متن کامل نگلیسی ساینو-ایرانیکا اثر جاوید برتولت لوفر

IRANIAN PRECIOUS STONES EMERALD, TURQUOIS 519

Co ken lu, written in I366. 1 The Dictionary in Four Languages 2 writes
this word tsie-mu-lu IB. l& $&. This is a transcription of Persian
zumurrud.

The word itself is of Semitic origin. In Assyrian it has been traced
in the form barraktu in a Babylonian text dated in the thirty-fifth year
of Artaxerxes I (464-424 B.C.). 3 In Hebrew it is bdreket or barkat, in
Syriac borko, in Arabic zummurud, in Armenian zemruoot; in Russian
izumrud. The Greek maragdos or smaragdos is borrowed from Semitic;
and Sanskrit marakata is derived from Greek, Tibetan mar-gad from
Sanskrit. 4 The Arabic-Persian zummurud appears to be based directly
on the Greek form with initial sibilant.

87. In regard to turquois I shall be brief. The Persian turquois,
both that of Nisapur and Kirman, is first mentioned under the name
tien-tse 'fcj -f in the Co ken lu of 1366. This does not mean that the
Chinese were not acquainted with the Persian turquois at a somewhat
earlier date. It is even possible that the Kitan were already acquainted
with turquois. 5 I do not believe that pi-lu 8 5& represents a transcrip-
tion of Persian firuza ("turquois"), as proposed by WATTERS C without
indicating any source for the alleged Chinese word, which, if it exists,
may be restricted to the modern colloquial language. I have not yet
traced it in literature. 7 As early as 1290 turquoises were mined in Hui-
'wan, Yun-nan. 8 The Geography of the Ming dynasty indicates a
turquois-mine in Nan-nin Sou ;c #1 in the prefecture of Yun-nan,

1 Ch. 7, p. 5 b; Wu li siao Si, Ch. 7, p. 14. The author of this work cites the
writing of the Yuan work as the correct one, adding tsu-mu-lu, which he says is at
present in vogue, as an erroneous form. It is due to an adjustment suggested by
popular etymology, the character lu ("green") referring to the green color of the
stone, whose common designation is lit pao Si $jfc jt 5 ("green precious stone");
see GEERTS, Produits, p. 481.

2 Ch. 22, p. 66.

3 C. FOSSEY, Etudes assyriennes (Journal asiatique, 1917, I, p. 473).

4 Cf. Notes on Turquois, p. 55; T'oung Pao, 1916, p. 465. MUSS-ARNOLT
(Transactions Am. Phil. Assoc., Vol. XXIII, 1892, p. 139) states erroneously that
both the Greek and the Semitic words are independently derived from Sanskrit.
In the attempt to trace the history of loan-words it is first of all necessary to ascer-
tain the history of the objects.

5 As intimated by me in American Anthropologist, 1916, p. 589. Tien-tse as the
product of Pan-ta-li are mentioned in the Tao i ci lio, written in 1349 by Wan Ta-
yuan (ROCKHILL, T'oung Pao, 1915, p. 464).

6 Essays on the Chinese Language, p. 352.

7 In the Pen ts'ao kan mu (Ch. 8, p. 17 b) is mentioned a stone p'iao pi lu Hf
H $&, explained as a precious stone (pao Si) of pi ^ color. This is possibly the
foundation of Watters' statement.

8 Yuan Si, Ch. 16, p. 10 b. See, further, Notes on Turquois, pp. 58-59.



520 SlNO-lRANICA

Yun-nan Province. 1 In this text, the term pi t'ien-tse it *0| -f- is em-
ployed. T'an Ts'ui 2 says that turquoises (pi t'ien) are produced in the
Mori-van t'u-se : ^ 3 of Yun-nan. In the Hin-nan fu ci R ^
fl\f S, 3 the gazetteer of the prefecture of Hiii-iian in southern Sen-si,
it is said that pi Vien (written Hi) were formerly a product of this lo-
cality, and mined under the T'ang and Sung, the mines being closed in
the beginning of the Ming. This notice is suspicious, as we hear of
pi-tien or tien-tse neither under the T'ang nor the Sung; the term comes
into existence under the Yuan. 4

88. & fit kin tsih (" essence of gold") appears to have been the term
for lapis lazuli during the T'ang period. The stone came from the
famous mines of Badaxsan. 5

At the time of the Yuan or Mongol dynasty a new word for lapis
lazuli springs up in the form lan-fri S3 Jfc. The Chinese traveller C'an
Te, who was despatched in 1259 as envoy by the Mongol Emperor
Mangu to his brother Hulagu, King of Persia, and whose diary, the
Si U ki, was edited by Liu Yu in 1263, reports that a stone of that name
is found on the rocks of the mountains in the south-western countries
of Persia. The word Ian-Pi is written with two characters meaning
"orchid" and "red," which yields no sense; and BRETSCHNEiDER 6 is
therefore right in concluding that the two elements represent the tran-
scription of a foreign name. He is inclined to think that "it is the same
as landshiwer, the Arabic name for lapis lazuli." In New Persian it is
la&vard or Idjvard (Arabic lazvard). Another Arabic word is Unej, by
which the cyanos of Dioscorides is translated. 7 An Arabic form lanjiver
is not known to me.

"There is also in the same country [Badashan] another mountain,
in which azure is found; 'tis the finest in the world, and is got in a vein
like silver. There are also other mountains which contain a great
amount of silver ore, so that the country is a very rich one." Thus runs

1 Ta Min i t'un ci, Ch. 86, p. 8.

2 Tien hai yil hen ci, 1799, Ch. I, p. 6 b (ed. of Wen yin lou yii ti ts'un $u). See
above, p. 228. T'u-se are districts under a native chieftain, who himself is subject to
Chinese authority.

3 Ch. ii, p. ii b (ed. of 1788).

4 The turquois has not been recognized in a text of the Wei si wen kien ki of
1769 by G. SOULI (Bull, de I'Ecole fran$ aise, Vol. VIII, p. 372), where the question
is of coral and turquois used by the Ku-tsun (a Tibetan tribe) women as ornaments;
instead of yuan-song, as there transcribed, read lii sun Si %Jk ffi ^.

6 CHAVANNES, Documents sur les Tou-kiue, p. 159; and T'oung Pao, 1904,
p. 66.

6 Chinese Recorder, Vol. VI, p. 16; or Mediaeval Researches, Vol. I, p. 151.

7 LECLERC, Traite* des simples, Vol. Ill, p. 254.



IRANIAN PRECIOUS STONES LAPIS LAZULI 521

Marco Polo's account. 1 YULE comments as follows: "The mines of
Lajwurd (whence 1'Azur and Lazuli) have been, like the ruby mines,
celebrated for ages. They He in the upper valley of the Kokcha, called
Koran, within the tract called Yamgan, of which the popular etymology
is Hamah-Kan, or ' All-Mines,' and were visited by Wood in iS^S. 2
The produce now is said to be of very inferior quality, and in quantity
from thirty to sixty pud (thirty-six Ibs. each) annually. The best
quality sells at Bokhara at thirty to sixty tillas, or 12 /. to 24 I. the pud
(Manphul)." 3 In the Dictionary of Four Languages, 4 lapis lazuli is
styled ts'in kin $i W & ^; in Tibetan mu-men, Mongol and Manchu
nomin.

The diamond is likewise attributed by the Chinese to Sasanian
Persia, and I have formerly shown that several Iranian tribes were
acquainted with this precious stone in the beginning of our era. 5 Dia-
mond-points were imported from Persia into China under the T'ang
dynasty. 6

89. The first mention of amber in Chinese records is the reference
to amber in Ki-pin (Kashmir). 7 Then we receive notice of the occurrence
of amber in Ta Ts'in (the Hellenistic Orient) 8 and in Sasanian Persia. 9
The correctness of the latter account is confirmed by the Bundahisn, in
which the Pahlavi term for amber, kahrupai, is transmitted. 10 This word
corresponds to New Persian kahrubd, a compound formed with kdh
("straw") and rubd ("to lift, to attract"). 11 The Arabs derived their
kahrubd (first in Ibn el- Abbas) from the Persians; and between the

1 YULE'S edition, Vol. I, p. 157. /

2 This refers to WOOD, Journey to the Oxus, p. 263.

3 See, further, M. BAUER, Precious Stones, p. 442.

4 Ch. 22, p. 65.

5 The Diamond, p. 53.

6 Ta Tan leu tien, Ch. 22, p. 8.

7 Ts'ien Han $u, Ch. 96 A, p. 5.

8 In the Wei Ho and Hou Han $u (cf. CHAVANNES, T'oung Pao, 1907, p. 182).

9 Nan si, Ch. 79, p. 8; Wei su, Ch. 102, p. 5 a; Sui $u, Ch. 83, p. 7 b. The Sui
Su has altered the name hu-p'o into $ou-p*o Hfc S|, in order to observe the tabu
of the name Hu in Li Hu $: jj^, the father of the founder of the T'ang dynasty.
Amber (also coral and silver) is attributed to Mount Ni / \\1 in the country Fu-lu-ni
W lit JB to the north of Persia, also to the country Hu-se-mi Pf j[ Jg (Wei Su,
Ch. 102, p. 6 b).

10 WEST, Pahlavi Texts, Vol. I, p. 273.

11 Analogies occur in all languages: Chinese U-kiai f* 3f* ("attracting mustard-
seeds"); Sanskrit Ir.inagrahin ("attracting straw"); Tibetan sbur len or sbur Ion,
of the same meaning: French (obsolete) tire-paille. Another Persian word for amber
is saihbari.



522 SlNO-lRANICA

ninth and the tenth century, the word penetrated from the Arabic into
Syriac. 1 In Armenian it is kahribd and kahribar. The same word
migrated westward: Spanish carabe, Portuguese carabe or charabe,
Italian carabe, French carabe; Byzantine Kepafit', Cumanian charabar.
Under the Ming, amber is listed as a product of Herat, Khotan, and
Samarkand. 2 A peculiar variety styled "gold amber" (kin p'o & 59)
is assigned to Arabia (T'ien-faii). 3

The question arises, From what sources did the Persians derive their
amber? G. JACOB, 4 from a study of Arabic sources, has reached the
conclusion that the Arabs obtained amber from the Baltic. The great
importance of Baltic amber in the history of trade is well known, but,
in my estimation, has been somewhat exaggerated by the specialists,
whereas the fact is easily overlooked that amber is found in many parts
of the world. I do not deny that a great deal of amber secured by the
Arabs may be credited to the Baltic sources of supply, but I fail to see
that this theory (for it is no more) follows directly from the data of
Arabic writers. These refer merely to the countries of the Rus and Bui-
gar as the places of provenience, but who will guarantee that the amber
of the Russians hailed exclusively from the Baltic? We know surely
enough that amber occurs in southern Russia and in Rumania. Again,
Ibn al-Baitar knows nothing about Rus and Bulgar in this connection,
but, with reference to al-Jafiki, speaks of two kinds of amber, one
coming from Greece and the Orient, the other being found on the littoral
and underground in the western portion of Spain. 5 Pliny informs us
that, according to Philemon, amber is a fossil substance, and that
it is found in Scythia in two localities, one white and of waxen color,
styled electrum; while in the other place it is red, and is called suali-
ternicum* This Scythian or South-Russian amber may have been traded
by the Iranian Scythians to Iran. In order to settle definitely the
question of the provenience of ancient Persian and Arabic amber, it
would be necessary, first of all, to obtain a certain number of authentic,
ancient Persian and Arabic ambers, and to subject them to a chemical
analysis. We know also that several ancient amber supplies were

1 Cf. E. SEIDEL, Mechithar, p. 146; and G. JACOB, ZDMG, Vol. XLIII, 1889,
P- 359-

3 Ta Min i t'un li, Ch. 89, pp. 23, 24 b, 25 (ed. of 1461).

* Ibid., Ch. 91, p. 20.

4 L. c., and Arabische Handelsartikel, p. 63.

5 LECLERC, Traite" des simples, Vol. Ill, p. 209.

6 Philemon fossile esse et in Scythia erui duobus locis, candidum atque cerei
colons quod vocaretur electrum, in alio fulvum quod appellaretur sualiternicum
(xxxvn, ii, 33).



IRANIAN MINERALS AMBER 523

exhausted long ago. Thus Pliny and the ancient Chinese agree on the
fact that amber was a product of India, while no amber-mines are
known there at present. 1 Amber was formerly found in the
district of Yun-6'an in Yun-nan, and even on the sacred Hwa-San in
Sen-si. 2

G. JACOB 3 has called attention to the fact that the supposition of a
derivation of the Chinese word from Pahlavi kahrupdl is confronted
with unsurmountable difficulties of a chronological character. The
phonetic difficulties are still more aggravating; for Chinese hu-p'o % ffi
was anciently *gu-bak, and any alleged resemblance between the two
words vanishes. Still less can Greek harpax* come into question as the
foundation of the Chinese word, which, in my opinion, comes from an
ancient San or T'ai language of Yun-nan, whence the Chinese received
a kind of amber as early at least as the first century A.D. Of the same
origin, I am inclined to think, is the word tun-mou ^ ^ for amber,
first and exclusively used by the philosopher Wan C'un. 5

Uigur kubik is not the original of the Chinese word, as assumed by
Klaproth; but the Uigur, on the contrary (like Korean xobag), is a
transcription of the Chinese word. Mongol %uba and Manchu xdba
are likewise so, except that these forms were borrowed at a later period,
when the final consonant of Chinese bak or bek was silent. 6

90. Coral is a substance of animal origin; but, as it has always been
conceived in the Orient as a precious stone, 7 a brief notice of it, as far
as Sino-Persian relations are concerned, may be added here. The

1 Cf. Ts'ien Han Su, Ch. 96 A, p. 5 (amber of Kashmir); Nan Si, Ch. 78, p. 7.

2 Cf. Hwa yoli ^ ^ jg, Ch. 3, p. i (ed. of 1831).

3 L. c., p. 355.

4 Proposed by HIRTH, China and the Roman Orient, p. 245. This was merely
a local Syriac name, derived from Greek dpTrdfw (In Syria quoque f eminas verticillos
inde facere et vocare harpaga, quia folia paleasque et vestium fimbrias rapiat.
Pliny, xxxvn, n, 37).

6 Cf. A. FORKE, Lun-heng, pt. II, p. 350. This is not the place for a discussion
of this problem, which I have taken up in a study entitled "Ancient Remains from
the Languages of the Nan Man."

6 For further information on amber, the reader may be referred to my Historical
Jottings on Amber in Asia (Memoirs Am. Anthr. Assoc., Vol. I, pt. 3). I hope to come
back to this subject in greater detail in the course of my Sino-Hellenistic studies,
where it will be shown that the Chinese tradition regarding the origin and properties
of amber is largely influenced by the theories of the ancients.

7 The proof of the animal character of coral is a recent achievement of our
science. Peyssonel was the first to demonstrate in 1727 that the alleged coral-
flowers are real animals; Pallas then described the coral as Isis nobilis; and Lamarck
formed a special genus under the name Corallium rubrum (cf. LACAZE-DUTHIERS,
Histoire naturelle du corail, Paris, 1864; GUIBOURT, Histoire naturelle des drogues,
Vol. IV, p. 378). The common notion in Asia was that coral is a marine tree.



524 SlNO-lRANICA

Chinese learned of the genuine coral through their intercourse with
the Hellenistic Orient : as we are informed by the Wei lio and the Han
Annals, 1 Ta Ts'in produced coral; and the substance was so common,
that the inhabitants used it for making the king-posts of their habita-
tions. The T'ang Annals 2 then describe how the marine product is fished
in the coral islands by men seated in large craft and using nets of iron
wire. When the corals begin to grow on the rocks, they are white like
mushrooms; after a year they turn yellow, and when three years have
elapsed, they change into red. Their branches then begin to intertwine,
and grow to a height of three or four feet. 3 Hirth may be right in
supposing that this fishing took place in the Red Sea, and that the
"Coral Sea" of the Nestorian inscription and the "sea producing
corals and genuine pearls" of the Wei lio are apparently identical with
the latter. 4 But it may have been the Persian Gulf as well, or even the
Mediterranean. Pliny 5 is not very enthusiastic about the Red-Sea
coral; and the Periplus speaks of the importation of coral into India,
which W. H. ScnoFF 6 seems to me to identify correctly with the Medi-
terranean coral. Moreover, the Chinese themselves correlate the above
account of coral-fishing with Persia, for the Yi wu ci H $7 ;S is cited
in the Cen lei pen ts*ao 7 as saying that coral is produced in Persia, being
considered by the people there as their mosjt precious jewel; and the
Pen ts'ao yen i speaks of a coral-island in the sea of Persia, 8 going on to
tell the same story regarding coral-fishing as the T'ang Annals with
reference to Fu-lin (Syria). Su Kuii of the T'ang states that coral grows
in the Southern Sea, but likewise comes from Persia and Ceylon, the
latter statement being repeated by the T'u kin pen ts*ao of the Sung.
It is interesting that the Pen ts'ao of the T'ang insists on the holes in
coral, a characteristic which in the Orient is still regarded (and justly
so) as a mark of authenticity. Under the T'ang, coral was first intro-
duced into the materia medica. In the Annals, coral is ascribed to

1 HIRTH, China and the Roman Orient, pp. 41, 73.

2 Ibid., p. 44-
* Ibid., p. 59.

4 Ibid., p. 246.

B xxxii, ii.

6 The Periplus of the Erythraean Sea, p. 128.

' Ch. 4, p. 37.

8 Ch. 5, p. 7 (ed. of Lu Sin-yuan). The coral island where the coral- tree grows
is also mentioned by an Arabic author, who wrote about A.D. 1000 (G. FERRAND,
Textes relatifs a 1'Extr^me-Orient, Vol. I, p. 147). See, further, E. WIEDEMANN,
Zur Mineralogie im Islam, p. 244.



IRANIAN MINERALS CORAL, BEZOAR 525

Sasanian Persia; 1 and it is stated in the T'ang Annals that Persia pro-
duces coral not higher than three feet. 2 There is no doubt that Persian
corals have found their way all over Asia; and many of them may still
be preserved by Tibetans, who prize above all coral, amber, and tur-
quois. The coral encountered by the Chinese in Ki-pin (Kashmir) 3
may also have been of Persian origin. Unfortunately we have no
information on the subject from ancient Iranian sources, tor do we
know an ancient Iranian name for coral. Solinus inform, us that
Zoroaster attributed to coral a certain power and salubrious effects; 4
and what Pliny says about coral endowed with sacred properties and
being a preservative against all dangers, sounds very much like an
idea emanating from Persia. Persian infants still wear a piece of coral
on the abdomen as a talisman to ward off harm; 5 and, according to
Pliny, this was the practice at his time, only that the branches of coral
were hung at the infant's neck.

The Chinese word for coral, M $8 $an-hu, *san-gu (Japanese
san-go), possibly is of foreign origin, but possibly it is not/ 1 For the
present there is no word in any West-Asiatic or Iranian language with
which it could be correlated. In Hebrew it is ra 'mot, which the Seventy
transcribes pa^oQ or translates /-terewpa. The common word in New
Persian is marjdn (hence Russian marZan); other designations are
birbdl, xuruhak or xurohak, bussad or bissad (Arabic bessed or bussad).
In Armenian it is bust. 7

91. The identification of Chinese H l p'o-so (*bwa-sa) with Persian
pdzdhr or pddzahr* ("bezoar," literally, "antidote"), first proposed by
HiRTH, 9 in my opinion, is not tenable, although it has been indorsed

1 Cou $u, Ch. 50, p. 6; Sui $u, Ch. 83, p. 7 b; regarding coral in Fu-lu-ni, see
above, p. 52 1 , note 9.

2 Tan $u, Ch. 221 B, p. 6 b. The Lian Su (Ch. 54, p. 14 b) attributes to Persia
coral-trees one or two feet high.

3 Ts'ien Han su, Ch. 96 A, p. 5. This passage (not Hou Han su, Ch. 1 18, as stated
by HIRTH, Chau Ju-kua, p. 226, after Bretschneider) contains the earliest mention
of the word san-hu.

4 Habet enim, ut Zoroastres ait, materia haec quandam potestatem, ac propterea
quidquid inde sit, ducitur inter salutaria (n, 39, 42).

5 SCHLIMMER, Terminologie, p. 166.

6 According to BRETSCHNEIDER (Chinese Recorder, Vol. VI, p. 16), "it seems not
to be a Chinese name."

7 Cf . PATKANOV, The Precious Stones according to the Notions of the Armenians
(in Russian), p. 52.

8 Pazand padazahar (see HUBSCHMANN, Persische Studien, p. 193). STEINGASS
gives also pdnzahr. The derivation from bad "wind" (H. FUHNER, Janus, Vol. VI,
I 9 OI P- 317) is not correct.

9 Lander des Islam, p. 45.



526 SlNO-lRANICA

by PELLiOT. 1 Pelliot, however, noticed well that what the Chinese
describe as p'o-so or mo-so IS c is not bezoar, and that the tran-
scription is anomalous. 2 This being the case, it is preferable to reject
the identification, and there are other weighty reasons prompting us
to do so. There is no Chinese account that tells us that Persia had
bezoars or traded bezoars to China. The Chinese were (and are) well
acquainted with the bezoar 3 (I gathered several in China myself), and
bezoars are easy to determine. Now, if p*o-so or mo-so were to repre-
sent Persian pdzahr and a Persian bezoar, the Chinese would not for
a moment fail to inform us that p'o-so is the Po-se niu-hwan or Persian
bezoar; but they say nothing to this effect. On the contrary, the texts
cited under this heading in the Pen ts'ao kan mu* do not make any
mention of Persia, but agree in pointing to the Malay Archipelago as
the provenience of the p'o-so stone. Ma Ci of the Sung assigns it to
the Southern Sea (Nan Hai). Li Si-Sen points to the Ken sin yil ts'e
J^J ^ 3 flfl", written about 1430, as saying that the stone comes from
San-fu-ts'i (Palembang on Sumatra). 5 F. DE MELY designates it only
as a "pierre d'epreuve," and refers to an identification with aventurine,
proposed by Remusat. 6 Bezoar is a calculus concretion found in the
stomachs of a number of mammals, and Oriental literatures abound in
stories regarding such stones extracted from animals. Not only do the
Chinese not say that the p'o-so stone is of animal origin, but, on the con-
trary, they state explicitly that it is of mineral origin. The Ken sin yu ts*e
relates how mariners passing by a certain mountain on Sumatra break
this stone with axes out of the rock, and that the stone when burnt
emits a sulphurous odor. Ma Ci describes this stone as being green
in color and without speckles; those with gold stars, and when rubbed
yielding a milky juice, are the best. All this does not fit the bezoar.
Also the description in the Pen ts'ao yen i 1 refers only to a stone of
mineral origin.

1 Toung Pao, 1912, p. 438.

2 The initial of the Persian word would require a labial surd in Chinese. Whether
the p'o-sa | of the Pei hu lu belongs here is doubtful to me; it is not explained
what this stone is. As admitted in the Pen ts'ao yen i (Ch. 4, p. 4 b), the form mo-so
is secondary.

3 It is first mentioned in the ancient work Pie lu, then in the Wu Si pen ts'ao
of the third century, and by T'ao Hun-kin.

4 Ch. 10, p. 10 b.

6 This text is cited in the same manner in the Tun si yan k'ao of 1618 (Ch. 3,
p. 10). Cf. F. DE MLY, Lapidaire chinois, p. 120.

6 Ibid., pp. LXIV, 260.

7 Ch. 4, p. 4 b (ed. of Lu Sin-yuan).



IRANIAN MINERALS BEZOAR 527

Even as early as the T'ang period, the term p'o-so merely denotes
a stone. It is mentioned in a colophon to the P*ih ts*uan $an ku ts*ao mu
fo^&tllS^/Ktfiby Li Te-yu 3 s IS IS (A.D. 787-849) as a curious
stone preserved in the P'o-so Pavilion south of the C'an-tien & K in
Ho-nan.

Yada or jada, as justly said by Pelliot, is a bezoar; but what at-
tracted the Chinese to this Turkish-Mongol word was not its char-
acter as a bezoar, but its role in magic as a rain-producing stone. Li
Si-cen l has devoted a separate article to it under the name fcfc ^ ca-ta,
and has recognized it as a kind of bezoar; in fact, it follows immediately
his article on the Chinese bezoar (nin-hwan) . 2

The Persian word was brought to China as late as the seventeenth
century by the Jesuits. Pantoja and Aleni, in their geography of the
world, entitled Cifan wai ki? and published in 1623, mention an animal
of Borneo resembling a sheep and a deer, called pa-tsa'r JC H Bl, 4 in
the abdomen of which grows a stone capable of curing all diseases, and
highly prized by the Westerners. The Chinese recognized that this was
a bezoar. 5 Bezoars are obtained on Borneo, but chiefly from a monkey
(Simia longumanis, Dayak buhi) and hedgehog. The Malayan name
for bezoar is gullga; and, as far as I know, the Persian word is not used
by the Malayans. 6 The Chinese Gazetteer of Macao mentions "an
animal like a sheep or goat, in whose belly is produced a stone capable



1 Pen ts'ao kan mu, Ch. 50 B, p. 15 b.

1 There is an extensive literature on the subject of the rain-stone. The earliest
Chinese source known to me, and not mentioned by Pelliot, is the K'ai yuan t'ien
pao i fi M j. 3? X ife 4( J* Wan Zen-yu 3 C IS of the T'ang (p. 20 b).
Cf. also the Sit K'ien su 0f Jfr |J, written by Can Cu jjjt $f in 1805 (Ch. 6, p. 8,
ed. of Yiie ya fan ts'un Su). The Yakut know this stone as sata (BOEHTLINGK, Jakut.
WCrterbuch, p. 153); Pallas gives a Kalmuk form sadan. See, further, W. W. ROCK-
HILL, Rubruck, p. 195; F. v. ERDMANN, Temudschin, p. 94; G. OPPERT, Presbyter
Johannes, p. 102; J. RUSKA, Steinbuch des Qazwlnl, p. 19, and Der Islam, Vol. IV,
1913, pp. 26-30 (it is of especial interest that, according to the Persian mineralogical
treatise of Mohammed Ben Mansur, the rain-stone comes from mines on the frontier
of China, or is taken from the nest of a large water-bird, called surxab, on the frontier
of China; thus, after all, the Turks may have obtained their bezoars from China);
VAMBRY, Primitive Cultur, p. 249; POTANIN, Tangutsko-Tibetskaya Okraina
Kitaya, Vol. II, p. 352, where further literature is cited.

1 Ch. i, p. ii (see above, p. 433).

4 This form comes very near to the pajar of Barbosa in 1516.

8 Cf. the Lu can kun Si k'i (above, p. 346), p. 48.

6 Regarding the Malayan beliefs in bezoars, see, for instance, L. BOUCHAL in
Mitt. Anthr. Ges. Wien, 1900, pp. 179-180; BECCARI, Wanderings in the Great
Forests of Borneo, p. 327; KREEMER in Bijdr. taal- land- en volkenkunde, 1914,
p. 38; etc.



528 SlNO-lRANICA

of curing any disease, and called pa-tsa'r" (written as above); 1 cf.
Portuguese bazar y bazodr, bezoar.

On the other hand, bezoars became universal in the early middle
ages, and the Arabs also list bezoars from China and India. 2 From the
Persian word fddaj, explained as "a stone from China, bezoar," it
appears also that Chinese bezoars were traded to Persia. In Persia, as
is well known, bezoars are highly prized as remedies and talismans. 3

1 Ao-men li lio, Ch. B, p. 37.

2 J. RUSKA, Steinbuch des Aristoteles, p. 148.

8 C. ACOSTA (Tractado de las drogas, pp. 153-160, Burgos, 1578), E. KAEMPFER
(Amoenitates exoticae, pp. 402-403), GUIBOURT (Histoire naturelle des drogues
simples, Vol. IV, pp. 106 et seq.), and G. F. KUNZ (Magic of Jewels and Charms,
pp. 203-220) give a great deal of interesting information on the subject. See also
YULE, Hobson-Jobson, p. 90; E. WIEDEMANN, Zur Mineralogie im Islam, p. 228;
D. HOOPER, Journal As. Soc. Bengal, Vol. VI, 1910, p. 519.



TITLES OF THE SASANIAN GOVERNMENT

92. SI J? sa-pao, *saS(sar)-pav. Title of the official in charge of
the affairs of the Persian religion in Si-nan, an office dating back to the
time when temples of the celestial god of fire were erected there, about
A.D. 621. In an excellent article PELLIOT has assembled all texts relative
to this function. 1 I do not believe, however, that we are justified in
accepting Deveria's theory that the Chinese transcription should render
Syriac saba ("old man"). This plainly conflicts with the laws of tran-
scription so rigorously expounded and upheld by Pelliot himself: it is
necessary to account for the final dental or liquid in the character sa,
which regularly appears in the T'ang transcriptions. It would be
strange also if the Persians should have applied a Syriac word to a
sacred institution of their own. It is evident that the Chinese tran-
scription corresponds to a Middle-Persian form traceable to Old Persian
x$aOra-pavan (x$$pava, x$a$apdva), which resulted in Assyrian axSadar-
apan or axSadrapdn, Hebrew axaSdarfnim? Greek crarpctTnys (Armenian
Sahapand, Sanskrit k$atrapa). The Middle-Persian form from which the
Chinese transcription was very exactly made must have been *sa0-pav
or *xsa0-pav. The character sa renders also Middle and New Persian
sar ("head, chief"). 3

93. J5 Si ftl K'u-sa-ho, *Ku-sa5(r)-7wa, was the title. 3r* of the
kings of Parsa (Persia). 4 This transcription appears to be based on an
Iranian xtadva or xZarva, corresponding to Old Iranian *xsayavan-,
*xsaivan, Sogdian x$evan (" king ") . 5 It is notable that the initial spirant
x is plainly and aptly expressed in Chinese by the element k'u, G while
in the preceding transcription it is suppressed. The differentiation in
time may possibly account for this phenomenon: the transcription
sa-pao comes down from about A.D. 621; while K'u-sa-ho, being con-

1 Le Sa-pao, Bull, de I'Ecole fran$ aise, Vol. Ill, pp. 665-671.
z H. POGNON, Journal asiatique, 1917, I, p. 395.

3 R. GAUTHIOT, Journal asiatique, 1911, II, p. 60.

4 Sui $u, Ch. 83, p. 7 b.

5 R. GAUTHIOT, Essa", sur le vocalisme du sogdien, p. 97. See also the note of
ANDREAS in A. Christens ^n, L'Empire des Sassanides, p. 113. I am unable to see
how the Chinese transcription could correspond to the name Khosrou, as proposed
by several scholars (CHAVANNES, Documents sur les Tou-kiue occidentaux, p. 171;
and HIRTH, Journal Am. Or. Soc., Vol. XXXIII, 1913, p. 197).

6 In the Manichasan transcriptions it is expressed by P$ *xu (hu) ; see CHA-
VANNES and PELLIOT, Traite" maniche"en, p. 25.

529



53 SlNO-lRANICA

tained in the Sui Annals, belongs to the latter part of the sixth century.
According to SALEMANN, 1 Iranian initial xS- develops into Middle-
Persian ; solely the most ancient Armenian loan-words show aSx- for
x$-, otherwise appears regularly save that $x takes the place of inter-
vocalic xL z In view of our Sino-Iranian form, this rule should perhaps
be reconsidered, but this must remain for the discussion of Iranian
scholars.

94. i 5? $a-ye, *sat(sa5)-ya. Title of the sons of the king of
Persia (Wei $u, Ch. 102, p. 6; T*ai fin hwan yu ki, Ch. 185, p. 17).
It corresponds to Avestan xSaBrya ("lord, ruler"). 3 The princes of
the Sasanian empire were styled sa0ra5aran. 4 According to Sasanian
custom, the sons of kings ruled provinces as "kings." 5 Regarding $c
in transcriptions of Iranian names, cf. the name of the river Yaxartes
H$t (Sui hi, Ch. 83, p. 4 b) Yao-sa, that is *Yak-s"a5(sar). As the
Middle-Persian name is Xsart or Asart (Pazend Asard), 6 we are bound
to assume that the prototype of the Chinese transcription was *Axart
or *Yaxart.

95. H p lt i-tsan, but, as thefan-ts'ie of the last character is indicated
by ^ 9J, the proper reading is i-ts'at, *i-dza5, i-dza5, designation of the
king of Parsa ( A $ or II V Pg: Wei ht, Ch. 102, p. 6; Tai
fin hwan yu ki, Ch. 185, p. 17). The Chinese name apparently repre-
sents a transcription of IxseS, the Ixsidh of al-Beruni, title of the
kings of Sogd and Fergana, a dialectic form of Old Persian xSdyaBiya. 7
IxseS is the Avestan x$aeta ("brilliant"), a later form being Sedah.
It must be borne in mind that Sogdian was the lingua franca and
international language of Central Asia, and even the vehicle of civiliza-

1 Grundriss der iran. Phil., Vol. I, pt. I, p. 262.
s Cf. also GAUTHIOT, op. cit., p. 54, 61.

3 K. Hori's identification with New Persian $ah (Spiegel Memorial Volume,
p. 248) must be rejected. The time of the Wei $u plainly refers to Sasanian Persia;
that is, to the Middle-Persian language.

4 A. CHRISTENSEN, op. cit., p. 20. Cf. Old Persian xsc,m, xsa$am ("royalty,
kingdom"), Avestan xSadrem, Sanskrit ksatram (A. MEILLET, Grammaire du vieux
perse, p. 143); xsadrya corresponds to Sanskrit kfatriya.

6 N&LDEKE, Tabari, p. 49; Grundriss, Vol. II, p. 171. I think that H. POGNON
(Journal asiatique, 1917, I, p. 397) is right in assuming that "satrap" was a purely
honorific title granted by the king not only to the governors of the provinces, but
also to many high functionaries.

6 WEST, Pahlavi Texts, Vol. I, p. 80.

7 See SACHAU, Chronology of Ancient Nations, p. 109; F. JUSTI, Iranisches
Namenbuch, p. 141; A. MEILLET, Grammaire du vieux perse, pp. 77, 167 (xsayaBiya
parsaiy, "king in Persia"); F. W. K. MULLER, Ein Doppelblatt aus einem mani-
chaischen Hymnenbuch, p. 31.



TITLES or THE SASANIAN GOVERNMENT 531

tion. 1 The suggestion offered by K. HoRi, 2 that the Chinese transcrip-
tion should represent the Persian word izad ("god"), is not acceptable:
first, New Persian cannot come into question, but only Middle Persian;
second, it is not proved that izad was ever a title of the kings of Persia.
On the contrary, as stated by NoLDEKE, 3 the Sasanians applied to them-
selves the word bag ("god"), but not yazdan, which was the proper word
for "god" even at that time.

96. W$?^ fan-pu-$wai, *pwan-bu-zwi5, designation of the queen
of Parsa (Wei $u, Ch. 102, p. 6; T'ai p'in hwan yu ki, Ch. 185, p. 17).
The foundation of this transcription is presented by Middle Persian
bdnbufn, bdnbiSn (Armenian bambiSn), "consort of the king of Persia." 4
The Iranian prototype of the Chinese transcription seems to have been
*banbuzwi5. The latter element may bear some relation to Sogdian
wdbu or wydySth ("consort"). 5

97. 81 $1 }J[ mo-hu-t*an, *mak-ku(mag-gu)-dan. Officials of
Persia in charge of the judicial department ^ ft ^ f& (Wei lu,
Ch. 102, p. 6). K. HoRi 6 has overlooked the fact that the element
fan forms part of the transcription, and has simply equalized mo-hu with
Avestan moyu. The transcription *mak-ku (mag-gu) is obviously found-
ed on Middle Persian magu, and therefore is perfectly exact. The later
transcription 8 H *muk-gu (mu-hu) is based on New Persian muy,
moy. 7 The ending dan reminds one of such formations as herbeddn
("judge") and mobeddn mobed ("chief of the Magi"), the latter being
Old Persian magupati, Armenian mogpet, Pahlavi maupat, New Persian
mubid (which, according to the Persian Dictionary of Steingass, means
also "one who administers justice, judge"). Above all, compare the
Armenian loan-word movpetan (also movpet, mogpet, mog). s Hence it

1 R. GAUTHIOT, Essai sur le vocalisme du sogdien, p. x; P. PELLIOT, Les in-
fluences iraniennes en Asie centrale et en Extreme-Orient, p. II.

2 Spiegel Memorial Volume, p. 248.

3 Tabari, p. 452.

4 HuBSCHMANN, Armen. Gram., p. 116. In his opinion, the form bdnbuSn,
judging from the Armenian, is wrong; but its authenticity is fully confirmed by the
Chinese transcription.

5 R. GAUTHIOT, Essai sur le vocalisme du sogdien, pp. 59, 112. The three afore-
mentioned titles had already been indicated by ABEL-RMUSAT (Nouvelles melanges
asiatiques, Vol. I, p. 249) after Ma Twan-lin, but partially in wrong transcription:
"Le roi a le titre de Yi-thso; la reine, celui de Tchi-sou, et les fils du roi, celui de
Cha-ye."

6 Spiegel Memorial Volume, p. 248.

7 CHAVANNES and PELLIOT, Traite" maniche"en, p. 170. Accordingly this example
cannot be invoked as proving that muk might transcribe also mak, as formerly
assumed by PELLIOT (Butt, de VEcole franqaise, Vol. IV, p. 312).

8 HORN, Neupersische Etymologic, No. 984; and HUBSCHMANN, Persische
Studien, p. 123.



532 SlNO-lRANICA

may justly be inferred that there was a Middle-Persian form *ma-
gutan or *magudan, from which the Chinese transcription was exactly
made.

98. JE ^ ff ni-hu-han, *ni-hwut-7an. Officials of Persia who have
charge of the Treasury (Wei $u, Ch. 102, p. 6). The word, in fact, is a
family-name or title written by the Greek authors Naxopayav, Naxoepyav,
'Zapvaxopyavr]* (prefixed by the word sar, "head, upper"). Firdausl
mentions repeatedly under the reign of Khosrau II a Naxwara, and
the treasurer of this king is styled "son of Naxwara." 1 The treasury
is named for him al-Naxirajan. The Chinese transcription is made
after the Pahlavi model *Nixur7an or Nexur7an; and, indeed, the
form Nixorakan is also found. 2

99. *& $ ^] ti-pei-p'o, *di-pi-bwi$(bir, wir). Officials of Persia
who have charge of official documents and all affairs (Cou lu, Ch. 50,
p. 5b). In the parallel passage of the Wei $u (Ch. 102, p. 6), the second
character is misprinted > tsao* *tsaw; *di-tsaw would not correspond
to any Iranian word. From the definition of the term it becomes
obvious that the above transcription *di-pi answers to dipi ("writing,
inscription"), 4 Middle Persian dijfir or dapir, New Persian dibUr or dabir
(Armenian dpir)', and that *di-pi-bwi5 corresponds to Middle Persian
dipivar, from *dipi-bara, the suffix -var (anciently bar a) meaning "carry-
ing, bearing." 5 The forms dipir and diblr are contractions from dipivar.
This word, as follows from the definition, appears to have comprised
also what was understood by devdn t the administrative chanceries of
the Sasr ,nian empire.

100. JH It M & no-lo-ho-ti, *at(ar)-la-ha-di. Officials of Persia
who superintended the inner affairs of the king (or the affairs of the
royal household Wei $u, Ch. 102, p. 6). Theophylactus Simocatta 6
gives the following information on the hereditary functions among
the seven high families in the Sasanian empire: "The family called
Artabides possesses the royal dignity, and has also the office of placing

1 NOLDEKE, Tabari, pp. 152-153, 439.

2 JUSTI, Iran. Namenbuch, p. 219. In Naxuraqan or Na%Irajan q and j represent
Pahlavi g. The reconstructions attempted by MODI (Spiegel Memorial Volume,
p. nx) of this and other Sino-Iranian words on the basis of the modern Chinese
pronunciation do not call for any discussion.

3 This misprint is not peculiar to the modern editions, but occurs in an edition
of this work printed in 1596, so that in all probability it was extant in the original
issue. It is easy to see how the two characters were confounded.

4 In the Old-Persian inscriptions, where it occurs in the accusative form dipim
and in the locative dipiya (A. MEILLET, Grammaire du vieux perse, pp. 147, 183).

6 C. SALEMAN, Grundriss iran. Phil., Vol. I, pt. i, pp. 272, 282.
8 m, 8.



TITLES OF THE SASANIAN GOVERNMENT 533

the crown on the king's head. Another family presides over military
affairs, another superintends civil affairs, another settles the litigations
of those who have a dispute and desire an arbiter. The fifth family com-
mands the cavalry, the sixth collects the taxes and supervises the
royal treasures, and the seventh takes care of armament and military
equipment." Artabides ('Apra^ldrjs), as observed by NOLDEKE,* should
be read Argabides ('Apyafildrjs), the equivalent of ArgabeS. There
is also a form apyaTre-nys in correspondence with Pahlavi arkpat. This
title originally designated the commandant of a castle (arg, "citadel"),
and subsequently a very high military rank. 2 In later Hebrew we find
this title in the forms alkafta, arkafta, or arkabta* The above tran-
scription is apparently based on the form *Argade ('Apyadrj) = Argabe5.

1 01. l $& ^8 sie-po-p'o, *sit-pwa-bwi5. Officials of Persia in
charge of the army (infantry and cavalry, pai7an and aswaran), of the
four quarters, the four patkos (pat, "province"; kos, "guarding")
^ ^^^: Wei $u, Ch. 102, p. 6. The Cou $u (Ch. 50, p. 5^)
has II *sat, sar, in the place of the first character. The word corresponds
to Middle Persian spdhbed ("general"); Pahlavi pat, New Persian -bad,
-bud ("master"). EranspahbeS was the title of the generalissimo of
the army of the Sasanian empire up to the time of Khusrau I. The
Pahlavi form is given as spahpat;* the Chinese transcription, however,
corresponds better to New Persian sipahbad, so that also a Middle-
Persian form *spahba5 (-bed or -bud) may be inferred.

102. 3L & 31 nu-se-ta, *u-se-da5, used in the Chinese inscription dated 1489
of the Jews of K'ai-fon fu in Ho-nan, in connection with the preceding name ^0 i$(
Lie-wei (Levi). 5 As justly recognized by G. DEVRIA, this transcription represents
Persian ustad,{ which means "teacher, master." 6 The Persian Jews availed them-
selves of this term for the rendering of the Hebrew title Rab (Rabbi), although
in Persian the name follows the title. The Chinese Jews simply adopted the Chinese
mode of expression, in which the family-name precedes the title, Ustad Lie-wei
meaning as much as "Rabbi Levi." The transcription itself appears to be of much
older date than the Ming, and was doubtless recorded at a time when the final
consonant of ta was still articulated. In a former article I have shown from the
data of the Jewish inscriptions that the Chinese Jews emigrated from Persia and
appeared in China not earlier than in the era of the Sung. This historical proof is
signally confirmed by a piece of linguistic evidence. In the Annals of the Yuan
Dynasty (Yuan Si, Ch. 33, p. 7 b; 43, p. n b) the Jews are styled Su-hu (Ju-hud)

1 Tabari, p. 5.

2 CHRISTENSEN, op. tit., p. 27; NOLDEKE, op. cit., p. 437; HUBSCHMANN, Per-

sische Studien, pp. 239, 240.

3 M. JASTROW, Dictionary of the Targumim, p. 73.

4 HUBSCHMANN, Armen. Gram., p. 240.

6 J. TOBAR, Inscriptions juives de K'ai-fong-fou, p. 44.

6 Regarding this word, see chiefly H. HUBSCHMANN, Persische Studien, p. 14.



534 SlNO-lRANICA

jffl ^ or Cu-wu ^ 7C- This form can have been transcribed only on the basis of
New Persian JuhuS or JahuS with initial palatal sonant. As is well known, the
change of initial y into j is peculiar to New Persian. 1 In Pahlavi we have Yah at,
as in Hebrew Yehudl and in Arabic Yahud. A Middle-Persian Yahut would have
been very easy for the Chinese to transcribe. The very form of their transcription
shows, however, that it was modelled on the New-Persian type, and that it cannot
be much older than the tenth century or the age of the Sung.

1 Cf. HORN, Grundr. iran. Phil., Vol. I, pt. 2, p. 73.



IRANO-SINICA

After dealing with the cultural elements derived by the Chinese
from the Iranians, it will be only just to look also at the reverse of the
medal and consider what the Iranians owe to the Chinese.

i. Some products of China had reached Iranian peoples long before
any Chinese set their foot on Iranian soil. When Can K'ien in 128 B.C.
reached Ta-hia (Bactria), he was amazed to see there staves or walking-
sticks made from bamboo of Kiufi flS 1t ^t 1 and cloth of Su (Se-S'wan)
18 3ft. What this textile exactly was is not known. 2 Both these articles
hailed from what is now Se-S'wan, Kiufi being situated in Zun cou IS $H
in the prefecture of Kia-tin, in the southern part of the province. When
the Chinese envoy inquired from the people of Ta-hia how they had
obtained these objects of his own country, they replied that they pur-
chased them in India. Hence Can K'ien concluded that India could
not be so far distant from Se-'wan. It is well known how this new
geographical notion subsequently led the Chinese to the discovery of
Yun-nan. There was accordingly an ancient trade-route running from
Se-5'wan through Yun-nan into north-eastern India; and, as India on
her north-west frontier was in connection with Iranian territory, Chinese
merchandise could thus reach Iran. The bamboo of Kiufi, also called
Sr, has been identified by the Chinese with the so-called square bamboo
(Bambusa or Phyllostachys quadrangularis) . 3 The cylindrical form is so
universal a feature in bamboo, that the report of the existence in China
and Japan of a bamboo with four-angled stems was first considered in
Europe a myth, or a pathological abnormity. It is now well assured
that it represents a regular and normal species, which grows wild in
the north-eastern portion of Yun-nan, and is cultivated chiefly as an
ornament in gardens and in temple-courts, the longer stems being used



1 He certainly did not see "a stick of bamboo," as understood by HIRTH (Journal
Am. Or. Soc., Vol. XXXVII, 1917, p. 98), but it was a finished product imported
in a larger quantity.

- Assuredly it was not silk, as arbitrarily inferred by P. v. RICHTHOFEN (China,
Vol. I, p. 465). The word pu never refers to silk materials.

3 For an interesting article on this subject, see D. J. MACGOWAN, Chinese Record-
er, Vol. XVI, 1885, pp. 141-142; further, the same journal, 1886, pp. 140-141. E.
SATOW, Cultivation of Bamboos in Japan, p. 92 (Tokyo, 1899). The square bamboo
(Japanese sikaku-dake) is said to have been introduced into Japan from Liukiu.
FORBES and HEMSLEY, Journal Linnean Soc., Vol. XXXVI, p. 443.

535



S3 6 SlNO-lRANICA

for staves, the smaller ones for tobacco-pipes. The shoots of this species
are prized above all other bamboo-shoots as an esculent.

The Pel hu lu l has the following notice on staves of the square
bamboo: "C'en cou $1 (in Kwan-si) produces the square bamboo.
Its trunk is as sharp as a knife, and is very strong. It can be made into
staves which will never break. These are the staves from the bamboo
of ICiun r, mentioned by Can K'ien. Such are produced also in Yuri
Sou & W, 2 the largest of these reaching several tens of feet in height.
According to the Cen $en tsi JE ^ 3ft, there are in the southern ter-
ritory square bamboo staves on which the white cicadas chirp, and
which C'en Cen-tsie K M 15 has extolled. Moreover, Hai-yen M H 3
produces rushes (lu JH, Phragmites communis) capable of being made
into staves for support. P'an 6ou M #I 4 produces thousand-years ferns
T ^ W, and walking-sticks which are small and resemble the palmyra
palm J| & (Borassus tftabelliformis') . There is, further, the su-tsie
bamboo J$ IB 1t, from which staves are abundantly made for the
Buddhist and Taoist clergy, all singular objects. According to the
Hui tsui if ft, the Vuh M bamboo from the Cen River K JI| is straight,
without knots in its upper parts, and hollow."

The Ko ku yao lun 5 states that the square bamboo is produced in
western Se-S'wan, and also grows on the mountain Fei-lai-fun 3$ ^ ^
on the West Lake in Ce-kian; the knots of this bamboo are prickly,
hence it is styled in Se-6'wan tse lu M 1t ("prickly bamboo").

According to the Min siao ki P3 /h IS, 6 written by Cou Liafi-kun
M J X in the latter part of the seventeenth century, square bamboo
and staves made from it are produced in the district of Yuri-tin ^C 3t
in the prefecture of T'in-c'ou and in the district of T'ai-niii ^ ^ in the
prefecture of Sao-wu, both in Fu-kien Province. 7

1 Ch. 3, p. 10 b (ed. of Lu Sin-yuan) ; see above, p. 268.

2 In the prefecture of Liu-ou, Kwan-si.

3 Explained in the commentary as the name of a locality, but its situation is
not indicated and is unknown to me.

4 The present Mou-min hi en, forming the pref ectural city of Kao-Sou f u, Kwan-tun.
6 Ch. 8, p. 9 (ed. of Si yin hilan ts'un Su).

6 Ed. of $wo lin, p. 17.

7 The San hai kin mentions the "narrow bamboo (hia lu ffi 1^) growing in
abundance on the Tortoise Mountain"; and Kwo P'o (A.D. 276-324), in his com-
mentary to this work, identifies with it the bamboo of Kiun. According to the
Kwan ci, the Kiun bamboo occurred in the districts of Nan-kwan [ff He (at present
Nan-k'i ffif j(f|) and Kiun-tu in Se-5'wan. The Memoirs of Mount Lo-fou (Lo-fou
San ki) in Kwan-tun state that the Kiun bamboo was originally produced on Mount
Kiun, being identical with that noticed by Can K'ien in Ta-hia, and that village-
elders use it as a staff. A treatise on bamboo therefore calls it the "bamboo support-
ing the old " ^ ^ Yf - These texts are cited in the T'ai p*in yu Ian (Ch. 963, p. 3).



IRANO-SINICA THE SQUARE BAMBOO, SILK 537

It is said to occur also in the prefecture of Ten-Sou ^ #1, San-tun
Province, where it is likewise made into walking-sticks. 1 The latter
being much in demand by Buddhist monks, the bamboo has received
the epithet "Lo-han bamboo" (bamboo of the Arhat). 2

It is perfectly manifest that what was exported from Se-c'wan by
way of Yun-nan into India, and thence forwarded to Bactria, was the
square bamboo in the form of walking-canes. India is immensely rich
in bamboos; and only a peculiar variety, which did not exist in India,
could have compensated for the trouble and cost which this long and
wearisome trade-route must have caused in those days. For years, I
must confess, it has been a source of wonder to me why Se-c'wan bamboo
should have been carried as far as Bactria, until I encountered the text
of the Pei hn IM, which gives a satisfactory solution of the problem. 3

2. The most important article by which the Chinese became
famously known in ancient times, of course, was silk. This subject is so
extensive, and has so frequently been treated in special monographs,
that it does not require recapitulation in this place. I shall only recall
the fact that the Chinese silk materials, after traversing Central Asia,
reached the Iranian Parthians, who acted as mediators in this trade
with the anterior Orient. 4 It is assumed that the introduction of seri-
culture into Persia, especially into Gilan, where it still flourishes, falls
in the latter part of the Sasanian epoch. It is very probable that the
acquaintance of the Khotanese with the rearing of silkworms, introduced
by a Chinese princess in A.D. 419, gave the impetus to a further growth
of this new industry in a western direction, gradually spreading to
Yarkand, Fergana, and Persia. 5 Chinese brocade (diba-i cm) is fre-
quently mentioned by Firdaus! as playing a prominent part in Persian
decorations. 6 He also speaks of a very fine and decorated Chinese silk
under the name parniydn, corresponding to Middle Persian parnlkan. 1
Iranian has a peculiar word for "silk," not yet satisfactorily explained:
Pahlavi *apresum, *aparesum; New Persian abreSum, abreSam (Arme-

1 San tun t*uh ci, Ch. 9, p. 6.

2 See K'ien su Jj^ ^jf , Ch. 4, p. 7 b (in Yue ya fan ts'un Su, t*ao 24) and Su K'ien
su, Ch. 7, p. 2 b (ibid.). Cf. also u p*u sian lu ft |g j^ f, written by Li K'an
:$: ffj in 1299 (Ch. 4, p. i b; ed. of Ci pu tsu cai ts'un su).

3 The speculations of J. MARQUART (Eransahr, pp. 319-320) in regard to this
bamboo necessarily fall to the ground. There is no misunderstanding on the part
of Can K'ien, and the account of the Si ki is perfectly correct and clear.

