Sec. v] BUDDHIST SITES DESCRIBED BY HSÜAN-TSANG 225
(Ultigh zdt), who had come from afar, is supposed to have sat at the spot, and evil would befall those who should touch the ground. The name of the saint is forgotten, and the villagers could not assert whether he rests under the mound or not. But the people of Somiya never pass without saying a prayer, and according to the testimony of Shâmi Sope and his forbears they have clung to this custom for the last two centuries. I take it as a sign of the antiquity of the tradition that no name is assigned to the saint whose memory lingers about the ' Döbe', whereas the names of his three companions, who are supposed to have disappeared beneath the ground at the spot now sanctified by the Mazâr, are still currently known to young and old as Mullah Yahyamutti, Mullah Allâmutti, Mullah Serekmutti 8. Nobody seemed to know of any other ` Döbe ' similarly surrounded by superstitious awe in the neighbourhood.
Considering the concordant evidence of the name and position of Somiya, I think it probable that the worship of this nameless mound is the last trace left of the So-mo jê Stûpa of Buddhist days. We may not be able to throw light on the character or origin of the name So-mo jê, but there is evidence of phonetic analogies to support the assumption that Somiya is the direct
derivative of the ancient name intended by the Hsi yü-chi's transcription. The character so
is attested elsewhere as a rendering of the syllable so 9. f mo as a rendering of Indian ma or Ind is also well known 10. Finally It jé, which in Julien's list appears as the representative of the Indian syllables jn"a, jnâ, na, nyd, nya, na, may very well have been used here to express a syllable of which the modern phonetic derivative would be ya, seeing that the ya of the local name Niya in the Hsi yü-chi's transcription is rendered by the character 0 jang, which is otherwise used to express the Indian sounds jna and na 11. The change from a form *Soman"a, such as may be conjecturally restored from the transcription So-mo je, into the present Somiya is slight, and can be fully accounted for by a well-known phonetic process '2.
The same day's search enabled me to identify in all probability another sacred site mentioned by Hsüan-tsang. In a convent known as the Ti-chic p`o fu-na Sanghârâma, and situated a little over 10 li to the south-west of the capital, the pilgrim was shown the statue of a standing Buddha which was supposed to have miraculously come to this spot from Kuchâ.l3. A Khotan minister exiled to the latter state had secured his repatriation by assiduous worship of this image, and as he continued to honour it after his return it came one night by itself. The minister then built for it this convent. The name in this case can no longer be traced 14, but exactly in the direction and at the distance indicated (about two miles) there lies the popular Ziârat of ` Bôwa-Kambar', visited by people from all parts of the Khotan district. I found it to consist of a large square cemetery enclosing the high mud-built tomb of the saint, who is believed to have acquired holiness by his devoted services as the groom of "Ali Pâdshâh'. The level of the cemetery near its centre lies fully twelve feet below the surrounding fields—a certain indication of its antiquity according to my previously detailed observations. A grove
9 I do not attempt to restore the orthodox Arabic or Turki forms of these names. The closing -mufti, also occasionally pronounced -mupti, probably represents the Arabic mufti.
9 See Julien, Mélhode pour transcrire, p. 186.
1° Comp. Julien, loc. cit., p. 154.
n See Julien, loc. cit., p. 115; for Ni-jang : Niya, see chap. x. sec. iii.
12 Julien, Mémoires, ii. p. 235, has assumed the Sanskrit form *Sanrajsia as the original of So-mo-jê; but quite apart from the question whether a Khotan locality in the seventh century is likely to have been known by a Sanskrit name, it
must be pointed out that, though samajta is a compound possible in grammar, it presents no meaning acceptable for a name,
" See Mémoires, ii. p. 23o; Si yu-ki, ii. p. 313 ; Ville de Khotan, p. 45.
14 Julien has tentatively restored Ti-chia-p`o-fu-na into Dirghabhâvana (recte °bhavana), but though p`o fu-na as a transcription of Skr. bhavana, ' residence, sacred habitation ' (common in the names of Buddhist shrines in Kashmir, see Reijat., II. pp. 339, 369) looks probable enough, I cannot find evidence for the rendering of Dirgha by 7Y-chia.
Convent of Tz=chia fu-na.