of the boundaries of China exhibits that country as bounded on the west by the Indo-Chinese wildernesses; on the south, by the seas; on the east, by the Eastern Ocean; on the north, by the land of Yajuj and Majuj, and other countries unknown.  Ibn Batuta, with less accurate geography in his head than Abulfeda, maugre his travels, asks about the Rampart of Gog and Magog (Sadd Yajuj wa Majuj) when he is at Sin Kalan, i.e.  Canton, and, as might be expected, gets little satisfaction.

[Illustration:  The Rampart of Gog and Magog]

Apart from this interesting point Marsden seems to be right in the general bearing of his explanation of the passage, and I conceive that the two classes of people whom Marco tries to identify with Gog and Magog do substantially represent the two genera or species, TURKS and MONGOLS, or, according to another nomenclature used by Rashiduddin, the White and Black Tartars.  To the latter class belonged Chinghiz and his MONGOLS proper, with a number of other tribes detailed by Rashiduddin, and these I take to be in a general way the MUNGUL of our text.  The Ung on the other hand, are the UNG-kut, the latter form being presumably only the Mongol plural of UNG.  The Ung-kut were a Turk tribe who were vassals of the Kin Emperors of Cathay, and were intrusted with the defence of the Wall of China, or an important portion of it, which was called by the Mongols Ungu, a name which some connect with that of the tribe. [See note pp. 288-9.] Erdmann indeed asserts that the wall by which the Ung-kut dwelt was not the Great Wall, but some other.  There are traces of other great ramparts in the steppes north of the present wall.  But Erdmann’s arguments seem to me weak in the extreme.

[Mr. Rockhill (Rubruck, p. 112) writes:  “The earliest mention I have found of the name Mongol in Oriental works occurs in the Chinese annals of the After T’ang period (A.D. 923-934), where it occurs in the form Meng-ku.  In the annals of the Liao Dynasty (A.D. 916-1125) it is found under the form Meng-ku-li.  The first occurrence of the name in the Tung chien kang mu is, however, in the 6th year Shao-hsing of Kao-tsung of the Sung (A.D. 1136).  It is just possible that we may trace the word back a little earlier than the After T’ang period, and that the Meng-wa (or ngo as this character may have been pronounced at the time), a branch of the Shih-wei, a Tungusic or Kitan people living around Lake Keule, to the east of the Baikal, and along the Kerulun, which empties into it, during the 7th and subsequent centuries, and referred to in the T’ang shu (Bk. 219), is the same as the later Meng-ku.  Though I have been unable to find, as stated by Howorth (History, i. pt.  I. 28), that the name Meng-ku occurs in the T’ang shu, his conclusion that the northern Shih-wei of that time constituted the Mongol nation