4 HIRTH, Chinesische Studien, p. 10.

5 SPIEGEL, Eranische Altertumskunde, Vol. I, p. 256.

6 J. J. MODI, Asiatic Papers, p. 254 (Bombay, 1905).

7 HUBSCHMANN, Persische Studien, p. 242.



538 SlNO-lRANICA

nian, loan-word from Persian, apribtm); hence Arabic ibarisam or
ibrisam; Pamir dialects war sum, warsiim, Sugni wre%om, etc.; Afghan
wresam. 1 Certain it is that we have here a type not related to any
Chinese word for "silk." In this connection I wish to register my utter
disbelief in the traditional opinion, inaugurated by KLAPROTH, that
Greek ser (" silk- worm "; hence Seres, Serica) should be connected with
Mongol sirgek and Manchu sirge ("silk"), the latter with Chinese se
M. 2 My reasons for rejecting this theory may be stated as briefly as
possible. I do not see how a Greek word can be explained from Mongol
or Manchu, languages which we merely know in their most recent
forms, Mongol from the thirteenth and Manchu from the sixteenth
century. Neither the Greek nor the Mongol-Manchu word can be
correlated with Chinese se. The latter was never provided with a final
consonant. Klaproth resorted to the hypothesis that in ancient dialects
of China along the borders of the empire a final r might (peut-ttre) have
existed. This, however, was assuredly not the case. We know that the
termination V JS, so frequently associated with nouns in Pekingese, is
of comparatively recent origin, and not older than the Yuan period
(thirteenth century) ; the beginnings of this usage may go back to the
end of the twelfth or even to the ninth century. 3 At any rate, it did not
exist in ancient times when the Greek ser came into being. Moreover,
this suffix 'r is not used arbitrarily: it joins certain words, while others
take the suffix tse -?, and others again do not allow any suffix. The
word se, however, has never been amalgamated with 'r. In all probabil-
ity, its ancient phonetic value was *si, sa. It is thus phonetically im-
possible to derive from it the Mongol-Manchu word or Korean sir,
added by Abel-Remusat. I do not deny that this series may have its
root in a Chinese word, but its parentage cannot be traced to se. I do

1 HUBSCHMANN, Arm. Gram., p. 107; HORN, Neupers. Etymologic, No. 65.
The derivation from Sanskrit k$auma is surely wrong. Bulgar ibrisim, Rumanian
ibriSin, are likewise connected with the Iranian series.

2 Cf. KLAPROTH, Conjecture sur 1'origine du nom de la soie chez les anciens
(Journal asiatique, Vol. I, 1822, pp. 243-245, with additions by ABEL-REMUSAT,
245-247); Asia polyglotta, p. 341; and Me"moires relatifs a 1'Asie, Vol. Ill, p. 264.
Klaproth's opinion has been generally, but thoughtlessly, accepted (HIRTH, op.
tit., p. 217; F. v. RICHTHOFEN, China, Vol. I, p. 443; SCHRADER, Reallexikon, p. 757).
PELLIOT (T'oung Pao, 1912, p. 741), I believe, was the first to point out that Chinese
se was never possessed of a final consonant.

3 See my note in T'oung Pao, 1916, p. 77; and H. MASPERO, Sur quelques textes
anciens de chinois parle", p. 12. Maspero encountered the word mao'r (" cat ") in a text
of the ninth century. It hardly makes any great difference whether we conceive V
as a diminutive or as a suffix. Originally it may have had the force of a diminutive,
and have gradually developed into a pure suffix. Cf. also P. SCHMIDT, K istorii
kitaiskago razgovornago yazyka, in Sbornik stat'ei professorov, p. 19 (Vladivostok,
1917).



I RANO-SiNiCA SILK, PEACH AND APRICOT 539

not believe, either, that Russian folk ("silk"), as is usually stated (even
by Dal'), is derived from Mongol sirgek: first of all, the alleged phonetic
coincidence is conspicuous by its absence; and, secondly, an ancient
Russian word cannot be directly associated with Mongol; it would be
necessary to trace the same or a similar word in Turkish, but there it
does not exist; "silk" in Turkish is ipak, torgu, torka, etc. It is more
probable that the Russian word (Old Slavic selk, Lithuanian $snlkc&),
in the same manner as our silk, is traceable to sericum. There is no
reason to assume that the Greek words ser, Sera, Seres, etc., have
their origin in Chinese. This series was first propagated by
Iranians, and, in my opinion, is of Iranian origin (cf. New Persian
sarah, "silk"; hence Arabic sarak).

Persian kimxaw or kamxab, kamxa, kimxd (Arabic kimxaw, Hin-
dustani kamxab), designating a "gold brocade," as I formerly ex-
plained, 1 may be derived from Chinese IS ffi kin-hwa, *kim-xwa.

3-4. Of fruits, the West is chiefly indebted to China for the peach
(Amygdalus persica) and the apricot (Prunus armeniaca). It is not
impossible that these two gifts were transmitted by the silk-dealers,
first to Iran (in the second or first century B.C.), and thence to Armenia,
Greece, and Rome (in the first century A.D. ) . In Rome the two trees appear
as late as the first century of the Imperium, being mentioned as Persica
and Armeniaca arbor by Pliny 2 and Columella. Neither tree is men-
tioned by Theophrastus, which is to say that they were not noted
in Asia by the staff of Alexander's expedition. 3 DE CANDOLLE has ably
pleaded for China as the home of the peach and apricot, and ENGLER 4
holds the same opinion. The zone of the wild apricot may well extend
from Russian Turkistan to Sungaria, south-eastern Mongolia, and the
Himalaya; but the historical fact remains that the Chinese have been -p^
the first to cultivate this fruit from ancient times. Previous authors
have justly connected the westward migration of peach and apricot
with the lively intercourse of China and western Asia following Can
K'ien's mission. 5 Persian has only descriptive names for these fruits,
the peach being termed saft-alu ("large plum"), the apricot zard-dlu

1 T'oung Pao, 1916, p. 477; YULE, Hobson-Jobson, p. 484.

2 xv, n, 13.

3 DE CANDOLLE (Origin of Cultivated Plants, p. 222) is mistaken in crediting
Theophrastus with the knowledge of the peach. JORET (Plantes dans 1'antiquite",
p. 79) has already pointed out this error, and it is here restated for the benefit of
those botanists who still depend on de Candolle's book.

4 In Hehn, Kulturpflanzen, p. 433.

5 JORET, op. cit., p. 81; SCHRADER in Hehn, p. 434.



540 SlNO-lRANICA

("yellow plum"). 1 Both fruits are referred to in Pahlavi literature
(above, pp. 192, 193).

As to the transplantation of the Chinese peach into India, we have
an interesting bit of information in the memoirs of the Chinese pilgrim
Htian Tsafi. 2 At the time of the great Indo-Scythian king Kaniska,
whose fame spread all over the neighboring countries, the tribes west of
the Yellow River (Ho-si in Kan-su) dreaded his power, and sent hostages
to him. Kaniska treated them with marked attention, and assigned to
them special mansions and guards of honor. The country where the
hostages resided in the winter received the name Cmabhukti ("China
allotment," in the eastern Panjab). In this kingdom and throughout
India there existed neither pear nor peach. These were planted by the
hostages. The peach therefore was called cmanl ("Chinese fruit");
and the pear, cmarajaputra ("crown-prince of China"). These names
are still prevalent. 3 Although Hiian Tsafi recorded in A.D. 630 an oral
tradition overheard by him in India, and relative to a time lying back
over half a millennium, his well-tested trustworthiness cannot be
doubted in this case: the story thus existed in India, and may indeed
be traceable to an event that took place under the reign of Kaniska,
the exact date of which is still controversial. 4 There are mainly two rea-
sons which prompt me to accept Huan Tsafi's account. From a botani-
cal point of view, the peach is not a native of India. It occurs there only

1 In the Pamir languages we meet a common name for the apricot, Minjan
eri, WaxI ciwan or loan (but Sariqoll no$, Signi na&). The same type occurs in the
Dardu languages (jui or ji for the tree, jarote or jorote for the fruit, and juru for
the ripe fruit) and in Kacmlii (tser, tser-kul) ; further, in West-Tibetan cu-li or co-li,
Balti su-ri, Kanaurl lul (other Tibetan words for "apricot "are k'am-bu, a-u, and
Sa-rag, the last-named being dried apricots with little pulp and almost as hard as
a stone). KLAPROTH (Journal asiatique, Vol. II, 1823, p. 159) has recorded in Bu-
khara a word for the apricot in the form iserduli. It is not easy to determine how this
type has migrated. TOMASCHEK (Pamir-Dialekte, p. 791) is inclined to think that
originally it might have been Tibetan, as Baltistan furnishes the best apricots.
For my part, I have derived the Tibetan from the Pamir languages (T'oung Pao,
1916, p. 82). The word is decidedly not Tibetan; and as to its origin, I should
hesitate only between the Pamir and Dardu languages.

