________________
330
THE INDIAN ANTIQUARY.
of the instances a ἅπαξ λεγόμενον. In connection with these statistics are considered certain attendant phenomena helping to prove that the occurrence of ia or -ya is not arbitrary, but depends on the organism of the word. Thus ia is found with very few exceptions wherever the suffix has the circumflex. Further, a long syllable is followed in 189 simple words (in 41 of which the suffix has the circumflex) by -ia, and in only 37 simple words (all without the circumflex) by -ya. A short syllable is followed in 85 simple words (in 45 of which the suffix has the circumflex) by ia, and in 98 simple words (in 12 of which the suffix has the circumflex) by ya. Finally, in regard to their derivation, words clearly derived from a theme in a (as gania from gana) take almost invariably the suffix-form ia. Of 110 such derivatives, there are only 8 absolute exceptions to the rule. Next in order, all the other formative suffixes containing an i or y or u or v are considered, and it is shown that the concurrent phenomena of a preceding long or short syllable and a subsequent vowel or semivowel respectively is clearly traceable everywhere, more absolutely so in regard to u, v than in regard to i, y. In declensional endings the vowel i is found altogether 128 times, and the semi-vowel y nearly 4800 times. Even here the preservative influence of a preceding long syllable seems traceable, the vowel occurring in 123 instances (out of 128) after a long syllable. In verb-inflection, the semi-vowels are found almost exclusively, the exceptions being-ia as a class-sign five times, as a tense-sign once, in optatives 95 times; u(a) in a class-sign six times, in a personal ending 12 times, and in various forms of the root dhanv (perhaps as a class-sign) 14 times.
[NOVEMBER, 1882.
which witnessed the conquest of his native land by the Mongols under Kublai Khan; and he was busy writing his Antiquarian Researches while Marco Polo was travelling about the country (A.D. 1275 to 1295) in the service of the Grand Khan. The date of the deaths of these two men was about the year 1324. Ma Twan-lin's work is arranged in twenty-five books, the last one named "Researches into the Four Frontiers." Out of its 250 sections, only nine describe the maritime countries on the east. The account of Japan is too long to translate and would add nothing to clear up the question as to the identification of Fu-sang. The other eight are translated from the original text, in the order in which they stand, since this order has a bearing upon the position of Fu-sang. They are the following:
Sect. XVI.Hia-i. The land of the Crab Barbarians or Foreigners.
Sect. XVII-Fu-sang. The kingdom of Fusang.
Sect. XVIII.-Nü Kwoh. The kingdom of Women.
Sect. XIX.-Wan Shăn. The kingdom of Pictured Bodies.
Sect. XX.-Ta Han. The kingdom of Great Han.
Sect. XXI.-Chi Jü Kwoh. The kingdom of Dwarfs.
Sect. XXII.-Chang-jin Kwoh. The kingdom of Giants.
Sect. XXIII.-Liu-kiu. The kingdom of Lewchew.
The first of the eight is known to refer to the island of Yezo, and the Chinese still call the region by that name. The next country, Fu-sang, is not described by Ma Twan-lin himself; he merely quotes the narrative of the Shaman or Buddhist priest Hwui-shin, who returned from Fu-sang in A.D. 499. This man reported that it lay twenty thousand li (about 7,000 miles) east of China, and was famous for its fu-sang trees, whence it derived its name. The people made paper from the bark of this tree, and also spun thread of which they manufactured cloth and brocade for dresses. They knew how to write, and had an established government. Hwui-shin's account contains several other particulars, which were first made use of by the learned orientalist De Guignes in 1761 to prove that the land thus described was Mexico. This view has been criticised by Klaproth, supported by Neumann, and in China made the subject of papers by Bretschneider and Sampson, who opposed the view of De Guignes. All their arguments were reviewed by Leland in a small volume published in 1875, in which he upheld the original
Ma-Twan-lin flourished in the troubled times opinion of De Guignes. His conclusion has since
Finally, the occurrence of the vowels or semivowels in question in roots or in the radical part of words is considered; and it was shown that the semi-vowels are found with comparatively few exceptions (chiefly tua- and súar) in the great mass of such words. The exceptions are-i in verb-roots 27 times, in pronominal roots 29 times, in more uncertain combinations 71 times; v once doubtfully in the verb-root suad, in pronominals several hundred times, owing to the frequency of the form tua-, in more uncertain combinations 244 times, mainly in the word súar (233 times) and its compounds.
The fourth article, by Prof. S. Wells Williams, of Yale College, refers to the " Notices of Fu-sang and other countries lying east of China, given in the Antiquarian Researches of Ma twan-lin," of which we extract the following rérumé from the Proceedings:
