÁvÁ?a was at the height of its power, Indians of TientÇo, that is Sindhu, brought presents by sea to China (Journal Royal Asiatic Society for January 1896 page 9). In a.d.166 (perhaps the same as the preceding) the Roman emperor Marcus Aurelius sent by sea to China ambassadors with ivory rhinoceros’ horn and other articles apparently the produce of Western India (DeGuignes’ Huns, I. [Part I.] 32). In the third century a.d.247 the Periplus (McCrindle, 17, 52, 64, 96, 109) notices large Hindu ships in the east African Arab and Persian ports and Hindu settlements on the north coast of Sokotra. About a century later occurs the doubtful reference (Wilford in Asiatic Researches, IX. 224) to the Diveni or pirates of Diu who had to send hostages to Constantine the Great (a.d.320–340) one of whom was Theophilus afterwards a Christian bishop. Though it seems probable that the Kshatrapas (a.d.70–400) ruled by sea as well as by land fresh seafaring energy seems to have marked the arrival on the Sindh and KÁthiÁvÁ? coasts of the Juan-Juan or Avars (a.d.390–450) and of the White HÚ?as (a.d.450–550). During the fifth and sixth centuries the ports of Sindh and GujarÁt appear among the chief centres of naval enterprise in the east. How the sea ruled the religion of the newcomers is shown by the fame which gathered round the new or revised gods Siva the Poseidon of SomnÁth and K?ish?a the Apollo or St. Nicholas of DwÁrka. (Compare Tod’s Annals of RÁjasthÁn, I. 525.) In the fifth century (Yule’s Cathay, I. lxxviii.) according to Hamza of IspahÁn, at Hira near Kufa on the Euphrates the ships of India and China were constantly moored. In the early sixth century (a.d.518–519) a Persian ambassador went by sea to China (Ditto, I. lxxiv.) About the same time (a.d.526) Cosmas (Ditto, I. clxxviii.) describes Sindhu or Debal and Orhota that is Soratha or VerÁval as leading places of trade with Ceylon. In the sixth century, apparently driven out by the White HÚ?as and the Mihiras, the Jats from the Indus and Kachh occupied the islands in the Bahrein gulf, and perhaps manned the fleet with which about a.d.570 NaushiravÁn the great Sassanian (a.d.531–574) is said to have invaded the lower Indus and perhaps Ceylon.16 About the same time (Fergusson Architecture, III. 612) AmrÁvati at the K?ish?a mouth was superseded as the port for the Golden Chersonese by the direct voyage from GujarÁt and the west coast of India. In a.d.630 Hiuen Tsiang (Beal’s Buddhist Records, II. 269) describes the people of SurÁsh?ra as deriving their livelihood from the sea, engaging in commerce, and exchanging commodities. He further notices that in the chief cities of Persia Hindus were settled enjoying the full practice of their religion (Reinaud’s Abulfeda, ccclxxxv.) That the Jat not the Arab was the moving spirit in the early (a.d.637–770) Muhammadan sea raids against the GujarÁt and Konkan coasts is made probable by the fact that these seafaring ventures began not in Arabia but in the Jat-settled shores of the Persian Gulf, that for more than fifty years the Arab heads of the state forbad them, and that in the Mediterranean where they had no Jat element the Arab was powerless at sea. (Compare Elliot, I. 416, 417.) That during the seventh and eighth centuries when the chief migrations by sea from GujarÁt to Java and Cambodia seem to have taken place, Chinese fleets visited Diu (Yule’s Cathay, lxxix.), and that in a.d.759 Arabs and Persians besieged Canton and pillaged the storehouses going and returning by sea (DeGuignes’ Huns, I. [Pt. II.] 503) suggest that the Jats were pilots as well as pirates.17 On the Sindh Kachh and GujarÁt coasts besides the Jats several of the new-come northern tribes showed notable energy at sea. It is to be remembered that as detailed in the Statistical Account of ThÁna (Bombay Gazetteer, XIII. Part II. 433) this remarkable outburst of sea enterprise may have been due not only to the vigour of the new-come northerners but to the fact that some of them, perhaps the famous iron-working Turks (a.d.580–680), brought with them the knowledge of the magnet, and that the local BrÁhman, with religious skill and secrecy, shaped the bar into a divine fish-machine or machiyantra, which, floating in a basin of oil, he consulted in some private quarter of the ship and when the stars were hid guided the pilot in what direction to steer. Among new seafaring classes were, on the MakrÁn and Sindh coasts the Bodhas Kerks and Meds and along the shores of Kachh and KÁthiÁvÁ?a the closely connected Meds and Gurjjaras. In the seventh and eighth centuries the Gurjjaras, chiefly of the ChÁpa or ChÁva?Á clan, both in DwÁrka and SomnÁth and also inland, rose to power, a change which, as already noticed, may explain the efforts of the Jats to settle along the Persian Gulf and the Red Sea. About a.d.740 the ChÁpas or ChÁva?Ás, who had for a century and a half been in command in DwÁrka and SomnÁth, established themselves at A?ahilavÁ?a Pattan. According to their tradition king VanarÁja (a.d.720–780) and his successor YogarÁja (a.d.806–841) made great efforts to put down piracy. YogarÁja’s sons plundered some Bengal or Bot ships which stress of weather forced into VerÁval. The king said ‘My sons with labour we were raising ourselves to be ChÁva?Ás of princely rank; your greed throws us back on our old nickname of Choras or thieves.’ YogarÁja refused to be comforted and mounted the funeral pyre. Dr. BhagvÁnlÁl’s History, 154. This tale seems to be a parable. YogarÁja’s efforts to put down piracy seem to have driven large bodies of Jats from the GujarÁt coasts. In a.d.834–35, according to Ibn Alathyr (a.d.834), a fleet manned by Djaths or Jats made a descent on the Tigris. The whole strength of the KhilÁfat had to be set in motion to stop them. Those who fell into the hands of the Moslems were sent to Anararbe on the borders of the Greek empire (Renaud’s Fragments, 201–2). As in the legend, the ChÁva?