230. FEMELES (ISLAND OF WOMEN) 707
NINGHAM, but neither its original form, nor that of San-po-ho, has been restored with certainty. Yet, Hsüan-tsang's transcription Mo-lo-so can be based only on *Mrâ-sa, *Mar-sa or *Maràsa, and I think that the balance weighs in favour of A. H. FRANCKE'S hypothesis (JRAS, 1908, 189), according to which the original name would be Tib. *Mar-sa, « Low-land ». For San-po-ho, VIVIEN DE SAINT-MARTIN'S assertion (in JULIEN, Mém., II, 334) that « it certainly represents Champàka, which is the Sanskrit form of Chamba » is egregiously erroneous. ROCKHILL (The Land of the Lamas, 340) has stated without any qualification that it was « Yaru tsang-po », i. e. the upper Brahmaputra. This,
coming from a man who knew some Chinese, can be explained only by confusion between ho, a
character of transcription, and rpr ho, « river », the remaining san-po being taken as the equivalent of «tsang-po». The curious fact is that a similar solution has been proffered from a non-sinological side: FRANCKE (JRAS, 1908, 189) says that San-po-ho « is the Chinese attempt to represent ytsang-po, 'river', this being the ordinary name of the Indus in Ladakh ». But *Sâm-puâ-xâ supposes an original *Sampaha, which can have nothing to do either with gcan po itself, or with a sanskritized form based on gcari po. THOMAS (Tibetan Texts and Documents, I, 149) speaks of the « Sampaha and the other three Sàkya youths, connected with Sàmbi, Udyàna, Himatala, and Bàmiàn », and, on p. 152, in reference to the location of Suvarnagotra, unreservedly gives «Sam-pa-ha (Sàmbi) ». But I know of no
àkya called « Sampaha ». ROCKHILL, Life of the Buddha, 118, to whom THOMAS gives a reference, actually speaks of « King Shampaka », not « Sampaha ». As to « ambi », which THOMAS took from BEAL (Buddhist Records, II, 21), who himself copied JULIEN (Mém., I, 318), it is an erroneous restoration. The text says that the four àkya became respectively kings of Uddiyàna, *Bâmyana (Bamyan, Bàmiyàn), Himatala (< Hephtalites), and if 4;4 Shang-mi. Shang-mi, restored by JULIEN into *Sàmbi, can render only *Syànmi (< *yami) or *8ànmÎ) (; *Sàmi). As a matter of fact, the name is also written. Shê-mi, which is either *Syàmi or *Sàmi, and Hui-ch'ao says that its king was called *Syàmaràja (or *8àmaràja); *Syàmi seems to me to be more probable. But the main fact is that this country is well known and does not at all correspond to Ladakh. It is Chitral, to the northwest of the Indus (cf. CHAVANNES, Doc. sur les Tou-kiue, 129; FUCHS, Huei-ch'ao's Pilgerreise, 447), and there is not the slightest possibility that it should be connected with *Sampaha. So the name remains unaccounted for, and we must be content with the admission that *Sampaha, alias *Mar-sa, which was at the western frontier of Suvarnagotra, is Ladakh. To account for the double name, we might suppose that *Mar-sa (sanskritized to *Maràsa ?) was the form used in Kashmir, while *Sampaha, heard in Kulûta, would be the Kulûta name of Ladakh.
The question of * jJ Yang-t'ung (*Jang-dung) is no less obscure. The original form is likely to be Tibetan, but remains unknown. One might think of one of the Tibetan names ending in gdon, « face » (cf. BEFEO, y, 291) ; and, from the phonetic point of view, *gYan-gdon, « Auspicious Faces », or Yarns-gdon, « Broad Faces », would be quite satisfactory, but, as far as I know, no such name is attested. Hui-ch'ao writes the name 3, jpJ Yang-t'ung (FucHs, 443, 445), which is phonetically identical; both characters yang were interchangeable in proper names during the T'ang dynasty (cf. BEFEO, iv, 370, 1100-1103). The Shih-chia fang-chih says that Suvarnagotra was the same as Ta Yang-t'ung, 'Great Yang-t'ung', and it is a fact that T'ang texts speak of a 'Great Yang-t'ung' (Ta Yang-t'ung) and a 'Lesser Yang-t'ung' (Hsiao Yang-t'ung).
The only special notice on Yang-t'ung occurs in T'ung tien, 190, 5 b (cf. BUSHELL, in JRAS,
IIIF Color HighRes
