the Minister of Sumuta state, when visiting Quilon on business, met the Chinese envoy there; heard of Kublai's commands, and, on behalf of his master, Takur, sent envoys to submit an address, and to offer presents of a ring, some check-pattern silk or piece-goods, and some embroidered quilts. The Chinese envoy on his way home called at Sumutula, and persuaded the lord of the state, Tuan Pati, to send to China two envoys named Hassan and Suleiman. In 1284 the Fuh Kien Government (ie., “Zaitun ") sent an officer" to summon Nan-wu-li and three other states to do homage." It is quite certain that Lambri is here meant, as will shortly be shown. In 1285 Sumutula sends an envoy to the Chinese Court; and in 1286 both Sumutula and Nan-wu-lih send envoys. In another part of the Mongol records it is said that Sümênna and Nan-wu-lih sent envoys in 1286. Owing to the ill-judged attempt of the Manchu Emperor Kienlung to "reform " Mongol spelling during the last century, the later editions of the Mongol Records are full of inconsistencies such as these. In 1294 the envoys of Sumutala and Nan-wu-li, who had been detained in China awaiting the result of Kublai's war with Java, were sent back with presents and an official safe-conduct tally. It will be noticed that in both names there are slight variations of syllabic spelling, which seem to me to point in the one instance to a short and weak initial syllable followed by an accented second syllable and a final slur— in short, the sound Smûdra, or something like it.

There is only one more entry in the Mongol Records which seems to me likely to refer to Sumatra. It is in 1309, when, immediately after the arrival of a mission from Chanpah (Ciampa) with elephants, three special envoys (with Mongol or Hindoo names) were dispatched to Ciampa, Puh-lin-pa (Palembang) and Pah-sih (Pasei). Puh-lin-pang is the form used in a work cited by M. Groeneveldt and dating from 1416.

It will be remembered that Marco Polo speaks of leaving Pentam (Bantam) for the kingdoms of Little Java (Sumatra),

namely, Ferlech, Basman, Samara, Dagroian, and Lambri. Now in 1323 the Mongol Records mention missions to the Emperor Shotepala from Pintan and Chao-wa (Bantam and Java), and Colonel Yule identifies Basman with the Paçem of the Portuguese and Pasei of the Malays. He also quotes the Malay Chronicle to show that Pasei was founded by the first Mussulman Sovereign of Samudra, whose two sons, were reigning in Samudra and Pasei in 1346 when the Arab Ibn Batuta was there. I notice that, in the translation I have consulted, Ibn Batuta writes the word Shumutra, and speaks of its camphor. Here, then, we have ample primâfacie evidence to show that the Mongols had an official knowledge of at least two states in the island before Marco Polo was in Samara (Samudra); that the Mongol Records are the first to use this last word, which they first pronounce as Ibn Batuta pronounces it; and that they summon Pasei to do duty as soon as ever they hear of its existence.

As soon as ever the Mongols were well out of China, the founder of the new Chinese dynasty of Ming, in 1370, sent an officer to San-foh-ts'i or Sarbaza, to demand the usual submission; no mention whatever is made of any other Sumatran state just yet. The King, Mahârâjâ Palapu, at once responded, and his envoys landed at Ts'üan-chou. It seems that three Kings were reigning at the time, and moreover, a great part of the country had already been conquered by Java, which bestowed a new name (Kewkiang or "Old Haven") upon the chief port, which name, according to Mr. Groeneveldt, is still in current use. The parts which Java was unable to occupy effectively fell a prey to Chinese adventurers, who, as petty Kings themselves, sent tribute to Peking. The anarchy resulting from these political changes led to the blocking of the ocean highways. China invited Siam to use her influence with Java to induce the latter to keep her vassal quiet; and when Malacca put in a claim to part of Sarbaza, alleging the authority of China for it, the Emperor wrote to Java to

disclaim any such idea, so that it is plain China recognised the superior claims of Java. The Cantonese piratical rulers seem also to have recognised the suzerainty of Java, whilst at the same time sending tribute missions of their own to Peking; the last that is heard of the place is that a Chinese pirate named Chang Lien was in charge, of the trading port at least, in 1566, most of his subjects being Fuh Kien men from the two rival Zaitun cities; the pirate collected duties on merchandise, and seemed quite able to preserve some sort of order.