Á king’s sons, that is the Chauras Mers and Gurjjaras, proved not less dangerous pirates than the Jats whom they had driven out.18 About fifty years later, in a.d.892, Al-BilÁduri describes as pirates who scoured the seas the Meds and the people of SaurÁsh?ra that is Devpatan or SomnÁth who were Choras or Gurjjaras.19 BilÁduri (Reinaud Sur L’Inde, 169) further notices that the Jats and other Indians had formed the same type of settlement in Persia which the Persians and Arabs had formed in India. During the ninth and tenth centuries the GujarÁt kingdom which had been established in Java was at the height of its power. (Ditto, Abulfeda, ccclxxxviii.) Early in the tenth century (a.d.915–930) Masudi (Yule’s Marco Polo, II. 344; Elliot, I. 65) describes Sokotra as a noted haunt of the Indian corsairs called BawÁrij which chase Arab ships bound for India and China. The merchant fleets of the early tenth century were not Arab alone. The Chauras of A?ahilavÁ?a sent fleets to Bhot and Chin (RÁs MÁla, I. 11). Nor were Mers and Chauras the only pirates. Towards the end of the tenth century (a.d.980) GrahÁri the ChÚ?ÁsamÁ, known in story as Graharipu the Ahir of Sorath and GirnÁr, so passed and repassed the ocean that no one was safe (Ditto, I. 11). In the eleventh century (a.d.1021) Alberuni (Sachau, II. 104) notes that the BawÁrij, who take their name from their boats called behra or bira, were Meds a seafaring people of Kachh and of SomnÁth a great place of call for merchants trading between Sofala in east Africa and China. About the same time (a.d.1025) when they despaired of withstanding MÁhmud of Ghazni the defenders of SomnÁth prepared to escape by sea,20 and after his victory MÁhmud is said to have planned an expedition by sea to conquer Ceylon (Tod’s RajasthÁn, I. 108). In the twelfth century IdrÍsi (a.d.1135) notices that Tatariya dirhams, that is the Gupta (a.d.319–500) and White HÚ?a (a.d.500–580) coinage of Sindh and GujarÁt, were in use both in Madagascar and in the Malaya islands (Reinaud’s MÉmoires, 236), and that the merchants of Java could understand the people of Madagascar (Ditto, Abulfeda, cdxxii).21 With the decline of the power of A?ahilavÁ?a (a.d.1250–1300) its fleet ceased to keep order at sea. In a.d.1290 Marco Polo (Yule’s Ed. II. 325, 328, 341) found the people of GujarÁt the most desperate pirates in existence. More than a hundred corsair vessels went forth every year taking their wives and children with them and staying out the whole summer. They joined in fleets of twenty to thirty and made a sea cordon five or six miles apart. Sokotra was infested by multitudes of Hindu pirates who encamped there and put up their plunder to sale. Ibn Batuta (in Elliot, I. 344–345) fifty years later makes the same complaint. MusalmÁn ascendancy had driven RÁjput chiefs to the coast and turned them into pirates. The most notable addition was the Gohils who under MokherÁji Gohil, from his castle on Piram island, ruled the sea till his power was broken by Muhammad Tughlak in a.d.1345 (RÁs MÁla, I. 318). Before their overthrow by the Muhammadans what large vessels the RÁjput sailors of GujarÁt managed is shown by Friar Oderic, who about a.d.1321 (Stevenson in Kerr’s Voyages, XVIII. 324) crossed the Indian ocean in a ship that carried 700 people. How far the RÁjputs went is shown by the mention in a.d.1270 (Yule’s Cathay, 57 in Howorth’s Mongols, I. 247) of ships sailing between Sumena or SomnÁth and China. Till the arrival of the Portuguese (a.d.1500–1508) the AhmedÁbÁd SultÁns maintained their position as lords of the sea.22 In the fifteenth century Java appears in the state list of foreign bandars which paid tribute (Bird’s GujarÁt, 131), the tribute probably being a cess or ship tax paid by GujarÁt traders with Java in return for the protection of the royal navy.23 In east Africa, in a.d.1498 (J. As. Soc. of Bengal, V. 784) Vasco da Gama found sailors from Cambay and other parts of India who guided themselves by the help of the stars in the north and south and had nautical instruments of their own. In a.d.1510 Albuquerque found a strong Hindu element in Java and Malacca. Sumatra was ruled by Parameshwara a Hindu whose son by a Chinese mother was called RÁjput (Commentaries, II. 63; III. 73–79). After the rule of the sea had passed to the European, GujarÁt Hindus continued to show marked courage and skill as merchants seamen and pirates. In the seventeenth century the French traveller Mandelslo (a.d.1638, Travels 101, 108) found Achin in north Sumatra a great centre of trade with GujarÁt. During the seventeenth and eighteenth centuries the Sanganians or Sangar RÁjputs of MÁndvi in Kachh and of NavÁnagar in north KÁthiÁvÁ?a were much dreaded. In a.d.1750 Grose describes the small cruisers of the Sanganians troubling boats going to the Persian Gulf, though they seldom attacked large ships. Between a.d.1803 and 1808 (Low’s Indian Navy, I. 274) pirates from Bet established themselves in the ruined temple at SomnÁth. In 1820, when the English took Bet and DwÁrka from the WÁghels, among the pirates besides WÁghels were Badhels a branch of RÁhtors, Bhattis, KhÁrwÁs, LohÁnÁs, MakwÁnÁs, RÁhtors, and Wagharis. A trace of the Chauras remained in the neighbouring chief of Aramra.24 Nor had the old love of seafaring deserted the KÁthiÁvÁ?a chiefs. In the beginning of the present century (a.d.1825) Tod (Western India, 452; compare RÁs MÁla, I. 245) tells how with Biji Singh of BhÁvnagar his port was his grand hobby and shipbuilding his chief interest and pleasure; also how RÁo Ghor of Kachh (a.d.1760–1778) built equipped and manned a ship at MÁndvi which without European or other outside assistance safely made the voyage to England and back to the MalabÁr Coast where arriving during the south-west monsoon the vessel seems to have been wrecked.25?