2 Ta Tan Si yil ki t Ch. 4, p. 5.

8 There are a few other Indian names of products formed with "China":
cinapitfa ("minium"), cinaka ("Panicum miliaceum, fennel, a kind of camphor"),
cinakarpura ("a kind of camphor"), cinavanga ("lead").

4 Cf. V. A. SMITH, Early History of India, 3d ed., p. 263 (I do not believe with
Smith that "the territory of the ruler to whose family the hostages belonged seems
to have been not very distant from Kashgar"; the Chinese term Ho-si, at the time
of the Han, comprised the present province of Kan-su from Lan-c"ou to An-si);
T. WAITERS, On Yuan Chwang's Travels, Vol. I, pp. 292-293 (his comments on
the story of the peach miss the mark, and his notes on the name Clna are erroneous;
see also PELLIOT, Bull, de VEcolefranQaise, Vol. V, p. 457).



IRANO-SINICA PEACH, CINNAMON 541

in a cultivated state, and does not even succeed well, the fruit being
mediocre and acid. 1 There is no ancient Sanskrit name for the tree; nor
does it play any rdle in the folk-lore of India, as it does in China. Fur-
ther, as regards the time of the introduction, whether the reign of
Kaniska be placed in the first century before or after our era, it is
singularly synchronous with the transplantation of the tree into western
Asia.

5. As indicated by the Persian name ddr-cml or dar-cm ("Chinese
wood" or "bark"; Arabic ddr sml), cinnamon was obtained by the c<^
Persians and Arabs from China. 2 Ibn Khordadzbeh, who wrote between
A.D. 844 and 848, is the first Arabic author who enumerates cinnamon
among the products exported from China. 3 The Chinese export cannot
have assumed large dimensions: it is not alluded to in Chinese records,
Cao Zu-kwa is reticent about it. 4 Ceylon was always the main seat of ^Tj
cinnamon production, and the tree (Cinnamomum zeylanicum) is a native
of the Ceylon forests. 5 The bark of this tree is also called dar-cmi. It
is well known that cassia and cinnamon are mentioned by classical
authors, and have given rise to many sensational speculations as to the
origin of the cinnamon of the ancients. Herodotus 6 places cinnamon in
Arabia, and tells a wondrous story as to how it is gathered. Theo-
phrastus 7 seeks the home of cassia and cinnamomum, together with
frankincense and myrrh, in the Arabian peninsula about Saba, Had-
ramyt, Kitibaina, and Mamali. Strabo 8 locates it in the land of the
Sabaeans, in Arabia, also in Ethiopia and southern India; finally he has
a "cinnamon-bearing country" at the end of the habitable countries
of the south, on the shore of the Indian ocean. 9 Pliny 10 has cinnamomum
or cinnamum grow in the country of the Ethiopians, and it is carried
over sea on rafts by the Troglodytae.

1 C. JORET, Plantes dans l'antiquit<, Vol. II, p. 281.

2 LECLERC, Traite" des simples, Vol. II, pp. 68, 272. The loan-word daricenik
in Armenian proves that the word was known in Middle Persian (*dar-i c"enik) ; cf .
HUBSCHMANN, Armen. Gram., p. 137.

8 G. FERRAND, Textes relatifs a l f Extreme-Orient, p. 31.

4 SCHOFF (Periplus, p. 83) asserts that between the third and sixth centuries
there was an active sea-trade in this article in Chinese ships from China to Persia.
No reference is given. I wonder from what source this is derived.

5 DE CANDOLLE, Origin of Cultivated Plants, p. 146; WATT, Commercial Prod-
ucts of India, p. 313.

8 in, 107, in.

7 Hist, plant., IX. iv, 2.

3 XV. iv, 19; XVI. iv, 25; XV. I, 22.
9 1. iv, 2.

10 xii, 42.



542 SlNO-lRANICA

The descriptions given of cinnamon and cassia by Theophrastus 1
show that the ancients did not exactly agree on the identity of these
plants, and Theophrastus himself speaks from hearsay ("In regard to
cinnamon and cassia they say the following: both are shrubs, it is said,
and not of large size. . . . Such is the account given by some. Others
say that cinnamon is shrubby or rather like an under-brush, and that
there are two kinds, one black, the other white"). The difference be-
tween cinnamon and cassia seems to have been that the latter possessed
stouter branches, was very fibrous, and difficult to strip off the bark.
This bark was used; it was bitter, and had a pungent odor. 2

Certain it is that the two words are of Semitic origin. 3 The fact that
there is no cinnamon in Arabia and Ethiopia was already known to
GARCIA DA ORTA. 4 An unfortunate attempt has been made to trace
the cinnamon of the ancients to the Chinese. 5 This theory has thus
been formulated by Muss-ARNOLx: 6 "This spice was imported by
Phoenician merchants from Egypt, where it is called khisi-t. The
Egyptians, again, brought it from the land of Punt, to which it was
imported from Japan, where we have it under the form kei-chi ('branch
of the cinnamon-tree'), or better kei-shin ('heart of the cinnamon')
[read sin, *sim]. The Japanese itself is again borrowed from the Chinese
kei-& [?]. The -t in the Egyptian represents the feminine suffix." As
may be seen from O. SCHRADER, ? this strange hypothesis was first put
forward in 1883 by C. SCHUMANN. Schrader himself feels somewhat
sceptic about it, and regards the appearance of Chinese merchandise on
the markets of Egypt at such an early date as hardly probable. From a
sinological viewpoint, this speculation must be wholly rejected, both
in its linguistic and its historical bearings. Japan was not in existence
in 1500 B.C., when cinnamon-wood of the country Punt is spoken of in
the Egyptian inscriptions; and China was then a small agrarian inland
community restricted to the northern part of the present empire, and

1 Hist, plant., IX. v, 1-3.

2 Theophrastus, IX. v, 3.

3 Greek /ccurJa is derived from Hebrew qesi'a, perhaps related to Assyrian kasu,
kasiya (POGNON, Journal asiatique, 1917, I, p. 400). Greek kinnamomon is traced
to Hebrew qinnamon (Exodus, xxx, 23).

4 MARKHAM, Colloquies, pp. 119-120.

6 Thus also FLUCKIGER and HANBURY (Pharmacographia, p. 520), whose
argumentation is not sound, as it lacks all sense of chronology. The Persian term
dar-clnl, for instance, is strictly of mediaeval origin, and cannot be invoked as evidence
for the supposition that cinnamon was exported from China many centuries before
Christ.

6 Transactions Am. Phil. Assoc., Vol. XXIII, 1892, p. 115.

7 Reallexikon, p. 989.



IRANO-SINICA CINNAMON 543

not acquainted with any Cassia trees of the south. Certainly there was
no Chinese navigation and sea-trade at that time. The Chinese word
kwei ft (*kwai, kwi) occurs at an early date, but it is a generic term for
Lauraceae; and there are about thirteen species of Cassia, and about
sixteen species of Cinnamomum ,'m China. The essential point is that the
ancient texts maintain silence as to cinnamon; that is, the product from
the bark of the tree. Cinnamomum cassia is a native of Kwaii-si, Kwan-
tun, and Indo-China ; and the Chinese made its first acquaintance under
the Han, when they began to colonize and to absorb southern China.
The first description of this species is contained in the Nan fan ts*ao
mu cwan of the third century. 1 This work speaks of large forests of this
tree covering the mountains of Kwan-tun, and of its cultivation in
gardens of Kiao-ci (Tonking) . It was not the Chinese, but non-Chinese
peoples of Indo-China, who first brought the tree into cultivation, which,
like all other southern cultivations, was simply adopted by the con-
quering Chinese. The medicinal employment of the bark (kwei p*i
&) is first mentioned by T'ao Hun-kin (A.D. 451-536), and probably
was not known much earlier. It must be positively denied, however,
that the Chinese or any nation of Indo-China had any share in the
trade which brought cinnamon to the Semites, Egyptians, or Greeks
at the time of Herodotus or earlier. The earliest date we may assume
for any navigation from the coasts of Indo-China into the Indian Ocean
is the second century B.C. 2 The solution of the cinnamon problem of
the ancients seems simpler to me than to my predecessors. First, there
is no valid reason to assume that what our modern botany understands
by Cassia and Cinnamomum must be strictly identical with the products
so named by the ancients. Several different species are evidently in-
volved. It is perfectly conceivable that in ancient times there was a
fragrant bark supplied by a certain tree of Ethiopia or Arabia or both,
which is either extinct or unknown to us, or, as Fee inclines to think,
a species of Amyris. It is further legitimate to conclude, without forc-
ing the evidence, that the greater part of the cinnamon supply came from
Ceylon and India, 3 India being expressly included by Strabo. This, at
least,' is infinitely more reasonable than acquiescing in the wild fantasies
of a Schumann or Muss-Arnolt, who lack the most elementary knowl-
edge of East-Asiatic history.

6. The word " China " in the names of Persian and Arabic products,

1 The more important texts relative to the subject are accessible in BRET-
SCHNEIDER, Bot. Sin., pt. Ill, No. 303.