The neglect of the ocean states to send their duty missions led to the despatch in 1405 of the celebrated eunuch Chêng Ho, who took with him a strong escort, amounting almost to an army, and an adequate staff of interpreters. He made seven voyages to the Indian Ocean between 1405 and 1431, but on his first venture he seems to have only visited in person the one single Sumatran state of Samudra, contenting himself with sending lieutenants to the other minor states of the same island. However, on his return. journey from Samudra (now for the first time called Sumên-ta-la, and corresponding with Friar Odoric's Sumoltra) he had a narrow escape of capture at the hands of the Chinese pirate chief of Sarbaza, who was in the end betrayed to the eunuch by another Cantonese adventurer, carried to Peking, and executed there.

Some of the eunuch's account of Sarbaza is admittedly a mere repetition of what was stated in the Sung history; thus, we are told its empire extended over fifteen islands, which probably means that it included Jambi, Kampar, and the islands between Carimon and Banca. The ruler is stated to be styled chan-pi, to which statement the later account adds, "afterwards the residence of the great chief was called chan-pi state, and the old capital was changed to Kewkiang," clearly a confusion in terms, and referring to the removal of the old princely house from the port to the interior. Among the new statements are the one that a large

part of the population lived on boats or rafts, only the ruling classes living on terra firma. According to Kollewijn's account of the Dutch possessions, Bra Vijaya was ruling in Java at the time of the Palembang conquest, and the son of the Javan Governor of Palembang rebelled against him and defeated him in 1478, driving him to Bali. This Bra Vijaya is evidently the "Pala-wu" or "Bra the military" of the Ming Records, the King of Java who sent tribute in 1452, as I shall explain when I approach the subject of Java.

Sumoltra is stated to be nine whole days' sail with a favourable wind westwards from Malacca, as Sarbaza is eight westwards from Java, two precise statements which go far towards establishing their exact positions. Subsequent to the Javan conquest of Sarbaza in 1377, and the partition of that state between Chinese adventurers as already related, but before the first Chinese envoy visited Sumêntala in 1403 to notify that and other states of the usurping Emperor's accession, an envoy named Ambar had been sent to the Chinese Court by King Shutan Maleh uh Ta-fên (or Ta-pan-pronounced both ways) of Sü-wên-ta-na, This was in 1383. Again, in 1436 the envoys from Ciampa complain that Siam has been molesting her mission to Sü-wên-ta-na. The syllables Si and na are the same as those of Sumênna in 1286; the syllable wên in the Fuh Kien and Kwang Tung dialects becomes mên, and (as in the word Lambri) the initials / and n are confused all over China. Of course Sumoltra or Samudra is meant; in fact

the Ming Records say: "Some think that Suwêntana is simply Sumêntala, as changed during the reign of Hung-wu (1368-1398); but the King's names differ. It is impossible to find this point out."

Fortunately, however, the Chinese have said enough to make it quite certain what place is meant, and at the same time to confirm the accuracy of Colonel Yule's admirable researches. It is clear from the well-known Malay syllables

Tuan and pati that in 1282 the Mussulmans were not yet in power. When Ibn Batuta visited the city of Sămăthrah (or Shumutra as Lee's translation writes it) in 1347 (Heg. year 748), he found "El Malik El Zahir Samāl Oddū" reigning, and all his neighbours paying him tribute; the said King gave Ibn Batuta a junk for China (Zaitun). Ibn Batuta found China in a state of rebellion, and the Emperor Firun (ie., Toghon Timur, the last Mongol) far away at Karakorum. It is a remarkable coincidence that the Ming Records specially mention the arrival in Fuh Kien of Nekulun the Frank (fifty years too soon for Nicolo Conti), and also of a mission from Java, just at the time when the Mongols were collapsing, and they mention no others at that time; both were received and sent safely back through the China seas. Colonel Yule spells the Sumatra King's name Malik-al-Dháhir. The Chinese syllables are clearly intended for Sultan Malek ud' Dháfir; in fact, uh-ting is often used in such words-for instance, in Nasr u'din.

It is interesting to mark that Marco Polo notices the absence of wheat in Samara, as the Chinese do in Samudra and Sarbaza; also that he observes in Samara the same. absence of fermented wine in favour of palm and date wines which the Chinese remark in Sarbaza.

When the eunuch Chêng Ho arrived to summon the King in 1405, Tsai-nu-li-a-pi-ting was reigning. He had already in anticipation sent submissive envoys to China. Mr. Groeneveldt takes the syllable i, "already," to be part of the name, and suggests the native title petinggi. Probably some such name as Senur Abu 'din is meant, but of course that is mere conjecture: his father had been killed in war with the "tattooed faces," or Nagur, and, according to the Chinese story, the widow swore to marry the first man who would avenge her. An old fisherman, succeeding in this exploit, had married the Queen, and became "the old King"; but the legitimate son, on attaining years of discretion, had killed him, driving "the old King's" brother Sukanla to the mountains, whence he waged a harassing