15 Alexander built his own boats on the Indus. (McCrindle’s Alexander, 77.) He carried (pages 93 and 131) these boats to the Hydaspes: on the Jhelum (134 note 1) where he found some country boats he built a flotilla of gallies with thirty oars: he made dockyards (pages 156–157): his crews were Phoenikians, Cyprians, Karians, and Egyptians.?

16 Reinaud’s MÉmoire Sur L’Inde, 125. The statement that NaushiravÁn received KarÁchi from the king of Seringdip (Elliot’s History, I. 407: Tabari, II. 221) throws doubt on this expedition to Ceylon. At the close of the sixth century KarÁchi or Diul Sindhi cannot have been in the gift of the king of Ceylon. It was in the possession of the SÁharÁi kings of Aror in Upper Sindh perhaps of ShÁhi Tegin Devaja shortened to Shahindev. (Compare Cunningham Oriental Congress, I. 242.) According to Garrez (J. As. Ser. VI. Tom. XIII. 182 note 2) this Serendip is Surandeb that is Syria and Antioch places which NaushiravÁn is known to have taken. Several other references that seem to imply a close connection between GujarÁt and Ceylon are equally doubtful. In the MahÁbhÁrata (a.d.100–300?) the Sinhalas bring vaidÚryas (rubies?) elephants’ housings and heaps of pearls. The meaning of Sainhalaka in Samudragupta’s inscription (a.d.395) Early GujarÁt History page 64 and note 5 is uncertain. Neither Mihirakula’s (a.d.530) nor LalitÁditya’s (a.d.700) conquest of Ceylon can be historical. In a.d.1005 when Abul Fatha the Carmatian ruler of MultÁn was attacked by MÁhmud of Ghazni he retired to Ceylon. (Reinaud’s MÉmoire, 225). When SomnÁth was taken (a.d.1025) the people embarked for Ceylon (Ditto, 270).?

17 Compare at a later period (a.d.1342) Ibn Batuta’s great ship sailing from KandahÁr (GandhÁr north of Broach) to China with its guard of Abyssinians as a defence against pirates. Reinaud’s Abulfeda, cdxxv.?

