国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

Khotanese language. Contents not identified. No. 3,
Hoernle Reg. 4′0″×10″. Pl. CXLVII.

Ch. 0044. Roll, practically complete; dirty dun-coloured
paper, much damaged in the upper portion (c. 10 ft.) and
along middle and R. edge. Inscr. on obv. with Chin.; on
rev. with 70 ll. in Cursive Gupta chars. and bad Skr. lan-
guage; remaining space, c. 18 ft., blank.
Contents: Kaśiakṛt-prajñā-pāramitā. See J.R.A.S., 1911,
p. 473. No. 4, Hoernle Reg. 23′10″×10″. Pl. CXLVII.

Ch. 0045. Roll, incomplete; on dun-coloured paper; fr.
torn off a roll. On rev. portions of 12 ll. in Cursive Gupta
chars. and Khotanese language. No. 5, Hoernle Reg. 15⅜″
×4⅔″.

Ch. 0046. Roll, three sets of small frs. apparently from
three different rolls. (1) Three torn frs. of dirty brown
colour, inscr. on rev. with Cursive Gupta chars., in Khotanese.
(2) Three torn frs. of dun-coloured paper; each inscr. on
rev. with identical remarks in Khotanese and Cursive Gupta.
(3) One large and one small fr. inscr. on rev. with Cursive
Gupta alphabetic table, but not in usual order. Quoted in
J.R.A.S., 1911, p. 458, and shown, ib., Pl. IV. No. 5 (con.),
Hoernle Reg.

Ch. 0047. Roll, practically complete but lower end dam-
aged. Inscr. on rev. with 35 ll. in fair Cursive Gupta script,
partly in Khotanese, partly in corrupt Skr.
Contents: text commences with 14 ll. in Khotanese, not
identified. Then follow 10 verses in corrupt Skr., and finally,
short sentence in Khotanese. No. 6, Hoernle Reg. 2′33″×
12¼″.

Ch. 0048. Roll, complete; dun-coloured paper, slightly
damaged along L. edge; inscr. on obv. with Chin.; on rev.,
upper portion, with 71 ll. in crude Cursive Gupta and Kho-
tanese language; bottom, about 15″, blank. Along top line
of writing is faint impression of two seals, apparently Chinese.
Contents: Khotanese text commencing with a date, quoted
in J.R.A.S., 1911, p. 470. Then follows Buddhist story
with introduction; both quoted, ibid., p. 474. No. 7, Hoernle
Reg. 7′11″×13⅜″.

Ch. 0049. Roll, shapeless fr. torn off larger roll; at
bottom, a pasted-on strip, 9⅔″×1⅔″. Inscr. with 7 ll. in
Khotanese and Cursive Gupta. In top line traces of a date.
Also another small fr. No. 8, Hoernle Reg. 8⅜″×5″ to 6″.

Ch. 0070. a. Palm-leaf Pōthi, on 69 fols. R. and L.
margins of fols. bounded by inked lines. String-hole at 6⅓″
from L. end (or at ⅓ of whole leaf), and between two inked
lines crossing from upper to lower edge of leaf; the space
between these two lines divides the surface of the fol. into two
inscr. areas, the lines of which are, however, continuous.
Fol. No. on L. margin of obv. Leaves on the whole well
preserved, except some which are mutilated, e.g. fol. 63,
which has lost about 6″ at R. end; in five fols. L. margin is
gone, together with fol. numbers. Six ll. writing on page.
Script an early form of Nepalese, rather than a late form
of Cursive Indian Gupta. A marked peculiarity is the ex-
aggerated knot attached to foot, and R. side of top of such
letters as ga, la, dha, tha, etc. (cf. Nos. 1702 and 1049 of
the Table of Letters in Bendall, Catalogue of Buddhist Sans-

krit MSS.). This points to the eighth or ninth century as
date of Pōthi.
The preserved fol. Nos., arranged in centuries, are as
follows:—

Total.
1st cent. 13, 25, 41, 57, 60, 62, 63, 82, 85,
94, 97 ... ... ... ... ... 11 fols.
2nd cent. 101, 102, 120, 124, 129, 140, 160,
167, 169, 180, 182, 195 ... ... 12 ,,
3rd cent. 205, 221, 242, 273, 279, 284, 291,
295 ... ... ... ... ... ... 8 ,,
4th cent. 310, 317, 343, 364, 373, 395 ... 6 ,,
5th cent. 404, 405, 407, 460, 461, 463, 465,
468, 492, 493 ... ... ... ... 10 ,,
6th cent. 524, 553, 560, 564, 575, 599 ... 6 ,,
7th cent. ... ... ... ... ... ... ... 0 ,,
8th cent. 721, 771 ... ... ... ... ... 2 ,,
9th cent. 811, 813, 860, 865, 866, 893, 894,
895 ... ... ... ... ... ... 8 ,,
10th cent. 913 ... ... ... ... ... ... 1 ,,
Total 64 ,,

In four of existing fols. (310 bls, 395 bls, 895 al, clsl) there
occur numeral figures, marking the end of a section, but
without mentioning the name of the work.
From the non-consecutive order of figures (28, 77, 65) it
is clear that they do not indicate sections or chapters, but
either subsections or a number of verses in the subsections.
Bendall (loc. cit., p. 2, l. 21) says with reference to the Aṣṭa-
sāhasrikā that the figures refer to the number of verses in
each chapter; and that, no doubt, is correct with regard to
the Aṣṭasāhasrikā redaction; for the total of figures amounts
to 8,164. All the redactions of the Prajñā-pāramitā are
written in prose; and the Indian method of estimating the
extent of prose writing is to count the number of Ślokas
(aggregate of 32 Akṣaras) contained within it. On the other
hand, a similar remark of Bendall (loc. cit., p. 146, l. 6) leaves
it doubtful whether he took the figures to refer to 'sub-
divisions' or to the verses contained in a 'subdivision'.
This doubt can be set at rest only by referring to an actual
complete MS. of the Śatasāhasrikā, such as is not accessible
to me.
From colophon of unnumbered fol. (clsl) it is seen that the
surviving fols. belong to one of the redactions of the Prajñā-
pāramitā. This may be seen also from many other passages:
e.g. on fol. 25 al.
Seeing that the last extant fol. is numbered 913, and that
there are 12 ll. on each fol., with about 96 Akṣaras (or 3
Ślokas) on each line, i.e. 36 Ślokas on each fol., it follows
that 913 fols. yield a total of 32,868 Ślokas. Now there are
five redactions of the Prajñā-pāramitā, of 125,000, 100,000,
25,000, 10,000, and 8,000 Ślokas respectively. The total of
32,868 excludes the last three redactions. It can belong only
to either the redaction of 125,000 or to that of 100,000. The
latter is the better known redaction, and our MS. probably be-
longed to it. On that premise, about two-thirds of the Pōthi are
missing; for three times 32,868 comes to 98,604 Ślokas. Now,

1374
8 Y