18 As an example of the readiness with which an inland race of northerners conquer seamanship compare the Franks of the Pontus who about a.d.279 passed in a few years from the Pontus to the Mediterranean ports and leaving behind them Malta the limit of Greek voyages sailed through Gibraltar to the Baltic. Gibbon, I. 404–405.?

19 Reinaud’s MÉmoire Sur L’Inde, 200. The traders of ChorwÁr, that is of the old Chaura or ChÁpa country near VirÁval and Mangrul, are now known in Bombay as ChÁpadias. The received explanation of ChÁpadia is the roofed men it is said in derisive allusion to their large and heavy headdress. But as the Porbandar headdress is neither specially large nor ungraceful the common explanation can be hardly more than a pun. This suggests that the name ChÁpadia is a trace of the early ChÁpa tribe of Gurjjaras who also gave their name to ChÁpanir. Tod’s (Western India, 250, 256) description of the Chauras race with traditions of having come from the Red Sea and as a nautical Arabia is the result of taking for Sokotra SankodwÁra that is Bet to the north of DwÁrka.?

20 According to Abulfeda a.d.1334 (Reinaud’s Abulfeda, cccxlix.) some of the besieged fled to Ceylon. Farishtah (Briggs’ Muhammadan Powers, I. 75) records that after the fall of SomnÁth MÁhmud intended to fit out a fleet to conquer Ceylon and Pegu. According to Bird (MirÁt-i-Ahmedi, 146) Ceylon or Sirandip remained a dependency of SomnÁth till a.d.1290 when the king VijayabÁhu became independent.?

21 The common element in the two languages may have been the result of GujarÁt settlements in Madagascar as well as in Java and Cambodia. This is however doubtful as the common element may be either Arabic or Polynesian.?

22 When in a.d.1535 he secured BahÁdur’s splendid jewelled belt HumÁyÚn said These are the trappings of the lord of the sea. Bayley’s GujarÁt, 386.?

23 Compare in Bombay Public Diary 10, pages 197–207 of 1736–37, the revenue headings Surat and Cambay with entries of two per cent on all goods imported and exported from either of these places by traders under the Honourable Company’s protection.?

24 These Badhels seem to be Hamilton’s (a.d.1720) Warels of Chance (New Account, I. 141). This Chance is ChÁch near Diu apparently the place from which the BhÁtiÁs get their Bombay name of ChÁchiÁs. Towards the close of the eighteenth century BhÁtiÁs from ChÁch seem to have formed a pirate settlement near DÁhÁnu on the ThÁna coast. Major Price (Memoirs of a Field Officer, 322) notes (a.d.1792 June) the cautionary speed with which in travelling from Surat to Bombay by land they passed DÁhÁnu through the ChÁnsiÁh jungle the district of a piratical community of that name.?

25 According to Sir. A. Burnes (Jl. Bombay Geog. Soc. VI. (1835) 27, 28) the special skill of the people of Kachh in navigation and ship-building was due to a young RÁjput of Kachh. RÁmsingh MÁlani, who about a century earlier had gone to Holland and learned those arts. See Bombay Gazetteer, V. 116 note 2.?

26 Crawford (a.d.1820) held that all Hindu influence in Java came from Kalinga or north-east Madras. Fergusson (Ind. Arch. 103, Ed. 1876) says: The splendid remains at AmrÁvati show that from the mouths of the K?ish?a and GodÁvari the Buddhist of north and north-west India colonised Pegu, Cambodia, and eventually the Island of Java. Compare Tavernier (a.d.1666: Ball’s Translation, I. 174.) Masulipatam is the only place in the Bay of Bengal from which vessels sail eastwards for Bengal, Arrakan, Pegu, Siam, Sumatra, Cochin China, and the Manillas and west to Hormuz, Makha, and Madagascar. Inscriptions (Indian Antiquary, V. 314; VI. 356) bear out the correctness of the connection between the Kalinga coast and Java which Java legends have preserved. As explained in Dr. Bhandarkar’s interesting article on the eastern passage of the Sakas (Jour. B. B. R. A. S. XVII.) certain inscriptions also show a Magadhi element which may have reached Java from Sumatra and Sumatra from the coast either of Bengal or of Orissa. Later information tends to increase the east and south Indian share. Compare Notices et Extraits des Manuscripts de la Bibliotheque Nationale Vol. XXVII. (Partie II) 2 Fasicule page 350.?

27 Compare Hiuen Tsiang in Beal’s Buddhist Records, II. 222 note 102. TÁhia may be Tochara that is Baktria, but the PanjÁb seems more likely. Compare Beal’s Life of Hiuen Tsiang, 136 note 2.?

28 IdrÍsi a.d.1135 (Elliot, I. 92) has a Romala a middling town on the borders of the desert between MultÁn and SeistÁn. Cunningham (Ancient Geog. 252) has a Romaka Bazaar near where the NÁra the old Indus enters the Ran of Kachh.?

29 Cunningham’s Num. Chron. 3rd Ser. VIII. 241. The MahÁbhÁrata Romakas (Wilson’s Works, VII. 176: Cunningham’s Anc. Geog. 187) may have taken their name from one of these salt stretches. Ibn KhurdÁdbah (a.d.912) mentions RumÁla (Elliot, I. 14, 87, 92, 93) as one of the countries of Sindh. In connection with the town Romala Al IdrÍsi a.d.1153 (Elliot, I. 74, 93) has a district three days’ journey from Kalbata.?

30 Cunningham’s Numismatic Chronicle 3rd Ser. VIII. 236. The date of KÁrur is uncertain. Fergusson (Arch. III. 746) puts it at a.d.544. It was apparently earlier as in an inscription of a.d.532 Yasodharmman king of MÁlwa claims to hold lands which were never held by either Guptas or HÚ?as. Cunningham Num. Chron. 3rd Ser. VIII. 236. Compare History Text, 76, 77.?

31 Jour. As. Soc. Bl. VII. (Plate I.) 298; Burnes’ BokhÁra, III. 76; Elliot’s History, I. 405. Diu which is specially mentioned as a SÁharÁi port was during the seventh and eighth centuries a place of call for China ships. Yule’s Cathay, I. lxxix.?

32 Phra like the PanjÁb Porus of the embassy to Augustus in b.c.30 (though this Porus may be so called merely because he ruled the lands of Alexander’s Porus) may seem to be the favourite Parthian name Phraates. But no instance of the name Phraates is noted among White HÚ?a chiefs and the use of Phra as in Phra Bot or Lord Buddha seems ground for holding that the Phra Thong of the Cambodia legend means Great Lord.?

33 Epigraphia Indica, I. 67.?

34 In a.d.637 raiders attacked ThÁna from Oman and Broach and Sindh from Bahrein. Reinaud’s MÉmoire Sur L’Inde, 170, 176.?

35 The passage of a Chinese army from Magadha to the GandhÁra river about a.d.650 seems beyond question. The emperor sent an ambassador Ouang-h-wuentse to SrÍ Harsha. Before Ouang-h-wuentse arrived SrÍ Harsha was dead (died a.d.642), and his place taken by an usurping minister (Se-na-fu-ti) Alana-chun. The usurper drove off the envoy, who retired to Tibet then under the great Songbtsan. With help from Tibet and from the RÁja of NepÁl Ouang returned, defeated Alana, and pursued him to the GandhÁra river (Khien-to-wei). The passage was forced, the army captured, the king queen and king’s sons were led prisoners to China, and 580 cities surrendered, the magistrates proclaimed the victory in the temple of the ancients and the emperor raised Ouang to the rank of Tch’ao-sau-ta-fore. Journal Asiatique Ser. IV. Tom. X. pages 81–121. The translator thinks the whole war was in the east of India and that the mention of the GandhÁra river is a mistake. The correctness of this view is doubtful. It is to be remembered that this was a time of the widest spread of Chinese power. They held Balk and probably Bamian. Yule’s Cathay, I. lxviii. Compare Julien in Jour. As. Soc. Ser. IV. Tom. X. 289–291.?

36 Regarding these disturbances see Beal’s Life of Hiuen Tsiang, 155; Max MÜller’s India, 286. The Arab writers (a.d.713) notice to what a degraded state Chach had reduced the Jats. In comparing the relative importance of the western and eastern Indian strains in Java it is to be remembered that the western element has been overlaid by a late Bengal and Kalinga layer of fugitives from the Tibetan conquest of Bengal in the eighth century, the Babu with the Gurkha at his heels, and during the ninth and later centuries by bands of Buddhists withdrawing from a land where their religion was no longer honoured.?

37 In a.d.116 after the capture of Babylon and Ctesiphon Hadrian sailed down the Tigris and the Persian Gulf, embarked on the waters of the South Sea, made inquiries about India and regretted he was too old to get there. Rawlinson’s Ancient Monarchies, VI. 313.?

38 Reinaud’s Abulfeda, cccxc.?

39 The origin of the name KÁmboja seems to be KÁmbojÁpura an old name of KÁbul preserved almost in its present form in Ptolemy’s (a.d.160) Kaboura. The word is doubtfully connected with the AchÆmenian Kambyses (b.c.529–521) the Kambujiya of the Behistun inscription. In the fifth of the Asoka edicts (b.c.240) KÁmboja holds the middle distance between GandhÁra or PeshÁwar and Yona or Baktria. According to YÁska, whose uncertain date varies from b.c.500 to b.c.200, the Kambojas spoke Sanskrit (Muir’s Sanskrit Texts, II. 355 note 145). In the last battle of the MahÁbhÁrata, a.d.100 to 300 (Jl. Roy. As. Soc. [1842] VII. 139–140), apparently from near Bamian the Kambojas ranked as Mlechchhas with Sakas Daradas and HÚ?as. One account (Fergusson, III. 665) places the original site of the Kambojas in the country round Taxila east of the Indus. This is probably incorrect. A trace of the Kambojas in their original seat seems to remain in the Kaumojas of the Hindu Kush.?

40 See Hunter’s Orissa, I. 310.?

41 Yavana to the south-west of Siam. Beal’s Life of Hiuen Tsiang, xxxii.?

42 Quoted in Bunbury’s Ancient Geography, II. 659. Bunbury suggests that Pausanias may have gained his information from Marcus Aurelius’ (a.d.166) ambassador to China.?

43 Jour. Bengal Soc. VII. (I.) 317.?

44 Remusat Nouveaux Melanges Asiatiques, I. 77 in Jour. Asiatique Series, VI. Tom. XIX. page 199 note 1; Fergusson’s Architecture, III. 678.?

45 Barth in Journal Asiatique Ser. VI. Tom. XIX. page 150.?

46 Barth in Journal Asiatique, X. 57.?

47 Barth in Jour. As. Ser. VI. Tom. XIX. page 190; Journal Royal Asiatic Society, XIV. (1882) cii.?

48 Barth in Journal Asiatique Ser. VI. Tom. XIX. pages 181, 186.?

49 Mr. Fergusson (Architecture page 666) and Colonel Yule (Ency. Brit. Cambodia) accept the local Buddhist rendering of Nakhonwat as the City Settlement. Against this it is to be noted (Ditto ditto) that nagara city corrupts locally into Angkor. Nagara therefore can hardly also be the origin of the local Nakhon. Farther as the local Buddhists claim the temple for Buddha they were bound to find in Nakhon some source other than its original meaning of Snake. The change finds a close parallel in the NÁga that is snake or Skythian now NÁgara or city BrÁhman of GujarÁt.?

50 Barth in Journal Asiatique Ser. VI. Tom. XIX. 190.?

51 Yule’s Marco Polo, II. 108; Reinaud’s Abulfeda, cdxvi.?

52 Barth in Journal Asiatique Ser. VI. Tom. XIX. 174.?

53 Mr. Fergusson at first suggested the fourth century as the period of migration to Cambodia. He afterwards came to the conclusion that the settlers must have been much the same as the GujarÁt conquerors of Java. Architecture, III. 665–678.?

54 Fergusson, Architecture, 665. Compare Tree and Serpent Worship, 49, 50. The people of Cambodia seem Indian serpent worshippers: they seem to have come from Taxila.?

55 The name Khmer has been adopted as the technical term for the early literature and arts of the peninsula. Compare Barth J. As. Ser. VI. Tom. XIX. 193; Renan in ditto page 75 note 3 and Ser. VII. Tom. VIII. page 68; Yule in EncyclopÆdia Britannica Art. Cambodia. The resemblance of Cambodian and KÁbul valley work recalls the praise by Chinese writers of the Han (b.c.206–a.d.24) and Wei (a.d.386–556) dynasties of the craftsmen of Kipin, that is Kophene or Kamboja the KÁbul valley, whose skill was not less remarkable in sculpturing and chiselling stone than in working gold silver copper and tin into vases and other articles. Specht in Journal Asiatique, II. (1883), 333 and note 3. A ninth century inscription mentions the architect Achyuta son of RÁma of KÁmboja. Epigraphia Indica, I. 243.?

56 Reinaud’s Abulfeda, cdxxi.; Sachau’s Alberuni, I. 210.?

57 Fergusson’s Architecture, III. 666.?

58 For the joint Kedarite-Ephthalite rule in Kashmir see Cunningham’s Ninth Oriental Congress, I. 231–2. The sameness of names, if not an identity of rulers, shows how close was the union between the Ephthalites and the Kedarites. The coins preserve one difference depicting the Yuechi or Kedarite ruler with bushy and the White HÚ?a or Ephthalite ruler with cropped hair.?

59 About a.d.700 Urumtsi Kashgar Khoten and Kuche in the Tarim valley became Tibetan for a few years. Parker’s Thousand Years of the Tartars, 243. In a.d.691 the western Turks who for some years had been declining and divided were broken by the great eastern Turk conqueror Mercho. The following passage from MasÚdi (Prairies D’Or, I. 289) supports the establishment of White HÚ?a or Mihira power in Tibet. The sons of AmÚr (a general phrase for Turks) mixed with the people of India. They founded a kingdom in Tibet the capital of which they called Med.?

60 EncyclopÆdia Britannica Articles Tibet and Turkestan.?

61 Both Ibn Haukal and Al Istakhri (a.d.950) call the Bay of Bengal the sea of Tibet. Compare Reinaud’s Abulfeda, ccclviii.; EncyclopÆdia Britannica Article Tibet page 345.?

62 Yule’s Cathay, I. lxxxi.?

63 Ency. Brit. China, 646.?

64 Thisrong besides spreading the power of Tibet (he was important enough to join with MÁmÚn the son of the great Harun-ar-Rashid (a.d.788–809) in a league against the Hindus) brought many learned Hindus into Tibet, had Sanskrit books translated, settled Lamaism, and built many temples. It is remarkable that (so far as inscriptions are read) the series of Nakhonwat temples was begun during Thisrong’s reign (a.d.803–845).?

65 Yule’s Marco Polo, II. 39–42; J. R. A. Soc. I. 355.?

66 Yule Jour. R. A. Soc. (N. S.) I. 356.?

67 Compare Yule in Jour. R. A. S. (N. S.) I. 355. KandahÁr in south-west AfghanistÁn is another example of the Kedarite or Little Yuechi fondness for giving to their colonies the name of their parent country.?

68 Compare Yule’s Marco Polo, II. 82–84.?

69 Yule in Ency. Brit. Art. Cambodia, 724, 725, 726.?

70 Fa Hian (a.d.400) about fifty miles north-west of Kanauj found a dragon chapel (Beal’s Buddhist Records, I. 40) of which a white-eared dragon was the patron. The dragon, he notes, gives seasonable showers and keeps off all plagues and calamities. At the end of the rains the dragon turns into a little white-eared serpent and the priests feed him. At the deserted Kapilavastu in Tirhut Fa Hian was shown a tank and in it a dragon who, he says, constantly guards and protects a tower to Buddha and worships there night and morning (Ditto, I. 50).

Sung-Yun (a.d.519) notices (Beal’s Buddhist Records, I. 69) in SwÁt (UdyÁna) a tank and a temple with fifty priests called the temple of the NÁga RÁja because the NÁga supplies it with funds. In another passage (Ditto, 92) he notices that in a narrow land on the border of Posse (Fars) a dragon had taken his residence and was stopping the rain and piling the snow. Hiuen Tsiang (Ditto, I. 20) notes that in Kucha, north of the Tarim river east of the Bolor mountains, the Shen horses are half dragon horses and the Shen men half dragon men. In Aksu, 150 miles west of Kucha, fierce dragons molest travellers with storms of flying sand and gravel (Ditto, 25); the hot lake or Johai, 100 miles north-east of Aksu, is jointly inhabited by dragons and fish; scaly monsters rise to the surface and travellers pray to them (Ditto, 26). An Arhat (page 63) prays that he may become a NÁgarÁja. He becomes a NÁgarÁja, kills the real NÁgarÁja, takes his palace, attaches the NÁgas to him, and raises winds and tempests; Kanishka comes against him and the Arhat takes the form of a BrÁhman and knocks down Kanishka’s towers. A great merit-flame bursts from Kanishka’s shoulders and the BrÁhman NÁgarÁja apologises. His evil and passionate spirit, the fruit of evil deeds in a former birth, had made the Arhat pray to be a NÁgarÁja. If clouds gathered the monks knew that the NÁgarÁja meant mischief. The convent gong was beaten and the NÁgarÁja pacified (or scared) Ditto, 64–66. NÁgas were powerful brutes, cloud-riding wind-driving water-walking brutes, still only brutes. The account of the NÁga or dragon of JelalÁbÁd (in Kambojia) is excellent. In Buddha’s time the dragon had been Buddha’s milkman. He lost his temper, laid flowers at the Dragon’s cave, prayed he might become a dragon, and leaped over the cliff. He laid the country waste and did so much harm that TathÁgata (or Buddha) converted him. The NÁga asked Buddha to take his cave. Buddha said No. I will leave my shadow. If you get angry look at my shadow and it will quiet you (Ditto, 94). Another typical dragon is ApalÁla of the SwÁt river (Ditto, 68). In the time of Kasyapa Buddha ApalÁla was a weaver of spells named Gangi. Gangi’s spells kept the dragons quiet and saved the crops. But the people were thankless and paid no tithes. May I be born a dragon, cursed Gangi, poisonous and ruinous. He was born the dragon of the SwÁt valley, ApalÁla, who belched forth a salt stream and burned the crops. The ruin of the fair and pious valley of SwÁt reached Sakya’s (Buddha’s) ears. He passed to Mangala and beat the mountain side with Indra’s mace. ApalÁla came forth was lectured and converted. He agreed to do no more mischief on condition that once in twelve years he might ruin the crops. (Ditto, 122.) In a lake about seven miles west of TakshasilÁ, a spot dear to the exiled Kambojan, lived ElÁpatra the NÁgarÁja, a Bhikshu or ascetic who in a former life had destroyed a tree. When the crops wanted rain or fair weather, the Shamans or medicine-men led the people to pray at ElÁpatra’s tank (page 137). In Kashmir, perhaps the place of halt of the Kambojan in his conquests eastwards, in old times the country was a dragon lake.71 Madhyantika drove out the waters but left one small part as a house for the NÁga king (I. 150). What sense have these tales? In a hilly land where the people live in valleys the river is at once the most whimsical and the most dangerous force. Few seasons pass in which the river does not either damage with its floods or with its failure and at times glaciers and landslips stop the entire flow and the valley is ruined. So great and so strange an evil as the complete drying of a river must be the result of some one’s will, of some one’s temper. The Dragon is angry he wants a sacrifice. Again the river ponds into a lake, the lake tops the earth bank and rushes in a flood wasting as only a dragon can waste. For generations after so awful a proof of power all doubts regarding dragons are dead. (Compare Drew’s Cashmere and Jummoo, 414–421.) In India the Chinese dragon turns into a cobra. In China the cobra is unknown: in India than the cobra no power is more dreaded. How can the mighty unwieldy dragon be the little silent cobra. How not? Can the dragon be worshipful if he is unable to change his shape. To the spirit not to the form is worship due. Again the worshipped dragon becomes the guardian. The great earth Bodhisattva transforms himself into a NÁgarÁja and dwells in lake Anavatapta whose flow of cool water enriches the world (Buddhist Records, II. 11). In a fane in SwÁt Buddha takes the form of a dragon and the people live on him (125). A pestilence wasted SwÁt. Buddha becomes the serpent Suma, all who taste his flesh are healed of the plague (126). A NÁga maiden, who for her sins has been born in serpent shape and lives in a pool, loves Buddha who was then a Sakya chief. Buddha’s merit regains for the girl her lost human form. He goes into the pool slays the girl’s snake-kin and marries her. Not even by marriage with the Sakya is her serpent spirit driven out of the maiden. At night from her head issues a nine-crested NÁga. Sakya strikes off the nine crests and ever since that blow the royal family has suffered from headaches (132). This last tale shows how Buddhism works on the coarser and fiercer tribes who accept its teaching. The converts rise to be men though a snake-head may peep out to show that not all of the old leaven is dead. In other stories Buddha as the sacramental snake shows the moral advance in Buddhism from fiend to guardian worship. The rest of the tales illustrate the corresponding intellectual progress from force worship to man, that is mind, worship. The water force sometimes kindly and enriching sometimes fierce and wasting becomes a Bodhisattva always kindly though his goodwill may have to give way to the rage of evil powers. So BrÁhmanism turns NÁrÁyana the sea into Siva or SomnÁth the sea ruler. In this as in other phases religion passes from the worship of the forces of Nature to which in his beginnings man has to bow to the worship of Man or conscious Mind whose growth in skill and in knowledge has made him the Lord of the forces. These higher ideals are to a great extent a veneer. The Buddhist evangelist may dry the lake; he is careful to leave a pool for the NÁgarÁja. In times of trouble among the fierce struggles of pioneers and settlers the spirit of Buddha withdraws and leaves the empty shrine to the earlier and the more immortal spirit of Force, the NÁgarÁja who has lived on in the pool which for the sake of peace Buddha refrained from drying.?

71 Kashmir has still a trace of GandhÁra. Compare (Ency. Brit. Art. Kashmir page 13: The races of Kashmir are GandhÁras, KhasÁs, and Daradas.)?

72 Mr. Fergusson (Architecture, 219) places the KÁshmir temples between a.d.600 and 1200 and allots MÁrtand the greatest to about a.d.750. The classical element, he says, cannot be mistaken. The shafts are fluted Grecian Doric probably taken from the GandhÁra monasteries of the fourth and fifth centuries. Fergusson was satisfied (Ditto, 289) that the religion of the builders of the KÁshmir temples was NÁga worship. In Cambodia the BrÁhman remains were like those of Java (Ditto, 667). But the connection between the Nakhonwat series and the KÁshmir temples was unmistakeable (Ditto, 297, 665). NÁga worship was the object of both (Ditto, 677–679). Imperfect information forced Fergusson to date the Nakhonwat not earlier than the thirteenth century (Ditto, 660, 679). The evidence of the inscriptions which (J. As. Ser. VI. Tom. XIX. page 190) brings back the date of this the latest of a long series of temples to the ninth and tenth centuries adds greatly to the probability of some direct connection between the builders of the MÁrtand shrine in KÁshmir and of the great Nakhonwat temple at Angkor.?

73 Ency. Brit. Art. Tibet, 344.?

74 Ency. Brit. Art. Cambodia.?

75 Yule’s Marco Polo, II. 45, 47.?